Poster Session 3

G. Miserocchi¹, A. Farolfi¹, F. Merloni¹, C. Gianni¹, N. Gentili¹, M. Mariotti¹, G. Di Menna¹, O. Serra¹, C. Casadei¹, F. Rusconi², A. Andalò¹, S. Sabbioni¹, A. Carlini¹, D. Montanari¹, M. Sirico¹, R. Maltoni¹, L. Cecconetto¹, S. Sarti¹, M. Palleschi¹, A. Musolino¹; ¹Medical Oncology, Breast & GYN Unit, IRST IRCCS Dino Amadori, Meldola, ITALY, ²TrineTX, TrineTX, Cambridge, MA, USA, MA.

Impact of Medically Assisted Procreation and Pregnancy on Breast and Gynecologic Cancer Risk in Women Under 45: A Real-World, Propensity-Matched Analysis Using the TriNetX NetworkBackgroundThe potential oncologic risks associated with medically assisted procreation (MAP) remain a concern, particularly in young women undergoing hormonal stimulation. Although previous studies have suggested no significant increase in breast cancer (BC) risk after MAP, data remain limited for ovarian (OC) and endometrial (EC) cancers. Moreover, the protective role of pregnancy against BC is well established but less explored for other gynecologic malignancies. This study aimed to evaluate the risk of BC, OC, and EC in women under 45 undergoing MAP compared with women who had a delivery, and to assess the impact of pregnancy compared with nulliparity.MethodsThis was a retrospective analysis using real-world data from the TriNetX Global Collaborative Network. Two propensity-matched comparisons were performed: (1) MAP vs. a cohort of parous women undergoing mammographic screening (n=878 per group); (2) parous vs. nulliparous women undergoing mammographic screening (n=1,322 per group). Matching was performed on age, race, body mass index (BMI), genetic predisposition to malignancies, and previous oral contraceptive use. Outcomes included the incidence of BC, OC, and EC. Risk ratios (RR), odds ratios (OR), hazard ratios (HR), and cumulative incidence curves were calculated. The observation period extended up to 20 years.ResultsIn the first analysis, no significant difference was found in BC risk (3.6% vs. 4.3%; RR=0.84; HR=1.005; p=0.985). OC cases were absent in both cohorts. EC incidence was significantly higher in the MAP group, although the number of cases was low (1.1% vs. none; p=0.002). In the second analysis, parous women had a significantly lower BC incidence (3.3% vs. 7.3%; RR=0.45; HR=0.42; p=0.010). OC occurred only in nulliparous women (0.8% vs. none; p=0.002). Conversely, EC was higher in parous women (0.8% vs. none; p=0.002).ConclusionsThis real-world study suggests that MAP in women under 45 does not increase the risk of BC or OC compared with parous women. However, a possible increased risk of EC, although based on few cases, warrants attention. Moreover, pregnancy appears to confer a significant protective effect against BC and OC compared with nulliparity. Pending further long-term studies, these data support the oncologic safety of MAP for BC and OC, while highlighting the importance of individualized risk counseling regarding EC. Keywords breast cancer, ovarian cancer, endometrial cancer, medically assisted procreation, fertility treatments, pregnancy, nulliparity, TriNetX, real-world data, cancer risk.

T. Simmons, E. Hughes, M. Kucera, A. Gutin; Biostatistics, Myriad Genetics, Inc., Salt Lake City, UT.

Background Polygenic risk scores (PRSs) combine information from single-nucleotide polymorphisms (SNPs) across the genome to explain a substantial portion of genetic breast cancer (BC) susceptibility. In previous studies, a multiple-ancestry PRS (MA-385) based on 56 ancestry-informative and 329 BC-associated SNPs was accurate for diverse populations and ranked among the most significant factors affecting the risk of BC development. However, medical guidelines discourage the routine clinical use of PRS, citing a need for studies to evaluate potential interactions between SNPs and environmental and hormonal risk factors in diverse populations. Here, we investigate interactions of MA-385 and individual BC-associated SNPs with clinical risk factors in the widely used Tyrer-Cuzick (TC) model in a large cohort of self-reported Black/African women. Methods We examined clinical and genetic records from self-reported Black/African women who were referred for hereditary cancer testing with a multigene panel and tested negative for pathogenic variants in BC-associated genes. Multivariable logistic regression models adjusted for age, personal/family cancer history, and genetic ancestry were used to test individual TC risk factors, MA-385, and 5 individual BC SNPs with the greatest impact on BC risk for women of Black/African ancestry. Effect modification of MA-385 and individual SNPs by TC risk factors was evaluated by including interaction terms in the models. Odds ratios (ORs) and 95% confidence intervals (CIs) are reported per standard deviation of MA-385. Significance tests were performed using two-sided p-values based on likelihood-ratio test statistics. Results We identified 39,817 women who met study eligibility criteria. 8,802 (22.1%) women had been diagnosed with invasive BC prior to hereditary cancer testing. 17,529 women, independent of previous MA-385 development studies, were included in analyses of MA-385. Individual SNPs and clinical TC risk factors were examined in women with complete risk factor information. All 39,817 women had complete information for age and BC family history; sample sizes for other clinical risk factors ranged from 1,795 for breast density to 29,486 for height. The MA-385 was a highly significant predictor of invasive BC after accounting for clinical factors (OR=1.38; 95% CI 1.32-1.45, p=4.2 x 10^-45). We found no evidence of interaction of MA-385 with any individual TC factor (all Bonferroni-adjusted p>0.05). The strongest evidence of interaction was for hormone replacement therapy status (unadjusted p=0.04; Bonferroni-adjusted p=0.66). The 5 SNPs with the greatest impact on BC risk for women of Black/African ancestry were distinct from those previously identified as having the greatest impact on BC risk for women of European ancestry. We found no evidence of interaction between these individual SNPs and TC risk factors (all Bonferroni-adjusted p>0.05). The strongest evidence for interaction was between a SNP on chromosome 19 (rs2363956) and BMI category (unadjusted p=0.04; Bonferroni-adjusted p=1). Conclusions In a large cohort of Black/African women, MA-385 was a highly significant predictor of BC and substantially improved risk prediction over clinical risk factors. In contrast, interactions of MA-385 and individual BC SNPs with clinical risk factors were not statistically significant. These findings suggest that TC factors have a minimal, if any, impact on the strength of the association between the MA-385 PRS and invasive BC in women of Black/African ancestry.

H. Chiang¹, R. Araya¹, L. Qi², H. Arestake¹, S. Martinez¹, A. Sanchez¹, I. Kalia³, Y. Hu², A. Horvath¹, R. Li¹; ¹Biochemistry and Molecular Medicine, George Washington University, Washington, DC, ²Anatomy & Cell Biology, George Washington University, Washington, DC, ³Medicine, George Washington University-MFA, Washington, DC.

Women harboring germline mutations in BRCA1 have an approximately 80% lifetime risk of developing breast cancer. Extensive research on the breast epithelium, the cell of origin for BRCA1-related tumors, has shed light on the molecular functions of BRCA1 in tumor suppression. However, all somatic cells in a germline BRCA1 mutation carrier contain the same mutant allele, yet it remains unclear whether BRCA1 heterozygosity in stromal compartments could also contributes to the increased risk of BRCA1-associated cancer. To determine the impact of BRCA1 heterozygosity on T cells at the transcriptional and functional levels, we first investigated a cohort of T cells isolated from healthy individuals who are BRCA1 mutation carriers (BRCA1^mut/+) and non-carriers, using single cell RNA sequencing (scRNA-seq) and high-parametric spectral flow cytometry. Single-cell transcriptomic analysis revealed a significant enrichment of a gamma-delta (γδ) T cell subpopulation in BRCA1^mut/+ samples. A similar trend, albeit not statistically significant, was observed for a SOX4^high T cell subpopulation. Functionally, BRCA1^mut/+ T cells exhibited a delayed response to CD3/CD28 costimulation and showed an enrichment of naïve cells in the CD4 compartment. In addition, the production of cytokine IL-17 is significantly reduced in BRCA1^mut/+ T cells after 24 hours of activation. To determine a causal effect of T-cell BRCA1 heterozygosity on tumorigenesis, we next deleted one allele Brca1 in mouse T lymphocytes. Implanted mammary tumors grew more robustly in T cell-specific Brca1^-/+ mice than their Brca^+/+ counterparts, supporting the notion that Brca1 heterozygosity in mouse T lymphocytes compromises the antitumor immunity. Furthermore, mechanistic studies indicate that BRCA1 regulates T cell-related gene expression in conjunction with other transcriptional coregulators. In summary, our data from clinical samples and preclinical models suggest that the heterozygous BRCA1 mutation status leads to T cell-intrinsic functional alterations and compromised antitumor immunity. We propose that germline BRCA1 mutations exert a “double whammy” effect—contributing to elevated cancer risk through their presence in both the cell of origin and the surrounding stromal compartments.

B. Arun¹, S. Rivero-Hinojosa², R. Green², V. N. Aushev², A. Antonopolous², A. Rodriguez², M. C. Liu², J. E. Garber³; ¹Breast Medical Oncology, Division of Cancer Medicine, The University of Texas MD Anderson Cancer Center, Houston, TX, ²Oncology, Natera, Inc., Austin, TX, ³Division for Cancer Genetics and Prevention, Dana-Farber Cancer Institute and Harvard Medical School, Boston, MA.

Background: Germline mutations in CDH1 are well-established as risk factors for hereditary lobular breast cancer (BC). However, the frequency and phenotypic landscape of CDH1 germline mutations in unselected BC populations remain incompletely defined. Here, we investigated the prevalence of CDH1 pathogenic/likely pathogenic (P/LP) germline variants in a large real-world BC cohort. Methods: We analyzed germline whole-exome sequencing (WES) data from 53,607 BC patients in Natera’s Real-World Database. CDH1 variants were annotated using Ensembl Variant Effect Predictor (VEP). Variants classified as “LOW” impact (synonymous or non-functional splice region variants) were excluded. Germline P/LP classification was sourced from ClinVar. In the absence of ClinVar annotation, truncating variants were designated as P/LP. Ancestry inference was determined using the EthSeq algorithm with four groups: African (AFR), East Asian (EAS), European (EUR), and South Asian (SAS). Global variant frequencies were also compared using 1000 Genomes Project data. Results: Among the 53,607 BC patients analyzed, the majority had ductal histology (68.55%), followed by lobular (10.36%), mixed (2.78%), and not available (18.31%). The median age in the cohort was 57 [range: 20-102]. Prevalence of germline CDH1 variants meeting the inclusion criteria was 11.0% (5892/53607) globally, with 10.9%, 11.9%, 7.7%, and 8.8% prevalence in EUR, AFR, EAS, and SAS populations, respectively. By histology, 11.1% of lobular and 11.7% of ductal cases presented a CDH1 germline variant. We observed 367 unique CDH1 allele variants, with 19 classified as P/LP. These included three splice acceptor site variants, twelve variants within the extracellular protein domain, one in the transmembrane domain, and three within the cytoplasmic domain. P/LP CDH1 variants were found in 23 individuals, representing 0.39% of patients with CDH1 germline variants and 0.25% of all patients with lobular breast cancer (0.042% of all BC patients analyzed). Among the patients with P/LP variants, 14 had lobular, 7 ductal, and 2 unknown histologies, with a median age of 54 [range: 38-71]. The majority of pathogenic variants were observed in patients of European ancestry. The variant spectrum included known loss-of-function alleles, such as truncating and splice-site alterations, as well as selected missense changes with predicted deleterious impact, indicating a range of functional consequences relevant to tumor biology. Conclusions: In this large real-world BC cohort, CDH1 germline variants were observed in 11% of patients with variable prevalence by genetic ancestry. We identified 19 P/LP CDH1 variants in 23 BC patients. These germline variants remain strongly enriched in lobular BC histology, consistent with prior literature, with 0.25% of patients with lobular histology carrying variants classified as P/LP. The majority of P/LP variants localized to functionally important regions of the E-cadherin protein, including the extracellular and cytoplasmic domains, underscoring their likely disruption of cell-cell adhesion. These findings support broader consideration of CDH1 in germline testing panels for BC risk assessment. Overall, this study reinforces the clinical value of CDH1 germline analysis, particularly in lobular BC.

A. Spata¹, V. Goldbarg¹, M. Brichard², M. Ibanez-Alda³, I. Valle⁴, F. Giugliano⁵, T. Grinda¹, A. Viansone¹, C. Poisson¹, C. Bousrih¹, S. Akla¹, C. Roussel-Simonin¹, E. Rassy¹, A. Sabau¹, L. Antoun¹, C. Nicotra⁶, M. NgoCamus⁶, B. Verret⁷, A. Bayle⁶, E. Rouleau⁸, B. Pistilli¹, A. Italiano⁹, O. Caron¹, J. M. Ribeiro¹; ¹Department of Medical Oncology, “Gustave Roussy” Institute, Villejuif, FRANCE, ²Department of Medical Oncology, Cliniques universitaires Saint-Luc, Bruxelles, BELGIUM, ³Department of Medical Oncology, Hospital Universitario Marqués de Valdecilla, Santander, SPAIN, ⁴Clinical Oncology Unit, Careggi University Hospital, Firenze, ITALY, ⁵Unit 981, Molecular Predictors and Novel Targets in Oncology, “Gustave Roussy” Institute, Villejuif, FRANCE, ⁶Drug Development Department (DITEP), “Gustave Roussy” Institute, Villejuif, FRANCE, ⁷INSERM Unit U981, “Gustave Roussy” Institute, Villejuif, FRANCE, ⁸Medical Biology and Pathology Department, “Gustave Roussy” Institute, Villejuif, FRANCE, ⁹Department of Early Phase Trial Unit, Institut Bergonié Comprehensive Cancer Centre, Bordeaux, FRANCE.

Background: In France, access to PARP-inhibitors for metastatic breast cancer (mBC) is restricted to BRCA1, BRCA2 and PALB2 germline variant carriers. Genomic tumor sequencing or circulating tumor DNA assays (ctDNA or liquid biopsy (LB)) can incidentally reveal potential germline alterations. Currently, LB assays are nor designed nor validated for germline variant detection or interpretation, and it is not known whether their performances are good enough to replace a regular germline test. Our aim is to quantify the concordance between germline and LB in BC, and to explore the potential of LB to serve as a proxy for germline testing.Methods: Among patients (pts) with mBC enrolled between February 2021 and March 2025 in the STING molecular screening program (NCT04932525), pts with a pathogenic/likely pathogenic variant (P/LPVs) in BRCA1, BRCA2, and PALB2 genes in ctDNA were retrieved through the study database. LB in STING was carried out by using the FoundationOne Liquid CDx assay covering the analysis of 324 genes. We checked all the selected pts medical records to look for germline test results. We first performed descriptive analyses of clinical and pathological variables of the selected pts. We then assessed the positive predictive value (PPV) and false discovery rate (FDR) of ctDNA. Variant allele frequency (VAF) values were explored to evaluate their potential as a surrogate marker of germline origin, with a threshold of 40% VAF prospectively examined based on prior biological rationale.Results: Among 1,325 pts enrolled in STING program, 119 (9.0%) LB showed P/LPVs in BRCA1 (n=33; 27.7%), BRCA2 (n=71; 59.7%), or PALB2 (n=15; 12.6%). Germline genetic testing was performed in 93 of these pts (79.8%: 27 BRCA1, 56 BRCA2, and 12 PALB2).Among the entire cohort of 93 pts with BRCA1, BRCA2, or PALB2 mutations detected by LB and who underwent germline testing, 67 pts had confirmed germline P/LPV, resulting in an overall PPV of 72 %. Conversely, 26 pts were germline negative. In this subcohort, median age at diagnosis was 43.5 years (SD ± 12.1) for germline P/LPV carriers versus 45.6 years (SD ± 13.1) for somatic pts.Among the 27 LB BRCA1 pts, 21 (77.78%) pts showed concordance, all with a VAF >40%. LB FDR for BRCA1 variants was 22.2%. Among the 6 discordant pts, 5 pts harboured mutations with VAF <40% (median 4.9%), consistent with somatic origin, while one case involved a rearrangement intron 4 with a VAF of 42%.Among 54 LB BRCA2 pts, 39 (57.4%) showed concordance. Of these, 37 (94.9%) had VAF >40%, while 2 pts exhibited VAF between 26-38%. LB FDR for BRCA2 variants was 27.7%. Among the 15 discordant pts, 14 exhibited low VAF consistent with somatic origin; for one pt with exon 2 loss alteration, VAF was unavailable.Among 12 LB PALB2, 7 (58.3%) pts showed concordance, all with VAF >40%. LB FDR for PALB2 variants was 41.7%. Among the 5 discordant pts, all presented mutations with VAF <40% (median 0.28%) on LB, suggesting somatic variants. Interestingly, in our population, the PALB2 mutation rate was higher than previously reported.Conclusion: Our findings highlight a substantial concordance (PPV 98.5%) between germline and LB results for BRCA1, BRCA2, and PALB2 P/LPVs in mBC pts enrolled in the STING program, when applying a 40% VAF threshold. Although the overall FDR appears elevated, 92% of discordant cases involved variants with VAFs below 40%, which would not typically raise suspicion for germline origin in clinical practice. These results suggest that 40% VAF threshold could be a practical and reliable surrogate marker for germline status provided it is confirmed in larger independent cohorts. Data on LB false negative rates will be presented at the conference, to measure the likelihood of a positive germline test when the LB shows no PV and thereby evaluate the safety of omitting germline testing in LB-negative cases.

M. Thomas¹, A. Woodworth², A. Hazra³, Y. Duron⁴, J. Miller⁵, A. Patel⁶, C. Allen⁷, L. Schlager⁸, L. Soltani⁹, L. Meier¹⁰; ¹National Breast Cancer Roundtable, American Cancer Society, Cincinnati, OH, ²Division of Breast & Soft Tissue Surgery, Henry Mayo Hospital Keck Medicine of USC, Santa Clarita, CA, ³Division of Preventive Medicine, Brigham and Women’s Hospital, Boston, MA, ⁴n/a, The Latino Cancer Institute, San Jose, CA, ⁵National Breast and Cervical Cancer Early Detection Program, Centers for Disease Control and Prevention, Rockville, MD, ⁶The Breast Care Center at Liberty Hospital, The University of Missouri-Kansas City School of Medicine, Liberty, MO, ⁷n/a, The Cancer Prevention and Research Institute of Texas, Austin, TX, ⁸Public Policy, Facing Our Risk of Cancer Empowered, Tampa, FL, ⁹n/a, El Rio Health Congress Health Center, Tucson, AZ, ¹⁰National Breast Cancer Roundtable, American Cancer Society, Atlanta, GA.

Title:Advancing Risk-Informed Breast Cancer Screening: Development of a Digital Toolkit for Primary Care and OB/GYN IntegrationAbstract:Background:Despite strong evidence supporting risk-informed breast cancer screening, implementation in primary care and OB/GYN settings remains limited. Barriers include lack of awareness, clinical time constraints, and inconsistent workflows for identifying and managing at-risk individuals. To address these challenges, the American Cancer Society National Breast Cancer Roundtable (ACS NBCRT) developed a digital Breast Cancer Risk Assessment Toolkit to support frontline providers in initiating, conducting, and acting on breast cancer risk assessment during routine care.Methods:The toolkit was developed over a 12-month period through a collaborative process led by the ACS NBCRT Early Action Priority Team and a national Advisory Group of experts in primary care, breast oncology, health equity, genetics, and public health. The group identified key barriers to implementation and defined practical tools needed to support guideline-concordant care. The resulting toolkit includes:1.Provider communication scripts2.Clinical workflow diagrams3.A risk assessment checklist4.A health system readiness assessment5.A comparison matrix of risk models6.Case studies from academic and community-based settingsContent was reviewed by ACS subject matter experts and aligned with ACS screening guidelines, while acknowledging variation in NCCN and USPSTF recommendations. Tools were designed to support both standalone and team-based care models, with consideration for feasibility in underserved and FQHC settings.Results:Toolkit reviewers—including clinicians and implementation scientists—reported high usability and clinical relevance. The materials supported patient-centered communication, flexible integration into EHR workflows, and strategic decision-making about referrals for genetic counseling or supplemental screening. Case studies highlighted implementation successes from both academic centers and bilingual, community-engaged programs serving Latina women. The model comparison table further supports providers in choosing risk tools that align with their population and system capabilities.Conclusions:The ACS NBCRT Risk Assessment Digital Toolkit offers a novel, digital-first approach to operationalizing breast cancer risk assessment in primary care and women’s health. By providing a streamlined, equity-informed package of tools, the initiative aims to improve early identification of high-risk individuals and expand access to personalized screening and prevention. Future directions include expansion of the case study library to feature additional models from rural and safety-net settings, as well as pilot testing the toolkit through practice-based dissemination and implementation research with primary care and OB/GYN users to assess feasibility, acceptability, and impact on clinical workflows. This scalable resource addresses a long-standing implementation gap and supports broad uptake of risk-informed breast cancer care.

J. B. Gibbons, V. Ro, J. E. McGuinness, R. Carlos, K. D. Crew; Department of Medicine, Division of Hematology / Oncology, Columbia University, New York, NY.

BackgroundCurrent breast cancer risk assessment tools tend to underestimate risk of invasive breast cancer (IBC) among Black or Hispanic women and include clinical risk factors, such as family history of breast cancer, that are not readily available or under-reported in the electronic health record (EHR). Allostatic load (AL)—a measure of cumulative physiological stress which is comprised of standard laboratory results and vitals—has been associated with IBC incidence and mortality. Our objective was to determine whether components of AL derived from the EHR can improve IBC risk prediction. MethodsWe conducted a retrospective analysis of women undergoing screening mammography at Columbia University Irving Medical Center from 11/1/14 to 4/7/25. We extracted EHR data on the earliest index screening mammogram, age at index scan, race, ethnicity, body mass index (BMI), mammographic density (MD, 4-category BIRADS classification), breast biopsy results, and family history of breast cancer. Patients with a diagnosis of IBC or ductal carcinoma in situ prior to the index scan were excluded. The following laboratory data and vitals were collected for AL biometrics representing major regulatory systems: metabolic (albumin, alkaline phosphatase, glucose), renal (blood urea nitrogen, creatinine), immune (white blood cell count), and cardiovascular (pulse, systolic and diastolic blood pressure). Patients with at least five of these components available in the two years preceding the index scan were included, and multiple imputation by chained equations was used to impute missing AL data. For modeling, AL-derived data were used as individual values and as a composite score (0-9), with one point assigned for each biometric in the highest quartile except albumin which was the lowest quartile. The primary outcome measure was development of IBC, defined as histologically-confirmed IBC at least six months after the index scan. Multiple logistic regression models were constructed, and area under the receiver operating characteristic curve (AUC) values were compared using DeLong’s test to a baseline model constructed from established IBC risk factors readily extracted from the EHR. ResultsAmong 27,155 evaluable patients, mean age was 59 years (SD 12) and 40% were Hispanic, 29% non-Hispanic White, and 14% non-Hispanic Black. A total of 135 (0.5%) patients had a subsequent IBC diagnosis. The baseline IBC risk prediction model of age, race / ethnicity, BMI, and MD achieved an AUC of 0.642; adding AL composite score or including breast biopsy results or family history did not significantly improve risk prediction. However, adding all nine individual AL-derived variables improved the AUC to 0.663 (p=0.011). Among non-Hispanic White women, the baseline model with AUC of 0.643 improved with the addition of breast biopsy results and family history (AUC of 0.656, p<0.001) and the AL-derived variables (AUC of 0.691, p=0.037). Among non-Hispanic Black women, the baseline model had an AUC of 0.634, which improved with the addition of breast biopsy results and family history to an AUC of 0.649 (p<0.001). Among Hispanic women, including AL-derived variables significantly improved the baseline model with an AUC of 0.611 to an AUC of 0.714 (p=0.002). Including the composite AL score did not improve prediction for any racial / ethnic group. DiscussionWe demonstrated that addition of AL-derived laboratory data and vitals extracted from the EHR significantly improved IBC risk prediction, particularly among Hispanic women, achieving AUCs comparable to existing well-validated breast cancer risk assessment tools. With further validation, incorporation of AL variables may facilitate efficient IBC risk prediction in the clinical setting to inform breast cancer screening and risk-reducing strategies among diverse women.

G. Pourali¹, L. Lyu², X. Guo², A. Toriola¹; ¹Division of Public Health Sciences, Department of Surgery, Washington University School of Medicine, Saint Louis, MO, ²Division of Epidemiology, Department of Medicine, Vanderbilt Epidemiology Center, Vanderbilt-Ingram Cancer Center, Vanderbilt University Medical Center, Nashville, TN.

Background: Genome-wide association studies (GWAS) have identified genetic loci associated with mammographic breast density (MBD) and breast cancer risk, but the mechanisms underlying these associations remain unclear. Metabolite quantitative trait loci (mQTL) analyses can map genetic loci associated with metabolites, potentially offering insight into mechanisms involved in risk, but there is limited data on how genetically predicted metabolites are associated with MBD and breast cancer risk in premenopausal women. To address this, we performed mQTL analyses in non-Hispanic White (NHW) and non-Hispanic Black (NHB) women and examined their overlap with established GWAS loci for MBD and breast cancer. Methods: Our analysis included 494 NHW and 163 NHB women who had screening mammograms at Washington University in St. Louis. Plasma metabolomic profiling (Metabolon®) identified 1,074 metabolites, and genotyping (Infinium microarray, Illumina) covered 2,028,571 loci. Metabolite data were batch-normalized, metabolites with >10% missingness were excluded, and remaining missing values were imputed. We conducted mQTL mapping using linear regression in PLINK 2.0, stratified by race and adjusted for age, body mass index, alcohol use, oral contraceptive use, and first five genotype principal components. Significant mQTLs (FDR-p<0.05) were compared to GWAS loci for breast cancer, two fine mapping studies, and GWAS of MBD. Results: In NHW women, 30 mQTLs were associated with 54 metabolites overlapping breast cancer GWAS loci. Key pathways included xenobiotics (anti-inflammatory/immunosuppressant drugs, chemicals, tobacco metabolites), lipid (fatty acid, corticosteroid, bile acid metabolism), and amino acid (tryptophan, methionine/cysteine, tyrosine metabolism). In NHB women, 2 mQTLs were associated with 3 metabolites overlapping breast cancer GWAS loci: rs16866849 (diltiazem, ß =0.2, FDR-p=1.6×10⁻⁶, cardiovascular drug), and rs9397068 with two metabolites (2-hydroxyibuprofen, ß=5.3, FDR-p=9.6×10⁻³, analgesics; and 2,2′-methylenebis(6-tert-butyl-p-cresol), ß=14.6, FDR-p=9.6×10⁻³, chemical). Additionally, in NHW women, 5 mQTLs were associated with 5 metabolites overlapping GWAS loci of MBD. For dense area, these included rs150249911 (levulinate (4-oxovalerate), ß=6.5, FDR-p=2.3×10^-7, food component/plant), rs150249911 (N-acetyltyrosine, ß=5.0, FDR-p=3.3×10^-4, tyrosine metabolism), rs2042239 (5-acetylamino-6-formylamino-3-methyluracil, ß=1.3, FDR-p=3.3×10^-2, xanthine metabolism), and rs833472 (3-hydroxystachydrine, ß=11.0, FDR-p=2.5×10^-3, food component/plant). For percent mammographic density, rs61941038 (3-hydroxystachydrine, ß=10.4, FDR-p=7.4×10^-3, food component/plant) was associated. For non-dense area, rs78395856 (taurochenodeoxycholic acid 3-sulfate, ß=3.7, FDR-p=3.3×10^-2, bile acid metabolism) was associated. Conclusions: Our novel study identified genetic loci associated with metabolites that overlap with established MBD and breast cancer risk loci with differences in NHW and NHB women. These findings suggest biological mechanisms of genetic risk and provide a basis to prioritize metabolites and pathways for targeted prevention.

M. Bottosso¹, G. Griguolo¹, L. Trovò¹, S. Tognazzo², G. Faggioni³, E. Mioranza³, G. Vernaci³, D. Massa¹, V. Faso¹, D. Napetti¹, A. Menichetti³, S. Zovato², M. Dieci¹, V. Guarneri¹; ¹Department of Surgery, Oncology and Gastroenterology, University of Padova, Padova, ITALY, ²Hereditary Tumors Unit, Istituto Oncologico Veneto, Padova, ITALY, ³Medical Oncology 2, Istituto Oncologico Veneto, Padova, ITALY.

Background: Germline pathogenic variants (PVs) in BRCA1/2 are found in approximately 5% of patients with breast cancer (BC) and are associated with different clinical characteristics and subtype distributions. BC exhibits subtype-specific patterns of metastasis according to hormone receptor (HR) and HER2 status. Although limited, existing evidence suggests that germline BRCA1/2 PVs may be associated with a higher frequency of metastasis to specific sites, such as the central nervous system (CNS). However, these findings are often confounded by differences in subtype distribution among BRCA1/2 PV carriers. In this study, we investigated the association between germline BRCA1/2 mutation status and the patterns and frequency of metastasis across BC subtypes. Methods: Patients treated for HER2-metastatic BC at Istituto Oncologico Veneto between 2007 and 2025 and with known germline BRCA1/2 mutation status were retrospectively identified. Data on metastatic sites at initial diagnosis of metastatic BC and throughout the disease course (last follow-up) were collected. CNS involvement was defined as the presence of either brain metastases or leptomeningeal disease. Overall survival (OS) was defined as the time from the first diagnosis of metastatic BC to death from any cause or last follow-up. Results: A total of 184 patients were identified: 33 BRCA1 PV carriers, 28 BRCA2 PV carriers, and 123 noncarriers. Overall, 131 patients presented a HR+ BC and 53 a triple negative BC (TNBC). TNBC was more frequent among BRCA1 carriers (n=21, 63.3%), while HR+ BCs were more common among BRCA2 carriers (n=23, 82.1%) and noncarriers (n=96, 78.0 %; p<0.001). At initial diagnosis of metastatic BC, BRCA1/2 carriers with a HR+ BC presented a lower incidence of bone metastases compared to noncarriers (31.4% vs 51.0%, p=0.046). In contrast, no differences in metastatic site distribution at first metastatic BC diagnosis were observed among patients with TNBC. Median follow-up from stage IV disease diagnosis was 80.7 months (95%CI 53.5-108.0). At last follow-up, BRCA1/2 carriers with a HR+ BC presented a higher incidence of CNS involvement compared to noncarriers (31.4% vs 14.6%, p=0.030). Among BRCA1/2 carriers with HR+ BC, CNS involvement appeared relatively early, with 23.4% cumulative incidence at 24 months as compared to 5.2% in noncarriers (Gray’s p=0.041). At last follow-up, among patients with TNBC, BRCA1/2 carriers presented a lower incidence of liver metastases compared to noncarriers (34.6% vs 70.4%, p=0.009), while no significant difference in terms of CNS involvement was observed, with high rates in both subgroups (57.7% vs 37%, p=0.219). No significant difference in terms of OS was observed between BRCA1/2 carriers and noncarriers, both among patients with HR+ BC (39.2 vs 47.6 months, respectively; p=0.246) and among those with TNBC (25.8 vs 15.8 months, respectively; p=0.467). Conclusions: Germline BRCA1/2 PV are associated with distinct metastatic patterns among patients with HER2-negative BC. In the HR+ disease, carriers had relatively fewer bone metastases at diagnosis and presented a higher risk of CNS involvement over time. These findings may inform tailored radiological surveillance strategies.

A. Koric, Y. Park, J. Jiang, F. Boul, G. Colditz; Surgery, Washington University in Saint Louis, Saint Louis, MO.

Purpose: Existing breast cancer risk prediction models are ineffective for women with benign breast disease (BBD), especially in racially and ethnically diverse populations. We aimed to identify risk predictors for accurately estimating breast cancer development in a contemporary cohort of women from diverse racial backgrounds diagnosed with BBD. Patients and Methods: We identified 9,202 women diagnosed with histologically-confirmed benign lesions at the Joanne Knight Breast Health Siteman Cancer Center in St. Louis from 2010-2023. Risk modeling was performed with Cox regression using 95% confidence intervals (95%CI) and internally validated through a bootstrap resampling method with 100 replacements (i.e., corrected for overfitting). Discriminatory model performance was estimated by Harrell’s C-index. Calibration was assessed by estimating observed and predicted ratio (O/E) at 5-year BC predicted risk. Results: Among the 9,202 women with BBD, 64.3% were White, 32.5% were Black, and 3.2% were Asian women, 3.3% had a subsequent breast cancer at least six months following BBD. The final model predictors included: age at BBD diagnosis, family history of breast cancer, atypical hyperplasia (AH), proliferative disease without atypia (PDWA), breast density, and race. The model discrimination in the original sample had a Harrell’s C-index of 0.68 (95%CI 0.65, 0.71) and in the bootstrap sample of 0.67 (95%CI 0.63, 0.70). At the 5-year, the overall model calibration was good (O/E=0.98, 95%CI 0.86, 1.12). The model calibration was satisfactory across most risk-factor-defined subgroups; however, it tended to underestimate risk in the low-risk group and overestimate risk in the high-risk group (as defined by 20% quantiles). Conclusion: The final model outperformed existing breast cancer risk models among BBD women and future research will externally validate the model evaluating the risk among a subset of women with atypia.

K. Leggat-Barr¹, P. Giorgi Rossi², M. Guindy³, F. Gilbert⁴, M. Roman⁵, J. Burrion⁶, H. De Koning⁷, S. de Montgolfier⁸, L. Giordano⁹, D. Keatley¹⁰, J. Deleuze¹¹, E. Gauthier¹², S. Michiels¹³, C. Vissac-Sabatier¹⁴, H. Anton-Culver¹⁵, S. Borowsky¹⁶, S. Brain¹⁷, J. Esserman¹⁸, E. Ziv¹⁹, A. Fiscalini²⁰, D. Goodman-Gruen²¹, D. Heditsian¹⁷, R. Hiatt²², V. Lee¹⁷, D. Moorehead²³, A. Naiem²⁴, O. Olopade²⁵, H. Park²⁶, B. Parker²⁷, A. Petruse²⁸, M. Scheuner²⁹, L. van ‘t Veer³⁰, V. Arasu³¹, M. Eklund³², L. Madlensky³³, Y. Shieh³⁴, N. Wenger³⁵, J. Tice³⁶, C. Kaplan³⁷, A. Kaster³⁸, R. Lancaster³⁹, A. LaCroix⁴⁰, S. Delaloge⁴¹, L. Esserman²⁰; ¹Medical School, Tufts University School of Medicine, Boston, MA, ²Reggio Emilia, Azienda Unita Sanitaria Locale di Reggio Emilia, Reggio Emilia, ITALY, ³Medical School, HaBarzel, Tel Aviv, ISRAEL, ⁴Department of Radiology, University of Cambridge, Cambridge, UNITED KINGDOM, ⁵Hospital, Hospital Parc de Salud Del Mar, Barcelona, SPAIN, ⁶Medical School, Institut Jules Bordet, Bruxelles, BELGIUM, ⁷Department of Public Health, Erasmus, Rotterdam, NETHERLANDS, ⁸Aix Marseille Univ, INSERM, IRD, SESSTIM, Marseille, FRANCE, ⁹l’Epidemiologia e la Prevenzione Oncologica, Centro di Riferimento per l’Epidemiologia e la Prevenzione Oncologica, Torino, ITALY, ¹⁰Independent Cancer Patient Voice, Independent Cancer Patient Voice, London, UNITED KINGDOM, ¹¹Génomique Humaine, Centre National de Recherche en Génomique Humaine, Évry-Courcouronnes, FRANCE, ¹²Predilife, Villejuif France, Villejuif Cedex, FRANCE, ¹³Department of Biostatitics, Gustave Roussy, Villejuif, FRANCE, ¹⁴Unicancer, Unicancer, Paris, FRANCE, ¹⁵Department of Medicine, Genetic Epidemiology Research Institute, University of California, Irvine, Irvine, CA, USA., Irvine, CA, ¹⁶Department of Pathology, School of Medicine, University of California, Davis, Sacramento, CA, ¹⁷Breast Science Advocacy Core, Breast Oncology Program, University of California, San Francisco, San Francisco, CA, ¹⁸OB-GYN Department, Diagnostic Center of Miami, Miami FL USA, Miami, FL, ¹⁹Department of General and Internal Medicine, University of California San Francisco, San Francisco, CA, ²⁰Department of Surgery, University of California, San Francisco, San Francisco, CA, ²¹Department of Epidemiology and Biostatistics, University of California, Irvine, Irvine, CA, ²²UCSF Helen Diller Family Comprehensive Cancer Center, Department of Epidemiology & Biostatistics, University of California, San Francisco, San Francisco, CA, ²³Womens Cancer Resource Center, Womens Cancer Resource Center, Berkeley, CA, ²⁴Division of Hematology-Oncology, Department of Medicine, David Geffen School of Medicine, UCLA, Los Angeles, CA, ²⁵Center for Clinical Cancer Genetics and Global Health, Department of Medicine, The University of Chicago, Chicago, IL, ²⁶Department of Pathology and Laboratory Medicine, University of California, Irvine, Irvine, CA, ²⁷Moores Cancer Center, University of California San Diego, San Diego, CA, ²⁸Clinical and Translational Science Institute Office of Clinical Research, University of California Los Angeles, Los Angeles, CA, ²⁹Department of Medicine, University of California, San Francisco and San Francisco VA Health Care System, San Francisco, CA, ³⁰Department of Laboratory Medicine, University of California, San Francisco, San Francisco, CA, ³¹Division of Research, Kaiser Permanente, Pleasanton, CA, ³²Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, SWEDEN, ³³Department of Medicine, University of California, San Diego, San Diego, CA, ³⁴Department of Population Health Sciences, Weill Cornell Medicine, New York, NY, ³⁵Division of General Internal Medicine and Health Services Research, University of California, Los Angeles, Los Angeles, CA, ³⁶General Internal Medicine, School of Medicine, University of California San Francisco, San Francisco, CA, ³⁷Division of General Internal Medicine, Department of Medicine, University of California, San Francisco, San Francisco, CA., San Francisco, CA, ³⁸Edit Sanford Breast Center, Sanford Health, Fargo, SD, ³⁹Department of Surgery, University of Alabama at Birmingham, Birmingham, AL, ⁴⁰Medical School, University of California San Diego, San Diego, CA, ⁴¹Department of Cancer Medicine, Gustave Roussy, Villejuif, FRANCE.

Background WISDOM (Women Informed to Screen Depending on Measures of risk) (NCT02620852) and MyPeBS (My Personal Breast Screening) (NCT03672331) are two prominent risk-based breast cancer screening (RBS) trials based in the US and Europe & Israel, respectively. Both trials aim to assess whether RBS is as safe as current screening (no increase incidence of stage> II/IIB breast cancers) and can aid to enhance resources on those at higher risk, while minimizing harm to those at lowest risk. WISDOM and MyPeBS were conducted independently but with a planned joint analysis. Guidelines in the US range from annual (ACR, NCCN) to biennial screening starting at age 40 (USPSTF). In contrast, screening in MyPeBS countries is population-based by invitation every 2 years in (France, Belgium, Italy, Spain, Israel) and every 3 years in the UK starting at age 50. We will compare the baseline characteristics of both cohorts. Methods Eligibility included ages 40-74 in WISDOM and 40-70 in MyPeBS and no prior breast cancer history. Both are randomized 1:1 to RBS vs. country-based standard of care screening, though in WISDOM, women could choose their arm if they declined randomization (pragmatic, preference tolerant approach). Methods of recruitment reflect systematic/screening invitation (EU) vs. opportunistic (US). In WISDOM, the 5-year risk of invasive BC is estimated using the Breast Cancer Screening Consortium (BCSC) score, which includes clinical data & breast density, combined with a Polygenic Risk Score (PRS) (up to 126 single nucleotide polymorphisms (SNPs)), & a panel to identify germline mutations in 9 breast cancer risk genes. Four categories were used to classify risk (lowest, average, elevated, & highest), with corresponding screening assignments: age-based (mammogram (Mg) starting at 50), two-year (average), one-year moderate risk (MR), or Q6 (alternating Mg and MRI every 6 months). MR/Q6 includes breast health specialist consultation to discuss risk reduction and the Breast Health Decisions tool (automated risk/prevention recommendations, which is available to all patients). In MyPeBS, the 5-year risk is estimated using the Mammorisk™/BCSC model + PRS 313, or Tyrer-Cuzick (TC) + PRS 313 among women who have >1 first-degree family history of breast cancer (FDFH) with BC. Absolute risk cutoffs are defined as ‘low’ (1.66–6%), and ‘very high’ (≥6%) with corresponding screening assignments: next Mg at 4 years, Mg every 2 years, annually, or annually + MRI. Results WISDOM and MyPeBS enrollment was 46,289 and 53,143 women, respectively. Using the MyPeBS absolute risk cutoffs, the risk distribution for WISDOM and MyPeBS respectively were similar: ‘very high,’ (1% vs. 2%); ‘high’ (31% vs 33%), ‘average’ (26% vs 29%) & ‘low’ (40% vs 37%). Between the two trials, the screening recommendations for the low-risk group were the most different: every 2 years starting at 50 in WISDOM vs. next Mg at 4 years in MyPeBS. Median age was similar (54 vs. 55), though the percentage of women in their 40s was higher in WISDOM (38% to 23%). Baseline BMI was higher in WISDOM (27.6 vs 25.3). Differences (WISDOM vs MyPeBS) include: Education levels, with college degrees in 76% vs 52%; rates of prior biopsies 24% vs 14%; presence of a first-degree family member 24% vs 17%; and the use of hormonal replacement in the postmenopausal women 45% vs 17%. Chemoprevention use at baseline in both trials was 1% or less. Conclusion Within these two major trials, thresholds for risk assignment and proportion of patients in the personalized risk groups are very similar as are screening schemes except for low-risk women (every 2 yrs (WISDOM) vs. every 4 yrs (MyPeBS)). Variation provides the opportunity to learn from the effect of different practice patterns. Data emerging from both trials should be generalizable and have the potential to be practice-changing.

H. Esplen¹, B. Arun², C. DiNardo³, H. Abdel-Salam⁴, K. Richardson⁵, C. Peterson⁶, J. Corredor⁴; ¹Department of Pediatrics, Baylor College of Medicine, San Antonio, TX, ²Departments of Breast Medical Oncology and Clinical Cancer Prevention, The University of Texas MD Anderson Cancer Center, Houston, TX, ³Department of Leukemia, The University of Texas MD Anderson Cancer Center, Houston, TX, ⁴Department of Clinical Cancer Genetics, The University of Texas MD Anderson Cancer Center, Houston, TX, ⁵Department of Pediatrics, Division of Medical Genetics, McGovern Medical School at The University of Texas Health Science Center at Houston (UTHealth Houston) and Children’s Memorial Hermann Hospital, Houston, TX, ⁶Department of Biostatistics, The University of Texas MD Anderson Cancer Center, Houston, TX.

Background: Li-Fraumeni Syndrome (LFS), caused by pathogenic/likely pathogenic (P/LP) variants in TP53, is a cancer predisposition syndrome that increases the risk for numerous cancer types in both children and adults. Many studies have shown that cancer rates can vary between races and ethnicities. For example, the 2024 Cancer Statistics published by the American Cancer Society reported the incidence for breast cancer to be highest for White individuals and lowest for Hispanic/Latino individuals. One LFS study showed East Asian individuals were more frequently diagnosed with gastric cancer compared to North Americans and Europeans. Other studies identified P/LP TP53 variants associated with specific populations, such as the South and Southeast Brazil founder variant, p.Arg337His. However, there lacks extensive research on the variable expressivity of cancers within the LFS population, based on genotype, race, and/or ethnicity. This study aims to describe the specific TP53 germline variants, cancer rates, and cancer types among individuals of different racial and ethnic groups diagnosed with LFS. Methods: A retrospective chart review of individuals with germline P/LP TP53 variants at the University of Texas MD Anderson Cancer Center was completed. Data collected includes race, ethnicity, number of cancer diagnoses, cancer types, and location of the TP53 variant (exon and codon). Individuals with a suspected somatic TP53 variant were excluded from this study. Racial groupings included Asian, Black or African American, White, and Other/Unspecified. Ethnicity was categorized as Hispanic, Non-Hispanic, and Other/Unspecified. There were also six cancer type categories: adrenocortical carcinoma, brain cancer, breast cancer, osteosarcoma, soft tissue sarcoma, and non-core LFS cancers. Descriptive and statistical analyses, including chi-squared or Fisher’s exact tests and pairwise comparisons, were performed using R software (v4.4.1). Results: Two hundred and fifty individuals with a confirmed P/LP TP53 variant were included in the analysis. Breast cancer was the most common cancer type diagnosed for all ethnic and racial groups, except Asian individuals, who most frequently had non-core LFS cancer diagnoses. There was a significant difference in the rate of breast cancer across racial groups (p = 0.037), with Black or African American individuals with LFS having the highest rate of breast cancer at 76.5%. Additionally, individuals who identified as Black or African American were 1.87 times more likely to develop breast cancer in comparison to those who identified as White. For cancer rates, most individuals in each racial and ethnic group were diagnosed with one primary cancer. Statistical analysis did not identify a significant difference in cancer rates across groups defined by race or ethnicity. Regarding the location of variants within the TP53 gene, there was a significant difference in the rate of variants in exon four (p = 0.021) and exon six (p = 0.016) across racial groups, with variants in exon four being more common in Asian individuals (33.3%) and variants in exon six being more common in Black or African American individuals (35.3%). Conclusion: This study highlights similarities in the rates and types of cancers seen across racial and ethnic groups within a cohort of individuals with LFS. However, this study also identified potential differences in the rates of breast cancer and variant location across racial populations within the cohort. The results of this study provide information that can lead to more personalized counseling for patients and train risk models to accurately predict cancer risk for individuals with LFS of differing races and ethnicities.

F. P. Cavalcante¹, M. Antonini², A. L. Frasson³, I. de Oliveira-Junior⁴, F. P. Brenelli⁵, A. Berretini Junior⁶, G. N. Garcia⁷, F. Bagnoli⁸, F. P. Zerwes⁹, E. C. Millen¹⁰, A. Mattar¹¹, J. P. Reis¹², R. Fenile¹³, R. A. da Costa Vieira¹⁴, H. R. de Oliveira Filho¹⁵, M. Lichtenfels¹⁶, R. Sandoval¹⁷, P. L. Quidute⁴, R. Z. Torresan⁵, L. M. Zuliani¹⁸, M. M. Bortoluzzi⁹, A. D. Lima¹⁹, R. S. Guindalini²⁰, N. Polidorio²¹; ¹Breast Service, Hospital Geral de Fortaleza, fortaleza, BRAZIL, ²Breast Service, Hospital do Servidor Público Estadual Francisco Morato de Oliveira, Sao Paulo, BRAZIL, ³Breast Service, Hospital Israelita Albert Einstein, Sao Paulo, BRAZIL, ⁴Breast Service, Barretos Cancer Hospital, Barretos, BRAZIL, ⁵Breast Service, Universidade Estadual de Campinas, Campinas, BRAZIL, ⁶Breast Service, Hapvida NotreDame Intermédica, Sao Paulo, BRAZIL, ⁷Breast Service, Oncoclinicas, Sao Paulo, BRAZIL, ⁸Breast Service, Santa casa de misericórdia de São Paulo, Sao Paulo, BRAZIL, ⁹Breast Service, Pontifícia Universidade Católica do Rio Grande do Sul, Porto Alegre, BRAZIL, ¹⁰Breast Service, Hospital Geral de Fortaleza, Rio de Janeiro, BRAZIL, ¹¹Breast Service, Hospital da Mulher, Sao Paulo, BRAZIL, ¹²Breast Service, Grupo Orizonti, Belo Horizonte, BRAZIL, ¹³Breast Service, Hospital Ipiranga, São Paulo, BRAZIL, ¹⁴Breast Service, Hospital do Câncer de Muriaé, Muriae, BRAZIL, ¹⁵Breast Service, School of Medicine, Federal University of Paraná, Curitiba, BRAZIL, ¹⁶Breast Service, Centro de Mama Santa Casa e Oncologia Hospital Nora Teixeira/Rede Einstein, Porto Alegre, BRAZIL, ¹⁷Breast Service, Medical Oncology Center, Hospital Sírio-Libanês, Brasilia, BRAZIL, ¹⁸Breast Service, Grupo Américas, Hospital Paulistano, Sao Paulo, BRAZIL, ¹⁹Breast Service, Universidade Federal do Rio de Janeiro, Rio de Janeiro, BRAZIL, ²⁰Breast Service, Instituto D’or de Ensino e Pesquisa, Salvador, BRAZIL, ²¹Breast Service, Rede Americas Oncology, Sao Paulo, BRAZIL.

Background: Nipple-sparing mastectomy (NSM) is increasingly adopted as a risk-reducing surgical approach in women with germline pathogenic variants (PVs). Although oncologic outcomes are well established for risk-reducing mastectomies in BRCA1/2 carriers, data are limited for NSM as well as for individuals with other high- and moderate-penetrance PVs. This study aimed to evaluate the incidence of new breast cancer, cosmetic outcomes and surgical complications following risk-reducing NSM in PV carriers. Methods: We conducted a multicenter retrospective cohort study of risk-reducing NSM performed between 2008 and 2025 across 12 institutions. Asymptomatic patients (bilateral risk-reducing mastectomy [BRRM]) or individuals with unilateral breast cancer (contralateral risk-reducing mastectomy [CRRM]) were included. Patients with occult malignancy detected at NSM were excluded. Crude incidence rates were used to report new breast cancer diagnoses and complications. Cosmetic outcomes were assessed using the Harvard four-point scale. Risk factors for complications were analyzed using univariable (UVA) and multivariable (MVA) analyses. Results: A total of 524 patients underwent risk-reducing NSM: 219 BRRM and 305 CRRM (743 mastectomies). Most patients carried BRCA1 (46.0%) or BRCA2 (25.4%) PVs, followed by TP53 (15.1%). Median age was 40 years (range 20-68). At a median follow-up of 41.2 months (IQR 13.1-60.3), three patients (0.6%) developed breast cancer: one nodal recurrence post-BRRM, and two flap recurrences post-CRRM (BRCA1 [n=2] and BRCA2 [n=1]). Breast cancer-related death occurred in three patients (0.6%), all following CRRM. Cosmetic outcomes were favorable, with 90% rated as excellent or good. Surgical complications occurred in 101 patients (19.3%), more frequently after CRRM (24.6%) than BRRM (11.9%) (p<.001). On UVA, complications in BRRM were associated with obesity (OR 9.15, 95% CI 2.08-40.2, p=0.003) and Wise-pattern incision (OR 35.0, 95% CI 2.57-47.3, p=0.008). In CRRM, comorbidities (OR 2.76, 95% CI 1.38-5.54, p=0.004) and Wise-pattern incision (OR 7.33, 95% CI 1.65-32.5, p=0.009) were significant. On MVA, comorbidities (OR 2.59, 95% CI 1.23-5.47, p=0.012) and Wise-pattern incision (OR 7.83, 95% CI 1.73-35.5, p=0.008) remained independently associated with complications in CRRM; no independent predictors were found in BRRM. Conclusion: In this multicenter cohort of women with high- and moderate-penetrance germline PVs, risk-reducing NSM was associated with a very low incidence of new breast cancer (0.6%), supporting its oncologic safety. Most patients achieved favorable cosmetic outcomes. Surgical complication rates were higher after CRRM and associated with comorbidities and Wise-pattern incisions. These findings may guide surgical planning and patient counseling for risk-reducing NSM in genetically predisposed populations.

K. D. Crew¹, G. L. Anderson², K. B. Arnold³, A. Michel⁴, M. T. DeLucie⁴, C. W. Law⁴, S. Pruthi⁵, A. C. Sandoval Leon⁶, R. Shirley⁷, M. T. Grosse Perdekamp⁸, S. V. Colonna⁹, S. L. Krisher¹⁰, T. King¹¹, L. D. Yee¹², T. J. Ballinger¹³, C. Braun-Inglis¹⁴, D. A. Mangino¹⁵, K. B. Wisinski¹⁶, C. A. DeYoung¹⁷, M. Ross¹⁸, J. D. Floyd¹⁹, A. Kaster²⁰, L. VanderWalde²¹, T. Saphner²², C. Zarwan²³, S. Lo²⁴, C. Graham²⁵, A. K. Conlin²⁶, K. J. Yost²⁷, D. M. Agnese²⁸, C. Jernigan²⁹, D. L. Hershman¹, M. L. Neuhouser³⁰, J. A. Zell³¹, B. Arun³², R. Kukafka³³; ¹Medicine and Epidemiology, Columbia University Irving Medical Center, New York, NY, ²Public Health Sciences, SWOG Statistics and Data Management Center, Seattle, WA, ³Statistics, SWOG Statistics and Data Management Center, Seattle, WA, ⁴Medicine, Columbia University Irving Medical Center, New York, NY, ⁵General Internal Medicine, Mayo Clinic, Rochester, MN, ⁶Medicine, Miami Cancer Institute at Baptist Health South Florida, Miami, FL, ⁷Surgery, Good Samaritan Hospital Corvallis, Corvallis, OR, ⁸Clinical Sciences, Carle Cancer Center, Urbana, IL, ⁹Medicine, Huntsman Cancer Institute / University of Utah Medical Center, Salt Lake City, UT, ¹⁰Surgery, Holy Redeemer Hospital and Medical Center, Meadowbrook, PA, ¹¹Surgery, Dana-Farber Brigham Cancer Center, Brigham and Women’s Hospital, Boston, MA, ¹²Surgery, City of Hope Comprehensive Cancer Center, Duarte, CA, ¹³Medicine, Indiana University Simon Comprehensive Cancer Center, Indianapolis, IN, ¹⁴Cancer Biology, University of Hawaii Cancer Center, Honolulu, HI, ¹⁵Medicine, Memorial Sloan Kettering Cancer Center, New York, NY, ¹⁶Medicine, University of Wisconsin Carbone Cancer Center, Madison, WI, ¹⁷Medicine, Kaiser Permanente NCORP, Vallejo, CA, ¹⁸Medicine, Virginia Commonwealth University, Richmond, VA, ¹⁹Medicine, Cancer Care Specialists of Illinois, Heartland NCORP, Decatur, IL, ²⁰Family Medicine, Sanford Roger Maris Cancer Center, Fargo, ND, ²¹Surgery, Baptist Memorial Health Care, Memphis, TN, ²²Medicine, Aurora NCORP, Milwaukee, WI, ²³Medicine, Lahey Hospital & Medical Center, Burlington, MA, ²⁴Medicine, Loyola University Stritch School of Medicine, Maywood, IL, ²⁵Surgery, Emory University Hospital/Winship Cancer Institute, Atlanta, GA, ²⁶Medicine, Providence Cancer Institute, Portland, OR, ²⁷Medicine, Cancer Research Consortium of West Michigan NCORP, Kalamazoo, MI, ²⁸Surgery, The Ohio State University Comprehensive Cancer Center, Columbus, OH, ²⁹Patient Research Advocate, SWOG Prevention, Screening, & Surveillance Committee, Kansas City, MO, ³⁰Epidemiology, Fred Hutchinson Cancer Center, Seattle, WA, ³¹Medicine, University of California, Irvine, Irvine, CA, ³²Breast Medical Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX, ³³Biomedical Informatics, Columbia University Irving Medical Center, New York, NY.

Background: Breast cancer (BC) chemoprevention with selective estrogen receptor modulators (SERMs) and aromatase inhibitors (AIs) remains underutilized among high-risk women. Improving knowledge about BC risk and chemoprevention among high-risk patients and their healthcare providers may enhance informed decision-making about this BC risk reduction strategy. Methods: We conducted a cluster randomized controlled trial to evaluate the effectiveness of web-based patient and provider decision support, RealRisks and BNAV, to improve chemoprevention informed choice among women with atypical hyperplasia (AH) or lobular carcinoma in situ (LCIS). Thirty-one sites, including academic and community practices across the U.S., were randomized to standard patient educational materials alone or combined with RealRisks and BNAV. A total of 412 patients and 210 healthcare providers were enrolled. Patient surveys were administered at baseline, 6 and 12 months. The primary outcome at 6 months was chemoprevention informed choice, defined as adequate chemoprevention knowledge and attitudes congruent with decision. Secondary endpoints included BC risk perceptions, worry, chemoprevention knowledge, decision conflict and chemoprevention decision. Assuming 10% loss to follow up (LTFU), roughly equal accrual across sites, and an intraclass coefficient (ICC) of 0.02, we had 90% power to detect a 15% absolute difference in informed choice if the event rate in the control arm was 10%. Results: Among 24 evaluable sites, 14 were randomized to the intervention arm and 10 to the control arm. Among 210 enrolled providers, median age was 46 (range, 28-83) and 78% were female; 75% physicians and 25% advanced practice providers/nurses; 53% medical oncologists, 29% surgeons, and 10% gynecologists/family practitioners/internists; and 83% had previously prescribed chemoprevention. Among 412 enrolled patients, median age was 53 (range, 35-74), including 77% White, 8% Black, and 10% Asian race, and 14% Hispanic ethnicity. In terms of BC risk factors, 77% had AH and 23% LCIS, 34% had a first-degree family history of BC, and 65% had heterogeneously or extremely dense breasts on mammography. Thirty percent were unevaluable for the primary endpoint (1% ineligible, 26% LTFU, 3% missing data). Among 288 evaluable patients (148 at intervention sites, 140 at control sites), the proportion of patients exercising informed choice at 6 months in the intervention and control arms was 35% vs. 27%, respectively (odds ratio [OR]=1.44, 95% confidence interval [CI]=0.84-2.46, p=0.19; observed ICC=0.006). Compared to patients in the control arm at 6 months, those in the intervention arm were more likely to have accurate BC risk perceptions (OR=1.80, 95% CI=1.11-2.91) and adequate chemoprevention knowledge (OR=1.62, 95% CI=1.00-2.61). We observed no significant differences in BC worry, decision conflict, or chemoprevention decision between study arms. At 6 months, 49% of women in the control arm and 51% in the intervention arm decided to take a SERM or AI for BC chemoprevention (p=0.58), a rate higher than anticipated. Conclusions: We did not observe a significant increase in informed choice among patients and providers assigned to chemoprevention decision support compared to standard educational materials alone. Among women with AH or LCIS seen mainly by breast specialists who prescribed chemoprevention, about half decided to take a SERM or AI. Therefore, relatively high chemoprevention uptake may be achieved among women with high-risk breast lesions managed by breast specialists. Chemoprevention decision support may improve decision antecedents, such as accurate BC risk perceptions and chemoprevention knowledge, but may have a limited role in changing health behaviors.

P. Lin¹, Y. Chen², T. Yen¹, H. Chang¹, Y. Chu¹, Z. Liu¹, S. Kuo³, C. Huang⁴; ¹Medical Genetics, National Taiwan University Hospital, Taipei City, TAIWAN, ²Pediatrics, National Taiwan University Hospital, Taipei City, TAIWAN, ³Medical Oncology, National Taiwan University Hospital, Taipei City, TAIWAN, ⁴Surgery, National Taiwan University Hospital, Taipei City, TAIWAN.

Germline BRCA mutation carriers face a high risk of developing breast cancer, yet individual risk and age of onset vary significantly, influenced by both genetic and environmental factors. The gut microbiota is crucial for host health, regulating immunity, metabolism, and endocrine functions. Dysbiosis—an imbalance of gut microbiota—has been linked to diseases including cancer. We aimed to investigate whether specific gut microbes influence breast cancer risk in BRCA mutation carriers and to explore the underlying mechanisms. First, we studied 27 BRCA mutation carriers with early-onset breast cancer (60 years) without cancer. Fecal microbiota were analyzed using full-length 16S rRNA sequencing. The linear discriminant analysis (LDA) effect size (LEfSe) analysis showed that the Eggerthella lenta (E. lenta) and Dorea longicatena (D. longicatena) were the most significantly enriched bacteria in feces of the mutation carriers with early-onset breast cancer, compared to those in the old carriers without breast cancer. As these two bacteria are more common in older individuals, their presence in younger cancer patients suggests an association with cancer, independent of age. Second, we expanded the study to 102 BRCA mutation carriers (54 BRCA1, 48 BRCA2), 64 of whom had breast cancer. Mutation carriers with fecal E. lenta had a significantly earlier age of breast cancer than those without it (median age: 42.2 vs. 52.8 years-old, p<0.001), while D. longicatena showed no significant association. Subgroup analyses confirmed the link between E. lenta and early-onset breast cancer in both BRCA1 and BRCA2 groups (p<0.05). We then generated antibiotic-induced microbiome depletion (AIMD) model in the brca1-null mouse. The AIMD mice were gavaged with phosphate-buffered saline or E. lenta; mice gavaged with E. lenta developed breast tumors significantly earlier than those without E. lenta (p=0.022). We conducted untargeted metabolomics on urine samples from mutation carriers with early-onset breast cancer and from old cancer-free carriers (age >60). Tryptophan and its downstream metabolites—kynurenine, 3-hydroxykynurenine, 3-hydroxyanthranilic acid, and quinolinic acid—were significantly elevated in early-onset patients and positively correlated with E. lenta levels (p<0.05), compared to metabolites in the old cancer-free carriers. In mice fed E. lenta, serum levels of tryptophan/kynurenine pathway metabolites were elevated, too. To assess causality, brca1-null mice were intraperitoneal injected with quinolinic acid, a downstream end-product of the tryptophan/kynurenine pathway. The treated mice developed breast tumors significantly earlier than controls (p<0.01). Then, we tested the impact of quinolinic acid in the cells with heterozygous BRCA1 mutation. DNA fiber assay revealed significantly increased replication speed and impaired replication fork protection in the BRCA1 (185delAG/+) MCF10A cell line with co-culture of 10uM quinolinic acid, compared to those without quinolinic acid co-culture. We long-term (6 weeks) treated BRCA1 (185delAG/+) MCF10A cells with 10uM quinolinic acid. Whole exome sequencing revealed the treated cells harboring a significantly elevated homologous recombination deficiency score (by scarHRD software), compared to cells without quinolinic acid co-culture (p<0.05). These data suggested quinolinic acid enhanced the replication stress and caused the genomic instability in the human mammary cells harboring BRCA1 mutation. Our study demonstrated that dysbiosis is associated with early-onset breast cancer in BRCA mutation carriers. The E. lenta-derived tryptophan/kynurenine/quinolinic acid pathway induces the replication stress and genomic instability that drive breast cancer.

S. J. Winham¹, C. Wang¹, Y. Liu¹, B. M. McCauley¹, D. J. Tabrizi¹, D. L. Gehling², J. L. Fischer², L. R. Seymour², M. H. Solanki³, M. E. Sherman⁴, K. R. Lohani², T. L. Hoskin¹, D. C. Radisky⁵, A. C. Degnim²; ¹Department of Quantitative Health Sciences, Mayo Clinic, Rochester, MN, ²Department of Surgery, Mayo Clinic, Rochester, MN, ³Department of Laboratory Medicine and Pathology, Mayo Clinic, Rochester, MN, ⁴Department of Quantitative Health Sciences, Mayo Clinic, Jacksonville, FL, ⁵Department of Cancer Biology, Mayo Clinic, Jacksonville, FL.

Uptake of tamoxifen is limited to 5-15% of all eligible women due to concerns regarding medication side effects combined with the lack of ability to predict which women will benefit from the therapies. We used a unique retrospective sample of n=92 high-risk women with paired benign biopsies obtained before (PreTam) and during tamoxifen (OnTam) to discover tissue gene expression biomarkers of biologic response, defined as change (Δ) in lobular Ki67. Ki67 was digitally quantified in normal lobules from PreTam and OnTam benign tissue samples and samples with PreTam Ki67 = 0 were excluded. Data were analyzed from 51 women with both PreTam and OnTam Ki67 (median age = 49 years and median BMI = 25.5 kg/m² at PreTam biopsy). OnTam biopsies were collected a median of 11.5 months after PreTam biopsy (median 6.5 months after starting tamoxifen; dose 20 mg/day in 86%). Median Ki67 significantly decreased overall (Δ Ki67 = -0.29%, p=0.003), but the effect was restricted to women with baseline Ki67 ≥ 0.5% (n=38, Δ Ki67 = -0.78%, p=0.0002); no significant change occurred with baseline Ki67 14,600 genes), PreTam vs OnTam comparison across 33 pairs with baseline Ki67 ≥ 0.5% yielded 66 differentially expressed genes (fold-change > 2 and p<0.01). At p 0.50) and 14 genes with significant negative correlation (rho < -0.50) with Δ Ki67 (p<0.01). At p<0.01, ranked gene set enrichment analysis yielded 67 gene sets enriched with negative correlation (i.e. increased expression associated with a reduction in Ki67), including pathways related to ribosome biogenesis, DNA recombination, and nucleic acid metabolic processes; there were 62 gene sets enriched for positive correlation (i.e. decreased expression associated with a reduction in Ki67), including pathways such as VEGF and ERBB signaling. Our data support (i) baseline Ki67 ≥ 0.5 % and (ii) an 87‑gene baseline signature as complementary biomarkers of tamoxifen response. Prospective validation of this combined assay could target tamoxifen to the women most likely to benefit, potentially overcoming a major barrier to chemoprevention uptake.

M. Imran¹, N. M. Kettner¹, S. Hanash¹, O. Weaver², J. Leung², J. Dennison¹; ¹Clinical Cancer Prevention, The University of Texas MD Anderson Cancer Center, Houston, TX, ²Breast Imaging, The University of Texas MD Anderson Cancer Center, Houston, TX.

Background: The MERIT (Mammography, Early Detection, Risk Assessment, and Imaging Technologies) cohort was established in 2017 at MD Anderson to integrate clinical, imaging, and blood biomarker data for personalized risk assessment of breast and other cancers. This analysis evaluates breast cancer detection rates and staging outcomes among women undergoing annual screening mammography with or without supplemental imaging. Methods: Between 2017 and the present, 8,099 women receiving annual screening mammograms were enrolled in the MERIT cohort. Participants were categorized based on imaging modality: mammography (MG) alone or MG with supplemental imaging, defined as receipt of breast MRI or CEM within two years of enrollment. Clinical data, imaging outcomes, and cancer diagnoses were collected. Follow-up time was calculated from enrollment to the date of the last breast imaging exam or breast cancer diagnosis. Cancers were classified as ductal carcinoma in situ (DCIS) or invasive, with staging and receptor subtype data collected when available. Results: Over 28,153 person-years of follow-up, 236 breast cancers were diagnosed. Women with cancer had a mean follow-up time of 2.30 years (SD = 1.86), compared to 3.51 years (SD = 2.19) among women without cancer. Among 1,505 women who underwent supplemental imaging, 61 breast cancers were identified, resulting in an incidence rate of 1,046 per 100,000 person-years. In this group, 38 percent of cancers were DCIS and 62 percent were invasive. No stage III or IV cancers were detected among women who received supplemental imaging. In contrast, among 6,594 women who received screening mammography alone, 175 breast cancers were diagnosed, yielding an incidence rate of 784 per 100,000 person-years. In this group, 28 percent of cancers were DCIS and 72 percent were invasive. Eleven late-stage cancers were identified in the mammography-alone group, including five stage IIB, three stage IIIB, one stage IIIC, and two stage IV cancers. Notably, eight of these were triple-negative, and two stage IV cases were hormone receptor-positive. MRI utilization varied by both breast density and Gail model risk classification. Among high-risk women with dense breasts, 37 percent received MRI, compared to 19 percent of high-risk women with non-dense breasts. Among low-risk women, MRI use was 16 percent for those with dense breasts and 9 percent for those with non-dense breasts. Conclusions: Supplemental imaging was associated with a higher breast cancer detection rate and a greater proportion of early-stage disease compared to mammography alone. The absence of late-stage cancers in the supplemental imaging group, in contrast to 11 advanced-stage cases in the MG-alone group, suggests a potential benefit for earlier detection through supplemental imaging. However, supplemental MRI remains underutilized among high-risk women. These findings support the need for personalized, risk-aligned screening strategies that integrate both breast density and formal risk assessment to improve early detection of clinically significant breast cancer.

B. K. Arun¹, G. Gierach², M. Scoggins³, E. A. Bowles⁴, S. A. Khan⁵, T. B. Bevers⁶, S. S. Rao⁷, J. E. Garber⁸, S. Raza⁹, N. B. Kumar¹⁰, H. S. Han¹¹, J. Heine¹², B. Niell¹³, P. Chalasani¹⁴, K. Fitzpatrick¹⁵, L. G. Wilke¹⁶, A. Eowler¹⁷, H. C. Beckwith¹⁸, J. E. Kuehn-Haijder¹⁹, C. Mays⁶, L. A. Vornik⁶, O. Lee²⁰, E. Diamond²¹, M. Perloff²², W. Dong²³, D. Liu²⁴, J. J. Lee²⁴, F. W. Symmans²⁵, E. Vilar-Sanchez⁶, B. M. Heckman-Stoddard²⁶, P. H. Brown⁶; ¹Breast Medical Oncology and Clinical Cancer Genetics, UT MD Anderson Cancer Center, Houston, TX, ²Integrative Tumor Epidemiology, NCI, Rockville, MD, ³Radiology, UT Southwestern Medical Ctr, Dallas, TX, ⁴Kaiser Permanente Washington Health Research Institute, Kaiser Permanente Washington Health Research Institute, Seattle, WA, ⁵Breast Surgical Oncology, Northwestern University, Chicago, IL, ⁶Cancer Prevention, UT MD Anderson Cancer Center, Houston, TX, ⁷Genetics, Northwestern Medical Center, Chicago, IL, ⁸Breast Medical Oncology and Clinical Cancer Genetics, Dana Farber Cancer Institute, Boston, MA, ⁹Radiology, Brigham & Women’s Hospital, Boston, MA, ¹⁰Cancer Epidemiology, Mofitt Cancer Center, Tampa, FL, ¹¹Breast Medical Oncology, Moffit Cancer Center, Tampa, FL, ¹²Breast Medical Oncology and Clinical Cancer Genetics, Moffitt Cancer Center, Tampa, FL, ¹³Radiology, Moffitt Cancer Center, Tmpa, FL, ¹⁴Breast Medical Oncology and Clinical Cancer Genetics, University of Arizona Cancer Center, Tucson, AZ, ¹⁵Radiology, University of Arizona, Tucson, AZ, ¹⁶Surgery, University of Wisconsin, Madison, WI, ¹⁷Radiologycs, University of Wisconsin, madison, WI, ¹⁸Medical Oncology, University of Minnesota, Minneapolis, MN, ¹⁹Radiology, University of Wisconsin, Minneapolis, MN, ²⁰Breast Medical Oncology and Clinical Cancer Genetics, Northwestern University, Chicago, IL, ²¹Division of Cancer Prevention, NCI, Bethesda, MD, ²²Division of Cancer Prevention, NCI, Rockville, MD, ²³Bioistatistics, UT MD Anderson Cancer Center, Houston, TX, ²⁴Biostatistics, UT MD Anderson Cancer Center, Houston, TX, ²⁵Department of Pathology, UT MD Anderson Cancer Center, Houston, TX, ²⁶Division of Cancer Prevention, NCI, Houston, TX.

Background: Tamoxifen uptake for risk reduction has remained low due to concerns about toxicity despite available effectiveness data. Studies of tamoxifen in the adjuvant and preventive setting have demonstrated that a 10% or greater decline in mammographic density (MD) is associated with better outcomes; however, tamoxifen metabolism and associated MD declines are variable across patients. An alternative application of a bioactive tamoxifen metabolite may reduce side effects while maintaining efficacy. 4-hydroxytamoxifen topical gel (4-OHT) is a transdermal agent, shown in preliminary studies to be well-tolerated with similar decreases in Ki-67 to oral tamoxifen in presurgical DCIS studies and significant drug concentration in breast parenchyma but very low levels in the systemic circulation. Therefore, we conducted a multicenter clinical trial to evaluate changes in MD in women with dense breasts induced by 4-OHT versus placebo gel when applied to the breast for 12 months. Methods: Pre- and postmenopausal women between the ages 40-69 yrs, or less than 40 yrs with 5-year Gail risk greater than 1.66%, with heterogeneously or extremely dense breast tissue, were enrolled in this study. Participants were randomized (1:1): 2 mg 4-OHT gel or Placebo gel for 12 months (mo), applied daily to each breast. Participants underwent baseline mammograms, blood draws for toxicity and research biomarkers, and optional breast biopsy, all of which were repeated at the end of 12 mo. The primary endpoint was to estimate and compare the percent change in MD (measured on cranio-caudal mammograms using Cumulus) from baseline to mo 12 between 2 groups. With 64 women in each group, there is 80% power to detect a decrease in density of 6% in the 4-OHT gel vs 2% in the placebo group with a common SD of 8% using a two-sided t-test with a significance level of 0.05. Secondary endpoints included tolerability, measurement of drug metabolites in breast tissue and plasma, and research biomarkers. Results: This prospective, randomized, double-blind, placebo-controlled phase II study randomized 158 participants from Sept 2017 to October 2020 across 6 centers. 123 participants had both baseline and month 12 mammograms, 61 in 4-OHT and 62 in Placebo. The mean (standard deviation) breast density decrease over 12 month was 6.30% (9.11%) in the 4-OHT gel arm and 5.77% (11.24%) in the Placebo arm. The changes in breast density from baseline to month 12 were not significantly different between the two treatment arms (Hotelling’s T² test p-value = 0.09). For those with available research biopsies, in the placebo arm (N=18), all 4-OHT and metabolite levels were below the quantifiable level (BQL). In the treatment arm (N=12), at months 6 and 12, 10 patients had Z-4-OHT tissue levels above BQL with an average of 10.09 ng/g. No grade 3/4 toxicity was observed in either arm. Twenty-nine participants (thirteen 4-OHT and sixteen Placebo) had 41 gr 1 or 2 adverse events related to the application of gel. Drug exposure serum toxicity, including FVIII and vWB, was not different between treatment arms. IGF-1 levels decreased from baseline to month 12 in both arms; the difference was not statistically significant, whereas IGFBP-3 increased significantly at month 12 in the 4-OHT arm (p=0.04). Conclusion: This prospective double-blind randomized study did not demonstrate a significant difference in the change of MD from baseline to the 12-month treatment between the 4-OHT arm and the placebo arm. The 4-OHT gel was well tolerated, and drug and metabolites were measurable in the breast tissue. Serum IGFBP-3 was found to be increased in the 4-OHT treatment group. The effect of baseline MD characteristics by automated software measurements and other tissue biomarkers will be analyzed and reported separately.

K. Leggat-Barr¹, J. Tice², K. Badal³, R. Sayaman⁴, B. Parker⁵, S. Hartman⁶, H. Anton-Culver⁷, H. Park⁸, A. Ziogas⁹, A. Petruse¹⁰, A. Naiem¹¹, K. Malvin¹², L. Sabacan¹³, T. Lewis¹⁴, T. Glatt¹, S. Kapoor³, A. Verma¹⁵, P. Warner¹⁶, E. Tsopurashvili³, A. Griffin¹⁷, V. Lee¹⁸, J. McGuire¹⁹, S. Borowsky²⁰, WISDOM Study and Athena Breast Health NetworkInvestigators and Advocate Partners²¹, C. Kaplan²², R. Hiatt²³, A. Stover-Fiscalini³, K. Kerlikowske¹², Y. Shieh²⁴, L. Esserman³, L. van ‘t Veer⁴; ¹School of Medicine, Tufts University School of Medicine, Boston, MA, ²Department of General and Internal Medicine, University of California, San Francisco, San Francisco, CA, ³Department of Surgery, University of California, San Francisco, San Francisco, CA, ⁴Department of Laboratory Medicine, University of California, San Francisco, San Francisco, CA, ⁵Moores Cancer Center, University of California San Diego, San Diego, CA, ⁶Department of Family and Preventive Medicine, University of California, San Diego, San Diego, CA, ⁷Department of Medicine, Genetic Epidemiology Research Institute, University of California, Irvine, Irvine, CA, ⁸UC Irvine School of Medicine, Department of Pathology and Laboratory Medicine, Irvine, CA, ⁹Chao Family Comprehensive Cancer Center, University of California, Irvine, Orange, CA, ¹⁰Clinical and Translational Science Institute Office of Clinical Research, University of California Los Angeles, Los Angeles, CA, ¹¹Division of Hematology-Oncology, Department of Medicine, David Geffen School of Medicine at UCLA, Los Angeles, CA, ¹²Departments of Medicine and Epidemiology/Biostatistics, University of California, San Francisco, San Francisco, CA, ¹³Department of Surgery, University of California San Francisco, San Francisco, CA, ¹⁴School of Medicine, University of California, San Francisco, San Francisco, CA, ¹⁵School of Medicine, Pritzker School of Medicine, Uchicago, Chicago, IL, ¹⁶School of Medicine, Columbia University, New York, NY, ¹⁷Cancer Registry, UCSF Helen Diller Family Comprehensive Cancer Center, San Francisco, CA, ¹⁸Breast Science Advocacy Core, Breast Oncology Program, University of California, San Francisco, San Francisco, CA, ¹⁹Cancer Registry, University of California, San Francisco, San Francisco, CA, ²⁰Department of Pathology, School of Medicine, University of California Davis, Saramento, CA, ²¹²²Division of General Internal Medicine, Department of Medicine, University of California, San Francisco, San Francisco, CA., San Francisco, CA, ²³UCSF Helen Diller Family Comprehensive Cancer Center, Department of Epidemiology & Biostatistics, University of California, San Francisco, San Francisco, CA USA, San Francisco, CA, ²⁴Department of Population Health Sciences, Weill Cornell Medicine, New York, NY.

Background The Athena Breast Health Network (Athena) is a University of California (UC) and Sanford Health initiative integrating clinical care and research to drive improvements in breast screening. Here, we evaluated cancer occurrence and tumor type in our UC San Francisco (UCSF) Athena cohort. We assessed the association with 5-year invasive cancer risk level by the Breast Cancer Surveillance Consortium (BCSC) version 3 (BCSCv3) model and the development of aggressive or non-aggressive cancer. Methods BCSCv3 risk scores were calculated for 10,031 consented UCSF Athena participants enrolled 2011-2022 with a median follow-up of 6.9 years. The BCSCv3 model includes age, race/ethnicity, prior breast biopsy results, breast density, first and second-degree family history, body mass index, menopausal status, and age at first live birth. Risk information was gathered prior to breast cancer diagnosis. The ‘high’ risk threshold was previously set at >1.67% 5-year risk by the BCSCv3 model (PMID37976443). All others were classified as ‘low’ risk. To identify those who developed invasive breast cancer after intake, we performed a cancer linkage with the San Francisco Mammography Registry and the UCSF cancer registry (complete through Dec. 2018 and Jan. 2022 respectively). Tumor aggressiveness was defined in two ways. Advanced or non-advanced cancer was classified by the AJCC prognostic pathologic stage (>stage II is advanced, <stage II is non-advanced). Faster and slower growing cancer considered tumor biology only, where one or more of the following criteria defined faster: HR-negative, HER2-positive, and/or grade 3, all other were slower growing cancers. We used logistic regressions to examine associations between BCSCv3 5-year risk score and breast cancer incidence and tumor aggressiveness. Results Of the 10,031 participants, 2838 (28%) participants were designated as ‘high’ risk, and 7193 (72%) participants were ‘low’ risk. Overall, 213 (2%) developed breast cancer. 34 (16%) developed advanced cancer and 179 (84%) developed non-advanced cancer. Among women with breast cancer, 60 (28%) developed faster-growing cancers and 153 (72%) developed slower-growing cancers. The median 5-year BCSCv3 risk score for women with and without breast cancer was 1.73% and 1.38% respectively (p<0.001). The average 5-year BCSCv3 risk score for women with advanced and non-advanced cancers was 1.67% and 1.74%, respectively. The average 5-year BCSCv3 risk score for women with faster and slower growing cancers was 1.65% and 1.76%, respectively. In logistic regression analysis adjusted for age, women in the high-risk group had significantly higher odds of developing non-advanced cancer compared to healthy controls (OR = 3.08, 95% CI 1.87-5.08, p<0.001), whereas the association for advanced cancer was not statistically significant (OR = 1.24, 95% CI 0.56-2.77, p=0.59). Similarly, when analyzing slow vs. fast growing cancer, in logistic regression analysis adjusted for age, women in the high-risk group had significantly higher odds of developing slower-growing cancer compared to healthy controls (OR = 2.02, 95% CI 1.39-2.92, p<0.0001), whereas the association for faster-growing cancer was not statistically significant (OR = 1.62, 95% CI 0.88-2.99, p=0.12). Conclusions This study confirms that higher 5-year BCSCv3 risk scores are associated with higher overall breast cancer development. Participants with the highest 5-year BCSCv3 risk scores are more likely to develop cancers with less aggressive features, suggesting the BCSCv3 model may preferentially predict less aggressive tumors. Those with the highest 5-year BCSCv3 risk may be more likely to benefit from endocrine risk reduction therapy. The Analysis using the recently developed BCSC advanced cancer model is ongoing and may offer better discrimination for those at risk for advanced cancers.

R. Abou Zeidane¹, C. Pisano¹, F. Hussain², W. Dong², T. Lal³, L. Vu², N. Mehta⁴, S. Koroukian², J. Rose², C. Speers¹; ¹Radiation Oncology, University Hospitals Cleveland Medical Center and Case Western Reserve University School of Medicine, Cleveland, OH, ²Population and Quantitative Health Sciences, Case Western Reserve University School of Medicine, Cleveland, OH, ³Surgery, division of Surgical Oncology, University Hospitals Cleveland Medical Center and Case Western Reserve University School of Medicine, Cleveland, OH, ⁴Family Medicine, University Hospitals Cleveland Medical Center and Case Western Reserve University School of Medicine, Cleveland, OH.

Background: Triple-negative breast cancer (TNBC) is an aggressive subtype with disproportionate mortality among underserved populations. Mammographic screening reduces breast cancer mortality, but disparities in access and utilization persist, particularly in African American (AA) communities and socioeconomically vulnerable areas. This study aims to quantify the relationship between mammography uptake, demographic factors, and social determinants of health (SDOH) with advanced-stage TNBC (AS-TNBC) across diverse U.S. populations. Methods: We conducted a cross-sectional analysis of 593 counties across multiple states using data from the CDC PLACES database (2020 release, representing 2018 data), the American Community Survey (ACS 2014-2018), and county-level SDOH indicators. Communities were aggregated using the Max-P Regions method to balance TNBC case counts and socioeconomic indices (adjusted for area deprivation index, ADI). Descriptive statistics were used to summarize key characteristics across regions. Then, multivariable Poisson and negative binomial regression models were used to evaluate associations between AS-TNBC diagnoses mammographic screening uptake, demographic (% AA, % Hispanic, % aged <45 years), and social vulnerabilities, including limited English proficiency, rurality, and insurance status. Results: The median percentage of AS-TNBC diagnoses across counties was 38.8% (IQR: 34.9-43.6%). Counties with higher AS-TNBC incidence exhibited significantly lower mammography use (71.5% vs. 72.6%, p 0.001), higher uninsured rates (18.7% vs. 15.7%, p <0.001), and increased proportions of AA residents (14.0% vs. 7.9%, p <0.001), limited English proficiency (1.7% vs. 1.5%, p 0.005), higher socioeconomic deprivation (ADI top quartile: 33.7% vs. 16.4%, p <0.001), and greater rurality (41.5% vs. 30.2%, p 0.006). Multivariable Poisson regression modeling demonstrated a protective effect of mammographic screening, where a 10% increase in screening uptake correlated with a 5.2% decrease in AS-TNBC diagnoses (RR 0.95, 95% CI 0.91-0.99, p 0.007). Conversely, higher proportions of AA residents (RR 1.04, p <0.001), rural populations (RR 1.05, p 0.02), and Hispanic residents (RR 1.002, p <0.001) significantly correlated with increased AS-TNBC rates. Negative binomial regression (adjusting for total breast cancer cases) reaffirmed the robust protective role of screening (RR 0.91, 95% CI 0.89-0.93, p <0.001). In the negative binomial model (adjusted for total breast cancer cases), mammographic screening remained a strong protective factor (RR 0.91, 95% CI: 0.89-0.93, p <0.001), indicating a 9.1% reduction in AS diagnosis per 10% increase in screening. Higher proportions of AA residents (RR 1.04, p <0.001), uninsured individuals (RR 1.04, p <0.001), and those in the most deprived areas (ADI top quartile, RR 1.04, p <0.001) were significantly associated with higher AS diagnosis rates, as well as younger population composition (age 18-44; RR 1.03, p 0.004). Conclusions: Mammographic screening disparities and adverse social determinants significantly influence the prevalence of advanced-stage TNBC. High-risk populations, particularly AA communities, rural areas, and regions with elevated socioeconomic vulnerabilities, face disproportionate impacts. These findings highlight the urgent need for targeted interventions aimed at enhancing access to culturally sensitive and linguistically appropriate screening and healthcare services to mitigate disparities and improve breast cancer outcomes.

K. Badal¹, J. Staib², J. A. Tice³, M. Kim⁴, M. Eklund⁵, L. Wilson⁶, S. Dacosta Byfield², K. Catlett⁷, L. Maffey², R. Soonavala¹, A. Stover Fiscalini¹, L. Esserman¹; ¹Department of Surgery, University of California San Francisco, San Francisco, CA, ²Optum Labs, Optum Labs, Eden Prairie, MN, ³Department of Medicine, University of California San Francisco, San Francisco, CA, ⁴Department of Epidemiology and Biostatistics, University of California San Francisco, San Francisco, CA, ⁵Department of Medical Epidemiology and Biostatistics, Karolinska Institute, Solna, SWEDEN, ⁶Department of Medicine and Department of Clinical Pharmacy, University of California San Francisco, San Francisco, CA, ⁷Optum Center for Research and Innovation, UnitedHealth Group, Minnetonka, MN.

Background: The WISDOM (Women Informed to Screen Depending on Measures of risk) Study is testing the safety and effectiveness of risk-based breast cancer screening. We aimed to estimate the total, national, yearly cost of screening in the USA if WISDOM’s screening algorithm were used, with comparison to current screening guidelines. Methods: We used Optum Labs Data Warehouse (OLDW) to estimate 2D and 3D- mammograms, and MRI screening and recall costs and rates for commercial and Medicare Advantage insurance. OLDW contains longitudinal de-identified administrative, medical, and pharmacy claims for over 200 million commercial and Medicare Advantage enrollees and patients. Screening and recall costs were normalized to calendar year 2020. We used real-world participation rates by age of 54-78%. The risk-based strategy included a cost of USD$270 for genetic testing and risk assessment. We modeled three guidelines: 2021/2023 American College of Radiology (ACR), 2023 American Cancer Society (ACS), and 2024 United States Preventative Services Task Force (USPSTF). The percentage of high-risk/average-risk women per guideline were estimated using the literature. Total cost of screening was calculated as the sum of screening and recall costs. The average lifetime cost of screening per woman until 74 years was also calculated. One-way deterministic sensitivity analysis was used to determine the impact of inputs on the aggregate cost. Results: Median mammogram costs ranged from $139-$360, MRI costs ranged from $545-$2,439, while recall rates ranged from 9.2%-20.9%. The projected national cost of screening was $30B for ACR, $18B for ACS, $8B for USPSTF, and $7B for WISDOM. The average lifetime cost through age 74 years to screen a high-risk/average-risk woman was: $89,725/$13,416 for ACR, $89,725/$7,946 for ACS, $6,931/$6,931 for USPSTF, and $66,451/$6,618 for WISDOM. For USPSTF, the cost of screening high-risk/average-risk women were the same because these women were screened the same. Participation rates, proportion of women with a lifetime risk >20%, and commercial MRI and 3D costs had the largest impact on total costs. Conclusion: WISDOM’s risk-based screening algorithm can reduce national cost substantially, even with the added cost of genetic testing, while maintaining intensive screening for the highest-risk women. The resources saved can be used to improve screening for women at high risk for fast-growing cancers, which are often identified between screens, and to encourage improved adherence.

S. J. Friedman¹, R. H. Pugh Yi², M. Dean³, P. L. Welcsh⁴, K. N. Owens⁴, J. D. Rogers¹, D. B. Rose⁵, E. Kuhn⁶; ¹Research and Education, Facing Our Risk of Cancer Empowered (FORCE), Tampa, FL, ²Program Evaluation, Akeso Consulting, LLC, Vienna, VA, ³Department of Communication, University of South Florida, Tampa, FL, ⁴Education, Facing Our Risk of Cancer Empowered (FORCE), Tampa, FL, ⁵Research, Facing Our Risk of Cancer Empowered (FORCE), Tampa, FL, ⁶Education, Susan G. Komen, Dallas, TX.

Background New cancer treatment and prevention strategies have led to an increase in clinical trials enrolling people with germline mutations linked to cancer risk. Yet, referral of patients remains low. In a FORCE survey, 14% of respondents at high risk for breast cancer were told by their healthcare providers about clinical research opportunities, while 75% expressed interest¹. In a survey of genetic counselors, most had never referred breast cancer survivors (63%) or metastatic patients (69%) to clinical trials, and 55% listed research jargon/health literacy as barriers to referral². We developed a training for genetic counseling students consisting of a presentation, activities and group discussions on factors that impede patients’ ability to find, understand, and enroll in hereditary cancer research, and ways that genetic counselors can address these barriers. We introduced tools to help people find research studies, including FORCE’s “Search and Enroll Tool” that provides plain-language descriptions of studies enrolling people with inherited mutations. Methods We conducted 17 workshops for genetic counselor training programs, and trained 256 students. Almost half completed feedback surveys. Three training program directors participated in post-training interviews. Results Students indicated that they plan to apply what they had learned in practice. They appreciated the plain-language features of FORCE’s tool and how it prioritizes studies enrolling people with germline mutations, and they plan to use the tool to connect patients to research opportunities. Course directors agreed that the training was appropriate to integrate into their curricula. They valued the fact that the training helps students recognize the importance of sharing clinical research opportunities with patients, teaches students how to find this information and present it to patients in easy-to-understand terms, and emphasizes the importance of representativeness in research. Students and directors recommended retaining the content, flow and presentation approach, activities and discussions, and emphasis on the practical application of material presented. Suggestions included limiting the training to 90 minutes, providing more hands-on experience with finding and discussing studies with patients, and developing a version that students can complete on their own. Discussion Students and faculty found the training to be informative, relevant, and engaging. Genetic counseling training programs could benefit from training on this critical topic that may not be offered elsewhere in their curricula. In addition to learning information and skills related to finding and communicating with patients about hereditary cancer research, students and faculty agreed that FORCE’s Search and Enroll Tool is useful for matching patients to relevant research opportunities.

O. Somuncu¹, B. Lamarre¹, S. Mukkavalli¹, T. B. Branigan², R. Ravindranathan¹, N. V. Wietmarschen³, H. Nguyen¹, N. Ashton⁴, K. Parmar¹, G. Shapiro¹, D. A. Dillon⁵, D. Chowdhury³, P. A. Konstantinopoulos³, J. E. Garber³, A. D. D’Andrea¹; ¹Center for DNA Damage and Repair, Dana Farber Cancer Institute, Boston, MA, ²Department of Medical Oncology, Dana Farber Cancer Institute, Boston, MA, ³Center for BRCA and Related Genes, Dana Farber Cancer Institute, Boston, MA, ⁴Department of Radiation Oncology, Dana Farber Cancer Institute, Boston, MA, ⁵Department of Pathology, Dana Farber Cancer Institute, Boston, MA.

Replication stress as a driver of breast tumorigenesis in BRCA1 heterozygous carrier-derived breast organoids Ozge S. Somuncu, Benjamin Lamarre, Sirisha Mukkavalli, Timothy B. Branigan, Ramya Ravindranathan, Niek Van Wietmarschen, Huy Nguyen, Nicholas Ashton, Kalindi Parmar, Geoffrey Shapiro, Deborah A. Dillon, Dipanjan Chowdhury, Panagiotis A. Konstantinopoulos, Judy E. Garber, Alan D. D’Andrea Germline mutations in BRCA1 significantly increase the risk of BRCA1^(+/-) heterozygous carriers in developing early and rapid breast and ovarian cancers. However, the physiological alterations in heterozygous BRCA1^(+/-) breast tissue necessary for malignant transformation, and the specific features predisposing this tissue to loss of heterozygosity (LOH) are poorly understood. BRCA1^(+/-) carrier breast tissue is reported to contain a small population of cells with copy number alterations (CNAs) (1q gain and 16q loss), p53 mutations, and LOH of BRCA1. A previous study showed that BRCA1^(+/-) carrier breast tissue fibroblasts and mammary epithelial cell lines have several features of replication stress (R/S), which may contribute to genomic instability and progression to cancer. Luminal progenitor cells (LPs) appear to be the primary cells leading to these cancers but the contribution of other subpopulations in breast tissue remains to be elucidated. Patient-derived BRCA1^(+/-) carrier organoids are a living, physiological system that allows for this study of R/S, BRCA1 LOH, and acquired CNAs in all breast subpopulations in parallel. To further examine the phenotype of BRCA1^(+/-) breast tissue, we generated viably-growing carrier organoids derived from breast tissues of seventeen BRCA1^(+/-) women, who underwent prophylactic mastectomy, and from fifteen BRCA1^(+/+) women who underwent reduction surgery as controls. Hydroxyurea induced R/S, as indicated by increased levels of pKAP1, yH2AX, and ELF3, in all subpopulations of the BRCA1^(+/-) carrier organoids in comparison to the BRCA1^(+/+) organoids. Consequently, BRCA1^(+/-) carrier breast tissues and organoids exhibited shortened telomere length and high levels of ELF5, a premature aging marker, suggesting that BRCA1^(+/-) carriers are predisposed to altered telomere replication leading to R/S. BRCA1^(+/-) carrier organoids and tissues are enriched for G2/M markers where homologous recombination (HR) is a critical DNA damage pathway that resolves R/S. A BRCA1^(+/-) carrier with PARPi sensitivity, indicative of HR deficiency, exhibited some of the common CNAs previously detected in BRCA1^(-/-) breast tumors (1q gain and 16q loss) suggesting LOH in small fraction of cells. Overall, patient-derived organoid models allow for the evaluation of the BRCA1^(+/-) pathologic phenotype and testing drugs for preventing the outgrowth of BRCA1^(-/-) malignant clones. Acknowledgments We’d like to thank to patients for donating tissue, as well as Alicia Pollaci (CTO Sr Research Project Mgr) and Claire Capaldi (Research Lab Tech) for helping on the tissue collections. In addition, we would like to thank MacLean C. Sellars MD, PhD for the flow cytometry analysis.

J. Ragaz¹, K. S. Wilson², H. Wong¹, H. Qian¹, S. Shakeraneh³, J. J. Spinelli⁴; ¹School of Population and Public Health, University of British Columbia, Vancouver, BC, CANADA, ²Division of Medical Oncology, BC Cancer Agency, Victoria, BC, CANADA, ³Infection Prevention and Control, Providence Health Care, Vancouver, BC, CANADA, ⁴Department of Statistics, BC Cancer Agency, Vancouver, BC, CANADA.

Objective: To determine the value of estrogen in overall management of breast cancer. Background: Despite the prior consensus of Estrogen [E] Breast Cancer [BrCa] harm, recent 20-year follow-up data from the double-blind randomized Women’s Health Initiative Trial 2 [WHI-2], show E-alone based Hormone Replacement Therapy [E-HRT] reducing BrCa incidence by 22%, and, unexpectedly, BrCa mortality by 40%.¹ For comparison, the Tamoxifen randomized breast cancer prevention trial [IBIS-1] also reported a lower BrCa incidence [RR=0.71], but no mortality reduction [RR = 1.19, 95% CI].^2,3 (Table 1). Comments: These prevention outcomes show both E-HRT and Tamoxifen equally effective in reducing incidence rates. The striking difference, however, is the significant mortality reduction with E-HRT, but not with Tamoxifen. Furthermore, E-HRT offers major benefits for quality-of-life [QOL] and for other rate reductions: Alzheimer dementia mortality: -26%; women age <60 at E-HRT start, both myocardial infarctions [-41%] and the all-cause mortality [-26%].^3,4 In addition to these prevention benefit outcomes, estrogen emerged much more effective than Tamoxifen in stage IV BrCa [5 year overall survival: 35% vs 16%, adjusted p = 0.039].⁵ Conclusion: We believe the observed benefits of E-HRT* are sufficiently compelling to seek its guideline approval both in breast cancer prevention and in adjuvant settings, with or without anti-estrogens, with expectations of tens of thousands more avoided deaths per year, just in USA alone,⁴ while QOL is dramatically improved. *E-HRT: Estrogen alone for women with hysterectomy; Estrogen + progesterone for women with retained uterus References: 1. Chlebowski RT, Anderson GL, Aragaki AK, et al. Association of Menopausal Hormone Therapy With Breast Cancer Incidence and Mortality During Long-term Follow-up of the Women’s Health Initiative Randomized Clinical Trials. Jama. Jul 28 2020;324(4):369-380. 2. Cuzick J, Sestak I, Cawthorn S, et al. Tamoxifen for prevention of breast cancer: extended long-term follow-up of the IBIS-I breast cancer prevention trial. The Lancet Oncology. Jan 2015;16(1):67-75. 3. Manson JE, Aragaki AK, Rossouw JE, et al. Menopausal Hormone Therapy and Long-term All-Cause and Cause-Specific Mortality: The Women’s Health Initiative Randomized Trials. Jama. Sep 12 2017;318(10):927-938, Supplement. 4. Ragaz J, Shakeraneh S, Qian H, Wong H, Spinelli JJ, Wilson KS. Abstract PS7-05: Estrogen-based hormone replacement therapy [E-HRT] reduces all-cause, breast cancer, and Alzheimer’s dementia mortality. Cancer Research. 2021;81(4_Supplement):PS7-05-PS7-05. 5. Peethambaram PP, Ingle JN, Suman VJ, Hartmann LC, Loprinzi CL. Randomized trial of diethylstilbestrol vs. tamoxifen in postmenopausal women with metastatic breast cancer. An updated analysis. Breast cancer research and treatment. Mar 1999;54(2):117-22.

V. Lope¹, P. Fernández¹, Á. Guerrero-Zotano², P. Sánchez-Rovira³, A. Antón-Torres⁴, M. Benavent-Viñuales⁵, J. Baena-Cañada⁶, S. Antolín⁷, M. Muñoz⁸, L. Paris⁹, J. Chacón¹⁰, C. Olier¹¹, S. González¹², J. García-Sáenz¹³, Á. Jimenez-Arranz¹⁴, A. Oltra¹⁵, J. Brunet¹⁶, M. Marin-Alcala¹⁷, A. De Juan¹⁸, B. Pérez-Gómez¹, R. Rincón¹⁹, R. Caballero¹⁹, B. Bermejo²⁰, M. Martín²¹, M. Pollan¹; ¹Environmental Epidemiology and Cancer Area, Unidad de Epidemiologia del Cancer y Ambiental, Departamento de Epidemiologia de Enfermedades Cronicas, Centro Nacional de Epidemiologia (CNE), Instituto de Salud Carlos III (ISCIII); GEICAM Spanish Breast Cancer Group, Madrid, SPAIN, ²Medical Oncology, Instituto Valenciano de Oncologia (IVO); GEICAM Spanish Breast Cancer Group, Valencia, SPAIN, ³Medical Oncology, Hospital Universitario de Jaén; GEICAM Spanish Breast Cancer Group, Jaén, SPAIN, ⁴Medical Oncology, Hospital Universitario Miguel Servet, Instituto Investigación Sanitaria Aragon, Departamento Medicina Universidad de Zaragoza; GEICAM Spanish Breast Cancer Group, Zaragoza, SPAIN, ⁵Medical Oncology, Hospital Universitario Virgen del Rocio, Instituto de Biomedicina de Sevilla; GEICAM Spanish Breast Cancer Group, Sevilla, SPAIN, ⁶Medical Oncology, Instituto de Investigación e Innovación Biomédica de Cádiz; GEICAM Spanish Breast Cancer Group, Cádiz, SPAIN, ⁷Medical Oncology, Complejo Hospitalario de La Coruña (CHUAC); GEICAM Spanish Breast Cancer Group, A Coruña, SPAIN, ⁸Medical Oncology, Hospital Clinic i Provincial de Barcelona, FRCB-Translational Genomics and Targeted Therapies in Solid Tumors, August Pi i Sunyer Biomedical Research Institute (IDIBAPS), Department of Medicine, University of Barcelona; GEICAM Spanish Breast Cancer Group, Barcelona, SPAIN, ⁹Medical Oncology, Centro Oncológico de Galicia; GEICAM Spanish Breast Cancer Group, A Coruña, SPAIN, ¹⁰Medical Oncology, Hospital Univresitario de Toledo; GEICAM Spanish Breast Cancer Group, Toledo, SPAIN, ¹¹Medical Oncology, Hospital Universitario Fundación Alcorcón, Universidad Rey Juan Carlos; GEICAM Spanish Breast Cancer Group, Madrid, SPAIN, ¹²Medical Oncology, Research Foundation Mútua Terrassa/University of Barcelona, Hospital of Mútua Terrassa; GEICAM Spanish Breast Cancer Group, Madrid, SPAIN, ¹³Medical Oncology, Department of Oncology and Instituto de Investigación Sanitaria Hospital Clinico San Carlos (IdISSC); GEICAM Spanish Breast Cancer Group, Madrid, SPAIN, ¹⁴Medical Oncology, Instituto Maimónides de Investigación Biomédica de Córdoba (IMIBIC)-Hospital Universitario Reina Sofía, Universidad de Córdoba, Córdoba, SPAIN, ¹⁵Medical Oncology, Hospital Virgen de los Lirios; GEICAM Spanish Breast Cancer Group, Alcoy, SPAIN, ¹⁶Medical Oncology, Institut Catala d’Oncologia, Hospital Josep Trueta, Oncogir, IDIBGI; GEICAM Spanish Breast Cancer Group, Girona, SPAIN, ¹⁷Medical Oncology, Consorci Sanitari de Terrassa; GEICAM Spanish Breast Cancer Group, Terrassa, SPAIN, ¹⁸Medical Oncology, Hospital Universitario Marques de Valdecilla; GEICAM Spanish Breast Cancer Group, Santander, SPAIN, ¹⁹Translational, GEICAM Spanish Breast Cancer Group, Madrid, SPAIN, ²⁰Medical Oncology, Medicine Department, Hospital Clinico Universitario de Valencia, Instituto de Investigacion Sanitaria (INCLIVA), Universidad de Valencia; CIBERONC-ISCIII; GEICAM Spanish Breast Cancer Group, Valencia, SPAIN, ²¹Medical Oncology, Department of Medical Oncology, Hospital General Universitario Gregorio Marañon Instituto de Investigacion Sanitaria Gregorio Marañon, Universidad Complutense; CIBERONC-ISCIII; GEICAM Spanish Breast Cancer Group, Madrid, SPAIN.

Background: There is growing evidence supporting the benefits of the Mediterranean diet in breast cancer (BC) prevention, but their contribution to prognosis remains unclear. Our objective is to investigate whether adherence to the Mediterranean dietary patterns before diagnosis is associated with the risk of BC recurrence and tumor proliferation in an exploratory cohort of the multicenter case-control EpiGEICAM study.Methods: Transcriptome-wide gene expression was assessed using the nCounter Breast Cancer 360 panel (NanoString). Formalin-fixed paraffin-embedded tumor tissues from an exploratory cohort of 89 pre- or perimenopausal BC patients from the EpiGEICAM study were used. The PAM50 gene expression signature was employed to classify tumors into the gold-standard BC intrinsic molecular subtypes, and to compute both the PAM50 risk of recurrence (ROR) and PAM50 proliferation scores. Dietary patterns over the five years prior to diagnosis were identified in the control population of the EpiGEICAM study, applying principal components analysis without rotation of the variance-covariance matrix over 26 inter-correlated food groups. The association between adherence to Mediterranean dietary patterns and both ROR and proliferation scores was analyzed using linear regression models with continuous outcomes. All models were adjusted for age and adherence to a Western dietary pattern, with additional adjustment for PAM50 molecular subtype in the proliferation score analysis. Results: The mean age was 44 years and 28% had a family history of BC. Regarding molecular subtypes, 49.4% corresponded to Luminal A, 22.5% to Luminal B, 13.5% to HER2-enriched and 14.6% to Basal-like subtype. High adherence to the Mediterranean dietary pattern was observed in 56% of participants, with higher prevalence among women with Luminal A and Luminal B subtypes. The mean ROR score was 51 (95%CI=46-57) in women with low adherence to the Mediterranean diet, and 46 (95%CI=40-51) in those with high adherence. Mean proliferation scores were 4.8 (95%CI=4.6-5.1) and 4.5 (95%CI=4.3-4.8), respectively. Although no statistically significant associations were observed, women with high adherence to the Mediterranean dietary pattern showed lower mean ROR score (β=-5.8; 95%CI=-14.0-2.5) and a reduced proliferation score (β=-0.14; 95%CI=-0.40-0.12) compared to those with low adherence. Conclusion: The present EpiGEICAM exploratory analysis showed that adherence to a Mediterranean dietary pattern prior to diagnosis may be associated with lower breast cancer PAM50 proliferation and risk of recurrence scores. Studies with larger sample sizes are needed to confirm these results and to clarify the effect of the Mediterranean diet on the biology and prognosis of this tumor.

C. J. Fabian¹, C. R. Beaver¹, K. R. Powers¹, L. Ranallo¹, A. L. Kreutzjans¹, K. Pittman¹, C. Altman¹, T. Metheny¹, R. Biswell¹, A. Zelenchuk¹, H. Pathak², A. Mitra², A. K. Godwin², D. P. Mudaranthakam³, E. D. Giles⁴, S. D. Hursting⁵, K. L. Cook⁶, B. F. Kimler⁷; ¹Internal Medicine, University of Kansas Medical Center, Kansas City, KS, ²Pathology and Laboratory Medicine, University of Kansas Medical Center, Kansas City, KS, ³Biostatistics and Data Science, University of Kansas Medical Center, Kansas City, KS, ⁴Movement Science, University of Michigan, Ann Arbor, MI, ⁵Nutrition, University of North Carolina at Chapel Hill, Chapel Hill, NC, ⁶Cancer Biology and Surgery, Wake Forest University School of Medicine, Winston-Salem, NC, ⁷Radiation Oncology, University of Kansas Medical Center, Kansas City, KS.

Background: Achieving and sustaining the ≥10% weight loss sufficient to reduce risk for breast cancer has been difficult with diet and exercise alone. Tirzepatide is a dual agonist of glucagon-like peptide-1 (GLP-1) and glucose-dependent insulinotropic polypeptide (GIP) receptors. It is FDA-approved for obesity treatment with average weight loss of 14% at 24 weeks. Genetic variants of the GIP receptor (GIPR) that reduce GIPR signaling have been associated with increased breast cancer risk; however, little is known about the presence or localization of GIPR within breast tissue. High mammographic density, breast adipocyte size, breast epithelial Ki-67 ≥2%, and visceral fat of >1kg (iDXA) are all associated with increased risk of breast cancer. With the exception of BMI and insulin resistance, there is little information on the effect of tirzepatide on risk biomarkers for breast cancer. We sought to assess feasibility of a convenient clinical cohort trial model and gain preliminary information on biomarker modulation in a pilot study. Methods: Women with obesity and additional risk factors for breast cancer seen in the KUMC Weight Management Clinic (WMC) and Breast Cancer Prevention Clinic (BCPC), considering clinical use of tirzepatide, were approached for study participation. Procedures for assessment of risk biomarkers were performed by BCPC personnel before and after 6 months of tirzepatide. Inclusion criteria were age 35-64, BMI ≥30 kg/m², ≥2x population risk for breast cancer, and self-pay or insurance coverage for tirzepatide. Exclusion criteria included prior invasive cancer or current insulin or antiestrogens. Prior to tirzepatide initiation, women underwent a fasting blood draw, 3D mammography, iDXA, random periareolar fine needle aspiration, and optional stool for microbiome. Tirzepatide was started at a dose of 2.5 mg/week, escalated every 4 weeks to a maximum of 15 mg under the guidance of the WMC. Key imaging risk biomarkers are absolute fibroglandular volume (FGV) via Volpara™ and visceral fat via iDXA. Tissue biomarkers performed in the BCPC or Biomarker Validation Core laboratories include Ki-67, GIPR, and estrogen response gene expression. Systemic markers include estradiol, estrone, SHBG, progesterone, insulin, adipocytokines, circulating microRNA, and Klotho. Breast adipocyte size is measured at the University of Michigan and microbiome analyzed at Wake Forest University. Objectives: 1) Feasibility as defined by ≥1 enrollee/month and >70% completion of study including biomarker procurement; 2) Proportion of women with detected GIPR in breast epithelium; and 3) Identification of reliably modulated biomarkers of breast cancer risk. Results: 23 women have been enrolled in 9 months with median age 49, BMI 35 kg/m², IBIS 10-year breast cancer risk 6.3%, and visceral fat mass 1.7 kg. 13 are pre or perimenopausal and 10 postmenopausal. Of the 12 women completing study to date, median relative change in BMI is -15%, visceral fat -19%, lean mass -8%, FGV -22%, and HOMA-IR -31%. Median relative change in Ki-67 for the seven women with baseline Ki-67 ≥ 2% is -50%. Five of seven initial entrants have shown evidence at baseline of low, but detectable, levels of GIPR in breast epithelium by spatial genomics. Conclusions: Rapidity of recruitment, high retention, and completion of off-study biomarker assessments support expansion to a larger phase II trial where the directionality and magnitude of biomarker change can be evaluated with greater precision and confidence.

M. Wood¹, M. I. Elsaid², L. J. Grimm³, H. Liu⁴, C. Oswold⁵, C. S. Kuzma⁶, I. Bedrosian⁷, J. Ligibel⁸, P. J. Goodwin⁹; ¹Medical Oncology, University of Colorado Anschutz Medical Campus, Aurora, CO, ²Biomedical Informatics and Internal Medicine, The Ohio State University College of Medicine, Columbus, OH, ³Radiology, Duke University, Durham, NC, ⁴Quantitative Health Sciences, Mayo Clinic, Rochester, MN, ⁵Department of Quantitative Health Sciences, Mayo Clinic, Rochester, MN, ⁶Hematology/Oncology, First Health of the Carolinas, Pinehurst, NC, ⁷Breast Surgical Oncology, University of Texas MD Anderson Cancer Center, Houston, TX, ⁸Medical Oncology, Dana-Farber Cancer Institute, Aurora, MA, ⁹Medical Oncology, University of Toronto, toronto, ON, CANADA.

Background: Metformin may impact tumorigenesis directly, through effects on AMP kinase, and indirectly, through reduction in circulating growth factors such as insulin. To examine the potential preventive effects of metformin, we analyzed MD, a known risk factor for breast cancer, in participants enrolled in the NCIC MA.32 trial, a randomized, placebo-controlled trial investigating the effects of 5 years of metformin on invasive disease-free survival in women with early breast cancer. The primary aim of this companion to MA.32 (Alliance A211201) was to examine the effects of 1 year of metformin vs placebo on MD in women treated for estrogen receptor negative (ER-) breast cancer. Methods: Eligible women were enrolled on MA.32, had ER- breast cancer, had >25% MD at baseline, and had an unaffected contralateral breast. MD was calculated on the contralateral breast using Cumulus software and analyzed at baseline, 1 and 2 years. The aim was to achieve a sample size of 274 (137 per arm), which would have at least 85% power to detect a 4% difference between study arms, i.e., 5% change from baseline to one year of treatment for metformin vs. 1% for placebo, using a two-sided two-sample t-test at a significance level of 5%. The current study was activated in the United States on 8/22/2012. Enrollment was halted on 10/13/2017 with 177 participants enrolled. Results: 146 women were evaluable (66 on metformin and 80 on placebo). The mean age of evaluable women was 53.6 years (standard deviation [SD] 9.4), 88.4% were white, 31.5% had a body mass index greater than or equal to 30 kg/m², 99.3% of participants received chemotherapy. The proportions of white women and HER2+ participants in the metformin group were higher compared to placebo ([white 92.4% vs. 85.0%] and [HER2+ 21.2% vs. 11.3%], respectively). The mean MD at baseline was 22.7% (SD 14.1%) for the group as a whole (24.5% (SD 14.5%) for the metformin group and 21.2% (SD 13.6%) for the placebo group. This is lower than the eligibility criteria (>25%MD), due to difference in methods of calculation of MD (BIRADs for eligibility and Cumulus for study analysis). At 1 year, MD decreased by 0.7% (SD 7.6%) in the metformin group and 0.2% (SD 7.3%) in the placebo group (p = 0.58). At 2 years, MD decreased by 0.7% (SD 8.0%) in the metformin group and 0.8% (SD 7.1%) in the placebo group (p=0.60). Conclusions: In a cohort of women with early ER- breast cancer, there was no significant difference in the change in MD in women treated with metformin vs. placebo. There was a non-significant trend toward decreased MD while on metformin after 1 year of treatment. The arms were well balanced, though with more white and HER2+ participants in the metformin arm. The sample size for this study was about half of what was intended thus limiting power to detect a significant difference. Additional limitations include low mean MD, which may have limited our ability to detect the benefit in those at the highest risk, and further, the Cumulus method of MD determination can be imprecise. Interestingly, trends toward decreased MD with metformin were seen in Alliance A211102, and it might be reasonable to consider further study of metformin in a select population (highest MD or atypical hyperplasia). Support: UG1CA189823, U10CA180863, CCS 707213; https://acknowledgments.alliancefound.org. Clinicaltrials.gov ID: NCT01666171.

M. Alsuhaibani¹, A. Melanson¹, A. Viezel-Mathieu², T. Dionisopoulos², M. Basik¹, J. Boileau¹, M. Martel¹, I. Prakash¹, S. Meterissian¹, J. Vorstenbosch², W. Foulkes³, S. Wong¹; ¹General Surgery, McGill University, Montreal, QC, CANADA, ²Plastic Surgery, McGill University, Montreal, QC, CANADA, ³Oncology, McGill University, Montreal, QC, CANADA.

Introduction: Risk-reducing mastectomy (RRM) a well-established preventive intervention in germline pathogenic variants (GPVs) carriers. Historically, RRMs were performed as total mastectomy (TM) or skin-sparing mastectomy (SSM), although nipple-sparing mastectomy (NSM) has recently emerged as a RRM option. In this study, we assessed temporal trends in NSM uptake. Methods: This retrospective cohort study included all female patients with confirmed BRCA1, BRCA2, PALB2, or other GPVs carriers at two academic institutions in Montreal, Canada, between 2003 and 2024. The primary outcome was the proportion of GPV carriers NSM compared to SSM and TM over time as evaluated by Mantel Haenszel test for trend. Results: After exclusions, 441 GPV carriers were included in the cohort, 260 (59%) affected carriers and 181 (41%) unaffected carriers. The median age at RRM was 44 years old (range 22-76). Most were BRCA1 carriers (45.5%) or BRCA2 carriers (44.8%) while the remaining 9.8% were PALB2 and other GPV carriers. Overall, 269 (61%) underwent NSM, 118 (26.8%) underwent SSM, and 54 (12.2%) underwent TM. NSM uptake was highest among patients aged 40-49 years (68.6%) and lowest in those ≥60 years (45.8%, p=0.04). Relative to affected carriers, unaffected carriers were more likely to undergo NSM (72.4%, Vs.53.1% p<0.001). Prior radiation significantly decreased the likelihood of NSM (38.4% vs. 68.7%, p<0.001). A clear increase in NSM occurred over time, from 12.5% of GPV carriers pre-2012 to 77.5% after 2022 (p<0.001). Multivariable analysis adjusting for age, GPV, personal history of breast cancer and prior radiation identified year of surgery as the strongest predictor of NSM (OR 28.9, 95% CI 10.1-77.0) post-2019 compared to pre-2012 (p<0.001). Conclusion: There has been a significant increase in NSM uptake among GPV carriers over the last two decades. Understanding these trends supports informed decision-making and standardization of preventive care in high-risk populations.

Y. Wang, J. Zhan, S. Liu, W. Cai, Y. Qiu, P. He, M. Chen, L. Chen, F. Fu, C. Wang; Breast Surgery, Fujian Medical University Union Hospital, Fuzhou, CHINA.

Background: Patients with breast cancer are susceptible to coronavirus disease (COVID-19), which affects cancer treatment efficacy and prognosis. However, its effects on neoadjuvant chemotherapy (NAC) response and its genetic correlation with breast cancer remain unclear. We aimed to clarify these effects and investigate potential genetic associations and shared pathogenic mechanisms. Μethods: In this ambidirectional (retrospective and prospective) cohort study involving 856 breast cancer patients receiving NAC, with and without COVID-19, we evaluated the impact of COVID-19 on NAC response using multivariable logistic regression and three matching methods: propensity score matching, inverse probability of treatment weighting, and overlap weighting. Using single-cell and bulk transcriptome data from the Gene Expression Omnibus and genetic variants from the Genome-Wide Association Study databases, we conducted Mendelian randomization (MR) analysis to explore the core cellular subsets, shared genes, causal relationships, and common pathogenic mechanisms between COVID-19 and breast cancer. Results: Patients with breast cancer and COVID-19 showed poor NAC responses, including a low objective response rate and reduced likelihood of achieving RCB class 0-I in the HR+/HER2+ subgroup (OR=0.46, P=0.028, P for interaction=0.024). The consistent findings from the three matching models support these results. Integrating single-cell and bulk transcriptome data with MR analyses revealed genetic correlations between COVID-19 and breast cancer and identified ABLIM1 and GZMM as potential genes linked to NAC resistance. Conclusions: SARS-CoV-2 infection during NAC may compromise therapeutic efficacy in patients with breast cancer. ABLIM1 and GZMM may be causal genes in the genetic correlation between COVID-19 and breast cancer.

M. Schmalzle¹, R. Radigan¹, S. Singh², W. Liu³, K. Deng², M. Fabrikant², R. Shah⁴, S. L. Fu², J. Kao⁴; ¹Biomedical Sciences, New York Institute of Technology College of Osteopathic Medicine, Old Westbury, NY, ²The Cancer Institute, Good Samaritan University Hospital, West Islip, NY, ³Department of Pathology, Good Samaritan University Hospital, Hauppauge, NY, ⁴Department of Radiation Oncology, Good Samaritan University Hospital, West Islip, NY.

Introduction: Older women with early-stage luminal A breast cancer tend to have an excellent prognosis with a 5-year relative survival exceeding 99%. Ongoing research on curtailing overdiagnosis and overtreatment have challenged long established standards of care by reducing the use of whole breast radiotherapy and selectively omitting sentinel lymph node biopsy. In light of emerging data suggesting that endocrine therapy can be safely omitted in older women with ultra-low risk luminal A breast cancer, the marginal utility of Oncotype Dx genomic testing was explored. Materials and Methods: We queried the Good Samaritan University Hospital Cancer Registry to identify consecutive breast cancer patients treated from 2014 through 2022. Based on extensive research on radiation omission and selecting patients for partial breast irradiation, we defined an ultra-low risk breast cancer population as age ≥60 years, estrogen receptor ≥1% positive, progesterone receptor >20% positive, HER2-negative, pathologic stage T1b-T2N0, well to moderately differentiated, no lymphovascular invasion, and Ki-67 ≤13.25%. The primary endpoint was the proportion of patients with low (0-18) or intermediate (19-25) risk Oncotype scores, for which chemotherapy is not indicated. Secondary endpoints included local recurrence and overall survival. Results: Out of 1711 patients, 91 met all ultra-low risk eligibility criteria. Median age was 70 years (IQR 65-74; range 60-85) and median breast tumor size was 10 mm (IQR 7-15). Most patients were diagnosed with invasive ductal carcinoma (71.4%) or lobular carcinoma (16.5%). All patients underwent surgical intervention, with 79.1% receiving a lumpectomy and 20.9% receiving a mastectomy. Sixty-nine (75.8%) patients received post-operative radiotherapy and no patients received chemotherapy. The median follow-up was 5.14 years. Only 1 of 91 patients (1.1%) had a high risk Oncotype Dx score of 26, while 12 (13.2%) had an Oncotype Dx score of 19 to 25 and 78 (85.7%) scored 0 to 18. Five-year local control and overall survival was 100% and 93%, respectively. Nine deaths occurred, none related to breast cancer, and no patients developed distant metastases. Conclusion: We describe a pragmatic method of using clinical and pathological features routinely available in a community hospital breast program to define an ultra-low risk cohort. Older luminal A patients with favorable pathology have a very low probability of a high Oncotype Dx genomic score and have an excellent prognosis with standard adjuvant therapy. Future research for ultra-low risk breast cancer patients should focus on treatment de-escalation beyond radiotherapy.

C. Wang¹, L. Jing¹, Y. Lei¹, Y. Zhang¹, X. Wang¹, Y. Wang², H. Jia¹; ¹Department of Breast Surgery, First Hospital of Shanxi Medical University, Taiyuan, CHINA, ²Department of Basic Medicine, Shanxi Medical University, Taiyuan, CHINA.

Background: Breast cancer (BC) has been associated with psychiatric disorders. However, the shared genetic determinants underlying these associations remain unevaluated. This study aims to investigate the shared genetic architecture between psychiatric disorders and BC, and to identify the involved shared risk loci, potential key tissues, and underlying genetic mechanisms.Methods: Based on the summary statistics from genome-wide association studies (GWAS), the genetic correlation and overlap between psychiatric disorders and BC were investigated. Furthermore, shared pleiotropic loci and genes were identified through cross-trait analyses. Additionally, a series of functional annotations and tissue-specific enrichments were performed to determine potential associations between each trait pair. Pathway enrichment analysis was conducted to assess possible associated mechanisms. Finally, the causal relationship at the protein level was also evaluated.Results: The results demonstrated shared genetic mechanisms between eight psychiatric disorders and two BC subtypes. After Bonferroni correction,4 of the 16 trait pairs exhibited significant genetic correlations (p < 3.13 × 10⁻³). A total of 7,274 SNPs were identified at a genome-wide significance threshold of p < 5. 00 × 10⁻⁸ threshold. Furthermore, annotation analysis identified 156 genomic risk loci, of which 19 were analyzed using causal colocalization. Gene-level analysis revealed a series of pleiotropic genes, with MTRFA and FGFR2 being identified in the majority of trait pairs. Tissue enrichment analysis indicated that pleiotropic mechanisms play a significant role in the ovary, breast mammary tissue, uterus, and certain brain regions, including the caudate nucleus, cerebellar hemisphere, and putamen. Pathway analysis showed that most trait pairs were enriched for positive regulation of RNA metabolism. Additionally, protein-level causal correlation analysis identified 5 proteins significantly associated with BC risk and 36 proteins significantly associated with the risk of psychiatric disorders.Conclusion: This study explored the shared genetic architecture between psychiatric disorders and BC and identified several potential genetic mechanisms that are critical for the development of novel therapeutic strategies in clinical practice.

J. de la Haba-Rodríguez¹, R. Peña-Enríquez², D. Ponferrada¹, R. Bautista-Moreno³, S. Romero-Martín⁴, E. Contreras⁴, A. Díaz-Chacón², C. Morales¹, B. Rodríguez¹, P. Sanchez-Maurinho¹, A. Martinez-López⁵, C. León⁶, J. Cejas-Arjona⁶, P. Notario-Fernández⁶, P. Rioja-Torres⁶, J. Raya-Povedano⁴, M. Álvarez⁴, S. Guil-Luna²; ¹Medical Oncology Department, Reina Sofía University Hospital, Córdoba, SPAIN, ²New therapies in cancer group (CG06), Maimonides Institute of Biomedical Research of Cordoba (IMIBIC), University of Córdoba, Córdoba, SPAIN, ³Bioinformatics Unit, Supercomputing and Bioinformatics Centre (SCBI), University of Malaga, Malaga, SPAIN, ⁴Radiodiagnosis Department, Reina Sofía University Hospital, Córdoba, SPAIN, ⁵Anatomical Pathology Department, Reina Sofía University Hospital, Córdoba, SPAIN, ⁶Breast Surgery Department, Reina Sofía University Hospital, Córdoba, SPAIN.

Background: Breast cancer (BC) in young women under 45 is a growing public health concern, with around 50% of cases occurring in mothers within 5-10 years after childbirth. This entity, known as postpartum breast cancer (PPBC), is characterised by aggressive clinical behaviour and higher metastasis and mortality rates compared to age-matched BC patients. Despite its recognised poor prognosis, the molecular mechanisms driving PPBC remain poorly understood, and this knowledge gap limits the ability to identify women at higher risk and develop prevention strategies. This study aimed to characterize the transcriptomic alterations in both tumoral and histologically-normal adjacent breast tissue from PPBC patients, compared to non-PPBC and healthy controls, with a focus on identifying early risk signatures. Methods: RNA-seq analysis was performed on fresh breast tissue samples collected from 30 young women (aged ≤ 45 years) at Reina Sofia University Hospital of Córdoba: paired tumor and adjacent normal tissue from PPBC patients (n=12), tumor tissue from non-PPBC patients (n=8), and normal breast tissue from healthy nulliparous/parous women (n=10). Transcriptomic data were analysed for differential gene expression using DESeq2, and pathway enrichment analysis was performed using GSEA querying the Hallmarks (MSigDB) and 100 customized gene sets. Significantly enriched pathways were identified based on FDR q-values 1.5 and FDR p-value <0.05. The shared overexpressed genes in PPBC tumour and adjacent tissue versus healthy controls were evaluated as potential biomarkers using multivariate ROC analysis. Results: PPBC tumors exhibited a distinct, aggressive profile compared to non-PPBC patients with significant enrichment of cell cycle activation pathways (E2F, MYC targets and G2M checkpoint), reflecting high proliferative activity, as well as dysregulation of the WNT/β-catenin signalling pathway (FDR q-value < 0.1). Notably, adjacent histologically-normal tissue of PPBC patients also exhibited a unique transcriptomic profile, distinct from healthy breast tissue. This included enrichment of stromal-fibroblastic remodeling signatures, such as TGF-β signaling, epithelial-mesenchymal transition (EMT), and markers related with cancer-associated fibroblasts (CAFs) and endothelial cells. Interestingly, a robust post-lactational involution-like signature, characterized by upregulation of genes linked to mammary gland regression, late parity, and a transient postpartum immune state (FDR qvalue  0.9). These genes were predominantly involved in pathways related to profibrotic and cytokine signaling, with top gene markers such as SOCS4, TMED7 and DYNLT3. Conclusions: Our findings reveal that molecular alterations in PPBC extend beyond the tumor itself, implicating the surrounding normal tissue in a tumor-promoting, stromal fibroblastic remodeling and post-lactational microenvironment. This transcriptomic profile supports the hypothesis that PPBC may develop within a mammary microenvironment already altered by post-gestational involution and stromal processes that could favour normal cells to malignancy. These insights highlight the significant potential for improving risk stratification and guiding early intervention strategies in postpartum women.

K. Woods¹, S. Cummings¹, F. Ademuyiwa²; ¹Oncology Medical Affairs, Myriad Genetics, Inc., Salt Lake City, UT, ²Division of Oncology, Washington University School of Medicine in St. Louis, St. Louis, MO.

Background: Numerous studies have examined germline variant rates in individuals with breast cancer (BC), consistently reporting that about 5-10% carry a germline pathogenic variant (PV). Although multiple PVs in a single patient have been documented, such cases are rare and reported in small cohorts. As a result, the clinical implications of multiple PVs remain poorly understood. To address this gap we analyzed a large dataset of patients with BC who underwent germline genetic testing. Methods: The Myriad Collaborative Research Registry, containing clinical, genetic, and genomic information on over 1.2 million individuals with cancer, was queried. The registry includes data on 795,165 patients with a BC diagnosis who completed testing at a commercial laboratory between 1996 and 2025. Data was assessed using descriptive statistics to understand the prevalence and characteristics of patients with multiple PVs. Results: Of the 795,165 patients with BC, 75,491 (9.5%) patients had one PV, 1,600 (0.2%) carried two PVs, and 34 patients carried more than 2 PVs. Of the 1,600 patients with two PVs, the most common second primary diagnoses included ovarian cancer (84), colon adenomas/polyps (68), breast cancer (45), and endometrial cancer (39). In patients with two PVs, the median age at cancer diagnosis was 49 years; most patients were female (1,561) and White/Non-Hispanic (1,197). Patients with two PVs were grouped by PV risk categories in BC: high-risk, moderate-risk, and other-risk. See Table 1 for patient groupings based on risk categories. The largest subgroup (28.3%) had one high-risk and one other-risk PV. The median age at diagnosis was 48 years across most groups, increasing to 54 in the other/other-risk group. BC was the most reported family history across all groups. In groups with at least one high-risk PV, ovarian cancer was the second most reported family history. Colon cancer was the second most reported family history in all groups with at least one other-risk PV. BRCA1/BRCA2 made up 73.7% of pairings in the high/high-risk group, and ATM/CHEK2 accounted for 69.2% in the moderate/moderate-risk group. The frequency of these pairs was markedly higher than the highest frequency pairings in other groups. Notably, one-third of all patients with two PVs had a PV in MUTYH. Conclusions: Assessing patients with two germline PVs presents clinical and interpretive challenges. This study identified two PVs at a frequency of 1,600 in 795,165 (0.2%) patients with BC and highlights patterns suggesting that detailed personal and family history may aid in understanding these complex genetic profiles. Continued data collection is essential to identify subtle trends and improve risk assessment. Future research is needed on specific gene pairings to clarify phenotypic differences and refine risk stratification. The registry enabled identification of a large cohort with two PVs, laying a foundation for deeper investigation into their clinical significance.

M. Bustamante Eduardo¹, A. B. Islam², M. Zappia², A. Z. Samuel³, C. W. McCloskey⁴, M. V. Frolov², R. W. Khokha⁴, E. V. Benevolenskaya², S. A. Khan¹, S. E. Clare¹; ¹Department of Surgery, Northwestern University, Chicago, IL, ²Department of Biochemistry and Molecular Genetics, University of Illinois at Chicago, Chicago, IL, ³Department of Bioengineering, University of Illinois at Urbana-Champaign, Urbana, IL, ⁴Department of Medical Biophysics, Princess Margaret Cancer Centre, Toronto, ON, CANADA.

Introduction. The local breast environment is an important source for identifying the etiological and biological factors that contribute to the development of breast cancer (BC). We identified a lipid metabolism gene signature associated with Estrogen Receptor negative BC (ERnegBC) in high-risk benign breast tissue. We then exposed non-transformed breast cells and breast microstructures in vitro and ex vivo to fatty acids (FA) and observed a metabolic shift toward the serine, one-carbon, glycine and methionine pathways, engendering epigenetic plasticity, increasing reactive oxygen species (ROS), and promoting cell survival. This promted the question: are there in vivo context(s) where lipid alterations resemble those of FA exposure in vitro/ex vivo? Since epigenetic plasticity and increased ROS are also reported in the aged mammary gland, we hypothesize that FA-induced metabolic reprogramming increases in the aged mammary gland, contributing to pro-tumorigenic alterations observed during aging. Methods. MCF-10A cells were exposed to the medium-chain FA octanoic acid (OA) for proteomics. Breast tissue-derived microstructures exposed to ± OA were utilized for single-cell RNA-seq. Breast microstructures and 3D mammary spheres (formed from mammary cells in Matrigel) were exposed to ± OA for 7 days, stained for luminal/basal markers, F-actin, and nuclei, and imaged using confocal microscopy. Raman spectroscopy was used to characterize the lipid content in pre- versus post-menopausal breast tissue. Results. In vitro and ex vivo data on FA exposure in non-transformed mammary cells and breast microstructures revealed striking similarities to the aged mammary gland: 1) Significant (p < 0.01) aging-related changes in gene expression upon OA exposure in epithelial and non-epithelial compartments. Notably, up-regulation of GDF15 (related to EMT, invasion, lipolysis and aging), MDK (involved in neurogenesis, cancer progression, and aging) and the downregulation of MMP7, a gene for which reduced expression contributes to the accelerated aging of mammary epithelial cells. Moreover, OA led to significant downregulation (p < 0.01) of lineage markers. 2) Significant (p < 0.01) upregulation of components of the Senescence-Associated Secretory Phenotype in both epithelial and non-epithelial subtypes. Including those that promote reprogramming (AREG), cancer progression (EREG), EMT (CCL20) and migration (ANGPTL4). 3) Altered inter-cellular communications upon OA exposure. Cell-cell communication analysis indicated increased secreted signaling in OA, with AREG, GDF15, and MDK among the strongest signals. 4) Decreased cell-cell junctions and extracellular matrix-receptor (ECM-R) interactions. Differential proteomics revealed downregulation of proteins involved in ECM-R interaction and focal adhesion. Cell-cell communication analysis also revealed decreased ECM-R interactions following OA exposure, alongside reduced cell-cell adhesions. 5) OA led to aging-associated pro-tumorigenic changes that may increase susceptibility to BC: Ex vivo culture of breast microstructures and 3D mammary spheres revealed that OA exposure alters cell-cell adhesions, compromising the basal barrier and leading to luminal cell dissemination. Conclusions. The proportion of breast adipocytes releasing free FAs increases with age. We show that FA exposure causes changes typical of the aged mammary gland, such as elevated ROS, epigenetic alterations, downregulation of cell-cell junctions, altered ECM-receptor interactions, and aging associated gene changes. We propose that the release of free FAs from an increased number of breast adipocytes (potentially due to elevated GDF15-induced lipolysis during aging) is a factor that induces mammary gland remodeling, increasing vulnerability to BC.

A. S. Fiscalini¹, T. Glatt², S. Goodman¹, C. Kaplan³, A. Z. LaCroix⁴, L. van ‘t Veer⁵, M. Scheuner⁶, P. Sales⁷, A. Petruse⁸, A. Naeim⁹, A. Kaster¹⁰, J. Wernisch¹¹, O. I. Olopade¹², B. Gonzales¹³, H. L. Park¹⁴, A. D. Borowsky¹⁵, S. A. Raouf¹⁵, J. Atamer¹, K. Leggat-Barr¹⁶, J. Esserman¹⁷, I. Cabaleiro¹⁷, R. Lancaster¹⁸, L. Anderson¹⁹, D. Heditsian²⁰, S. Brain²⁰, V. Lee²⁰, D. Moorehead²¹, B. A. Parker²², A. Torres¹, L. Sabacan¹, L. Johansen⁸, X. Calderon²³, A. Rocha⁸, A. Verma²⁴, J. Cover²⁵, T. Miller²⁶, M. Che²⁷, N. Kim²⁸, T. Lewis²⁷, S. Kapoor¹, R. Soonavala²⁹, D. Goodman-Gruen³⁰, S. D. Stewart³¹, R. A. Hiatt³², N. Wenger³³, H. Anton-Culver³⁴, I. A. Soto¹, M. Eklund³⁵, K. F. Rhoads³⁶, L. J. Esserman¹; ¹Surgery, UCSF, San Francisco, CA, ²School of Medicine, Tufts University, Boston, MA, ³Division of General Internal Medicine, Department of Medicine, UCSF, San Francisco, CA, ⁴Family and Preventive Medicine, UCSD, San Diego, CA, ⁵Lab Medicine, UCSF, San Francisco, CA, ⁶Department of Medicine, UCSF, San Francisco, CA, ⁷Medicine Service, Hematology-Oncology Section, San Francisco VA Health Care System, San Francisco, CA, ⁸Clinical and Translational Science Institute Office of Clinical Research, UCLA, Los Angeles, CA, ⁹Division of Hematology-Oncology, Department of Medicine, David Geffen School of Medicine, UCLA, San Francisco, CA, ¹⁰Edith Sanford Breast Center, Sanford Health, Fargo, ND, ¹¹Edith Sanford Breast Center, Sanford Health, Sioux Falls, SD, ¹²Surgery, University of Chicago, Chicago, IL, ¹³Department of Radiology, University of Chicago, Chicago, IL, ¹⁴Department of Pathology and Laboratory Medicine, UCI, Irvine, CA, ¹⁵Pathology, UC Davis, Sacramento, CA, ¹⁶School of Medicine, Tufts, Boston, MA, ¹⁷OB-GYN, Diagnostic Center of Miami, Miami, FL, ¹⁸Department of Surgery, University of Alabama at Birmingham, Birmingham, AL, ¹⁹Division of General and Internal Medicine and Population Sciences, University of Alabama at Birmingham, Birmingham, AL, ²⁰Breast Science Advocacy Core, Breast Oncology Program, UCSF, San Francisco, CA, ²¹Community Engagement, Women’s Cancer Resource Center, Berkeley, CA, ²²Moores Cancer Center, UCSD, San Francisco, CA, ²³Surgery, UCSF, Los Angeles, CA, ²⁴Pritzker School of Medicine, University of Chicago, Chicago, IL, ²⁵Research Services, Northern California Institute for Research and Education, San Francisco, CA, ²⁶Department of Veteran’s Affairs, San Francisco VA Health Care System, San Francisco, CA, ²⁷School of Medicine, UCSF, San Francisco, CA, ²⁸School of Medicine, Georgetown University, Washington, DC, ²⁹David Geffen School of Medicine, UCLA, San Francisco, CA, ³⁰Department of Epidemiology and Biostatistics, UC Irvine, Irvine, CA, ³¹Department of Pathology, School of Medicine, UC Irvine, Sacramento, CA, ³²Helen Diller Family Comprehensive Cancer Center, Department of Epidemiology & Biostatistics, UCSF, San Francisco, CA, ³³Division of General Internal Medicine and Health Services Research, UCLA, Los Angeles, CA, ³⁴Department of Medicine, Genetic Epidemiology Research Institute, UCI, Irvine, CA, ³⁵Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, SWEDEN, ³⁶Helen Diller Family Comprehensive Cancer Center, UCSF, San Francisco, CA.

BackgroundThe Women Informed to Screen Depending on Measures of risk (WISDOM) Study 1.0 (2016-2023; NCT02620852) is testing a novel, personalized approach to breast cancer screening compared to annual mammography. As a large-scale screening trial, WISDOM aimed to enrolled individuals ages 40-74 with no personal history of breast cancer using a preference tolerant randomized trial design that aimed to reflect the heterogeneity of the U.S. population. The goal of this analysis is to evaluate the recruitment strategies employed during WISDOM 1.0 and assess their effectiveness in building a geographically and racially diverse cohort of U.S. women for breast cancer screening research.MethodsWe analyzed enrollment and recruitment data sources including participant-reported responses to the ‘How You Heard’ (HYH) survey question administered at enrollment. Recruitment channels were grouped into categories (e.g., email from doctor, flyer/brochure, social media, friend/family referral), and stratified analyses were performed by recruitment site and self-reported race/ethnicity. We conducted analyses to evaluate volume and proportion of participants recruited by each strategy, rates by outreach method over time, differences by geography and participant demographics. ResultsBetween August 2016 and February 2023, WISDOM 1.0 registered 70575 women, consented 55284, and fully enrolled 46289 participants. Initial recruitment was limited to the Athena Breast Health Network within the University of California system (2016 -2017) but expanded through collaborations with external institutions and health systems and later to nationwide virtual enrollment.The recruitment method that yielded the most participants during the trial was emails from medical centers, MyChart, and physicians (40% of total enrollment). This was followed by family/friend referrals (10%) and emails from a Veteran’s Affairs (VA) recruitment partnership (6%). During the VA outreach, this strategy accounted for 25% of enrollment driving upward spikes in participation particularly in women of color. Other strategies included health insurance plan (4%), news article/radio/TV (4.5%) and social media (2.5%). WISDOM’s cohort diversity improved over the course of the trial, led by efforts from our Community Leadership Advisory Board, new recruitment centers, nationwide collaborations, and partnership with the VA. Over the course of the trial, there was a reduction in white, Non-Hispanic participants from 83% pre-community engagement efforts to 60% after new strategies were implemented (average 75%). The VA contributed to 26% of Black/African American participant enrollment, 10% of Hispanic participants, and 9% each for American Indian/Alaska Native (AI/AN) and Native Hawaiian/Pacific Islanders (NH/PI). Family/friend referrals accounted for over 7% of NH/PI but less than 4% for Black/AA, and only 1.5% for AI/AN. NH/PI participants were less likely to report physician/medical center (31%) and had the highest proportion of other options such as community organizations (9%) and health insurance plan (9%).DiscussionWISDOM 1.0 demonstrated the feasibility of recruiting a large, nationwide cohort for a personalized breast cancer screening trial using multi-modal, decentralized outreach. Physician emails and MyChart emerged as the highest-yield strategy, leveraging pre-existing patient trust and communication channels. The VA collaboration stands out as a highly scalable and demographically impactful approach while friend/family referrals offered downstream enrollment benefits. WISDOM 2.0 launched in June 2023 leveraging WISDOM 1.0’s recruitment approaches with an emphasis on refining high-yield, low-barrier strategies, amplifying trusted voices, and integrating culturally responsive approaches.

E. Low, N. Nyamongo, A. Lackan, C. Gonzalez, K. Hoffman, N. Kunta, S. Sarlin, S. Bulsara, S. Hilsenbeck, T. Snow, C. Sullivan, G. Garza, A. Elkhanany, N. Chen, K. Osborne, M. F. Rimawi, S. Shah, J. Nangia; Dan L. Duncan Comprehensive Cancer Center, Baylor College of Medicine, Houston, TX.

Background: Pathologic complete response (pCR) after neoadjuvant chemotherapy in breast cancer (BC) has important prognostic implications. Understanding its associations with pathogenic/likely pathogenic variants (P/LPV) in hereditary cancer genes (HCG) adds further prognostic value. Minorities remain underrepresented in hereditary cancer research. In this retrospective chart review, we compare pCR rates among patients with P/LPV in HCG at a safety-net clinic comprised of majority Latino patients to those at an academic clinic with majority white patients. Methods: All patients at Baylor St. Luke’s Medical Center (BSLMC, academic clinic) and Smith Clinic (SC, safety-net clinic) with a diagnosis of stage I-III BC who were found to have a P/LPV in HCG from 10/1/09-9/30/19 and received neoadjuvant chemotherapy followed by surgery were included. Demographic, clinical characteristics, and pathologic data were collected. Results: 118 patients qualified for analysis (71 from SC, 47 from BSLMC). In SC 50 (70%) were Latino, 3 (4.2%) White, 11 (15.5%) Black, and 1 (1.4%) Asian. In BSLMC, 15 (32%) were Latino, 20 (42.6%) White, 8 (17%) Black, and 2 (4.3%) Asian. Median age was 42 for SC and 54 for BSLMC. SC comprised of 41 (58%) BRCA1+, 13 (18.3%) BRCA2+, and 12 (16.9%) PALB2+ patients. BSLMC comprised of 16 (34%) BRCA1+, 8 (17%) BRCA2+, 6 (12.8%) ATM+, 5 (10.6%) CHEK2, and 3 (6.4%) PALB2+. SC and BSLMC patients ranged from stage I-III, with 49% and 36.2% falling in stage III respectively. SC had 37 (52.1%) triple negative BC (TNBC) patients, 25 (35.2%) ER/PR+, HER2- (HR+), 7 (9.9%) ER+/HER2+, and 2 (2.8%) ER-/HER2+. BSLMC had 14 (29.8%) TNBC, 25 (53.2%) HR+, 7 (14.9%) ER+/HER2+, and 1 (2.1%) ER-/HER2+. 36 (51%) and 21 (44.7%) were treated with both anthracycline and taxanes in the SC and BSLMC populations respectively. In the SC population, pCR rates (detailed in Table 1) were highest among TNBC (16, 43.2%). Of the TNBC patients that had pCR, 15 (94%) had P/LPV in a high-risk HCG (BRCA1/2 or PALB2). Similarly, of the HR+ patients that had pCR, 6 (85%) had P/LPV in a high-risk HCG. In the BSLMC population, pCR was achieved in 4 (28.6%) TNBC patients. Of these patients, 4 (100%) carried P/LPV in a high-risk HCG. Of the HR+ patients that had pCR, 4 (80%) had P/LPV in a high-risk HCG. RCB class by hereditary P/LPV and biomarker status will be reported. Conclusions: Rates of P/LPV in high-risk HCG (BRCA 1/2 and PALB2) were higher in the SC population compared to BSLMC, which may be reflective of differences in ethnicities. Across both institutions, pCR rates were highest among patients with high-risk P/LPV, detailed in Table 1. When further divided by biomarker, HR+ patients had relatively high pCR rates. More data in patients with P/LPV in HCG may help us better determine prognosis and optimize treatment strategies. Given the high pCR rate in high-risk HCG, consider germline genetic testing in HR+ BC patients with pCR.

K. M. Veiga¹, P. F. Silva², W. D. Oliveira;³, D. C. Maia⁴, R. M. Rahal⁵, E. P. Silveira-Lacerda⁶; ¹Centro de Genetica Humana – ICB, Universidade Federal de Goiás, Goiânia, BRAZIL, ²Centro de Genética Humana – CEGH – ICB, Universidade Federal de Goiás, Goiânia, BRAZIL, ³Genética, Hospital das CLínicas, Universidade Federal de Goiás, Goiânia, BRAZIL, ⁴Oncologia, Universidade Federal do Ceará, Fortaleza, BRAZIL, ⁵CORA, Clinical Hospital, Hospital das CLínicas, Universidade Federal de Goiás, Goiânia, BRAZIL, ⁶Genética, Universidade Federal de Goiás, Goiânia, BRAZIL.

Cancer remains the second leading cause of death worldwide, with 5-10% of cases attributed to hereditary syndromes. Among these, Li-Fraumeni Syndrome (LFS) is closely associated with pathogenic germline variants in the TP53 gene, a critical tumor suppressor responsible for maintaining genomic stability and regulating cell cycle and apoptosis. In Brazil, the TP53-R337H variant is highly prevalent due to a well-documented founder effect, particularly in the central and southern regions. Recent studies have revealed the co-segregation of the truncating variant XAF1-E134* with TP53-R337H, forming a functional haplotype associated with increased tumor aggressiveness and earlier onset (Jung et al., 2020; Fortes et al., 2021).This study aimed to evaluate the frequency and clinical impact of this haplotype in a cohort of Brazilian individuals with suspected hereditary cancer syndromes, recruited through the Brazilian Unified Health System (SUS). A total of 527 participants (probands and relatives) were referred to the Human Genetics Center at the Federal University of Goiás (UFG), based on NCCN 2025 criteria for Hereditary Breast and Ovarian Cancer Syndrome (HBOC) and LFS. After informed consent, venous blood samples were collected for DNA extraction. Next-generation sequencing (NGS) was used to analyze the coding region of TP53, and genotyping assays were performed to detect the R337H and XAF1-E134*variants in both probands and their family members.Among the 527 participants, 198 (37.56%) were probands and 329 (62.44%) were family members. Of the probands, 7.58% (15/198) carried pathogenic TP53 variants, with 93.33% (14/15) being R337H. Thirteen probands were tested for XAF1-E134*, and 76.92% (10/13) tested positive. Of these, 90% (9/10) had a confirmed cancer diagnosis, with a mean age at diagnosis of 40.7 years. Furthermore, 30% (3/10) presented with multiple primary tumors, including prostate, osteosarcoma, sarcoma, and adenocarcinoma. Metastasis was observed in 10% (1/10). Among family members, 14.29% (47/329) carried TP53 variants. Of the 35 relatives tested for XAF1-E134*, 91.43% (32/35) were positive.These results provide strong evidence for co-inheritance of TP53-R337H and XAF1-E134*, forming a high-risk haplotype associated with more aggressive cancer phenotypes. Notably, in families where probands were negative for TP53-R337H, relatives were also negative for both variants, supporting the hypothesis of joint transmission. The clinical manifestations observed among carriers reinforce the functional synergy between these two variants and their potential contribution to early-onset and multifocal malignancies.This study highlights the relevance of including XAF1-E134* in genetic screening panels, particularly in Brazilian regions affected by the R337H founder mutation. Incorporating haplotype-based analysis into clinical genetics can improve risk stratification, facilitate early diagnosis, and guide personalized prevention and surveillance strategies in high-risk families. Further studies are essential to clarify the molecular mechanisms underlying this haplotype and to define its role in cancer pathogenesis and clinical management.

E. Kang, C. Lee, H. Cho, D. Shin, J. Jung, H. Kim, H. Lee, W. Han, H. Moon; General surgery, Seoul national university hospital, Seoul, KOREA, REPUBLIC OF.

BackgroundContralateral breast cancer (CBC) has traditionally been considered a newly developed, biologically independent tumor rather than a recurrence of the primary breast cancer. However, accumulating evidence suggests that some CBCs share molecular characteristics with the primary tumor, indicating possible biological relatedness. This study aimed to evaluate the biological association between primary and contralateral breast cancers by analyzing their subtype distribution and concordance patterns.MethodsWe analyzed 382 patients who underwent surgery for primary breast cancer at Seoul National University Hospital between 2001 and 2024 and subsequently developed contralateral breast cancer. Synchronous CBC was defined as contralateral breast cancer diagnosed simultaneously or within 6 months of primary invasive breast cancer diagnosis, while metachronous CBC was defined as contralateral breast cancer diagnosed thereafter. Patients with contralateral DCIS, unknown subtype, or initial distant metastasis were excluded. The study included 186 synchronous and 196 metachronous CBC cases. Subtype distributions of CBC were compared with those of sporadic breast cancers using multinomial exact tests. Additionally, subtype concordance between primary and contralateral tumors was evaluated using concordance rates.ResultsWe identified distinct clinical patterns between synchronous and metachronous CBCs. Synchronous CBCs were more frequently associated with hormone receptor (HR) positive and HER2 negative primary tumors, whereas metachronous CBCs were more often associated with primary triple-negative breast cancer (TNBC) and showed stronger associations with BRCA mutation status.The subtype distribution of both metachronous and synchronous CBCs differed significantly from that of sporadic breast cancers (p<0.001), with a relatively higher proportion of TNBC. Subtype concordance between primary and CBC tumors was significantly higher than expected under a model of random subtype distribution among all subtypes (p < 0.001, concordance rate vs. expected rate: 79.7% vs. 63.1%, 32.0% vs. 5.5%, 30.8% vs. 10.5%, and 61.7% vs. 20.9% for HR+/HER2-, HR+/HER2+, HR-/HER2+, and TNBC, respectively)ConclusionsContralateral breast cancer does not arise as a random second primary event but rather reflects subtype-specific characteristics of the primary tumor. These findings indicate that the pathogenesis of CBC might not be an independent one but rather be primary tumor-dependent or host factor-dependent which warrants further investigation.

T. McKamie¹, T. Hudson², L. Brzeskiewicz¹, S. Cummings¹; ¹Oncology Medical Affairs, Myriad Genetics, Inc., Salt Lake City, UT, ²Imaging, Singing River Health System, Ocean Springs, MS.

Background: To evaluate the effectiveness of using Tyrer-Cuzick version 8 (TCv8, herein referred to as TC) alone compared to a combined approach of a family history assessment, germline testing, and TC with a breast cancer (BC) multiple ancestry polygenic risk score (MA-PRS), in identifying patients at high risk for breast and other cancers at imaging centers. Here we compared approaches for identifying high-risk patients to improve personalized screening and management. Methods: In November 2020, Singing River Health System launched a comprehensive breast program using the TC risk model and germline testing with MA-PRS to identify high-risk patients. During mammogram visits, patients completed the TC risk model and were assessed for genetic testing eligibility based on NCCN guidelines using a cancer history questionnaire. Eligible patients received a comprehensive family history intake and education by telephone from a genetic counselor and, if consented, had blood drawn for genetic testing the same day. The testing included a laboratory’s, a laboratory’s multi-gene panel and MA-PRS to determine BC risk and hereditary cancer syndromes. Post-testing, a nurse navigator reviewed the results and counseled patients on appropriate next steps. We evaluated results to identify guideline-recommended management changes and to determine if high-risk patients would not have been identified without germline testing and MA-PRS. Results: From November 2020 through December 2023, a total of 2,434 patients completed TC and if they met the criteria for genetic testing at the time of their mammogram, they had multi-gene panel testing and MA-PRS. Within this cohort, 677 patients had a TC <20% lifetime BC risk, yet were candidates for medical management change based on their genetic testing, MA-PRS, and/or comprehensive family history results. Specifically, 70 patients with TC <20% met the criteria for genetic testing and were found to carry pathogenic variants in genes associated with hereditary cancer syndromes. Additionally, 133 patients with TC 20% due to the incorporation of MA-PRS. Furthermore, 474 patients with TC <20% were appropriate for modification in management strategy based on family history information obtained through patient education intake. Conclusions: A detailed risk assessment and testing program involving TC, a polygenic risk score model, and germline testing at a community imaging center identified more high-risk individuals and suggested alterations in screening and management compared to using TC alone. Without this approach, patients at elevated cancer risk requiring management changes may not have been recognized.

A. Hajirahimkhan¹, E. T. Bartom², S. M. Roy³, A. M. Eremin¹, R. Chen⁴, D. M. Watterson³, S. E. Clare¹, S. A. Khan¹; ¹Surgery, Northwestern University, Chicago, IL, ²Biochemistry and Molecular Genetics, Northwestern University, Chicago, IL, ³Pharmacology, Northwestern University, Chicago, IL, ⁴Preventive Medicine, Northwestern University, Chicago, IL.

L024, a promising candidate for the prevention and interception of triple negative breast cancer and its metastasis to the brainIntroduction: Estrogen receptor negative (ER-neg) breast cancer (BC) disproportionately affects younger women, women of African descent, and socioeconomically disadvantaged populations. Excluding high-risk groups, e.g. BRCA1 carriers, for whom there are surgical risk reduction strategies, no approved preventive agents exist for this subtype. ER-neg BCs, including triple-negative (TNBC) and HER2+ tumors, often metastasize to the brain at early stages, yet remain clinically undetected until neurologic symptoms appear, by which time prognosis is poor. Current therapies, including adjuvant drugs for ER-neg BC, offer limited efficacy against brain metastases due to poor brain penetration. Radio-surgical approaches to treat brain metastases carry significantly debilitating side effects with no significant enhancement in survival. There is an urgent need for safe, non-hormonal risk-reducing medications that can prevent both ER-neg BC and its dissemination to the brain.We have identified L024, a promising first-in-its-class, non-endocrine agent which shows strong efficacy against ER-neg BC. Mechanistic studies demonstrate that L024 acts through inhibition of SREBP1-driven lipogenesis, suppression of the PI3K-AKT axis, and downregulation of NF-κB-mediated inflammation, key pathways in ER-neg BC progression and metastasis. However, its potential to intercept or prevent metastatic dissemination, particularly to the brain, remains unexplored. We hypothesize that L024 suppresses ER-neg tumor growth and hinders early metastatic spread to the brain, representing a novel preventive and interceptive approach.Methods: We evaluated the antiproliferative effects of L024 across five ER-neg BC cell lines: MDA-MB-231 (human TNBC), HCC-1937 and HCC-3153 (BRCA1-mutant TNBC), 4T1 (mouse TNBC with high metastasis potential), and MDA-MB-231-Br (human brain-tropic TNBC) using IncuCyte under single and repeated dosing. We developed xenograft tumors of MDA-MB-231 cells in nude mice and treated the animals with daily doses of 80 mg/kg of L024 when the tumors reached a threshold size. We monitored changes in tumor size over 28 days and conducted RNA sequencing on the tumors upon the completion of the studies. Invasion and migration were assessed using Transwell assays with or without extracellular matrix following sublethal L024 treatment (5 μM, 2.5 μM). In silico ADME studies were conducted using ADC Lab software. Pharmacokinetics (PK) were evaluated in female CD-1 mice dosed with L024 (50 mg/kg). Drug levels in plasma, mammary, and brain tissue were measured and modeled using SAAM II.Results: L024 significantly inhibited malignant cell proliferation at concentrations dose-dependently, with effects sustained up to six days post-treatment. It also significantly reduced tumor growth and epithelial to mesenchymal transition in xenograft tumors, suggesting anticancer and antimetastatic effects. Sublethal doses reduced invasion and migration of 4T1 and MDA-MB-231-Br cells, suggesting further validating anti-metastatic potential. In silico profiling indicated favorable CNS penetration: low molecular weight (1 between 1–8 h post-dosing and clearance by 24 hours, indicating suitability for prevention and therapy. Conclusions: L024 reprograms metabolic and inflammatory pathways essential for ER-neg BC progression and early brain metastasis. CNS pharmacokinetics support further evaluation in immunocompetent models of TNBC with brain metastatic potential.

B. X. Zhang¹, T. Sella², Y. Zheng³, G. J. Kirkner⁴, K. J. Ruddy⁵, S. M. Rosenberg⁶, V. F. Borges⁷, L. Schapira⁸, S. E. Come⁹, J. Peppercorn¹⁰, A. H. Partridge⁴, K. D. Brantley⁴; ¹Department of Medicine, Brigham and Women’s Hospital, Boston, MA, ²Department of Oncology, Sheba Medical Center, Ramat Gan, ISRAEL, ³Department of Biostatistics, Dana-Farber Cancer Institute, Boston, MA, ⁴Department of Medical Oncology, Dana-Farber Cancer Institute, Boston, MA, ⁵Department of Medical Oncology, Mayo Clinic, Rochester, MN, ⁶Division of Epidemiology, Weill Cornell Medicine, New York, NY, ⁷Division of Medical Oncology, University of Colorado Cancer Center, Aurora, CO, ⁸Division of Medical Oncology, Stanford University, Stanford, CA, ⁹Department of Medical Oncology, Beth Israel Deaconness Medical Center, Boston, MA, ¹⁰Department of Medical Oncology, Massachusetts General Hospital, Boston, MA.

Background: Women younger than 40 generally fall outside the routine screening population for breast cancer (BC). We aimed to understand factors influencing diagnosis and treatment delays in these patients, and the impact of delays on disease-free survival (DFS). Methods: We included women aged ≤ 40 years at diagnosis of stage 0-IV BC enrolled in the Young Women’s BC Study from 2006-2016. Sociodemographics, initial BC symptoms, and disease status were collected via annual surveys. Medical record review was used to capture tumor characteristics and recurrence status. We used logistic regression to estimate associations between patient and tumor factors and three delay types: (1) self-delay (≥ 90 days from first symptom to presentation to a provider), (2) diagnosis delay (≥ 90 days from healthcare visit to diagnosis), and (3) treatment delay (≥ 60 days from diagnosis to first treatment). Odds ratios (OR) and 95 % confidence intervals (CI) were estimated in multivariable models after choosing factors via backwards selection (p≤0.05 retained). Hazard ratios (HR) and 95% CIs for the association between delays and DFS, overall and by stage, were estimated via Cox proportional hazard models, adjusted for tumor and treatment factors. Time to event was measured from diagnosis or treatment to recurrence, death, or last follow-up, whichever occurred first. Results: Of 1,297 patients, 1,071 who completed the long-form baseline survey were included. Median time from first symptom to presentation was 14 days (interquartile range (IQR)=2-45), presentation to diagnosis was 14 days (IQR=7-30), and diagnosis to treatment was 41 days (IQR=25-57). Overall, 14% (n=153) had a self-delay and 11% (n=120) had a diagnosis delay. Of 839 patients with self-detected tumors, 18% had a self-delay and 13% had a diagnosis delay. Among those with Stage 1-3 BC (N=928), 22% (n=205) had a treatment delay. In univariable models, income <50K (p=0.01), lower financial comfort (p=0.01), and not living with a partner (p=0.03) were associated with higher odds of self-delay, while pregnancy within 1 year pre-diagnosis (p=0.05) and first-degree family history of any cancer (p=0.04) or BC (p=0.03) were associated with lower odds. Self-delays were more likely for women diagnosed with stage 4 BC (p=0.004) or with HER2+ tumors (p=0.03) and less likely for those with triple negative (TN) BC (p<0.001). In the multivariable model, family history of BC (OR=0.40, 95% CI=0.17-0.80), income <50K (OR=1.97, 95% CI=1.22-3.13) and TN subtype (OR v. luminal A=0.30, 95% CI=0.13-0.60) remained associated. Odds of diagnosis delay were higher among women who were younger (p=0.01) or with Stage 4 BC (p=0.004), and lower for those with luminal B (p=0.04), TN (p=0.01), or high grade tumors (p=0.03). In the multivariable model, younger age (OR per 1-year=0.94, 95% CI=0.89-0.99), and subtype (luminal B v. A OR=0.60, 95% CI=0.38-0.95; TN v. luminal A OR=0.35, 95% CI=0.16-0.70) remained associated. Multivariable models for self-and diagnosis delays remained similar when restricted to self-detected patients.Treatment delays were less likely among unemployed individuals (p=0.001), and women with a positive germline BC risk variant (p=0.002), luminal B (p=0.005) or TN (p<0.001) subtypes, and high-grade tumors (p=0.03). Only tumor factors remained significant in the multivariable model, including subtype (luminal B v. A OR=0.59, 95% CI=0.41-0.86, TN v. luminal A OR=0.25, 95% CI=0.13-0.43), and stage (stage 3 v. 1 OR=0.37, 95% CI=0.20-0.65). Overall and by stage at diagnosis, delays were not statistically significantly associated with DFS before or after adjusting for tumor and treatment factors. Conclusions: Delays in breast cancer care are influenced by multiple psychosocial and clinical factors, and disparities exist by socioeconomic status. It is critical to decrease self- and diagnosis delays to prevent stage 4 breast cancer.

K. Meza¹, B. Arun², A. Ruiz Cuevas², K. Lee³, C. Amaram⁴, S. Shanker⁴, A. Vara⁴, D. Kizub²; ¹Department of Medicine, Baylor College of Medicine, Houston, TX, ²Department of Breast Medical Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX, ³Department of Breast Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX, ⁴Department of Clinical Cancer Genetics, The University of Texas MD Anderson Cancer Center, Houston, TX.

Introduction: Breast cancer (BC) remains the leading cause of cancer-related death among Hispanic women in Texas. While 5-10% of BC cases are linked to hereditary cancer syndromes, the uptake of genetic testing (GT) among eligible individuals remains low, especially in underserved populations. In our prior study using our validated one-page Cancer Genetic Risk Assessment (CGRA) administered at mammography visits, only 10% of 150 eligible underserved women completed GT. To address this gap, we implemented a 12-month, multi-level program in Southeast Texas that included: (1) culturally tailored genetic education, (2) on-site CGRA screening and GT, (3) financial navigation, and (4) telegenetic counseling. Here, we present barriers and facilitators to program implementation. Methods: This 12-month prospective program was conducted in Harris County (2023-2024) within mobile and safety-net clinics of the VALET mammography program. The primary outcome was GT completion. Women undergoing screening completed our validated one-page CGRA. If eligible, saliva-based GT kits were provided onsite or mailed. A bilingual coordinator assisted with consent and financial paperwork. Participants with pathogenic variants (PVs) or variants of uncertain significance (VUS) received telegenetic counseling; others were informed of negative results. From 2022-2023, 870 women were reached; 590 (87%) completed CGRA, 537 (91.0%) identified as Hispanic. Ninety-nine (15.8%) met criteria for GT, and 54.8% completed testing. completed it. 35 (64.8%) tested negative, 14 (25.9%) had a variant of unknown significance, and 5 (9.5%) had a pathogenic mutation in MUTYH, NF1, CHEK2, MSH3 genes. Program implementation was evaluated through semi-structured interviews guided by the RE-AIM (Reach, Effectiveness, Adoption, Implementation, Maintenance) framework. Interviews were transcribed, double-coded, and thematically analyzed. Results: In-depth interviews were conducted with 40 participants (34 patients, 6 staff). Among patients, program participation was motivated by a family history of cancer (73.7%), desire for greater knowledge (28.9%), commitment to one’s health (28.9%) and psychological relief (28.9%). Key facilitators of GT included the ease of testing (71.1%), prior GT knowledge (15.8%), and prior conversation with PCP (13.2%). Barriers included fear and uncertainty after a positive result (68.4%) and financial constraints (34.2%). Patients recommended expanding telehealth (10.5%), reducing costs (7.9%), increasing education (36.8%), and community outreach (21.1%). After participation, patients reported improved screening adherence (68.4%), encouraging family GT (31.6%), and lifestyle changes (18.4%). Notably, 18.4% said their PCP never discussed GT, even with strong family history. Staff were motivated by their desire to improve community health (100%). Facilitators included integration into clinic workflow and provider education (67%). Staff recommended more education (100%), cost reductions (50%), and provider follow-up (33.3%). Barriers included staff shortages (100%) and time constraints (75%). Conclusions: The qualitative data revealed that emotional factors, prior awareness, and ease of access significantly influenced participation. Patient narratives highlighted the importance of culturally responsive education and logistical support in promoting uptake. Staff insights underscored the value of integrating services into existing workflows and the need for sustained provider engagement. These findings offer actionable strategies to improve the design, delivery, and scalability of community-based genetic risk assessment programs.

A. Naeim¹, A. Stanton², K. Sepucha³, S. Vangala⁴, A. Petruse⁵, L. Madlensky⁶, A. S. Fiscalini⁷, L. J. Esserman⁷, M. Eklund⁸, K. Ross⁹, D. Goodman-Gruen¹⁰, J. Tice¹¹, E. Ziv¹², Y. Shieh¹³, L. van ‘t Veer¹⁴, A. Kaster¹⁵, A. M. Blanco⁹, O. I. Olopade¹⁶, M. Scheuner¹⁷, B. Tong¹⁸, L. Sabacan¹⁹, J. Atamer⁷, H. Harvey²⁰, R. A. Hiatt²¹, A. D. Borowsky²², J. Esserman²³, H. Anton-Culver²⁴, A. LaCroix²⁵, H. L. Park²⁶, B. A. Parker²⁷, R. Lancaster²⁸, WISDOM Investigators and Advocate Partners, N. Wenger²⁹; ¹Division of Hematology-Oncology, Department of Medicine, David Geffen School of Medicine at UCLA, Los Angeles, CA, ²Department of Psychology,, UCLA, Los Angeles, CA, ³Division of General Internal Medicine, Mass General, Boston, MA, ⁴Division of General Internal Medicine and Health Services Research, UCLA, Los Angeles, CA, ⁵Clinical and Translational Science Institute Office of Clinical Research, University of California Los Angeles, Los Angeles, CA, ⁶Department of Medicine,, University of California, San Diego,, San Diego, CA, ⁷Department of Surgery, UCSF, San Francisco, CA, ⁸Department of Medical Epidemiology and Biostatistics,, Karolinska Institutet, Stockholm, SWEDEN, ⁹Cancer Genetics and Prevention Program, University of California San Francisco, San Francisco, CA, ¹⁰Department of Epidemiology and Biostatistics, UC Irvine, Irvine, CA, ¹¹General Internal Medicine, School of Medicine, University of California San Francisco, San Francisco, CA, ¹²Department of General and Internal Medicine, University of California San Francisco, San Francisco, CA, ¹³Department of Population Health Sciences,, Weill Cornell Medicine, New York, NY, ¹⁴Lab Medicine, UCSF, San Francisco, CA, ¹⁵Edith Sanford Breast Center, Sanford Health, Fargo, ND, ¹⁶Center for Clinical Cancer Genetics and Global Health, Department of Medicine,, University of Chicago, Chicago, IL, ¹⁷Department of Medicine, University of California San Francisco and San Francisco VA Health Care System, San Francisco, CA, ¹⁸Cancer Genetics and Prevention Program, UCSF, San Francisco, CA, ¹⁹Department of Surgery, University of California San Francisco, San Francisco, CA, ²⁰Edith Sanford Breast Center, Sanford Health, Sioux Falls, SD, ²¹UCSF Helen Diller Family Comprehensive Cancer Center, Department of Epidemiology & Biostatistics,, UCSF, San Francisco, CA, ²²Department of Pathology, School of Medicine, UC Davis, Sacramento, CA, ²³OB-GYN, Diagnostic Center of Miami, Miami, FL, ²⁴Department of Medicine, Genetic Epidemiology Research Institute, UCI, Irvine, CA, ²⁵Department of Family and Preventive Medicine, Division of Epidemiology, UCSD, San Diego, CA, ²⁶Department of Pathology and Laboratory Medicine, Department of Epidemiology, UC Irvine School of Medicine, Irvine, CA, ²⁷Moores Cancer Center, University of California San Diego, San Diego, CA, ²⁸Department of Surgery, University of Alabama at Birmingham, Birmingham, AL, ²⁹Division of General Internal Medicine and Health Services Research,, University of California Los Angeles, Los Angeles, CA.

BackgroundThe Women Informed to Screen Depending on Measures of risk (WISDOM) Study 1.0 (2016-2023; NCT02620852) is testing a novel, personalized approach to breast cancer screening compared to annual mammography. WISDOM, a large-scale screening trial, enrolled women ages 40-74 with no personal history of breast cancer using a randomized, preference tolerant trial design, reflecting the heterogeneity of the U.S. population. The goal of this analysis is to evaluate whether women discussed WISDOM screening recommendations with their primary care provider, factors associated with discussion, and whether primary care providers (PCPS) agreed or disagreed with the screening recommendation.MethodsWe analyzed participant-reported responses among the first 15,000 WISDOM participants who completed patient-centered outcomes surveys at baseline, at the time of they received their screening recommendation, and at first annual follow-up. The annual survey asked about discussing the trial screening assignment with their PCPS and the reaction of their PCPS to the screening assignment (Likert 5 scale from Strongly Disagreed to Strongly Agreed). These items were compared between the randomized arms (annual versus personal risk-based) and across the study breast cancer screening assignments (6 months with MRI, Yearly with Risk-Reduction, Yearly, Two Years, Delayed Screening).ResultsParticipants in the risk-based screening arm were significantly more likely to report discussing the screening recommendation with their PCP (35.6%, 2708/7610) compared to those in the annual arm (19.0%, 1308/6888; p < 0.001). PCPS agreement with the screening recommendation also varied by arm. In the annual group, 75.7% of participants reported their PCPS agreed or strongly agreed with the recommendation and only 2.5% disagreed, In the risk-based group, 58.6% agreed and 16% disagreed. Among participants in the risk-based arm, discussion rates with PCPS and level of PCPS agreement differed by screening recommendation. The highest rate of discussion occurred in those assigned to 6 months with MRI (71.1%), followed by Yearly with risk-reduction (50.3%), Two Years (34.6%), Yearly (29.6%) and Delayed (32.4%). Disagreement was highest in those identified at lowest risk, Delayed Screening (24.2%) and those with the Two Year (16.1%) and disagreement was lowest with the higher screening recommendations wherein Yearly (1%) and 6 Month with MRI (5.6%) disagreed. Conclusion/DiscussionWomen often discuss breast cancer screening recommendations with their primary care providers, particularly risk-based recommendations. Data from WISDOM 1.0 show PCPS had stronger agreement more frequent screening recommendations and disagreement with less frequent intervals. These findings suggest that education to increase primary care physician familiarity with personalized screening approaches may enhance the implementation and acceptance of risk-based strategies in clinical practice.

L. A. Logie¹, A. Serrano¹, C. Campos Galván², A. Hurtado de Mendoza³, M. De Jesus⁴, A. Jimenez¹; ¹N/A, Nueva Vida, Alexandria, VA, ²Community Outreach and Engagement, George Washington University, District of Columbia, DC, ³Cancer Prevention and Control at Lombardi Comprehensive Cancer Center, Georgetown University, District of Columbia, DC, ⁴SIS | Environment, Development & Health, American University, District of Columbia, DC.

Background: An important step toward understanding the role of BRCA and other cancer genes in Latinas is to increase the number of individuals who receive genetic testing and counseling (GTC). Latinas have the 2nd highest prevalence of BRCA gene mutations after women of Ashkenazi Jewish ancestry and the lowest level of awareness about GTC for inherited cancer risk than all other racial/ethnic groups, and are 4-5 times less likely to get tested compared to Whites. Recognizing a lack of knowledge among Latina breast cancer survivors (BCS) and their daughters, Nueva Vida, developed an intergenerational intervention as an effective method to leverage cultural values and combat mistrust, laying the groundwork for learning and preventive action. The aim of De Madres a Hijas is to increase Latina mothers/daughters’ knowledge about Hereditary Breast Cancer (HBC) risk, enhance their talking about familial HBC risk with relatives, and develop an “action plan” to make educated decisions on their own health.Methods: We used an explanatory pre-post mixed methods design to assess how the intervention contributes to an increase in Latina mothers/daughters’ knowledge about HBC risk and promotes intent to participate in genetic counseling and testing. We used descriptive statistics to characterize the study sample of Latina BCS and their daughters (n=48). Participants were administered a 13-item pre-post survey to assess knowledge (i.e., having a genetic mutation does not mean you will get cancer T/F). An expanded Consolidated Framework for Implementation Research (CFIR) comprised of 15 domains (e.g., Difficulty Understanding Genetics, Actions to Learn about Family History) guided the qualitative data analysis. Four Coders utilized template analysis to analyze the interviews. Results: This first-of-its-kind, community-based initiative leverages a one-day, workshop to educate BCS and their daughters about HBC, their risks, preventive behaviors and GTC. These workshops show 91% of daughters improved understanding of genetic cancer, and 83% reported behavior changes toward healthier daily activities and prevention-focused habits. As stated by a participant, “We were told to be aware if we feel pain… I did my self-exam and I didn’t find anything, but I felt pain right next to my armpit and I felt more confident to make a follow up appointment and they found it (a lump).” Lastly, to extend the reach of the program, we have developed the Toolkit—a manual for other Latino-serving organizations hoping to facilitate these life-saving conversations in their communities. Discussion: Culturally tailored interventions are vital for ensuring that GTC services are accessible, effective, and equitable for all individuals, regardless of their ethnicity. Education plays a pivotal role in closing the gaps in GTC. Creating the opportunity for mothers and daughters to share this experience, provides not only knowledge building about GTC, but also a special bonding by listening and understanding each other’s lived experience. Second generation Latinas show better literacy levels than immigrant Latinas who deal even with their own language reading and writing barriers. For the Latina community, it’s crucial that we not only offer HBC education but do so in a way that resonates with their unique cultural perspectives.

A. L. Cicilini¹, J. C. Casali-da-Rocha², F. B. Makdissi³, S. M. Sanches⁴, S. M. Volc², R. Cagnacci-Neto³, F. P. Leite³, S. M. Castro³, M. Sonagli³, D. G. Rodrigues de Souza¹, P. V. de-Cerqueira¹, V. P. de Souza¹, L. R. Soares¹, V. C. Cordeiro de Lima⁴, A. A. Balieiro⁴, S. M. Caputo⁵, E. Santana dos Santos⁴; ¹Clinical Oncology, A.C. Camargo Cancer Center, São Paulo, BRAZIL, ²Oncogenetic Reference Center, A.C. Camargo Cancer Center, São Paulo, BRAZIL, ³Mastology- Breast Reference Center, A.C. Camargo Cancer Center, São Paulo, BRAZIL, ⁴Clinical Oncology – Breast Reference Center, A.C. Camargo Cancer Center, São Paulo, BRAZIL, ⁵Genetic Service, Institut Curie, Paris, FRANCE.

Background: Hereditary breast cancer represents 5-10% of all cases. While BRCA1/2 pathogenic variants (VP) are well-established high-risk factors, clinical implications of moderate-penetrance gene VP (PALB2, CHEK2, ATM, RAD51C, RAD51D, BARD1, BRIP1, FANCM) remains less explored. Global prevalence estimates for moderate-penetrance genes such as PALB2, CHEK2 and ATM range between 1-3%, underscoring the clinical relevance of tailored management. Definition of optimal surgical strategies and surveillance, particularly based on the prevalence of second primary tumors, are needed. We aimed to evaluate surgical management strategies and the prevalence of second primary tumors in patients with hereditary breast cancer associated with moderate-penetrance gene mutations, to inform clinical decision-making and surveillance recommendations. Methods: A retrospective cohort study was conducted at A.C. Camargo Cancer Center including women diagnosed with BC and VP in moderate-penetrance genes between 2016-2025. Descriptive statistics summarized clinical-pathological data. Cumulative incidence of second primary tumors and breast cancer (BC) recurrence was calculated. Five-year overall survival (OS) was estimated by the Kaplan-Meier method. Moderate-risk classification was defined according to current NCCN guidelines and a comprehensive hereditary cancer NGS panel. Results: Among 1,429 patients tested, 120 (8%) had VP in moderate-penetrance genes: 28 PALB2, 28 CHEK2, 28 ATM, 15 RAD51C, 4 RAD51D, 6 FANCM, 8 BRIP1, 3 BARD1. They were all female. Family history of BC was present in 73%. Most were diagnosed at stage I (54%). Predominant histology was invasive ductal carcinoma (79%), with SBR grade 3 in 63%. Subtypes included: Luminal B (24%), HER2-low (23%), HER2-overexpressing Luminal B (19%), TNBC (18%), Luminal A (10%) and HER2 -overexpressing (4%). Contralateral synchronous BC occurred in 3 patients (2%). Unilateral adenomastectomy was the most common surgical approach (51%), followed by segmental resection (45%). Four percent of patients underwent bilateral adenomastectomy. Sentinel lymph node biopsy was performed in 67%; axillary dissection in 33%. Over a median follow-up of 5 years, second primary tumors were observed in 16 patients (13.2%): 7 papillary thyroid carcinomas (5 CHEK2, 2 ATM), 6 ovarian cancers (1 ATM, 5 RAD51C), and single cases of melanoma, endometrial, lung, and colorectal cancer. BC recurrence occurred in 8 patients (6.6%). Prophylactic salpingo-oophorectomy was performed in 17%. The 5-year OS was 96.8%. Conclusion: Unilateral mastectomy with sentinel lymph node biopsy was the most frequent surgical strategy for moderate penetrance gene VP carriers. Despite excellent OS, patients with VP in moderate-penetrance genes—especially CHEK2, ATM, and RAD51C —faced a significant cumulative incidence of second primary tumors (13.2%), with thyroid and ovarian cancers being the most prevalent. These findings highlight the need for tailored surveillance strategies beyond breast-focused care in this genetically distinct population. This highlights the need for a multidisciplinary approach involving breast surgeons, genetic counselors, gynecologists, and endocrinologists to optimize long-term surveillance beyond breast-focused care. Breast surgeons should also consider bilateral prophylactic mastectomy in highly selected moderate-risk cases to further reduce long-term risk.

J. Moya¹, K. Ross², D. Goodman-Gruen³, L. Madlensky⁴, A. Blanco², B. Tong², H. Harvey⁵, A. S. Fiscalini¹, A. Kaster⁶, M. Scheuner⁷, L. Sabacan¹, J. Atamer¹, L. van ‘t Veer⁸, O. I. Olopade⁹, J. Tice¹⁰, E. Ziv¹¹, S. Kapoor¹, Y. Shieh¹², A. D. Borowsky¹³, A. Naeim¹⁴, H. L. Park¹⁵, A. LaCroix¹⁶, WISDOM Investigators and Advocate Partners, L. J. Esserman¹; ¹Department of Surgery, UCSF, San Francisco, CA, ²Cancer Genetics and Prevention Program, UCSF, San Francisco, CA, ³Department of Epidemiology and Biostatistics,, UC Irvine, Irvine, CA, ⁴Department of Medicine, UCSD, San Diego, CA, ⁵Edith Sanford Breast Center, Sanford Health, Sioux Falls, CA, ⁶Edith Sanford Breast Center, Sanford Health, Fargo, ND, ⁷Department of Medicine, University of California, San Francisco and San Francisco VA Health Care Sys, UCSF, San Francisco, CA, ⁸Department of Lab Medicine, UCSF, San Francisco, CA, ⁹Center for Clinical Cancer Genetics and Global Health, Department of Medicine, University of Chicago, Chicago, IL, ¹⁰General Internal Medicine, School of Medicine, UCSF, San Francisco, CA, ¹¹Department of General and Internal Medicine, UCSF, San Francisco, CA, ¹²Department of Surgery, Weill Cornell Medicine, New York, NY, ¹³Department of Pathology, School of Medicine, UC Davis, Sacramento, CA, ¹⁴Division of Hematology-Oncology, Department of Medicine, David Geffen School of Medicine, UCLA, Los Angeles, CA, ¹⁵Department of Pathology and Laboratory Medicine, Department of Epidemiology,, UCI, Irvine, CA, ¹⁶Department of Family and Preventive Medicine, Division of Epidemiology, UCSD, San Diego, CA.

TITLE: Completion Rates of Risk Counseling Consultations for Participants Identified as High Risk for Breast Cancer within the WISDOM 1.0 Pragmatic Screening Trial BACKGROUND: The WISDOM study is innovative in that it is the first, prospective large-scale trial of assessing risk-based mammographic screening recommendations versus annual screening. Women identified as high-risk were invited to complete a Breast Health Specialist consultation, which included use of an automated educational tool, Breast Health Decisions (BHD) tool, identifying their specific breast cancer risk factors, and how to promote uptake of risk-reducing options. During initial rollout, the preliminary data obtained demonstrated that racial and ethnic minority populations were underrepresented in the study demographics, thus, we created a program to incorporate a more diverse participant pool. We sought to understand whether completion rates of consultations with trained breast health specialist (BHS) varied among race/ethnicity, geography, and age. METHODS: We analyzed data from WISDOM 1.0 participants that completed the BHD tool from 2015-2023 that were considered being in the highest 2.5% breast cancer risk based on the Breast Cancer Surveillance Consortium. Of those participants, we compared completion of BHS consultations by comparing them by age (40-49, 50-59, 60-69, 70-79), race/ethnicity (White, Asian, Black, Hispanic, Other/Mixed Race), and geography (Northeast, Midwest, South, West). We compared each category by proportions of total completed versus not completed and obtained percentages of each. RESULTS: There were a total of 2883 participants identified as high-risk and offered a BHS consultation. Most of all participants were White (84%) and from the West region of the United States (66%). Completion rates were similar across age groups (52-71%), but lower for women ages 70-79 (52%). Within the race/ethnicity category, the proportions that completed consultations were: Other/Mixed Race 105/137 (77%); White 1666/2434 (68%); Hispanic 76/122 (62%); Asian 42/78 (54%); and Black 46/112 (41%). Lastly, the geographic region with the highest completion rate was the Midwest at 521/623 (84%) and the lowest in the South at 157/306 (51%). CONCLUSION: WISDOM identified women at high risk and provided virtual clinical consultations with breast health specialists to review their personalized risk and discuss breast cancer prevention strategies. There were notable differences in completion of clinician consultations rates seen across categories of age, geography, and race/ethnicity. Further assessment of scheduling availability, structural barriers, participant preferences, clinician contact methods, and other contributing factors is required to understand why completion rates differed so much across demographic groups. Based on these results and participant feedback, the WISDOM 2.0 study has changed consultation workflows, improved clinician availability and scheduling methods. We also plan to increase community engagement, tailor outreach methods based on community member feedback, and offer more options for participants to receive the personalized risk information than were offered in WISDOM 1.0. As a pragmatic and iteratively improving study, WISDOM 2.0 has been funded to create and implement these changes with the goal of providing a more equitable method of providing specialty consultations to high-risk participants.

E. Irajizad¹, S. Hanash², J. Fahrmann², A. Brewster²; ¹Biostatistics, MD Anderson Cancer Center, Houston, TX, ²Clinical Cancer Prevention, MD Anderson Cancer Center, Houston, TX.

There remains a need to establish tools to better tailor screening for women at elevated risk of breast cancer (BrCa). Alterations in metabolism are inherently intertwined with breast cancer development and progression. In the current study, we performed multi-assay metabolomic analyses on plasma samples collected from a total of 438 breast cancer cases and 311 female controls from two independent cohorts of high-risk women. Samples from one cohort were used for discovery and samples from the second cohort were used for initial pre-validation. Machine learning was applied to establish a 12-lipid marker panel and tertile-based thresholds for low-, intermediate-, and high-risk of harboring breast cancer subsequently established. The lipid panel and corresponding risk thresholds were further validated in an independent nested case-control cohort of women at elevated risk, using plasma samples prospectively collected from 81 women who developed invasive breast cancer and 191 age-matched female controls. The lipid panel achieved ORs [per unit increase] of 2.31, 2.37, 1.48, in the respective Discovery, Pre-Validation, and Validation Sets. In the combined datasets, the lipid panel yielded an overall OR of 2.02 and AUC of 0.67 for BrCa. Compared to women in the low-risk strata, women in the high-risk group based on the lipid panel were 3.48-times more likely to have BrCa. The lipid panel provides a potential tool for stratifying women at highest risk of developing or harboring breast cancer who would benefit from more intensive screening and interception strategies from those at lowest risk who may be spared overdiagnosis and overtreatment.

S. J. Friedman¹, D. B. Rose², J. D. Rogers¹, K. N. Owens³, P. L. Welcsh³, M. Dean⁴, R. H. Pugh Yi⁵; ¹Research and Education, Facing Our Risk of Cancer Empowered (FORCE), Tampa, FL, ²Research, Facing Our Risk of Cancer Empowered (FORCE), Tampa, FL, ³Education, Facing Our Risk of Cancer Empowered (FORCE), Tampa, FL, ⁴Department of Communication, University of South Florida, Tampa, FL, ⁵Program Evaluation, Akeso Consulting, LLC, Vienna, VA.

Background: Only 7% of adult cancer patients enroll in treatment-related clinical trials, with even lower rates for genetics and quality-of-life studies.¹ People with germline mutations are interested in participating in hereditary cancer research, but few are being offered the opportunity by their healthcare team. In a FORCE survey, only 14% of people at high risk for breast cancer reported hearing about clinical research opportunities from their healthcare providers, while 75% expressed interest.² Dedicated efforts are needed to ensure that the hereditary cancer community is aware of clinical research opportunities. In 2020 FORCE surveys, 46% of oncology nurses and 55% of certified genetic counselors identified research jargon and health literacy as major barriers to clinical trial referral.^3,4 FORCE and other members of the “Consistent Testing Terminology Working Group” found large gaps in patient understanding of commonly-used cancer research terms.⁵ In focus groups, FORCE asked women with or at high risk for hereditary breast cancer to review a cancer study listing on ClinicalTrials.gov. All agreed that the study description was difficult to understand due to complexity and jargon and that Clinicaltrials.gov required a lot of practice to use.⁶ Methods: In 2014, FORCE built a tool to help hereditary cancer patients find, understand, and enroll in relevant studies. The tool consists of a custom database with plain-language summaries of research studies relevant for people with inherited cancer risk. Users can search for studies listed in the FORCE database as well as on ClinicalTrials.gov by study type, cancer type, gene or biomarker, study location and key word. FORCE evaluated the tool through focus groups and a user survey. Focus group participants were asked to compare the ClinicalTrials.gov listing with the FORCE plain-language version of the same study. Results: Focus group participants all agreed that the FORCE listing was clearer and easier to understand than ClinicalTrials.gov.⁶ Key user survey results are included in Table 1. Of the respondents, 79% found the tool easy to use, 84% would use it again, 79% would refer others to the tool. Of respondents, 36% enrolled, and 43% planned to enroll in a research study. Conclusions: People with inherited cancer risk are highly motivated to participate in clinical research—but most are never informed about studies by their healthcare team. ClinicalTrials.gov is not patient friendly, and hereditary cancer patients find the site difficult to use and understand. FORCE’s “Search and Enroll Tool” bridges this gap by offering accessible, accurate, and easy-to-understand listings of studies enrolling hereditary cancer patients. Results from our focus groups and survey demonstrate that the tool meets a critical need. Participants reported high user satisfaction, comprehension, and intention to utilize the tool to enroll in future studies.

C. Pisano¹, R. Abou Zeidane¹, F. Hussain², W. Dong³, T. Lal⁴, L. Vu⁵, N. Mehta⁶, C. Speers⁷, S. M. Koroukian⁸, J. Rose⁹; ¹Radiation Oncology, Case Western Reserve University/University Hospitals Cleveland Medical Center, Cleveland, OH, ²Department of Population and Quantitative Health Sciences, Case Western Reserve University School of Medicine, Cleveland, OH, ³Department of Population and Quantitative Health Sciences , Case Western Reserve University, School of Medicine, Cleveland, OH, ⁴Surgery, Case Western Reserve University/University Hospitals Cleveland Medical Center, Cleveland, OH, ⁵Population Cancer Analytics Shared Resource, Case Western Reserve University, School of Medicine, Cleveland, OH, ⁶Preventive Medicine, Case Western Reserve University/University Hospitals Cleveland Medical Center, Cleveland, OH, ⁷Radiation Oncology, University of Alabama at Birmingham, Birmingham, AL, ⁸Department of Population and Quantitative Health Sciences, Case Western Reserve University, School of Medicine, Cleveland, OH, ⁹Population Cancer Analytics Shared Resource, Case Western Reserve University/University Hospitals Cleveland Medical Center, Cleveland, OH.

Introduction: Screening mammography has significantly reduced breast cancer mortality by enabling earlier diagnosis. Triple negative breast cancer (TNBC) is an aggressive subtype disproportionately affecting Black women. Given its aggressive nature and tendency to occur in younger women, understanding the role of screening in early detection of TNBC is critical. Accordingly, we sought to determine if common traits characterized communities where, despite high screening uptake, the proportion of patients presenting with late-stage TNBC (defined as presenting with regional or distant disease) remained elevated. By understanding characteristics of communities that appear to benefit less from screening, we hope to identify opportunities for intervention around improving the earlier diagnosis of TNBC. Methods: We identified women diagnosed with TNBC from 2010 to 2021 using SEER 22-Registry data (excluding Alaska). We calculated the proportion “late stage” in each county of the SEER regions. To ensure adequate sample sizes, we used the Max-P regionalization method to combine adjacent similar sociodemographic counties (determined by Area Deprivation Index) with ≤10 late-stage cases, creating composite counties (CC) as our unit of analysis. For each CC, we estimated the proportion of women up to date on mammography using CDC PLACES. We estimated community-level characteristics using the 2014-2018 American Community Survey 5-year data and CDC PLACES data including smoking, binge drinking, education, obesity. We used Classification and Regression Tree (CART) analysis to identify combinations of community characteristics associated with having a high (above median) proportion of late-stage TNBC despite having high (above median) mammography uptake. Results: The CART analysis delineated four distinct community profiles with varying rates of the outcome of interest. Communities with high proportions of Black residents (>25.8%) showed the highest rate of women with high screening uptake yet high late-stage TNBC at 63.2%. In contrast, communities with lower rates of binge drinking and lower Black population percentages showed the lowest rates (7.1%) of the same outcome. Conclusions: The study reveals that high screening uptake alone does not uniformly reduce late-stage TNBC, particularly in communities with higher percentages of Black residents or those with high-risk health behaviors (e.g. binge drinking). Future studies should assess the role that both system factors and racial differences in tumor biology play in this finding.

A. Maheswaran¹, C. Mauntel-Medici¹, J. Green¹, R. Hunt¹, N. Awati², S. Hay¹, T. Schmidlen³, E. Simmons⁴, M. Snir⁵, J. Lin¹; ¹Emergency Medicine, University of Illinois at Chicago- UI Health, Chicago, IL, ²Internal Medicine, University of Illinois at Chicago- UI Health, Chicago, IL, ³Clinical Development, Nest Genomics, Henderson, NV, ⁴Product Development, Nest Genomics, Henderson, NV, ⁵CEO, Nest Genomics, Henderson, NV.

Objectives: Breast cancer is the most common cancer in women, and those diagnosed with breast cancer early have a 93% higher 5-year survival rate compared to those diagnosed at a later stage. Presently, biennial mammogram screenings are recommended for women 40 yrs or older. In 2022, 67.8% of U.S. women aged 40 or older reported having had a mammogram in the past 2 years. However, mammography rates were lowest among women with no insurance or access to primary care. Emergency departments (EDs), which serve as a safety net for populations with poor access to care, offer a unique opportunity to reach these underserved groups with preventive health services. An urban ED in Chicago piloted an Electronic Health Record (EHR)-driven breast cancer screening initiative targeting females aged 40 years or older to improve access to preventive health services for underserved populations. Through a partnership with Nest Genomics, the program implemented a customized digital navigator, built on the Nest platform, to facilitate rapid breast symptom screening and hereditary cancer risk assessment (HCRA). The initiative used a multi-modal outreach strategy to engage with patients in-person, by phone, and via text and email. Two patient navigators contacted patients identified by the EHR algorithm and offered breast cancer education, the opportunity to complete the Nest symptom screener and HCRA, and linkage to screening mammography. Referrals for breast Magnetic Resonance Imaging (MRI), genetic testing, and/or chemoprevention were provided to qualifying patients. This retrospective quality improvement study evaluated outcomes from this initiative between 4/1/2024 and 6/30/2025. Methods: Descriptive statistics were used to summarize patient characteristics and the primary outcomes, including completion of risk assessment and linkage to mammogram. Results: Between 4/1/2024 and 06/30/2025, 303 patients aged 40 or above (mean age = 51) were reached (45% via phone outreach, 37% via text and email, and 18% via in-person outreach) through the ED Breast Cancer Screening initiative. Of those reached, the majority were Non-Latinx Black (n= 199, 66%) and insured through Medicaid (n= 136, 45%). Screening mammogram orders were placed for 65 patients, 56 screening mammograms were scheduled, and 26 screening mammograms were performed. For half (n=13, 50%) of the 26 patients completing a screening mammogram, it was their first ever mammogram. Among the 303 patients reached, 63% completed the HCRA (n= 191), 73 patients met national guidelines-based criteria for genetic testing, 13 met criteria for breast chemoprevention, and 11 met criteria for supplemental screening with Breast MRI. During the study period, 11 of 73 patients who met genetic testing criteria were scheduled for a genetic counseling appointment and no patients had yet attended the appointment. Two of the 13 patients eligible for chemoprevention were scheduled with the high-risk clinic, and 1 had attended and received a prescription. Four of the 11 patients eligible for supplemental breast MRI were referred to their PCP and 1 patient had a breast MRI ordered. Conclusions: Our ED initiative identified an underserved patient population that had not yet received recommended breast cancer screening in other venues. EHR-driven screening initiatives that leverage multi-modal outreach strategies and digital patient navigation can improve access to preventive health screenings for vulnerable patients seen in EDs. Further work is needed to address challenges in ensuring timely and effective linkage to recommended follow-up services (e.g. mammography, genetic counseling, high-risk clinic).

A. Saleem¹, A. Sanchez², E. Suskin², M. Lalla³, A. Moadel-Robblee⁴, D. Makower³, S. Klugman²; ¹Medicine, Albert Einstein College of Medicine, Bronx, NY, ²Department of Obstetrics and Gynecology and Women’s Health, Division of Reproductive and Medical Gen, Montefiore Medical Center, Bronx, NY, ³Montefiore Einstein Comprehensive Cancer Center, Montefiore Medical Center, Bronx, NY, ⁴Medicine, Albert Einstein College of Medicine / Montefiore Einstein Comprehensive Cancer Center, Bronx, NY.

Introduction Current national guidelines recommend consideration of genetic testing for all newly diagnosed breast cancer (BC) patients (pts). However, hereditary cancer predisposition genetic testing has been historically underutilized for many reasons, including limitations of prior National Comprehensive Cancer Network (NCCN) guidelines, insurance/financial barriers, and emotional burden of diagnosis and testing. Thus, many BC survivors did not receive current standard of care genetic testing at diagnosis. Benefits of germline genetic testing and identification of pathogenic variants (PV) in late BC survivorship include eligibility for enhanced screening for other malignancies and potential cascade testing of family members. We conducted a pilot study evaluating interest in and uptake of genetic testing among BC survivors. Methods Charts of pts seen at Montefiore Einstein Comprehensive Cancer Center BC survivorship program from January 1, 2023 to December 31, 2023 (n=551) were reviewed. All pts were at least 5 years from diagnosis and completed BC treatment. Pts who had not previously received multi-gene panel testing were contacted by telephone to determine interest in genetic counseling and testing. The benefits of genetic counseling were formally discussed with pts by either the clinician caring for them at their annual survivorship visit or by person distributing survey. Results 312 (57%) pts seen in survivorship had not previously completed genetic testing. 34 (6%) had been previously referred for counseling, but did not complete testing due to financial barriers, personal reasons, or guideline exclusion. Of pts who did not undergo testing, 95 (30%) met 2024 NCCN criteria for testing: 47 by age at diagnosis, 23 by diagnosis of triple negative BC, 12 by family history (FH) of ovarian cancer, 6 by FH of pancreatic cancer, and 7 by FH of BC and/or prostate cancer. 130 patients who had previously not completed testing were called for telephone survey, prioritizing those previously referred but not tested, and those who met current NCCN guidelines for testing. 20 pts declined the survey. Reasons for declining survey included other medical concerns (n=4), scheduling issues (n=10), and lack of interest (n=6). 64 pts were unable to be reached by telephone. 46 pts completed the survey (44 pts surveyed in English, 2 in Spanish). Of those surveyed, 29 (63%) expressed interest in genetic testing. Reasons for interest included obtaining information for family (n=2); obtaining information about prior diagnosis and cancer risk (n=3); reducing anxiety (n=2). Reasons for declining counseling included not desiring further invasive tests (n=3). Most pts did not endorse a reason for agreeing or declining counseling. 11 pts (38%) have completed testing to date. Results of testing include BRCA1 PV (n=1), TP53 PV (n=1), variants of uncertain significance (n=2), and negative results (n=7). 19 pts have been referred for genetic consultation and are awaiting testing. Conclusions While more than half of BC patients eligible for genetic testing at diagnosis did not receive it, 63% of those surveyed in survivorship expressed interest in undergoing testing, and 38% of those expressing interest have completed testing to date. Our study demonstrates feasibility of and interest in germline genetic testing in late BC survivorship. Clinicians caring for BC survivors who have not undergone genetic testing should discuss benefits of testing for patients and their families, as well as changes in guidelines and insurance coverage. Next steps include completing genetic consultation and testing for referred patients and continuing to reach out to all pts without prior genetic testing or with outdated testing.

R. Bernhisel¹, M. Kucera¹, S. Cummings², E. Smith², T. Simmons¹, E. Hughes¹; ¹Biostatistics, Myriad Genetics, Inc., Salt Lake City, UT, ²Medical Affairs, Myriad Genetics, Inc., Salt Lake City, UT.

Background Patients with breast cancer (BC) who harbor germline pathogenic variants (PVs) in hereditary cancer genes have improved survival when their surgical and treatment decisions are tailored to their specific genetic alterations. Additionally, the identification of germline PVs is crucial for family members, as it allows for interventions that can decrease cancer-related morbidity and mortality. Commonly utilized guidelines recommend genetic testing only under specific conditions: if cancer is diagnosed at a young age, typically at or before the age of 50; if there is a strong family history (FH) of cancer; or if there are specific tumor characteristics. Several studies have demonstrated that many patients who do not meet these criteria unknowingly carry PVs and may, therefore, be receiving suboptimal care. In the United States, Black and Hispanic women face disproportionately high rates of BC mortality, highlighting the importance of genetic testing in these populations. It is unclear whether guideline-based restrictions differentially impact women of certain ancestral backgrounds. Here, we explore this question by estimating PV prevalence among BC patients of Asian, Black/African, Hispanic, Multiple, and White/non-Hispanic ancestries who do not meet guidelines for genetic testing. Methods Multivariable logistic regression models were constructed to analyze trends in the prevalence of PVs based on age, personal/family history, and ancestry in a consecutive cohort of patients referred for hereditary cancer testing with a multigene panel. We report ancestry-specific model-based estimates of prevalence for patients diagnosed with BC at older ages with no FH and no prior personal history of cancer of any type. Prevalence is summarized overall for 25-48 hereditary cancer genes with guideline-based medical management recommendations, as well as for the genes most frequently implicated in BC. Results The study cohort comprised 245,669 female patients affected by BC. There were 8,025 (3.3%) Asian, 25,900 (10.5%) Black/African, 20,022 (8.1%) Hispanic, 141,279 (57.5%) White/non-Hispanic patients, and 16,689 (6.8%) patients with multiple self-reported ancestries. Estimates of PV prevalence for a patient with BC with no FH and no prior personal history of cancer of any type were similar across ancestries, with the highest prevalence among Hispanic women (Table). Overall prevalence estimates were 6.3% for those diagnosed at age 50 and 3.6% among those diagnosed at age 80. PVs were most frequently identified in BRCA1, BRCA2, CHEK2, ATM, and PALB2 (Table). Conclusions Among patients with BC of all ancestries, a substantial fraction of patients who do not meet guidelines for genetic testing may unknowingly carry PVs. Elimination of age-based restrictions on genetic testing could improve the survival of patients with BC and positively impact their family members.

l. kamal, E. Belal, a. Beddah, H. Shaheen; Clinical Oncology Department, Menoufia University, Shebin el-kom- Menoufia, EGYPT.

Male Breast Cancer: Genomic Landscape, Epidemiological Patterns, and Clinical Outcomes in The United States Background:Male breast cancer is a rare form of breast cancer that has its distinct biological characteristics. Because of the disease rarity, most data on genomic landscape, epidemiological patterns, and clinical outcomes originate from small studies with limited number of patients. In this study, we aimed to leverage the publicly available genomic and epidemiological datasets to provide the largest to date analysis of clinic-genomic characteristics of male breast cancer in the United States.Methods: We utilized the Surveillance, Epidemiology, and End Results (SEER) database to extract data on patients diagnosed with male breast cancer from the SEER 17 Reg. Nov 2024 submission dataset (including patients diagnosed 2000-2022). SEER collects cancer incidence data from population-based cancer registries covering approximately 45.9 percent of the U.S. population. We used the AACR project GENIE v17.0 database to extract data on genomics of patients with male breast cancer who had samples sequenced and results accessible as part of AACR project GENIE. To avoid duplication, only patients with one profiled sample were included in the analysis. Only genes which were sequenced in at least 50 samples were included in frequency calculations.Results: The epidemiology analysis set included 10,284 patients with male breast cancer. The median age was 68 (interquartile range, 18). Most patients were white (80.3%, n=8,256) and non-Hispanic non-Latino (92.7%, n=9,529). Almost 51.7% (n=5,314) had left-sided breast cancer. In patients with known receptor status, the majority had ER-positive (96.2%, n=8,908), PR-positive (88.2%, n=8,063), and HER2 negative disease (88.1%, n=5,218). Most cases had HR+/HER2- disease subtype (85.9%, n=5,073). Lower proportions had HR+/HER2+, HR-/HER-, or HR-/HER2+ disease (11.2%, n=662; 2.2%, n=128; and 0.8%, n=41; respectively). The age-adjusted incidence rate was 1.2 per 100,000 and the median OS in the overall cohort was 109 months (95% CI: 105.1-112.9).The genomic analysis set included 144 patients. The most frequently mutated genes were PIK3CA (38.2%, n=55/144), GATA3 (21.6%, n=29/134 profiled), KMT2C (14.1%, n=14/99), TP53 (13.9%, n=20/144), and BRCA2 (10.4%, n=14/135). Structural variants were most frequently observed in INPPL1 (3.4%, n=2/58), CDK12 (1.7%, n=2/121), FGFR2 (1.6%, n=2/128), FGFR1 (1.6%, n=2/128), and NSD3 genes (1.2%, n=1/81). The most frequent copy number alterations were amplifications in FGF19 (19.4%, n=14/72), FGF4 (19.4%, n=14/72), FGF3 (19.4%, n=14/72), CCND1 (17.8%, n=18/101), and FGFR1 (14.9%, n=15/101).Conclusions: Male breast cancer occurs at a rate of 1.2 per 100,000 and has a median OS of 9 years. Patients mostly present with ER+ HER- disease and a molecular profile is quite different from female breast cancer.

H. Almarzooqi¹, N. Cuillerier¹, S. Wong¹, V. Villareal-Corpuz², M. Basik¹; ¹General Surgery, McGill university, Montreal, QC, CANADA, ²Stroll Cancer Prevention Centre, Jewish General Hospital, Montreal, QC, CANADA.

INTRODUCTION: Women diagnosed with atypical hyperplasia (AH) and lobular carcinoma in situ (LCIS) are at higher risk of developing invasive breast cancer. This risk may be reduced by endocrine prevention, though patient uptake and compliance are low. We sought to evaluate factors associated with the uptake and initiation of endocrine prevention therapy among women with AH and LCIS within a real-world setting.METHODS: This retrospective cohort study included women with high-risk lesions (HRL) diagnosed on biopsies performed between 2014 and 2025 and referred to the Jewish Stroll Cancer Prevention Centre and Stroll Cancer Prevention Centre, including women with AH and classical LCIS. The primary outcome was uptake of endocrine prevention, defined as receipt of a prescription and endocrine prevention initiation in eligible women.RESULTS: Endocrine prevention was prescribed to 106 (56.4%) of 188 eligible patients, and 71 (67.0%) started treatment. The regimen most frequently prescribed was low-dose tamoxifen at 64.2%, Raloxifene at 19.8%, and standard-dose tamoxifen at 11.3%. The highest uptake was 73.4% for women aged 51-60 years compared to 56.5% of women aged 41- 50 years, and 41.5% of women aged above 60 years (p < 0.001). A high uptake rate was also observed in parous women (64.6% vs 44.0%, p=0.005) and in women with a history of oral contraceptive use (69.5% vs. 46.2%, p=0.001). Before 2019, prescription rates were similar between Raloxifene and standard dose tamoxifen (47.1% each). However, low-dose tamoxifen became the most commonly prescribed regimen after 2019 (76.4%, p<0.001).CONCLUSION: After 2019, low-dose tamoxifen became the most commonly prescribed endocrine prevention regimen in HRL patients who accepted a prescription.

N. Love¹, K. S. Hughes², M. Robson³, G. Kelly¹, T. Wallace¹, T. Cruse¹, K. Miller¹, K. A. Ziel¹, K. H. Pang¹, D. Paley¹; ¹Clinical Content Department, Research To Practice, Key Biscayne, FL, ²Department of Surgery, Medical University of South Carolina, Charleston, SC, ³Breast Medicine Service, Memorial Sloan Kettering Cancer Center, New York, NY.

BACKGROUND: Germline genetic testing has important implications for both the patient with breast cancer and family members. The first physician with the opportunity to consider this option is usually the surgeon. Current practice guidelines commonly consulted on this issue from the American Society of Clinical Oncology (ASCO), the National Comprehensive Cancer Network (NCCN) and the American Society of Breast Surgeons (ASBrS) provide differing recommendations regarding which patients should be tested. We sought to determine how age, ER/PR and HER2 status and other variables factor into surgeons’ recommendations for testing. METHODS: From February to March 2025, 50 US-based surgeons in community practice were recruited to complete a survey designed to identify clinical situations when germline genetic testing is recommended. For each case-based question, a number of factors (age, ER/PR and HER2 status, family history, gender) were varied. Ten surgeons at academic centers were also recruited to complete the survey. A modest honorarium was provided. RESULTS: For a 30-year-old patient without a family history (all 3 guidelines recommend testing), almost all respondents recommend testing regardless of ER/PR/HER2 status (Table). For a similar 55-year-old patient, almost all participants recommended testing in patients with TNBC (as is consistent with guidelines), but less than half would test patients with ER/PR-positive tumors regardless of HER2 status (testing recommended by ASCO and ASBrS and a consideration per NCCN). For a 70-year-old patient (only ASBrS recommending testing), again almost all respondents recommended testing for patients with TNBC, but far fewer respondents would test patients with ER/PR-positive tumors regardless of HER2 status. There was also agreement that all men with localized breast cancer should be tested. Interestingly, the responses from the community-based cohort largely aligned with those of their academic counterparts. CONCLUSION: The findings from this survey demonstrate that whether a patient with newly diagnosed localized breast cancer receives a recommendation for germline genetic testing depends largely on the practice pattern of the surgeon initially managing the case as many do not initiate testing for patients 55 years and older with ER/PR-positive tumors. While other physicians, including medical oncologists, may seek out this information at a later time as they weigh adjuvant therapy options, it is important to consider that the main downside of genetic testing is financial. Given the important implications for family members and the overall survival benefit associated with olaparib in the OlympiA trial, perhaps in the future genetic testing will join ER and HER2 assessment as a standard for all patients with localized disease.

J. Ahn; General surgery, Yonsei university, Seoul, KOREA, REPUBLIC OF.

Purpose There have been reported different characteristics of multigene assays (MGAs) to discriminate high and low risk groups in estrogen receptor (ER)-positive/HER2-negative breast cancer patients. In real world data, however, direct comparison between MGAs is not easy because clinical guidelines recommended against using two or more MGAs in a patient simultaneously. In this study, clinicopathological characteristics and results of Oncotype Dx (ODx) and EndoPredict (EP) were indirectly compared to investigate characteristics of each assay. Methods A total of 1,782 node-negative stage I/II patients with ER-positive/HER2-negative breast cancer were retrospectively selected between 2013 and 2025. Of them, 1,537 (86.3%) patients performed ODx and 245 (13.7%) examined EP. High risk of ODx was defined considering clinical risk and recurrence score. EPclin was used to determine high risk of EP. Clinicopathological characteristics and assay results were compared using logistic regression models. After 1:1 propensity score matching (PSM) adjusting clinicopathological characteristics, predictive factors for high risk assay result were re-analyzed. Results High risk of ODx and EP was 25.4% and 24.5%, respectively. EP patients showed older age, more obese, and higher mixed/lobular/special histology. ODx patients demonstrated larger tumor size, higher grade II/III disease, and higher Ki-67 levels. Allred score of ER and progesterone receptor (PR) and HER2 immunohistochemical score were not different. Correlation analysis showed ODx recurrence score was significantly associated with ER (Spearman’s rho, -0.079) and PR (Spearman’s rho, -0.425) Allred score negatively and with Ki-67 levels (Spearman’s rho, 0.251) positively. EP 12-gene molecular score was negatively associated with PR Allred score (Spearman’s rho, -0.290) and positively related to Ki-67 values (Spearman’s rho, 0.516). Multivariate analysis adjusting clinicopathological characteristics showed that EP had a higher probability of high risk [Hazard ratio, 1.783 (95% confidence interval, 1.330-2.390); p<0.001]. Subset analysis by each assay showed younger age, lower ER and PR Allred score, and HER2 2+ immunohistochemical score were major determinants of high risk in ODx patients. But larger T2-3 stage, histologic grade II/III, and Ki-67 ≥20% were significant factors of high risk in EP cases. After 1:1 PSM, assay method was not significant [Hazard ratio, 1.286 (95% confidence interval, 0.761-2.174); p=0.348] but T2-3 stage, grade II/III, low PR, and higher Ki-67 were independently significant predictors of high risk result. Conclusion Each MGA showed different features of major determinants for high risk group. Although definition of high risk considering clinical risk factors and gene score might be partly related, ODx was strongly dependent on ER-related signatures and EP was significantly determined by proliferative features of histologic grade and Ki-67 levels. Primary physician and patient could select one among available MGAs based on understanding of its distinct characteristics.

A. Koutras¹, K. Papazisis², I. Boukovinas³, R. Iosifidou⁴, I. Intzidis⁴, K. Potska⁵, C. Dogka⁵, A. Katseli⁵, D. Paranou⁵, A. Karavaggeli⁵, A. Meintani⁵, D. Bouzarelou⁵, D. Lachanas⁵, G. Rigas⁶, M. Kanara⁷, M. Müslümanoglu⁸, I. Bobolaki⁹, E. Biziota¹⁰, E. Papadopoulou⁵, G. Nasioulas⁵; ¹Medical School, University of Patras, Patras, GREECE, ²3rd Dpt of Oncology, Interbalcan European Medical Centre, Thessaloniki, GREECE, ³Oncology unit, Bioclinic, Thessaloniki, GREECE, ⁴Breast Surgical Oncology Clinic, Theageneio Anticancer Hospital, Thessaloniki, GREECE, ⁵Molecular Oncology, Genekor Medical SA, Attica, GREECE, ⁶Breast Clinic, Agios Savvas Cancer Hospital, Attica, GREECE, ⁷Breast Clinic, General Hospital of Trikala, Trikala, GREECE, ⁸Breast Clinic, Istanbul University School of Medicine, Instabul, TURKEY, ⁹Oncology Department, General Hospital of Chania, Crete, GREECE, ¹⁰Oncology Department,, Universal Hospital of Alexandroupoli, Alexandroupoli, GREECE.

Background: Neurofibromatosis type 1 (NF1), is a classic example of a complete penetrance disease with an estimated prevalence of 1 in 3000. Βroader application of multigene panel testing in candidate hereditary cancer predisposition cases, reveals pathogenic NF1 variants in Breast/ovarian/colorectal cancer patients without a clinical NF1 diagnosis. The scope of this study is to enlighten the contribution of NF1 to cancer risk in general population and the need of medical management guidelines in NF1 breast/ ovarian cancer patients lacking the relevant clinical phenotype. Methods: This retrospective observational study included patients with mutated germline NF1, referred at GENEKOR Medical SA, Athens, Greece between the period of 1^st Jan 2020 to 31^st Oct 2024. Ηigh risk cancer patients, with no neural crest derivant malignancies, underwent germline mutation testing, with a 52 gene panel. All the NF1 mutations were crosschecked via the ClinVar mutation database to be classified as pathogenic. Results: This retrospective observational study included 15 patients with incidentally identified mutated germline NF1, who were referred due to breast cancer (5/15, 33%), ovarian cancer (4/15,27 %), colorectal cancer (2/15,13%), multiple myeloma (1/15,7%) and glioblastoma (1/15,7%). Among the 12782 cancer patients, with no neural crest derivant malignancies, who underwent germline high-risk cancer gene susceptibility testing, 2616 (20.5%) cases had pathogenic mutations and 15 out of them (0.6%) were NF1 mutated. We identified an unexpectedly high prevalence (1 out of 852), of PVs in the NF1 gene, more than 3 times the rate expected given the reported prevalence of NF1. We identified 13 pLOF variants, 1 missense, 1 single amino acid deletion and 1 deletion involving the entire NF1 gene, all in heterozygosity, with the exception of 2 samples with blood mosaicism. Physical exam, medical and family history revealed no features consistent with NF1 diagnosis contrary to the complete penetrance of the disease. In 3 samples, additional PVs were identified, in BRCA2 (ovarian cancer), ATM and NTHL1 (2 colorectal cancer cases) genes. Conclusions: This study highlights that NF1 PVs are more frequent than previously estimated and often discovered in unaffected individuals. Incomplete penetrance and missed diagnoses probably explain this. The aforementioned data presented provides insights into the need of establishment of screening and management strategies in patients with incidental findings.

B. Citgez¹, B. Yigit², E. Namal³; ¹General Surgery, Halic University, Memorial Hospital, Istanbul, TURKEY, ²General Surgery, Bagcilar Training and Research Hospital, Istanbul, TURKEY, ³Oncology, Memorial Hospital, Istanbul, TURKEY.

Background: Hepatosteatosis is frequently observed in breast cancer patients due to shared risk factors such as obesity, metabolic syndrome, and advanced age. The aim of this study is to quantitatively evaluate the association between hepatosteatosis and early-stage breast cancer by examining routine abdominopelvic computed tomography images obtained for staging purposes in breast cancer patients.Methods: The imaging and histopathology reports of patients who presented to Memorial Hospital for breast screening or diagnostic imaging were reviewed retrospectively. Patients with advanced-stage breast cancer, additional malignancies, or significant comorbidities were excluded from the study. A total of 129 patients (Group 1) who underwent surgery for early-stage breast cancer between January 2017 and January 2025 were included in this retrospective analysis. A control group (Group 2) was formed from 103 patients without a history of cancer or significant comorbidities. Liver attenuation values on abdominopelvic computed tomography, as well as serum transaminases (AST, ALT), gamma-glutamyl transferase (GGT), alkaline phosphatase (ALP) and total bilirubin levels were recorded and compared with those of the control group. Hepatosteatosis was evaluated using a four-point scale ranging from 0 to 3, where higher grades reflect greater degrees of lipid accumulation in the liver tissue.Results: There were no statistically significant differences between Group 1 and Group 2 in terms of age, AST, ALT, GGT, ALP, or total bilirubin levels (all p > 0.05) (Table 1). The rate of patients with moderate to severe hepatosteatosis (grades 2-3) was significantly higher in Group 1 than in Group 2, whereas the rate of patients with grade 0-1 hepatosteatosis was significantly higher in Group 2 than in Group 1 (18.6% vs. 8.74% and 91.26% vs. 8.74%, respectively; p= 0.033).Conclusion: The study results indicated an increased prevalence of hepatosteatosis in early-stage breast cancer patients when compared to control group. Lifestyle modifications, weight loss, increased regular physical activity, and a healthy diet help reduce hepatosteatosis and insulin resistance, thereby preventing the progression of non-alcoholic fatty liver disease. Considering the evidence suggesting that hepatosteatosis may increase the risk of breast cancer development, clinicians should keep in mind the importance of planning and implementing preventive strategies in this patient group.

P. Mora¹, M. Acosta¹, J. Herzog², N. Valdiviezo³, L. Reynaga⁴, J. Abugattas⁵, Y. Sullcahuaman¹, T. Vidaurre³, S. Neciosup³, J. Dunstan⁵, H. Gómez⁶, I. Otoya³, Z. Morante³, V. Acuña⁷, A. López⁸, S. Gruber⁴; ¹Unidad Funcional de genética y biología molecular, Instituto Nacional de Enfermedades Neoplásicas, Lima, PERU, ²Division of Clinical Cancer Genomics Molecular Laboratory, City of Hope National Medical Center, Duarte, CA, ³Oncología Médica, Instituto Nacional de Enfermedades Neoplásicas, Lima, PERU, ⁴Center for Precision Medicine, City of Hope National Medical Center, Duarte, CA, ⁵Cirugía de Mamas y Tejidos Blandos, Instituto Nacional de Enfermedades Neoplásicas, Lima, PERU, ⁶Instituto de Investigaciones de Ciencias Biomédicas, Universidad Ricardo Palma, Lima, PERU, ⁷Facultad de Farmacia y Bioquímica, Universidad Nacional Mayor de San Marcos, Lima, PERU, ⁸Cirugía Ginecológica, Instituto Nacional de Enfermedades Neoplásicas, Lima, PERU.

Breast cancer (BC) is one of the most common cancers among women in both Peru and worldwide. Ovarian cancer (OC), due to the difficulty of early detection, also represents a significant public health concern in the country. In recent years, genetic testing has become essential for identifying individuals at high risk for hereditary cancer predisposition syndromes, guiding decisions related to prevention, early detection, and treatment. In Peru, the use of genetic testing is still developing. This work presents an inter-institutional collaborative experience that has facilitated access to genetic testing and contributed to the implementation of a multidisciplinary care system for managing high-risk patients. The main objective of this study is to identify the genomic profile of patients with breast and/or ovarian cancer and to evaluate the clinical impact of genetic testing on cancer management. Methods: This study includes individuals diagnosed with breast and/or ovarian cancer who were referred for Medical Genetics evaluation at the Instituto Nacional de Enfermedades Neoplásicas (INEN) between August 2013 and March 2020. Participants met criteria for inclusion in the inter-institutional protocol INEN13-48 on hereditary cancer and underwent genetic testing at City of Hope (IRB #96144).Results and Discussion: We recruited 1577 patients; 99.3% were females (n=1566/1577) with the initial diagnosis of BC (89.4%. n=1401/1566), OC (8.5%. n=133/1566) and BC/OC (2%. n=32/1566); 0.7% were males (n=11/1577) all BC. Of all the individuals, 46.8% (n=738/1577) had the first diagnosis under 40yo; 36.1% (n=570/1577) was between 41-50yo. 13.4% (n=212/1577) had a LP/P variant in an actionable gene with management guidelines for at least one cancer type; 0.8% (n=12/1577) had a LP/P variant in an non actionable gene; 34.5% (n=545/1577) had a BRCA1/BRCA2 panel with no variant detection and 51.2% (n=808/1577) had an extended panel with no variant detection. The most frequent LP/P variants in actionable genes were 53.8% (n=114/212) in BRCA1 and 24% (n=51/212) in BRCA2; 1% (n=2/212) had more than one LP/P variant (BRCA1/BRIP1 and BRCA1/BRCA2). The detection rate of LP/P variant in BC was 12% (n=170/1401) in BC, 19.5%(n=26/133) in OC and 50% (n=16/32) in B/OC. The most frequent LP/P variants were BRCA1(NM_007294.4):c.68_69del (p.Glu23ValfsTer17) (6%: n=13/212) and BRCA2(NM_000059.4):c.1219C>T (p.Gln407Ter) (4% n=8/212). 9% (n=1/11) of the male patients, had a variant in BRCA2 gene. All 212 patients were evaluated in clinical genetics. Over 10 years of follow-up, 72% (n=152) received recommendations for high-risk breast cancer (BC) management and 54% (n=115) for ovarian cancer (OC); the rest did not due to disease progression or geographic barriers. Among those at high BC risk, 34% (n=60) began imaging-based surveillance, 26% (n=40) underwent unilateral/bilateral risk-reducing mastectomy (RRM)—with 10% (n=4) diagnosed with new BC—and 34% (n=52) declined recommendations. In the high OC risk group, 18% (n=21) had prior ovarian surgery, 49% (n=56) underwent risk-reducing salpingo-oophorectomy (RRSO)—with 2 OC diagnoses—and 23% (n=27) declined. As of now, 36.7% (n=78) have died and 17% (n=37) are lost to follow-up.Conclusion: The importance of ensuring access to genetic studies, in addition to the recognition of genetic predisposition syndromes to cancer is paramount. In this cohort we identified that 13% of individuals had an LP/P in an actionable gene, similar to other cohorts. This information support the needs of long-term financing, not only to access to genetic testing, but helps us to improve the access to multidisciplinary high-risk management in our institution considering the geographical and economic barriers in our country.

Y. Lin¹, J. She¹, R. Zhao¹, Y. Zhang², A. Qiu¹, L. Zhang¹, Q. Yang¹; ¹Department of Breast Surgery, the First Affiliated Hospital of Xiamen University, School of Medicine, Xiamen University, Xiamen, CHINA, ²Medical college, Guangxi University, Nanning, CHINA.

Background: An increasing body of epidemiological evidence has suggested that breast cancer (BC) and major depressive disorder (MDD) are frequently comorbid, with MDD being particularly prevalent among BC patients during diagnosis and treatment. However, whether this relationship reflects a shared genetic predisposition or a causal biological link remains unclear. Here, we aimed to comprehensively investigate the shared genetic architecture and potential causal associations between BC and MDD using large-scale genome-wide association study (GWAS) data. Methods: We obtained publicly available GWAS summary statistics for BC and MDD. First, linkage disequilibrium score regression (LDSC) was used to estimate the genome-wide genetic correlation between the two traits. To identify shared genetic variants (pleiotropic loci), we conducted multiple cross-trait GWAS analyses including multi-trait analysis of GWAS (MTAG), cross-phenotype association (CPASSOC), and pleiotropic analysis under composite null hypothesis (PLACO). We then functionally annotated PLACO-identified SNPs using Functional Mapping and Annotation (FUMA) to identify lead variants and genomic risk loci. Gene-level associations were further evaluated using MAGMA to uncover shared susceptibility genes. Finally, we used both Generalized Summary-data-based Mendelian Randomization (GSMR) and two-sample Mendelian Randomization (TSMR) approaches to assess the direction and magnitude of potential causal effects between MDD and BC. Results: LDSC revealed a significant positive genetic correlation between BC and MDD (rg = 0.0545, SE = 0.0151, P = 3.0 × 10⁻⁴), indicating a modest but meaningful shared genetic basis. MTAG analysis identified 17 pleiotropic SNPs reaching genome-wide significance. CPASSOC and PLACO analyses revealed 16,470 and 775 pleiotropic SNPs, respectively, suggesting widespread cross-trait genetic overlap. FUMA-based annotation of PLACO hits yielded 22 independent genomic loci and MAGMA further identified 130 pleiotropic genes, highlighting biological convergence between psychiatric and oncologic pathways. GSMR (OR = 1.27, 95% CI: 1.11-1.46, P = 5.90 × 10⁻⁴) and TSMR (OR = 1.12, 95% CI: 1.02-1.22, P = 1.43 × 10⁻²) demonstrated significant causal effects of MDD on increased BC risk. The sensitivity analysis suggested that the results of TSMR were not affected by horizontal pleiotropy (P = 0.75). Conclusions: Our study charted the shared genetic architecture between BC and MDD and explored their potential causal relationship. The findings enhanced the understanding of the genetic pathways commonly involved in both BC and MDD, highlighting their complex interconnections. This research provided valuable insights for experimental studies aimed at developing new preventive measures and therapeutic interventions for BC and MDD.

W. Kuo¹, W. Shen², H. Chou¹, M. Peng³, S. Huang⁴, S. Chen⁵, Y. Lin³; ¹Breast Cancer Center, Chang Gung Memorial Hospital, Linkou, Tauoyuan, TAIWAN, ²2.Medical Oncology, Chang Gung Memorial Hospital, Linkou, Tauoyuan, TAIWAN, ³Medical Oncology, Chang Gung Memorial Hospital, Linkou, Tauoyuan, TAIWAN, ⁴Medical Oncology, Chang Gung Memorial Hospital, Kaohsiung, Tauoyuan, TAIWAN, ⁵General and Breast Surgery, Chang Gung Memorial Hospital, Linkou, Tauoyuan, TAIWAN.

Background: The incidence of breast cancer in East Asia has been steadily increasing, with cases presenting approximately a decade earlier than those observed in Western populations. This study utilizes whole-exome sequencing (WES) to investigate the prevalence of hereditary germline pathogenic or likely pathogenic variants (PV) in Taiwanese breast cancer patients identified as having moderate to high hereditary genetic risk. Method: Breast cancer patients with age at diagnosis less than 40 years, positive familial cancer history, bilateral cancer or multiple cancers, and triple negative breast cancer (TNBC) were enrolled from the Linkou, Taipei, and Kaohsiung hospital branches of Chang Gung Medical Foundation. BRCA1/2 likelihood was evaluated with the pathology-adjusted Manchester Scoring System (MSS3). (1) Whole-exome sequencing (WES) was conducted using the IDT Lotus DNA Library Prep Kit and xGen Exome Research Panel v2, with sequencing on the NovaSeq 6000 platform. Reads were aligned to the GRCh38 genome reference using DRAGEN v.3.9.3 with stringent quality metrics. Variants were filtered against databases such as dbSNP version 150 and Taiwan Biobank, annotated via wANNOVAR, and BRCA1/2 PVs were confirmed using Sanger sequencing. Results: A cohort of 525 breast cancer patients were enrolled, revealing 36 individuals (6.9%) carring confirmed germline BRCA1/2 PVs and another 28 patients (5.3%) carrying PVs in other non-BRCA hereditary breast and ovarian cancer (HBOC)-related genes (ATM, BARD1, BRIP1, CHEK2, MUTYH, PALB2, PMS2, RAD50, RAD51D, and TP53) (Table1). Among the risk groups classified by the Manchester Scoring System, 72.8% were categorized as low (20%) groups. The detection rate of BRCA PV was 1.6% (6/382), 11.3% (9/80), and 33.3% (21/63), respectively, in low to high-risk Manchester categories. For patients diagnosed with breast cancer under age 40 (n= 291,55%), the prevalence of BRCA1/2 PV was observed at 7.2% (21/291). Among patients with TNBC (n=79), the proportion of patients with BRCA1/2 PVs was notably higher, at 16.5% (n=13). In cases of bilateral breast cancer (n=47) (synchronous or metachronous), 12.8% (n=6) had BRCA1/2 PVs. Among the 16 male breast cancer patients, no BRCA 1/2 PVs were detected. For individuals with a high Manchester Score but no detected BRCA1/2 PVs through sequencing, further analysis using the MLPA assay remains ongoing.Conclusion: Whole-exome sequencing has provided valuable insights into the prevalence and distribution of germline pathogenic variants in HBOC-related genes among Taiwanese breast cancer patients with moderate to high hereditary risk. (2) The occurrence rates of BRCA1/2 pathogenic variants in the overall cohort and within triple-negative breast cancer patients align with trends observed in other breast cancer populations. (3) 1. Evans DG, Harkness EF, Plaskocinska I, Wallace AJ, Clancy T, Woodward ER, et al. Pathology update to the Manchester Scoring System based on testing in over 4000 families. (1468-6244 (Electronic)).2. Felicio PS, Grasel RS, Campacci N, de Paula AE, Galvão HCR, Torrezan GT, et al. Whole-exome sequencing of non-BRCA1/BRCA2 mutation carrier cases at high-risk for hereditary breast/ovarian cancer. Hum Mutat. 2021;42(3):290-9.3. Lee NY, Hum M, Amali AA, Lim WK, Wong M, Myint MK, et al. Whole-exome sequencing of BRCA-negative breast cancer patients and case-control analyses identify variants associated with breast cancer susceptibility. Hum Genomics. 2022;16(1):61.

J. CLAROS¹, I. SISO¹, M. ECHARRI¹, D. MORCHON², L. PIZARROSO¹, J. MESA¹, C. ORTEGA¹, A. GARRIDO¹, C. PERNAUT¹, H. FRAILE¹, A. PEÑALVER¹, B. ALONSO³, A. LOPEZ¹; ¹MEDICAL ONCOLOGY, HOSPITAL UNIVERSITARIO SEVERO OCHOA, MADRID, SPAIN, ²MEDICAL ONCOLOGY, COMPLEJO ASISTENCIAL UNIVERSITARIO DE SALAMANCA, SALAMANCA, SPAIN, ³INTERNAL MEDICINE, HOSPITAL UNIVERSITARIO SEVERO OCHOA, MADRID, SPAIN.

Background:Multi-gene panel testing has expanded the identification of hereditary breast cancer syndromes. ESMO Precision Oncology Working Group recently published recommendations advising against routine testing of certain genes suggesting minimal mortality benefit (ATM, CHEK2). We present the germline findings from a large, single institution cohort in Southern Europe, focusing on the diagnostic yield of high- and moderate-penetrance genes and the clinical context of variant identification.Methods:We conducted a retrospective analysis of patients (pts) referred to our Hereditary Cancer Genetic Counseling Unit between May 2023 and April 2025. Variant classification followed American College of Medical Genetics (ACMG) criteria.Results:From a total of 518 pts referred for counseling, 387 were tested and included in the analysis (median age 57). Overall, 91 pts (23.5%) carried a pathogenic/likely pathogenic (P/LP) variant (Class 4/5) and 34 (8.8%) had a variant of uncertain significance (VUS). Among the 151 (39%) patients with a personal history of BC, the specific diagnostic yield for a P/LP variant was 14.56% (n=22), while the VUS rate was 15.2% (n=23). Among P/LP carriers, 51 (56%) were in high-penetrance breast cancer genes: BRCA1 (n=24), BRCA2 (n=24) and PALB2, PTEN, CDH1 (n=1 each). 36 of 51 (70.5%) of these carriers were identified via cascade testing of healthy relatives. 18 of 91 pts (19.7%) P/LP variants were found in moderate-penetrance genes: ATM (n=9), CHEK2 (n=6), RAD51C (n=2), BRIP1 (n=1). Of note, 11 variants (61%) were identified through cascade testing rather than personal history of cancer. Conclusion:In this real-world cohort, nearly a quarter of individuals undergoing testing carried a clinically significant germline variant. While BRCA1/2 remain dominant, a substantial number of variants in genes like ATM, CHEK2, and RAD51C were discovered through cascade testing, mainly in unaffected carriers. This highlights the clinical challenges surrounding the moderate penetrance genes, including appropriate follow-up and counseling strategies.

L. P. Lacerda¹, P. F. Silva², K. M. Veiga², B. F. Nascimento², I. M. Calassa², W. D. Oliveira³, A. S. Limongi⁴, R. L. Mesquita⁵, K. J. Monteiro⁶, R. R. Santos Junior⁵, R. Taminato⁷, R. M. Rahal⁸, E. P. Silveira-Lacerda⁹; ¹Medicine, Universidade do Planalto Central, Brasília, BRAZIL, ²Genética, Universidade Federal de Goiás, Goiânia, BRAZIL, ³Genetic, Universidade Federal de Goiás, Goiânia, BRAZIL, ⁴Superintendência de Políticas e Atenção Integral à Saúde, Secretaria Estadual de Saúde, Goiânia, BRAZIL, ⁵SPAIS – Superintendência de Políticas e Atenção Integral à Saúde, SES – Secretaria Estadual de Saúde de Goiás, Goiânia, BRAZIL, ⁶Superintendência de Políticas e Atenção Integral à Saúde, Secretaria Estadual de Saúde de Goiás, Goiânia, BRAZIL, ⁷Farmácia, Hospital Araújo Jorge, Goiânia, BRAZIL, ⁸Centro Avançado de Diagnóstico de Mama, Universidade Federal de Goiás, Goiânia, BRAZIL, ⁹Centro de Genética Humana – ICB, Universidade Federal de Goiás, Goiânia, BRAZIL.

Approximately 10% of breast and ovarian cancer cases are hereditary. Identifying germline pathogenic variants and acting on this information through preventive strategies is a key pillar of precision oncology. In Brazil, the public health system (SUS) in Goiás has incorporated genetic risk screening and management, setting a precedent for accessible, population-level genomic medicine.This study aimed to evaluate the frequency of germline pathogenic variants in patients at risk for hereditary breast and ovarian cancer, and to track these variants among their relatives for early diagnosis and prevention. We evaluated 268 patients who met the 2025 National Comprehensive Cancer Network (NCCN) criteria for hereditary breast and ovarian cancer syndromes, alongside 268 parents. All participants were referred to the Human Genetics Center at UFG through the SUS network.Following pre-test counseling and informed consent, blood samples were collected and analyzed using next-generation sequencing (NGS) with the Oncomine™ BRCA Expanded Panel (Ion Torrent platform). Among the 268 index patients, 14% (38/268) had pathogenic germline variants. The distribution was: BRCA1 (50%, 19/38), TP53 (21%, 8/38), PALB2 (10.5%, 4/38), CHEK2 (2.9%), BRCA2 (5.3%), and ATM (5.3%).Further familial testing was conducted in 30 families. Among eight families with the TP53 c.1010G>A variant, 36 of 99 tested relatives (36.4%) were positive. In two BRCA2-positive families, 9 of 18 relatives (50%) carried the same variant. Altogether, 45 of 117 relatives (38.46%) tested positive for the familial variant.Crucially, all relatives with positive results were evaluated by a medical geneticist and referred for cancer surveillance and risk-reducing strategies according to NCCN 2025 guidelines. This included MRI, mammography, and risk-reducing surgery when indicated, ensuring translation from genetic finding to clinical action.The project established a structured, replicable workflow for managing hereditary cancer risk in families. This includes genetic counseling, testing, cascade screening, and follow-up within the public healthcare system. By integrating genetic risk management into routine SUS care, the program enables early-stage diagnosis and reduces cancer-related mortality, particularly in underserved populations.This initiative demonstrates that comprehensive genomic care and precision prevention are feasible within a public system and scalable to other regions. It highlights how transforming a genetic finding into preventive action can generate real-world impact bridging the gap between variant and value.

W. A. Mantilla¹, Y. T. Zambrano-O², L. Rey-Vargas², D. C. Sierra-Diaz³, M. Lema⁴, D. Y. Fonseca-Mendoza³, S. J. Serrano-Gomez⁵, C. M. Restrepo³, M. C. Sanabria-Salas⁵; ¹Breast Cancer Unit, Luis Carlos Sarmiento Angulo Cancer Treatment and Research Center, Bogota, COLOMBIA, ²Grupo de Investigacion Biologia del Cancer, Instituto Nacional de Cancerologia, Bogota, COLOMBIA, ³School of Medicine and Health Sciences, Center for Research in Genetics and Genomics (CIGGUR), Universidad del Rosario, Bogota, COLOMBIA, ⁴Medical Department, Clinica de Oncologia Astorga, Medellin, COLOMBIA, ⁵NA, NA, Bogota, COLOMBIA.

Background: Breast cancer (BC) susceptibility is driven by a complex interplay of modifiable and non-modifiable risk factors. While the impact of BRCA1 and BRCA2 pathogenic/likely pathogenic variants (PVs) is well documented, the contribution of genetic ancestry to the distribution of PVs in BRCA1/2 and other homologous recombination repair (HRR) genes remains understudied in genetically admixed populations such as Colombians. Methods: Whole-exome sequencing data from 358 unselected Colombian women with BC were analyzed for genetic ancestry. Among them, 24 individuals were identified as carriers of PVs in BRCA1 (n=3), BRCA2 (n=10), PALB2 (n=2), ATM (n=5), CHEK2 (n=2), BARD1 (n=1), and RAD51D (n=1), as classified by ACMG/AMP criteria. Raw BAM files underwent preprocessing, quality control, and joint genotyping, yielding 14.9 million variants. Quality filtering using PLINK v1.9 included genotype call rate ( 1e-6), minor allele frequency (>0.1), and linkage disequilibrium pruning. Samples with >25% missingness were excluded, resulting in a final dataset of 344 individuals and 95,671 variants (genotyping rate: 99.93%). Global ancestry proportions were estimated using principal component analysis with reference populations from the 1000 Genomes and Human Genome Diversity Projects, specifically African (AFR), European (EUR), and Native American (NAM) populations. Statistical comparisons of ancestry proportions were performed across clinical variables (e.g., age at diagnosis, family history), genetic status (PVs carriers vs non-carriers), and tumor subtypes. Results: Global ancestry was predominantly EUR (mean ~53%), followed by NAM (~36%) and AFR (~10%). Significant differences in ancestry distribution were observed across clinical subgroups. AFR ancestry was higher among cases born in the Caribbean and Pacific coast regions (18% and 11%, respectively; p < 0.001), while EUR ancestry was higher among those from the Andean and Plains regions (54% and 60%, respectively; p < 0.001). The NAM component was more homogeneous across regions (34-38%). Higher EUR ancestry was associated with lower rates of breastfeeding and higher reported oral contraceptive use (p < 0.05 for both). EUR ancestry was higher in individuals with ER-positive tumors compared to ER-negative tumors (54% vs. 51%, respectively, p < 0.05), whereas AFR ancestry was higher in ER-negative (8% vs. 6% in ER-positive, p < 0.05) and PR-negative tumors (8% vs. 5% in PR-positive, p < 0.05). A higher EUR component was observed among those with a positive family history (55%) (p = 0.057) and PVs carriers in BC-related genes (58%). Two cases with CHEK2PVs [c.349A>G and c.1100delC] had 71% EUR ancestry, significantly higher than non-carriers (54%, p < 0.05), consistent with their Northern/Central European founder origins. Among the 24 PV carriers, luminal subtypes predominated (63%), particularly in ATM (100%) and BRCA2 (80%) carriers. Four of five ATM PV carriers with ancestry estimates showed predominantly EUR ancestry (49-59%). The ATM splice-site variant c.5496+2_5496+5delTAAG was identified in three unrelated individuals from the Andean region—two of whom have predominantly European ancestry and one with missing ancestry data. Conclusions: Our findings demonstrate significant associations between global ancestry proportions and key clinical and genetic features in an admixed Colombian BC cohort. To further elucidate the genetic and demographic factors influencing the frequency, distribution, origin, and potential impact of PVs identified in this cohort, we have initiated ongoing local ancestry estimations and haplotype analyses. This study highlights the importance of ancestry-informed analyses to improve risk stratification and guide precision medicine approaches in admixed populations.

F. Gan, Y. Yang, Q. Liu; Breast Tumor Center, Sun Yat-sen Memorial Hospital, Sun Yat-sen University, Guangzhou, CHINA.

Background: ‌ BRCA1/2 germline mutations (gBRCAm) increase the risk of breast and ovarian cancers. Multidisciplinary collaboration between gynecological oncologists and breast specialists is crucial for the comprehensive assessment and personalized treatment of patients. Breast specialists often recommend gBRCAm breast cancer patients consult gynecological oncologists for ovarian cancer issues, but they rarely receive referrals from gynecologists for gBRCAm ovarian cancer patients to consult about breast cancer. This study aimed to investigate gynecological oncologists’ knowledge, attitudes, and practices (KAP) regarding breast cancer-related risks in ovarian cancer patients with gBRCAm.‌ Methods:‌ A cross-sectional online survey was conducted among gynecological oncologists from November 19 to December 7, 2024. The data were collected using a self-designed electronic questionnaire assessing general characteristics of respondents, KAP regarding breast cancer risks in ovarian cancer with gBRCAm. A total of 625 valid questionnaires were collected. Descriptive analysis and logistic regression analysis were performed on the collected data.‌ Results:‌ The participants’ mean age was 44.04 years, with 85.6% being female, and 64.5% held senior professional titles.‌ Regarding KAP on breast cancer risks in ovarian cancer with gBRCAm, 11.4% of participants reported being unfamiliarity of the breast cancer risk; 88.0% believed it necessary to refer ovarian cancer patients with gBRCAm to breast specialists; 61.6% had previously referred such patients. Primary reasons for non-referral included: prioritizing ovarian cancer treatment due to its high lethality (46.4%), lack of focus on breast cancer risk (43.4%), and concurrent breast evaluation during ovarian cancer management (5.4%). The annual number of ovarian cancer patients treated was ≤10 for 310 participants (49.6%), 11-50 for 243 (38.9%), and >50 for 72 (11.5%). 67.7% had treated ovarian cancer patients carrying gBRCAm. 334 participants (53.4%) had been consulted by breast cancer patients carrying gBRCAm regarding ovarian cancer risks. Multivariate logistic regression revealed that‌ understanding the breast cancer risk (P10 ovarian cancer patients annually (11-50 patients: P=0.041; >50 patients: P=0.003), having treated ovarian cancer patients with gBRCAm (P<0.001), and having been consulted by breast cancer patients carrying gBRCAm (P<0.001) were significantly associated with higher referral rates to breast specialists. Regarding opportunistic salpingectomy, 71.7% of participants agreed that all women should consider prophylactic salpingectomy concurrent with other pelvic surgery after completing childbearing. 70.6% agreed that prophylactic salpingectomy (preserving ovaries) could potentially reduce cancer risk while preserving ovarian function, and would consider this procedure for young carriers (≤40 years) with gBRCAm who had completed childbearing.‌ Conclusion:‌ Gynecologic oncologists’ knowledge regarding breast cancer risk in ovarian cancer patients with gBRCAm and their referral rates of these patients to breast specialists need improvement. Enhancing gynecologic oncologists’ understanding of BRCA1/2-related breast cancer risks will help increase referral rates to breast specialists, and promote multidisciplinary collaboration, and facilitate holistic healthcare for patients carrying gBRCAm.

C. Li¹, V. Patel¹, R. McShinsky², Z. Burningham², A. Chiba³, S. Bushan⁴, S. Trehan⁵, M. Kelley⁶, A. Halwani⁷, S. V. Colonna⁷; ¹Epidemiology, University of Utah / George E Wahlen VA, Salt Lake City, UT, ²Medical Oncology, University of Utah / George E Wahlen VA, Salt Lake City, UT, ³Surgical Oncology, Duke University / Durham VAMC / National Oncology Program, VA, Durham, NC, ⁴Medical Oncology, Baylor University / Michael E DeBakey VA / National Oncology Program, VA, Houston, TX, ⁵Medical Oncology, Tampa VA / National Oncology Program, VA, Tampa, FL, ⁶Medical Oncology, Duke University / Durham VAMC / National Oncology Program, VA, Durham, NC, ⁷Medical Oncology, Huntsman Cancer Institute / George E Wahlen VA / National Oncology Program, VA, Salt Lake City, UT.

Background Male breast cancer (MBC) accounts for ~1%of breast cancers, with ~2,800 new US cases in 2025. MBC has a higher germline mutation frequency than female breast cancer, mostly the result of a higher frequency of BRCA2 mutations in males (10-15%) compared to females (5-10%). NCCN guidelines and VA clinical pathways recommend genetic testing for all men with breast cancer given high mutation frequency and treatment implications. However, real-world genetic testing access and survival outcomes in veterans remain understudied. With >90% male patients, the Veteran Affairs (VA) represents the largest MBC cohort nationwide, providing a unique opportunity to evaluate genetic testing utilization and survival outcomes. Methods Veterans with MBC (2019-2025) were identified using VA Cancer Registry (site code C50). Patients receiving VA care were defined as having ≥2 visits with oncology services. Clinical notes were analyzed using natural language processing and large language model (Llama 3.3 70B) to extract germline genetic testing (GGT) access. GGT access proportion was compared for year 2019-2021 vs. 2022-2024. Overall survival (OS) analysis was conducted and outcomes by GGT access were compared. Results 523 Veterans with male breast cancer were identified. Mean age was 72 years, 61% were White, 28% were Black and 4% were Hispanic. 75% lived in urban areas. 520 (99.43%) were ER/PR+, 46 (8.8%) HER2+, and 3 (57%) triple negative. Of 523 veterans, 334 (63.9%) had access to germline genetic testing at VA over the course of their disease. Patient characteristics were similar between GGT groups. Five-year OS was 70.9% overall, with significant difference by GGT access (78.6% vs. 59.3%, p<0.001). Genetic testing access rates significantly increased from 56.5% (134/237) in 2019-2021 to 70.6% (192/272) in 2022-2024, representing a 14.1 percentage point improvement (p < 0.001, χ² test). Conclusion The vast majority of veterans with MBC receive GGT, with increasing access over time. GGT access is associated with improved overall survival, suggesting potential benefits of access to comprehensive genetic evaluation. A more detailed covariate analysis is needed to further explore.

A. R. cabello, C. Cabello, S. Ramalho, C. E. Alem, S. R. Teixeira, L. R. Silva, G. S. Paiva, B. N. Duarte, M. Souza, T. cabello, D. I. Silva, A. Viaro, T. Gaspar, L. C. Zeferino; Obstetric and Gynecology, State University of campinas – UNICAMP, campinas, BRAZIL.

Distribution of Molecular Subtypes of Hereditary Breast Cancer in Brazilian Patients According to Pathogenic or Likely Pathogenic Germline Genetic Variants. Ana Elisa Ribeiro da Silva Cabello ¹, Cesar Cabello ¹, Bárbara Narciso Duarte ¹; Christine Elisabete Rubio Alem ¹; Sandra Regina Campos Teixeira; ¹ Leonardo Roberto da Silva ¹, Geisilene Russano Paiva ¹, Susana Oliveira Botelho Ramalho ¹, Thiago Cabello¹ ,Márcio Lopes de Souza ¹, Thiago Gaspar, ¹ Daniel Imbassahy SBC Silva ¹, Ana Lidia Viaro ¹, Luiz Carlos Zeferino ¹ 1.State University of Campinas (UNICAMP)Corresponding author: cesar cabelloEmail: cabello@unicamp.brAddress: Alexander Fleming, 101-Cidade Universitaria, City: Campinas State; São Paulo, ZIP Code: 13083-881 Objective: To assess the distribution of molecular subtypes of hereditary breast cancers, according to the presence of pathogenic or likely pathogenic germline genetic variants detected by multigene panel testing. Methods: A prospective cohort study was conducted from November 2021 to October 2022, involving women under follow-up at the high-risk outpatient clinic at the University of Campinas (UNICAMP). Patients were required to meet the following criteria: a personal history of luminal or HER2-positive breast cancer diagnosed before age 45, or triple-negative breast cancer diagnosed before age 60, along with at least one first- or second-degree relative with breast, ovarian, and/or prostate cancer, or who met the NCCN high-risk criteria (NCCN version 3.2021). Patients completed a questionnaire, after which they were referred for genetic testing. Following the test results, they received counseling based on the findings. Results: A total of 184 patients and 198 breasts were studied. Pathogenic variants (PV) or likely pathogenic variants (LPV) were detected in 36% (67/184) of the patients. Approximately 28% (51/184) were under 40 years of age, 30% (74/184) were non- Caucasian, and 49% (90/184) were premenopausal. General Analysis: The molecular subtypes most associated with pathogenic/probably pathogenic variants (PV/LPV) were Luminal B HER2-negative “like” (OR=11.76; CI 4.19-33.01) and triple-negative (OR=8.4; CI 3.45-20.43). These associations remained consistent even when the analysis was restricted to BRCA1 and BRCA2 genes only (OR=9.33; CI 2.00-29.99) (OR=9.03; CI 3.37-24.20) or to non-BRCA genes (OR=7.75;CI 2.96-29.44) (OR=3.48; CI 1.32-9.20) for Luminal B HER2-negative “like” triple-negative subtypes. Conclusion: Two molecular subtypes were associated with the presence of pathogenic or likely pathogenic variants in genes associated with hereditary breast cancer: Luminal B HER2-negative and triple-negative subtypes. Keywords: High risk; Germline variants; Genetic testing; Breast cancer; Molecular subtypes

E. Gasparini¹, M. Rotolo¹, G. Moretti¹, A. Bologna¹, R. Di Cicilia¹, C. Casartelli¹, L. Mangone², F. Leoni², C. Pinto¹, M. Dominici³, L. Cortesi³; ¹Department of oncology and advanced technologies, Comprehensive Cancer Centre, AUSL-IRCCS Reggio Emilia, Reggio Emilia, ITALY, ²Epidemiology Unit, Comprehensive Cancer Centre, AUSL-IRCCS Reggio Emilia, Reggio Emilia, ITALY, ³Department of Medical and Surgical Sciences, University of Modena and Reggio Emilia, Modena, ITALY.

Background Hereditary breast cancer (HBC) accounts for 5-10% of all breast cancers (BC). Many high and moderate penetrance genes are responsible for HBC. The identification of HBC led to the development of preventive strategies for BC and ovarian cancer (OC). In recent years BRCA1/2 germline pathogenic variants (gPV) have been used as predictive role for PARP inhibitors treatment of BC, OC and prostate cancer improving overall survival. So far, criteria for selecting patients to test remain fundamental to identify mutation carriers to offer adequate treatment. Among those, second BC represents one of the typical characteristics to take in account, mostly in case of metachronous development. Our study aimed to analyze differences in second BC between BRCA1/2 and other mutation carriers, in terms of age, clinical characteristics and prognosis at the Reggio Emilia Hereditary Center. Methods The Reggio Emilia Hereditary Cancer Center was established in 2012, identifying patients at high HBC risk, based on personal/family history and Tyrer-Cuzick likelihood model. Among criteria utilized for multigene panel testing (MGPT), two BC arisen until 50 years alone, or at any age associated to first-degree BC, were considered sufficient to be analyzed with Sophia Genetic Hereditary Cancer Syndrome (HCS) panel, including 27 predisposing genes. From the clinical point of view, patients and BC characteristics were registered in hospital-based record charts. The record chart registered height, weight, BMI, smoking habit, hormonal status, pregnancies, vital status, histology, clinical and pathological status, grading, hormonal receptors, proliferative rate, HER2 expression, type of surgery, type of neoadjuvant/adjuvant treatments, other tumors, prophylactic surgeries and gPV in different genes. The comparisons between BRCA1/2 and other gPV carriers were made by Fisher’ exact test or by Chi-squared test when more than two variables were analyzed. Loco-regional breast cancer-free survival (LRBCFS) was defined as the time from diagnosis of BC to the date of the first local recurrence (including diagnosis of second BC and regional relapse). Overall survival (OS) was defined as the time from diagnosis of BC to death or last follow-up. OS and LRBCFS curves were calculated using Kaplan-Meier method. Results Among 220 multiple BC patients analyzed by MGPT, 44 (20%) resulted gPV carriers in predisposing genes (26 in BRCA1/2 genes and 18 in other ones). Out of 18 not BRCA1/2 genes, 5 MUTYH, 3 CHEK2, 3 PALB2, 3 Mismatch Repair genes, 2 ATM, 1 RAD50 and 1 TP53 gPV carriers were identified. Regarding the first BC, BRCA1/2 carriers showed a younger age at onset (45 vs. 53 years), more advanced disease, grading III and triple negative BC (TNBC) phenotype. The median LRBCFS was equal to 11.2 years (12.6 vs 10.9 in BRCA1/2 and other carriers, respectively). Of interest, 95% of BRCA1/2 carriers had contralateral BC, whereas 27% of other carriers had ipsilateral BC. The second BC characteristics reflect those of the first BC. Other gPV carriers had more non-second BC than BRCA1/2 (17.6% vs.9.5%). The bilateral salpingo-oophorectomy was performed in 62% of BRCA1/2 vs. 12% of not BRCA1/2. Finally, 92% of all patients were alive at 17.5 median follow-up years. Conclusions Second BC represents an important feature to consider for eligibility to MGPT. A high rate of gPV in predisposing genes (20%) was found in patients with two BC. More TNBC were shown, as expected, in BRCA1/2 gPV carriers, underlying the need to treat those patients with more aggressive regimens. The older age at the first BC onset and the lower rate of contralateral BC in not BRCA1/2 gPV carriers do not justify a bilateral mastectomy at the surgical time. However, the OS curves do not differ between the two groups suggesting that the second BC do not impact on the patients’ outcome

C. Tran¹, A. Deshmukh², T. Doan³, N. Kethireddy³, M. Elia³, S. Bhardwaj³, I. Kang³, J. Mortimer³; ¹Medical Oncology, City of Hope, Upland, CA, ²Breast Health, Hologic, San Diego, CA, ³Medical Oncology, City of Hope, Duarte, CA.

Background: The Breast Cancer Index® or BCI™ test is a gene-based test that analyzes two key aspects of early-stage, hormone receptor-positive breast cancer. In a first aspect, the HOXB13:IL17BR ratio (H/I) assesses how well the cancer is likely to respond to continued hormone therapy. In a second aspect, the molecular grade index (MGI) predicts the likelihood of recurrence. This combination of features helps inform the appropriate duration of endocrine therapy beyond traditional tests. Importantly, the BCI test is effective for both lymph node-negative and lymph node-positive patients. The BCI test was incorporated into the NCCN Clinical Practice Guidelines in Oncology (NCCN® Guidelines) in January 2021 and the ASCO® Clinical Practice Guideline (ASCO Guideline) in April 2022. Additionally, Epic® medical record software was recently updated to allow users to order the BCI test. We examined the frequency of BCI test orders as a quality improvement metric, both at the main campus of the City of Hope Cancer Center and our community network.  Enhancing guidelines for Breast Cancer Index (BCI) testing and improvement is essential for elevating its quality and effectiveness in clinical settings. Methods: Data was obtained from Biotheranostics, Inc., for the period January 2020 to May 31,2025. We analyzed the data for testing rates with the BCI test over four time periods: pre-NCCN Guidelines (Jan 2020-Jan 2021), post-NCCN Guidelines (Jan 2021- April 2022), post- ASCO Guideline (April 2022-May 2024) and post-Epic medical record software integration (June 2024 – May 2025).   Results: A total of 1456 women with early-stage breast cancer underwent testing with the BCI test during the study period. Following the NCCN Guidelines update in January 2021, there was a notable increase in adoption of testing with the BCI test, with utilization increasing by 83% (p=0.026).  Following endorsement of the BCI test by the ASCO Guideline update in April 2022, there was a further increase in test ordering by 30% (p=0.22). Increased utilization of the BCI test was comparable at the main campus and our community network. The test was ordered comparably for all age groups, LN- and LN+ disease, HER2- and HER2+ patients, and pre- and post-menopausal status. The Epic software enhancement, implemented one year ago, has led to a significant year-over-year increase in test ordering: 38% on the main campus and 110% at our community site. Conclusion: The implementation of NCCN and ASCO Guidelines significantly increased BCI test utilization across all City of Hope sites. Furthermore, an Epic software enhancement dramatically improved adherence to these guidelines, leading to an even greater uptake of BCI testing.

S. Majumder¹, K. G. Gray², N. Shinde¹, R. Mawalkar¹, A. Zhang¹, M. Basree¹, K. Ormiston¹, X. Zhang³, R. Ganju⁴, G. Sizemore⁵, B. Ramaswamy¹, D. G. Stover¹; ¹Internal Medicine, The Ohio State University, Coulmbus, OH, ²Internal Medicine, Dana Farber Cancer Institute, Boston, MA, ³Center for Biostatistics, The Ohio State University, Coulmbus, OH, ⁴Pathology, The Ohio State University, Coulmbus, OH, ⁵Radiation Oncology, The Ohio State University, Coulmbus, OH.

Background: Epidemiological studies link lack of or short-term breast feeding to increased risk of breast cancer (BC), including triple negative breast cancer (TNBC). However, few models allow interrogation of the underlying biology. To decipher this link, we developed a model of short-term breastfeeding (BF) leading to abrupt involution (AI) versus long-term BF leading to gradual involution (GI) in mice. In our prior work, we showed that AI led to significant increase in estrogen signaling and precancerous changes (hyperplasia, squamous metaplasia) in mammary glands (MG) (Basree, et al Br Can Res 2019). Further, AI glands are characterized by a notable increase in alveolar progenitor (AP) population, the cells of origin of TNBC. In this study, we addressed the hypothesis that the anti-estrogen tamoxifen (TAM) could revert the adverse effects of abrupt involution/lack of breast feeding by deep interrogation of shifts in cellular phenotypes through CyTOF single cell proteomics. Methods: FVB/N female mice (~8 week old) were bred and allowed to nurse six pups/dam at partum. All pups were removed on postpartum (PP) day 7 to induce AI. On PP day 8, 5mg-60day sustained release tamoxifen pellet or placebo pellet (PLAC) was implanted in the subscapular region of the AI mice. MGs were harvested on PP day 28 and 120 from three groups: 1) TAM-treated AI; 2) PLAC-treated AI; 3) age matched nulliparous mice. FFPE sections were used for histology and immunohistochemistry. Immune cell composition and epithelial subpopulations from PP day 28 and day 120 MGs were evaluated by FACS and subjected to CyTOF using an established 40-antibody murine mammary targeted proteome panel (Gray, et al Cell Rep 2024). The Ingest algorithm was used to integrate and analyze mass cytometry datasets. Results: Overall, TAM treatment for 60 days completely abrogated the hyperplastic changes induced by AI in MGs, including reduction in proliferation index, collagen deposition, macrophages, and alveolar AP cells in the AI glands. As seen in our prior work, at day 120, abrupt involution-placebo glands (relative to nulliparous MGs) demonstrated marked enrichment of AP (1.5-fold) and basal (BA 1.6-fold) cells at the expense of hormone sensing (HS) cells. Analyses of MGs from the abrupt involution-tamoxifen group at day 120 revealed a significant increase (3-fold) in hormone sensing (HS) epithelial cells and an attendant decrease in AP cells compared to placebo. Further, abrupt involution-placebo glands show upregulation of AP markers (CD61, TSPAN8, SSEA1, CD14) and basal markers (CD54, CD47, CD104, CD49F) in the AP cells and higher expression of BA markers and reduced expression of HS markers in the HS cells. TAM treatment reversed these aberrant changes in abrupt involution-placebo glands in fact inducing a shift in HS cells to an involution-like state. Integration of the AI-PLAC/TAM data with lineage trajectory of these three cell types across the gestation-lactation-involution cycle revealed that AI-AP cells exhibited a mix of post-involutional and gestational states at day 120. TAM led to a 4.1-fold higher proportion of post-involutional-like AP cells and a 10.2-fold lower proportion of the aberrant gestational-like cells at day 120. Similar changes were observed in BA cells but not in HS cells upon TAM treatment. Conclusion: Further understanding of the biological impact of short/no breast feeding is critical and this work suggests that TAM treatment offers marked protection against AI-induced precancerous changes in a murine model. TAM restored long-term lineage balance after AI, altering AP differentiation towards a typical post-involution state as opposed to AI-induced hyperplasia. The corrective effect of TAM provides a rationale for short-term TAM treatment for women who are unable to BF to reduce risk of BC.

L. Yang¹, H. Harsanyi², I. Bacha³, S. Uttamapinan³, A. Toriola³; ¹Oncology, University of Calgary, Calgary, AB, CANADA, ²Cancer epidemiology and prevention research, Alberta Health Services, Calgary, AB, CANADA, ³Surgery, Washington University in St. Louis, St. Louis, MO.

Background: Hormonal contraceptive use in premenopausal women has been associated with a small but meaningfully higher risk of breast cancer. However, contemporary patterns of hormonal contraceptive use among premenopausal women in the United States (US) are unclear. Objective: To evaluate trends in and correlates of hormonal contraceptive use among premenopausal women in the US. Methods: Serial, cross-sectional analyses of hormonal contraceptive use using US nationally representative data from the National Health and Nutrition Examination Survey (NHANES) were conducted. Participants included noninstitutionalized premenopausal women from ten NHANES study cycles (1999-2000 to 2017-March 2020 pre-pandemic). The prevalence of hormonal contraceptive use was extracted from prescription medication data collected during household interviews. Hormonal contraceptive formulations were determined by hormone type. Trends in the prevalence of hormonal contraceptive use across study cycles were examined using multivariable logistic regressions, adjusting for sociodemographic factors. Results: Data on 15531 premenopausal women from ten NHANES cycles were analysed. From 1999 to 2020, the prevalence (95% CI) of hormonal contraceptive use significant increased from 5.2% (2.2 to 11.5%) to 15.5% (11.2 to 21.1%) for young women aged 15-<20 years, was stable among women aged 20-<40 years (from 16.9% [13.3 to 21.3%] to 15.8% [13.0 to 19.0%]), and remained stably low among those aged 40-<60 years (from 3.8% [1.5 to 9.3%] to 7.0% [4.0 to 12.1%]). Compared to non-Hispanic White women, non-Hispanic Black and Hispanic women consistently had lower prevalence of use but exhibited greater relative increases (prevalence ratio [PR]=3, 95% CI: 1.0 to 8.5, PR=2, 95% CI: 0.9 to 4.1, respectively) from 1999 to 2020. Combined estrogen and progestin formulation accounted for over 95% of the hormonal contraceptive prescriptions in 1999-2020, yet progestin formulation increased approximately five-fold, accounting for 16.3% (10.8 to 2.8%) of all prescriptions in 2017-2020. The prevalence of contraceptive use varied by race/ethnicity and family income to poverty ratio in all age groups, as well as by educational attainment, health insurance coverage, marital status, and weight status among women 20-<40 years. Conclusions and Relevance: In the past two decades, hormonal contraceptive use in the US significantly increased among young women aged 15-<20 years and among non-Hispanic Black premenopausal women of all age groups. Combined estrogen and progestin formulations account for the majority of hormonal contraceptive use; however, prescriptions for progestin formulations are increasing. Future investigations are needed to elucidate the implications of these changes in hormonal contraceptive use on the incidence of breast cancer.

P. Mora¹, N. Valdiviezo², M. Acosta¹, J. Herzog³, L. Reynaga³, L. Taxa⁴, T. Vidaurre², Y. Sullcahuaman¹, C. Castañeda², C. Munive², B. Muñante², J. Abugattas⁵, H. Guerra⁶, I. Otoya², V. Acuña⁷, S. Gruber³; ¹Unidad Funcional de genética y biología molecular, Instituto Nacional de Enfermedades Neoplásicas, Lima, PERU, ²Oncología Médica, Instituto Nacional de Enfermedades Neoplásicas, Lima, PERU, ³Center for Precision Medicine, City of Hope National Medical Center, Duarte, CA, ⁴Equipo Funcional de Patología Quirúrgica, Instituto Nacional de Enfermedades Neoplásicas, Lima, PERU, ⁵Cirugía de Mamas y Tejidos Blandos, Instituto Nacional de Enfermedades Neoplásicas, Lima, PERU, ⁶Anatomía Patológica y Patología Oncológica, Patólogos Especializados SAC, Lima, PERU, ⁷Facultad de Farmacia y Bioquímica, Universidad Nacional Mayor de San Marcos, Lima, PERU.

Pathogenic and likely pathogenic (P/LP) variants in BRCA1/2 are key markers of hereditary cancer, guiding clinical and preventive care. In Peru, late-stage diagnoses and limited genetic services make breast and ovarian cancer major public health issues. This study aimed to describe the clinical and genomic characteristics of Peruvian patients with breast and/or ovarian cancer who underwent BRCA1/2 testing between 2013 and 2020. Ten-year overall survival was analyzed using univariate and multivariate Cox regression, with left truncation to minimize immortal time bias. METHODS This study includes individuals diagnosed with breast and/or ovarian cancer referred to a Medical Genetics evaluation at the Instituto Nacional de Enfermedades Neoplásicas (INEN) during the period of August 2013 to March 2020. Individuals who met clinical criteria to be included in the inter-institutional protocol INEN13-48: “Estudio del Cáncer Hereditario y Familiar en Población Peruana” as part of the “Molecular Genetic Studies in Cancer patients and their relatives”; IRB #96144 City of Hope (COH); accessed genetic testing (carried out at COH). In patients diagnosed with both breast and ovarian cancer, only the first diagnosed cancer was considered for classification. RESULTS A total of 1,576 patients were included in the study, with a median age at diagnosis of 41 years. The vast majority were female (99.3%), and only 0.7% (n = 11) were male. Family history of cancer was absent in 61% of cases; 18% reported first-degree relatives with breast and/or ovarian cancer, while 20% reported other types of cancer. Breast cancer was diagnosed in 89.5% of patients (n = 1,410), ovarian cancer in 8.5% (n = 134), and both in 2% (n = 32). In breast cancer cases, the most frequent tumor subtype was HR+/HER2− (39.6%), followed by triple-negative (31.3%) and HER2+ (28.1%). Overall, we identified pathogenic variants in BRCA1 were identified in 7.2% (n = 114) of the whole population, and in BRCA2 in 3.2% (n = 51). Specifically, 6% of breast cancer patients carried a BRCA1 pathogenic variant, and 3% had a BRCA2 variant. BRCA1-positive breast tumors were predominantly of the triple-negative subtype (85.4%), whereas HR+/HER2- (47.6%) tumors were the most frequent among BRCA2-positive cases. In patients with ovarian cancer, 82% were BRCA-negative, while 14% carried BRCA1 and 2.8% BRCA2 pathogenic variants. Clinical stage II was the most common at presentation (39.3%), followed by stage III (38.8%), stage I (11%), and stage IV (10%). Additionally, 15% of patients developed a second primary malignancy, most commonly a new breast cancer (52%), followed by other cancers (15%), thyroid (12%), ovarian (8%), and colorectal cancer (6%). Notably, ovarian cancer diagnosis, advanced clinical stage, and triple-negative breast cancer subtype were significantly associated with poorer overall survival (p < 0.001 for all). CONCLUSIONSPatients carrying pathogenic BRCA variants were found to have a higher risk of developing a second primary cancer, mainly a new breast cancer, reinforcing the need for long-term surveillance in this population. In the subgroup of breast cancer patients with BRCA1 mutations, a higher incidence of triple-negative tumors was observed, whereas BRCA2-mutated patients more commonly exhibited luminal subtypes. These findings underscore the importance of strengthening genetic counseling services and implementing risk-reduction strategies as well as precision oncology programs. Given the higher mortality observed in patients with BRCA1 mutations, it is essential to understand the clinical and genomic profiles of these individuals to optimize cancer care for high-risk populations in the country.

O. Lee, A. Eremin, S. A. Khan; Surgery, Feinberg School of Medicine, Northwestern University, Chicago, IL.

Background: Tamoxifen (TAM) is the only anti-estrogen approved for women with functioning ovaries who are advised to avoid pregnancy due to risk to the fetus but are cautioned against using hormonal contraceptives (HCs). This reluctance is based on concerns that HCs may counteract the efficacy of TAM and increase adverse effects when used concomitantly. However, these concerns lack direct supporting evidence, and the possibility of a neutral or even beneficial interaction has not been explored. We hypothesized that the most widely used combined hormonal contraceptive (17α-ethinylestradiol + levonorgestrel) would not reduce the tumor-preventive efficacy of oral TAM in a hormonally sensitive, ER-positive mammary tumorigenesis model in rats. Methods: Female Sprague Dawley rats (4-5 weeks old) received N-Methyl nitrosourea (50 mg/kg i.p.) to induce mammary tumors. Three weeks later, they were randomized (n=12-13 per group) to control, HC alone, TAM alone, or TAM+HC. HC pellets (17α-ethinylestradiol 0.06 mg + levonorgestrel 0.28 mg, 90-day release; Innovative Research of America) were implanted subcutaneously in the dorsal area and replaced at 90 days. TAM was incorporated into the diet (1.03 mg/kg/day in rats, equivalent to daily 7 mg in humans; Inotiv Inc.). Rats were monitored for 6 months for tumor formation. Tumor incidence, latency, weights, and volumes were measured. Wet uterine weight was recorded, adjusted for body weight, and processed for histology. Tumor-free survival was analyzed with the log-rank test; other comparisons used Kruskal-Wallis with Dunn’s post hoc tests. Results: In the control group, median tumor latency was 3 months with a final incidence of 92%, consistent with published studies. Tumor incidence was higher in the HC alone group (46%) than in the TAM alone (17%) and TAM+HC (15%) groups. Tumor weights and volumes were significantly lower in both TAM-treated groups compared to controls (all adjusted p  0.99). TAM significantly reduced body weight versus control (median difference = 30 g, adjusted p = 0.001), with no difference between TAM alone and TAM+HC groups (median difference = 16 g, adjusted p = 0.73). HC alone had no significant effect on uterine weight (adjusted p > 0.99). TAM significantly reduced uterine weight relative to controls (adjusted p < 0.001), with no difference between TAM alone and TAM+HC groups (adjusted p = 0.97). Histopathological assessment of mammary glands, uterus, and endometrial thickness is ongoing. Conclusions: Combined HC did not reduce the tumor-preventive efficacy of oral TAM in this ER-positive mammary tumorigenesis model. Ongoing analyses will assess uterine safety with this combination.

H. Abdel-Razeq, F. Tamimi, B. Sharaf, S. Abdel-Razeq, O. El Khatib, A. Sharaf, H. Bani Hani, A. Al-Atary, L. El Saket, Y. Talab, H. Khalil, M. Harb, S. Yousef, A. Nofal, A. Alghrabli, M. Abunasser; Internal Medicine, King Hussein Cancer Center, Amman, JORDAN.

Background: Invasive lobular carcinoma (ILC) is the second most common histological subtype of breast cancer after invasive ductal cancer (IDC). Previous studies have shown that 5-10% of all breast cancers are associated with cancer-predisposing germline mutation. The frequency of these mutations in ILC is scarce. Here, we report on the pattern and frequencies of germline genetic mutations among Arab patients with ILC. Methods: In this retrospective observational study, electronic medical records were reviewed for all breast cancer patients diagnosed at King Hussein Cancer Center (KHCC) from January 2016 to May 2024. Included patients had histologically confirmed ILC and underwent germline genetic testing using restricted or expanded panels. Descriptive analyses were performed, with categorical variables reported as percentages. Chi-squared and Fisher’s exact tests were used to compare proportions of categorical variables between patients with pathogenic or likely pathogenic (P/LP) mutations and those without, as well as between BRCA and non-BRCA patients. Results: Out of the 6,539 patients who met the inclusion criteria, 477 (7.3%) had pure ILC histology. The median age at diagnosis was 49 years and only 13% were ≥ 65 years. Except for 2, all were female. Most patients (95%) were ER-positive, 92% PR-positive and 91% HER2 negative. Multicentric disease was observed in 43% of patients, bilateral disease in 14%, and 19% presented with metastatic disease. A family history of breast cancer was reported in 39% of patients, including 25% who had a first-degree relative with the disease. The most common additional malignancy observed in these patients were breast cancer (9.2%), with colorectal and endometrial cancers each at 0.8%, and ovarian cancer in only one patient. Regarding genetic testing, 12.6% had limited gene panels and 87.4% had extended panels. Overall, 47 (9.9%) carried a P/LP variant, most frequently in BRCA2 (n=25, 53.2%), followed by ATM and PALB2 (n=6 each, 12.8%). BRCA1 pathogenic variant was detected in only one patient. BRCA1/2 mutations were significantly more common in patients younger than 40 years; 17.0% compared to only 4.2% in older patients, p<0.001. Patients with P/LP variants presented at younger ages, had more frequent first-degree breast cancer family history (p<0.001), and were more likely to undergo risk-reduction surgery (p<0.001). Both male patients also carried P/LP variants. Conclusions: We found that many Jordanian patients with invasive lobular breast cancer had germline mutations, especially in BRCA2. Younger age at diagnosis and having a first-degree relative with breast cancer, were linked to a higher chance of having a pathogenic mutation. These results show the importance of genetic testing to guide preventive measures and treatment personalization. Larger future studies are needed to improve genetic testing guidelines and outcomes for these patients.

K. Kim¹, E. Kang², S. Baek³, H. Lee², S. Ahn⁴, W. Park¹, D. Shin², H. Lee⁵, S. Nam¹, S. Kim¹, J. Yu¹, B. Chae¹, L. Se Kyung¹, J. Lee¹, J. Ryu¹; ¹Division of Breast Surgery, Department of Surgery, Samsung medical center, Seoul, KOREA, REPUBLIC OF, ²Department of Surgery, Seoul National University Hospital, Seoul, KOREA, REPUBLIC OF, ³Department of Surgery, Yongin Severance Hospital, Yongin, KOREA, REPUBLIC OF, ⁴Department of Surgery, Gangnam Severance Hospital, Seoul, KOREA, REPUBLIC OF, ⁵Department of Pathology and Translational Genomics, Samsung medical center, Seoul, KOREA, REPUBLIC OF.

Background BRCA1/2 mutations are known to significantly increase the risk of contralateral breast cancer (CBC). However, their influence on changes in estrogen receptor (ER) status between primary breast cancer and metachronous CBC remains unclear. This study investigates ER status alterations in CBC patients with BRCA1/2 mutations to provide insights into their clinical implications.Methods This multicenter retrospective study was conducted between 2004 and 2020. Patients with unilateral early breast cancer who underwent BRCA1/2 testing and developed metachronous CBC were grouped by BRCA1/2 mutation status. ER status of primary and CBC tumors was compared using Fisher’s exact test and logistic regression to assess the relationship between BRCA1/2 mutations and ER alterations.Results Among 423 CBC patients, those with BRCA1 mutation had a higher likelihood of ER negative primary tumors transitioning to ER negative CBC (70.7%) compared to BRCA1/2 negative patients (odds ratio [OR] 3.8, p < 0.001). Similarly, ER-negative primary tumors were more likely to remain ER negative in CBC among BRCA1 or BRCA2 mutation carriers (64.8%, OR 2.9, p = 0.002). These findings demonstrate a strong association between BRCA1 mutations and the development of ER negative CBC in ER negative primary breast cancer. Of the patients with primary ER-negative breast cancer, those who received adjuvant chemotherapy for both primary and contralateral breast cancers showed a significantly higher rate of CBC with ER-negative alteration (OR 4.23, p=0.001).Conclusion BRCA1 mutation carriers face a significantly higher risk of developing ER negative metachronous CBC, in ER negative primary tumor. These findings highlight the importance of genetic counseling and risk-reducing strategies, including prophylactic mastectomy, for BRCA1 mutation carriers, which typically requires more aggressive treatments.

K. P. Verma¹, L. J. Oshry¹, M. Rudderman¹, I. Arshad², B. Moy³, L. Wu¹; ¹Department of Medicine, Section of Hematology and Oncology, Boston Medical Center, Boston, MA, ²Boston University School of Public Health, Boston Medical Center, Boston, MA, ³Department of Medicine, Division of Hematology and Oncology, Massachusetts General Hospital, Boston, MA.

Background: Germline genetic testing is recommended for individuals with personal or family history of certain cancers, as pathogenic variants (PV) can guide cancer screening and management. Variants of uncertain significance (VUS) present clinical challenges as guidelines recommend that VUS alone should not determine cancer management or decisions about risk-reducing surgeries. VUS are disproportionately reported in non-white individuals, reflecting the historic underrepresentation in genetic reference databases. These disparities complicate genetic counseling and exacerbate inequities in cancer care. Methods: The High Risk Cancer Genetics Clinic (HRC) was established at Boston Medical Center, the region’s largest safety net hospital, to improve access to cancer screening and surveillance. We conducted a retrospective cohort study of patients with a VUS in hereditary breast and ovarian cancer (HBOC) genes and compared surgical decision-making among patients found to have VUS two years before and after HRC establishment from May 2021 to April 2025. Male patients and those with metastatic disease were excluded. Results: There were 438 patients with at least one VUS in BRCA1, BRCA2, BRIP1, TP53, PALB2, PTEN, CDH1, STK11, RAD51C, or RAD51D genes, for which PV can inform risk-reducing bilateral mastectomy (BLMa) and/or bilateral salpingo-oophorectomy (BSO). A total of 158 patients were included, and 69% had Medicaid, Medicare, or were uninsured, 78% self-identified as non-white, and 46% spoke a language other than English. After HRC, significantly more patients with HBOC VUS were referred to see genetic counselors (87% vs 96%; p=0.05) and HRC physicians (0 vs 10%, p=0.006). Before and after HRC establishment, there were no cases of BLMa among patients with a VUS who did not have breast cancer. Similarly, patients with a VUS who did not have ovarian cancer did not undergo BSO. A small proportion of breast cancer patients with a VUS had BLMa before and after HRC (5% vs 12%, p=0.64). Two patients with BRCA1 and PTEN VUS were reclassified as likely PV, and the HRC facilitated updated screening guidance that led to timely olaparib initiation and detection of localized breast cancer, respectively. Conclusion: At our safety-net hospital serving a diverse community, the management of patients with HBOC VUS has remained guideline concordant, with no risk-reducing surgeries performed solely based on a VUS result. Among breast cancer patients with an HBOC VUS who elected BLMa, all had documented counseling regarding the lack of survival benefit compared to breast conserving surgery, but chose BLMa for cosmetic or personal reasons. The integration of HRC has enabled timely re-evaluation and clinical action when a VUS is upgraded to PV. This demonstrates the value of multidisciplinary risk assessment in improving diagnostic and therapeutic outcomes for underserved populations.

M. Hou¹, F. Chen¹, C. Li¹, C. Chiang²; ¹Division of Breast Oncology and Surgery, Kaohsiung Medical University Hopsital, Kaohsiung, TAIWAN, ²Department of Biomedical Sciences, Chung Shan Medical University, Taichung, TAIWAN.

BackgroundMost breast cancer-related death is caused by recurrence and metastasis. Approximately 30% to 40% of patients will experience recurrence during their lifetime. Although the promising effect of endocrine therapy in ER-positive breast cancer, there are still about 30% of patients who ultimately die due to recurrence. In contrast to ER-negative, more than 50% of recurrences occur beyond five years (late recurrence, LR) in ER-positive breast cancer, which represents the proportion of early (ER, recurrence occur within five years) or even very early recurrence (VER, recurrence occur within one year in our preliminary results) are rare. In comparison with LR, the prognosis of ER and VER is poor and more aggressive; however, most studies focused on LR, including mechanisms, treatment, and genetic/immune profiles, few studies focused on ER, and even No studies concerned what we named VER. On the other hand, the introduction of a multigene prognostic test (MPT) can provide the dual role of prognostic and predictive to fulfill and strengthen the information in treatment decisions. However, there still existed some defects in the current MPT, including high cost, ethnic difference, and specificity for different time points of recurrence type. Thus, there is a huge gap in ER or even VER, and a critical unmet need to identify who is prone to recurrence for Asia breast cancer.MethodsFifty ER-positive patients with disease-free (DF, average follow-up 48 months), and 14 ER-positive patients with recurrence were enrolled in this study. Initially, three patients from 50 patients with DF and nine patients from 14 patients with recurrence were selected. In total of 12 patients (three patients for each group, including DF, VER, ER, and LR) were selected, and residual RNA samples from EndoPredict (EP) were analyzed by using the NanoString technique. Finally, raw data processing, normalization, and analysis were performed in the nSolver analysis software. Appropriate statistical analyses were performed using GraphPad Prism 9. ResultsIn our preliminary results, we collected the data with EndoPredict and long-term follow-up records. By using the NanoString technique, we found some novel and unique target genes in different groups of recurrence. The expression of SERPINA1 was specific to disease-free (DF, average follow-up 48 months), CEACAM6 was overexpressed in LR group (average PFS 45 months), TGFB2 was only observed in VER group (average PFS 8 months), and AREG was co-existed in the group of ER (average PFS 23 months) and VER. Moreover, we compared the scores from a low-cost platform of immunohistochemical 4 (IHC4) and EndoPredict within 64 patients (average follow-up 48 months) and found a positive correlation between the IHC4 risk score and EPClin score in prognosis prediction.ConclusionTo our knowledge, we provide the first preliminary results of novel gene targets at different time points of recurrence, especially for VER. We also found a positive correlation between the IHC4 risk score and EPClin score in prognosis prediction. The ability of prediction and mechanisms in recurrent breast cancer from these target genes will be validated in a large cohort and requires further investigations in the future.

B. Oh¹, S. Choi², S. Seo³, H. Park⁴, E. Connolly-Strong¹, T. Kim⁵, Y. Ju⁶, J. Hwang⁷, Y. Park⁸; ¹Medical Center, Inocras Inc,, San Diego, CA, ²Department of Internal Medicine, Division of Clinical Genomic Medicine(Hematology & Medical Oncology, Seoul National University Bundang Hospital, Seoul National University College of Medicine, Seongnam-si, Gyeonggi-do, KOREA, REPUBLIC OF, ³Department of Laboratory Medicine, Seoul National University Bundang Hospital, Seoul National University College of Medicine, Seongnam-si, Gyeonggi-do, KOREA, REPUBLIC OF, ⁴Genome Science, Bioinformatics, Inocras Inc,, San Diego, CA, ⁵Biology, Johns Hopkins University, Baltimore, MD, ⁶Graduate School of Medical Science and Engineering, Korea Advanced Institute of Science and Technology, Daejeon, KOREA, REPUBLIC OF, ⁷Department of Internal Medicine, Seoul National University Bundang Hospital, Seongnam-si, Gyeonggi-do, KOREA, REPUBLIC OF, ⁸Division of Hematology-Oncology, Department of Medicine, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, KOREA, REPUBLIC OF.

Introduction: Detecting pathogenic germline variants in BRCA1/2 is crucial for hereditary cancer risk assessment and guiding targeted therapies. While targeted NGS panels are effective for SNVs and indels, they often miss large structural variations (SVs), leading to false negatives. Whole-genome sequencing (WGS) offers a comprehensive solution by providing uniform, genome-wide coverage for robust SV detection. This report presents two clinical cases where WGS identified an identical, complex pathogenic BRCA1 SV missed by conventional panel testing.Methods: DNA was extracted from fresh-frozen tumor tissue and matched normal blood samples from two patients: a 40-year-old premenopausal female with breast and ovarian double primary cancers and a female in her 60s with metastatic pancreatic cancer (history of breast cancer). WGS analysis was performed using the CancerVision™ system.Results: Both patients initially had non-decisive hereditary cancer panel results. WGS revealed an identical, complex pathogenic germline BRCA1 variant in both: a large deletion of exons 12-14 coupled with an adjacent inversion. This BRCA1 deletion was successfully orthogonally confirmed by MLPA. The first patient’s breast tumor showed BRCA1 loss of heterozygosity (LOH), a high HRD score (0.93), 40.0% of base substitutions attributed to SBS3 (indicating flawed homologous recombination-based DNA repair), and a high prevalence of SV duplications (45.8%), suggesting a complete BRCA1 defect. The second patient’s tumor also showed BRCA1 LOH, 33.7% of base substitutions attributed to SBS3, and high SV duplications (33.7%), collectively suggesting impaired BRCA1 function. This patient achieved and maintained a complete response to platinum-based chemotherapy.Discussion: This study demonstrates WGS’s transformative power in resolving clinically important cases missed by conventional genomic approaches. The key finding is the identification of an identical, complex recurrent BRCA1 SV in two unrelated patients, providing definitive molecular diagnoses that significantly impacted clinical management. This recurrence suggests a potential underdiagnosed founder mutation. WGS’s unparalleled advantage lies in its uniform, genome-wide coverage, enabling simultaneous detection of diverse genomic variations, including complex SVs, ensuring a comprehensive and unbiased genetic landscape elucidation. WGS provides precise genomic breakpoint locations, allowing unequivocal determination of variant identity across patients and facilitating tracking of origins and population-specific prevalence. These cases underscore WGS’s critical role in identifying pathogenic SVs undetectable by standard methods, advocating for its integration into clinical diagnostics for high-risk patients to enhance accuracy and support personalized treatment strategies.

V. Ro¹, J. B. Gibbons¹, C. N. Ta², C. Liu², W. K. Chung³, C. Weng², K. D. Crew¹; ¹Hematology and Oncology, Columbia University Irving Medical Center, New York, NY, ²Biomedical Informatics, Columbia University Irving Medical Center, New York, NY, ³Pediatrics, Boston Children’s Hospital, Harvard Medical School, Boston, MA.

Introduction: Allostatic load (AL) is an index that accounts for cumulative stress on the body and is associated with increased breast cancer risk and breast cancer mortality. We aimed to determine the association between demographic, clinical and lifestyle factors and AL among women undergoing screening mammography.Methods: Patients were enrolled in the Electronic Medical Records and Genomics (eMERGE4) study at Columbia University Irving Medical Center and completed surveys on demographic, clinical and lifestyle factors, provided samples for germline BRCA1/2 and PALB2 genetic testing and polygenic risk score (PRS) for breast cancer (308-SNP model), and allowed access to electronic medical record (EMR) data. Women with at least one screening mammogram were included. To calculate AL score, we extracted the most recent EMR data on 9 biometrics representing major regulatory systems: cardiovascular (pulse, systolic and diastolic blood pressure), metabolic (albumin, alkaline phosphatase, glucose), renal (blood urea nitrogen, creatinine), and immune (white blood cells). One point was assigned to each biometric in the highest quartile except albumin which was the lowest quartile. Patients with at least 5 biometrics were included; missing values were imputed. High AL was defined as AL score above the cohort median. Demographic, clinical, and lifestyle factors from pre-test survey data included age, race/ethnicity, highest educational level, annual household income, health insurance, disability, body mass index (BMI), smoking and alcohol use, and physical activity level. We extracted breast density (4-category BIRADS classification) and prior breast biopsy results from the EMR. Univariable and multivariable logistic regression analyses were conducted to evaluate the association between high AL and demographic, clinical, and lifestyle factors.Results: Among 955 evaluable women, the mean age was 54.8 years (standard deviation [SD] 11.4), with 48.2% Hispanic, 34.9% non-Hispanic White, 9.1% non-Hispanic Black, and 7.9% Asian/other. Median AL score was 2 (range, 0-8), and 418 patients had high AL. Mean BMI was 27.7 kg/m² (SD 6.3). A total of 10.4% of the patients had a prior history of breast cancer. Most patients (61.7%) reported 0 days of vigorous exercise a week. Over half (54.6%) have a bachelor’s or advanced degree. Almost half (45.5%) of those with high AL have Medicaid versus 25.1% of those with low AL. In univariable analysis, older age, Black race, Hispanic ethnicity, higher breast cancer PRS, lower breast density, having a disability, lower educational attainment, lower income, public health insurance, higher BMI, less alcohol use, and less physical activity were associated with high AL. In multivariable analysis, older age (odds ratio [OR]=1.03, 95% confidence interval [CI]=1.02-1.05), higher BMI (OR=1.09, 95% CI=1.06-1.12), and having Medicaid vs. private health insurance (OR=2.21, 95% CI=1.30-3.75) were significantly associated with high AL score.Conclusions: In our diverse population of women undergoing screening mammography, we found a significant association between clinical, demographic, and socioeconomic factors and AL, which is a well-established breast cancer risk factor. In addition to age and BMI, having Medicaid compared to private health insurance was associated with a greater than 2-fold likelihood of having high AL. Although protective against breast cancer risk, lower breast density and less alcohol use were also associated with higher AL, warranting further investigation. These findings highlight the importance of addressing modifiable risk factors such as obesity for reducing AL and breast cancer risk and the need to address psychosocial stressors in vulnerable populations with Medicaid.

L. F. Martinez-Caudillo, M. S. Guzman-Garcia, V. Leitzelar-Bueso, J. E. Guzman-Garcia, D. Aguilar-Y-Mendez; Escuela de Medicina y Ciencias de la Salud, Tecnológico de Monterrey, Monterrey, MEXICO.

Introduction: Male breast cancer (MBC) accounts for 1.8% of all breast cancer cases worldwide. Compared to women, men tend to have poorer prognosis, mainly due to delayed diagnosis at an older age, when comorbidities are more prevalent. Risk factors can be divided into modifiable and non-modifiable. Among the non-modifiable, genetic predisposition -mutations in BRCA1 and BRCA2- play a major role. However others include: family history, Klinefelter Syndrome, and advanced age (>65 years). In contrast, modifiable risk factors include tobacco use, obesity, and alcohol. The mechanism for the latter remains unclear, but is found to be widely associated with MBC. Understanding key risk factors may aid in early diagnosis and diminish mortality rates. Given its low incidence, characterization of cohorts may help to identify relevant risk factors, improve early detection strategies and contribute to a better understanding of MBC. Methods: We conducted an observational retrospective study including eleven male patients diagnosed with breast cancer between 2016 and 2025 at the Breast Cancer Institute of the Zambrano Hellion Hospital, in northeastern Mexico. The following risk factors were evaluated, non-modifiable factors included: Klinefelter syndrome, advanced age, family history of breast cancer, breast cancer susceptibility genes (BRCA1, BRCA2, among others); modifiable: obesity (BMI > 30), alcohol and/or tobacco use, exogenous estrogen or testosterone use, ionizing radiation exposure, gynecomastia, bone fracture after the age of 45, cryptorchidism. Descriptive statistical analysis was applied. Results: Eleven male patients diagnosed with MBC were included in the analysis. Median age at diagnosis was 54 years with a SD of 13.92 and a range varying from 34-74. The most prevalent modifiable risk factors were alcohol consumption (75%) and tobacco use (63%), while obesity was present in 25% of patients. Among non-modifiable risk factors breast cancer susceptibility genes and advanced age prevailed with 38% and 27% respectively. Family history of cancer was found in 45%, more information in Table 1. Conclusions: This cohort-based analysis highlights important deviations from existing literature, particularly regarding age at diagnosis, having a median of 54 and cases as early as 34 years; despite the known association of MBC with advanced age (>65). Modifiable risk factors such as alcohol consumption and tobacco use were highly prevalent, detailing potential prevention opportunities. Moreover, a notable proportion presented family history of cancer, reinforcing the need for thorough familial assessment in clinical practice. Although limited by cohort size, this study contributes valuable insights to the scarce literature on MBC, and supports the need for larger, population-based research to better characterize risk profiles and detail early detection strategies.

M. S. Guzman-Garcia¹, J. E. Guzman-Garcia¹, V. Leitzelar-Bueso¹, L. F. Martinez-Caudillo¹, D. Aguilar-Y-Mendez², C. Villarreal-Garza²; ¹Escuela de Medicina y Ciencias de la Salud del Tecnologico de Monterrey, Tecnologico de Monterrey, Monterrey, MEXICO, ²Breast Cancer Institute, Hospital Zambrano Hellion, San Pedro Garza Garcia, MEXICO.

Introduction: Variants of uncertain significance (VUS) are variants in the genetic code for which the impact on the protein function is unclear. VUS estimated probability in genes associated with hereditary breast and ovarian cancer (HBOC) is 23%, and the rate of reclassification varies from 19-27% depending on the population studied. It represents a challenge to medical management and surveillance decisions. Methods: From 2018 to 2024 a total of 334 VUS have been reported among 1014 breast cancer patients at the Breast Cancer Institute of Zambrano Hellion Hospital. A retrospective clinical-genomic review was conducted, including only patients with breast cancer.Data were extracted from electronic medical records, including: age at diagnosis, tumor laterality, immunohistochemical (IHC) profile, VUS-related gene(s), and family history of cancer (e.g., breast, ovarian, pancreas). For each case, we documented the gene in which the VUS was identified, whether it underwent reclassification, the final classification (pathogenic, likely pathogenic, or benign), and the time elapsed (in months) until reclassification. Results: 19 VUS were identified across multiple cancer susceptibility genes (Table 1). Of these, 9 variants were reclassified: 55% as benign and 45% as likely pathogenic. The mean time to reclassification was 26.44 ± 12.6 months (range 7-52).Table 1 summarizes the number of VUS per patient and their corresponding reclassification time. A total of 9 female breast cancer patients carriers of at least one of these VUS reclassified with or without a pathogenic variant. The average age at diagnosis was 40.3 years (SD ±5.32). All BC diagnoses were unilateral. Four patients (44%) presented with triple-negative breast cancer (TNBC) by immunohistochemistry (IHC). The most frequently involved gene was NF1, accounting for 5 variants (55%), of which 4 were reclassified as benign and 1 as likely pathogenic. BRCA2 presented 2 variants (22%), with one benign and one likely pathogenic outcome. BRCA1 and POLE each had 1 variant (11%), both of which were reclassified as likely pathogenic. Conclusions: The observed rate of reclassification may be influenced by the underrepresentation of Hispanic populations in genetic databases and limited access to genetic testing. Given the potential clinical implications of VUS in HBOC-associated genes, further investigation is essential to improve variant classification and optimize genetic counseling. These variants are increasingly reclassified through the integration of population-specific data, functional assays, co-occurrence with known pathogenic mutations, and other lines of evidence.

H. Abdel-Razeq, L. El Saket, H. Bani Hani, S. Abdel-Razeq, F. Tamimi, B. Sharaf, Y. Talab, A. Al-Atary, M. El-Atrash, T. Al-Batsh, M. Horani, O. Othman, Y. Al-Masri, O. El Khatib, M. Abunasser; Internal Medicine, King Hussein Cancer Center, Amman, JORDAN.

Background: ATM gene alterations have increasingly been recognized for their role in breast cancer susceptibility and therapeutic implications. However, the phenotypic and molecular features of ATM-mutant breast cancers remain under-characterized in Middle Eastern populations. This study provides a comprehensive analysis of breast cancer patients harboring pathogenic or likely pathogenic (P/LP) ATM variants treated in Jordan, with emphasis on clinicopathologic features, genetic context, and therapeutic outcomes. While ATM mutations confer a moderate lifetime breast cancer risk (25%), their impact on treatment decisions remains under investigation. Methods: We retrospectively reviewed data on 52 female patients with breast cancer and confirmed pathogenic or likely pathogenic ATM mutations. Parameters analyzed included age at diagnosis, family history, histopathology, stage, hormone receptor status, Ki-67 proliferation index, and genetic testing eligibility based on NCCN/ASCO guidelines. Co-occurring variants in other genes and variant types (missense, nonsense, frameshift, splice site) were compiled. Results: The median (range) age at diagnosis was 48 (27-79) years, with 32 (61.5%) diagnosed before age 50. Invasive ductal carcinoma (IDC) was the predominant histologic subtype (n=41, 78.8%), followed by invasive lobular carcinoma (ILC). Hormone receptor (HR) positivity was high: ER-positive (92.3%), PR-positive (90.4%), and HER2-negative in most cases (n=36, 70.6%). High-grade tumors (GIII) comprised 39.2% of the cohort. The majority of ATM variants were deletions (frameshift, 46.2%), followed by nonsense (17.3%), missense (15.4%), and splice site alterations (11.5%). Common hotspots included c.2284_2285del and c.5755C>T. Co-occurrence with other susceptibility genes was notable; 6 (11.5%) had additional pathogenic variants in other genes: 3 with BRCA2, 1 with RAD50, 1 with MUTYH, and 1 with a second distinct ATM mutation. Furthermore, 16 of the 52 patients (30.8%) harbored additional variants of uncertain significance (VUS). Despite a high prevalence of early-stage disease 39 (75.0%), lymph node involvement was observed in 27 (51.9%) cases. Neoadjuvant chemotherapy was utilized in 16 (32.7%) patients with a modest rate of pathologic complete response (12.2%). Notably, 5 (9.6%) of patients did not meet testing criteria per conventional guidelines, suggesting gaps in current eligibility frameworks, particularly for patients under universal genetic testing. Conclusions: ATM-mutant breast cancers in this Jordanian cohort exhibited a predominance of HR-positive, high-grade tumors with frequent lymph node involvement. The diversity in variant types and co-existing mutations underlines the clinical complexity of these cases. A proportion of patients would have been missed by existing genetic testing guidelines, emphasizing the utility of universal testing in this setting. These findings advocate for broader germline screening in Middle Eastern populations to inform precision oncology strategies.

H. Abdel-Razeq, H. Bani Hani, F. Tamimi, L. El Saket, S. Abdel-Razeq, B. Sharaf, H. Abu-Jaish, Y. Talab, A. Al-Atary, M. El-Atrash, T. Al-Batsh, M. Horani, L. Barakat, Y. Al-Atary, O. El Khatib; Internal Medicine, King Hussein Cancer Center, Amman, JORDAN.

Background: The CHEK2 (Checkpoint Kinase 2) gene plays a central role in DNA damage repair and cell cycle regulation. Germline pathogenic or likely pathogenic (P/LP) CHEK2 variants confer a moderate risk of breast cancer, especially estrogen receptor (ER)-positive subtypes, and have been associated with other malignancies, including colorectal and thyroid cancers. Despite CHEK2 being one of the most frequently mutated genes identified in hereditary cancer panels, clinical data on variant distribution, tumor characteristics, and associated malignancies remain underreported in many populations. This analysis is the largest of its kind in a Middle Eastern population. Methods: We conducted a retrospective review of breast cancer patients with germline P/LP CHEK2 mutations identified through multigene panel testing (20-gene or 84-gene panels). Testing was done at one of three accredited international reference genomic labs. Clinical variables included demographics, tumor histology and grade, hormone receptor status, treatment modalities, family history, and the presence of second primary cancers. Genetic variant analysis and co-occurrence with additional mutations were also documented. Results: Fifty-nine female CHEK2-positive breast cancer patients were identified. Median age at diagnosis was 48 years (range: 28-77), with 59.3% diagnosed at or before age 50. The majority had invasive ductal carcinoma (n=48, 81.4%) and grade 1 or grade 2 tumors (73.1%). Hormone receptor status showed that all patients were estrogen receptor (ER)-positive, while progesterone receptor (PR) positivity was observed in 86.2% of cases (n=50). HER2 overexpression was observed in 23.2% of cases. Neoadjuvant therapy was administered in 48.1% of patients, and 52.5% underwent mastectomy. Family history of cancer was reported by 33 (91.5%) patients; 81.5% had at least one relative with breast cancer. The most frequent CHEK2 variant was c.592+3A>T (39.0%), followed by c.499G>A (20.3%). Notably, 5 patients (8.5%) carried additional germline P/LP variants: APC I1307K (n=3), BRCA2 (n=1), and co-occurring PALB2 + RAD51D (n=1). Second primary cancers were identified in 16 (27.1%) patients and mostly were breast cancer (n=12, 20.3%) and one case each of colorectal, endometrial, pancreatic and lymphoma. Among those who underwent screening colonoscopy (n=25), 3 patients (12.0%) were diagnosed with colorectal cancer, and 3 others (12.0%) had tubular adenomas and hyperplastic polyps. Conclusions: This study provides insight into the clinical and molecular landscape of CHEK2-positive breast cancer patients in a non-Western cohort. The high prevalence of the c.592+3A>T variant may suggest regional founder effects. The identification of additional actionable and the high rate of second primary cancers support the use of comprehensive multigene panels in hereditary cancer testing. These findings underscore the importance of tailored surveillance and risk-reduction strategies, including screening for second primary malignancies, in CHEK2 mutation carriers.

K. Papazisis¹, M. Paraskeva², S. Giassas³, M. Skondra⁴, R. Iosifidou⁵, C. Tolis⁶, G. Kesisis⁷, E. Zairi⁷, V. Venizelos⁸, C. Markopoulos⁹, I. Natsiopoulos¹⁰, E. Bleka¹¹, I. Xanthakis¹⁰, A. Ananiadis⁷, D. Matheos¹², D. Stefanou¹³, A. Adamidis¹⁰, A. Meintani¹⁴, G. Kapetsis¹⁴, D. Bouzarelou¹⁴, E. Papadopoulou¹⁴, G. Nasioulas¹⁴; ¹3rd Dpt of Oncology, Interbalcan European Medical Centre, Thessaloniki, GREECE, ²1st Department of Oncology, Rhodes General Hospital, Rhodes, GREECE, ³Dpt of Oncology, IASO Hospital, Athens, GREECE, ⁴Dpt of Oncology, Henry Dunant Hospital, Athens, GREECE, ⁵reast Surgical Oncology Clinic, Theageneio Anticancer Hospital, Thessaloniki, GREECE, ⁶Private practice, Private practice, Ioannina, GREECE, ⁷Dpt of Oncology, Agios Loukas Hospital, Thessaloniki, GREECE, ⁸Dpt of Oncology, Metropolitan Hospital – Persefs Ygionomiki Merimna AE, Piraeus, GREECE, ⁹Dpt of Oncology, Athens Medical center, Athens, GREECE, ¹⁰Dpt of Oncology, Interbalcan European Medical Centre, Thessaloniki, GREECE, ¹¹Dpt of Oncology, Euromedica Kyanous Stavros, Thessaloniki, GREECE, ¹²Dpt of Oncology, General Hospital of Alexandroupoli, Alexandroupoli, GREECE, ¹³Dpt of Oncology, Laiko Hospital, Athens, GREECE, ¹⁴Molecular Oncology, GENEKOR MEDICAL SA, Athens, GREECE.

Background Identifying pathogenic/likely pathogenic variants in high-risk cancer genes is key to cancer risk management.Prophylactic surgeries like RRM and RRSO are guideline-recommended options. This study aims to provide real-world evidence on their clinical implementation. Methods Between 2020 and 2025, 203 women referred for genetic testing at Genekor Medical S.A., were detected with P/LP variants in cancer susceptibility genes, for which risk-reducing surgical options are either recommended or considered for discussion. Referring clinicians were invited to complete a questionnaire to assess whether prophylactic surgical options had been discussed, if the patients had agreed to undergo such procedures, and whether the surgeries have already been performed or postponed. Results Among the individuals referred for genetic testing, 168(83%) were referred following a cancer diagnosis, while 35(17%) were unaffected individuals referred due to a family history. 153 individuals (76%) harbored BRCA1/2, 31(15%) harbored PALB2, PTEN, TP53 and 19 (9%) harbored BRIP1,RAD51C, RAD51D P/LP variants. RRSO was discussed with 78/153 (51%) of BRCA1/2 carriers and agreed in 60/78 (77%) cases. Addition of hysterectomy to RRSO was discussed in 16/78 (21%) of cases and agreed in 10/16 (63%). RRM was discussed with 113/203 (56%) of patients and agreed in 73/113(65%). In PALB2, PTEN, TP53 carriers, RRM was discussed in 21/31 (68%) cases, and accepted in 9/21 (43%), while in 5 cases both RRM/RRSO were discussed with 3 women agreeing. Finally, for 19 BRIP1,RAD51C, RAD51D carriers, in 9 cases the physician discussed RRSO as an option, with the examinees agreeing in 7 of them and in 5 cases RRM was discussed with agreement to proceed in 2 of them. In all cases, 22 individuals although agreed to proceed with RRM and/or RRSO, decided to postpone the procedure. Conclusions This study highlights the high awareness and acceptance of surgical management demonstrating the adherence to the recommendations. Age under 35 years and receiving therapy at the time of genetic testing were significant factors in lowering risk reduction RRSO and hysterectomy performance rates. Stage IV cancer was the main reason for not discussing surgical intervention.

F. Xu¹, K. Jiang¹, R. Chen², H. Wu¹, Y. Zhong¹, H. Liu², W. Lyu², S. Mo¹, H. Chen¹; ¹Medical Oncology, Sun Yat-sen University Cancer Center, Guangzhou, CHINA, ²Breast Surgery, Hainan General Hospital (Hainan Affiliated Hospital of Hainan Medical University), Haikou, CHINA.

Background: PIK3CA mutations are established therapeutic targets in breast cancer (BC). The population-level prevalence has been explored, but the distribution across key clinicopathological strata in Chinese patients (pts) remains to be elucidated. Methods: This retrospective study enrolled BC pts diagnosed between January 1997 and February 2025 with available PIK3CA testing (next-generation sequencing) results. Pts were divided based on PIK3CA mutation status into mutated and wild-type (WT) groups. Mutation status was defined as any positive test result during follow-up. Clinicopathological characteristics and prognosis were compared between the two groups. Results: This study enrolled 577 pts, divided into PIK3CA mutated group (n=226) and the WT group (n=351). PIK3CA mutation frequencies varied across molecular subtypes: highest in Luminal A (LA) (56.4%), followed in frequency by LB1 (45.8%), HER2-positive (45.5%), and Luminal B2 (LB2) (40.5%), but lower in triple-negative breast cancer (TNBC) (21.4%). Among mutated tumors, H1047R was predominant (42.5%), followed by E545K (17.7%) and E542K (11.5%). H1047R mutations showed higher prevalence in HER2+ (22.7% of HER2+ tumors) and LA (21.8%) subtypes compared to TNBC (11.7%). E545K mutations were enriched in LB2 (11.9%) and LA (11.5%) subtypes, while being rare in TNBC (2.8%). BRCA mutations were significantly less frequent in PIK3CA-mutated pts (mutated vs. WT: 4.0% vs. 16.2%, P0.05) (Table 1). After a median follow-up of 38.2 months, no significant survival difference was observed. Five-year overall survival (OS) rates were 93.1% (95%CI: 88.9-97.5%) for the PIK3CA mutated group compared to 90.2% (95%CI: 85.9-94.8%) for the PIK3CA WT group. A molecular subtype-stratified survival analysis revealed that pts with the LA subtype harboring PIK3CA mutations had a 5-year OS rate of 92.5% (95% CI: 82.9-100.0%), while PIK3CA wild-type LA patients achieved a 5-year OS rate of 100.0% (95% CI: 100.0-100.0%) (P=0.035). In the stratification analysis of E545K, the PIK3CA WT group had a 5-year OS rate of 90.2% (95% CI: 85.9-94.8%). In contrast, pts with other PIK3CA mutations (non-E545K, n=186) achieved a 5-year OS rate of 95.3% (95% CI: 91.6-99.1%). The E545K mutation group (n=40) showed the lowest 5-year OS rate at 84.4% (95% CI: 71.1-100.0%). Conclusions: This real-world study identifies distinct clinicopathological profiles linked to PIK3CA mutations in Chinese BC pts, characterized by LA subtype enrichment and H1047R dominance. Notably, PIK3CA mutations are associated with poor survival in LA tumors and E545K mutations. These findings underscore the significant impact of PIK3CA mutations on the survival of Chinese BC pts.

S. S. Rao¹, V. V. Shetty², V. Rajendra³; ¹Endocrine & Breast Surgery, Oncology, KS Hegde Medical Academy, Mangalore, INDIA, ²Medical Oncology, KS Hegde Medical Academy, Mangalore, INDIA, ³Surgical Oncology, KS Hegde Medical Academy, Mangalore, INDIA.

Background:Healthcare workers (HCWs) play a pivotal role in early detection, treatment, prevention and education strategies for breast cancer. However, there is limited data assessing their own levels of awareness and knowledge in resource-constraint settings. This survey aims to identify the gaps in breast cancer awareness among HCWs in Mangalore, a prominent healthcare hub in Southern India. Methods:A cross-sectional, questionnaire-based study was conducted from June to July 2025 among 45 HCWs, including doctors from non-oncological specialties, nurses, paramedical staff, interns, and resident doctors, across four tertiary care hospitals in Mangalore. The validated questionnaire assessed knowledge of breast cancer risk factors, symptoms, available treatment options and attitudes toward breast self-examination (BSE) and clinical breast examination (CBE). Data were analysed using SPSS v28; Chi-square tests determined statistical significance. Results:Majority of the participants were final year students and residents (65%). None of them were breast cancer survivors. Few of them reported family history of breast cancer (6.5%). In first and second degree relatives (6%). Surprisingly, only 11% of participants correctly identified all major risk factors, while 60% of them correctly understood the signs and symptoms. Only 40% of the participants in the survey were aware of all treatment modalities and 60% of them knew the routine investigations for follow up. Almost 88.8% reported to be aware of BSE monthly, and 35.5% of them believed mammogram could cause radiation exposure. Awareness scores were significantly higher among doctors compared to nursing and allied staff with the bias of difference in participation numbers(p<0.01). Conclusion:This study clearly demonstrates a critical and concerning gap in breast cancer awareness among healthcare providers — a gap that directly undermines early detection efforts and compromises patient outcomes. As frontline influencers of public health behavior, healthcare workers must not only advocate for cancer awareness but embody it. It is imperative that institutions implement structured breast health education, mandate annual clinical breast examinations (CBEs) for staff, and integrate awareness assessments into regular training programs. Investing in HCW education is investing in community cancer control.

H. Abdel-Razeq, T. Al-Batsh, F. Tamimi, B. Sharaf, H. Bani Hani, A. Al-Atary, L. El Saket, H. Khalil, Y. Talab, A. Issa, Z. Muhanna, O. Almuhaisen; Internal Medicine, King Hussein Cancer Center, Amman, JORDAN.

Introduction: Triple-Negative Breast Cancer (TNBC) is the most aggressive form of breast cancer. Despite advancements in medical treatments, TNBC continues to be associated with a high likelihood of recurrence and poor outcomes. Gaining a deeper understanding of genetic profiles of TNBC is essential for identifying potential biomarkers, improving prognostic tools, and developing targeted therapies. This study aims to explore the baseline characteristics, survival outcomes, and genetic alterations in TNBC among a population with distinct ethnic backgrounds. Methods: Data on TNBC patients diagnosed and treated at the King Hussein Cancer Center who underwent germline genetic testing between 2014-2024 were retrospectively collected. Testing was performed on peripheral blood, and sequencing was done at international reference labs. Based on the date of testing, some patients underwent limited testing with only BRCA1/2 ± PALB2 (n=150), or an expanded multigene panel (n=432). Mutations were classified as benign/likely benign (negative), pathogenic/likely pathogenic (P/LP) (positive), or variant of uncertain significance (VUS). Event-Free Survival (EFS) and Overall Survival (OS) were estimated using the Kaplan-Meier method to evaluate patient outcomes. Results: A total of 582 patients were included in the study, predominantly Jordanian (521, 89.5%), with a median age of 45 years (range: 37-54), and majority (548, 94.2%) presented with non-metastatic disease. Genetic testing revealed that 124 patients (21.3%) had P/LP variants, and 141 (24.2%) had VUS. The most frequent variants were BRCA1 (82, 66.1%) and BRCA2 (32, 22.8%). Other variants encountered (10, 8.1%) included PALB2, RAD51D, NF1, TP53, RAD51C, MLH1, BARD1, BRIP1, and MSH2. Rate of P/LP mutations were highest among younger patients aged 65 (3/36, 8.3%), p <0.001. Among the 124 patients with P/LP variants, 119 (96.0%) had a positive family history of cancer in a close relative. Of the 434 patients who underwent multigene panel testing, only 10 (2.3%) had mutations beyond BRCA1 and BRCA2, with just two in high-penetrance breast cancer-related genes (TP53 and PALB2). Risk-reducing contralateral mastectomy was performed by 79 (65.8%) patients with P/LP variants, while bilateral salpingo-oophorectomy was performed by 63 (52.5%). During follow-up, 28 patients (4.8%) experienced recurrence, and another breast cancer was diagnosed in 40 patients (6.8%), with a significantly higher rate observed among those with germline mutations (12.5% vs. 5.1%, p = 0.004). Conclusion: This study highlights the genetic characteristics of TNBC in an Arab population. The notably high prevalence of BRCA1 and BRCA2 pathogenic variants reinforces their central role in TNBC tumorigenesis. In resource-constrained settings, prioritizing genetic testing for these two genes offers a practical and cost-effective strategy. Although the prevalence of P/LP variants is significantly lower in patients over 65 years of age, it remains sufficiently substantial to justify continued testing in this group. Importantly, over half of mutation carriers opted for risk-reducing surgical interventions, reflecting the actionable value of genetic findings in clinical decision-making.

D. Guilherme de Albuquerque e Rodrigues de Sousa¹, L. Soares¹, T. Saruwatari¹, A. Cicilini¹, J. Casali da Rocha¹, S. Sanches¹, A. Gadelha¹, F. Makdissi², G. Baiochi², B. Tirapelli¹, F. Leite², M. Sonagli², R. Cagnacci², S. Caputo Durand³, A. Favarin⁴, B. Rossi⁴, G. Molin⁴, A. da Costa¹, E. dos Santos¹; ¹Clinical Oncology, A. C. Camargo Cancer Center, São Paulo, BRAZIL, ²Surgical oncology, A. C. Camargo Cancer Center, São Paulo, BRAZIL, ³Clinical Oncology, Institut Curie – Ensemble Hospitalier, Paris, FRANCE, ⁴Clinical Oncology, Hospital Beneficência Portuguesa, São Paulo, BRAZIL.

Introduction: Patients with ovarian cancer carrying germline BRCA1/2 mutations or pathogenic variants (gBRCA1/2mut) have a significantly increased lifetime risk of developing breast cancer. While the treatment of ovarian cancer is usually the primary clinical focus, proactive strategies to reduce future breast cancer risk – including intensified surveillance, risk-reducing mastectomy, and chemoprevention – are essential components of comprehensive care. Given the evolving therapeutic landscape and improving survival outcomes in ovarian cancer, individualized approaches to breast cancer risk management are becoming increasingly relevant in this high-risk population. Our objective was to evaluate and describe breast cancer risk management strategies in OC patients carrying gBRCA1/2mut. Methods: A retrospective review of medical records was conducted for patients with OC and gBRCA1/2mut, focusing on breast cancer risk – reduction strategies. Results: A total of 88 patients with OC and gBRCA1/2mut were identified. Among them, 66 (75%) had BRCA1mutations and 22 (25%) had BRCA2 mutations. Twenty-three patients (26%) had a previous history of breast cancer prior to the BRCA1/2 mutation diagnosis; of these, 16 (18%) were treated with lumpectomy, 6 (26%) with mastectomy, and 1 (4%) with adenomastectomy. Regarding axillary management, 11 (48%) underwent sentinel lymph node biopsy, 9 (39%) axillary dissection, and in 3 cases (13%) axillary status was unknown. Fifty-one patients (58%) had a family history of breast cancer.After a median follow-up time of 70 months, only 12 patients (13%) underwent risk-reducing breast surgery: 8 (67%) underwent bilateral adenomastectomy, 3 (25%) bilateral mastectomy, and for 1 patient the chosen procedure was not specified. Among 13/88 patients with stage I-II ovarian cancer, 5 underwent prophylactic mastectomy, compared to 7 out of 75/88 patients with stage III-IV disease. Of the 12 patients who underwent prophylactic mastectomy, only 3 had relapsed ovarian disease – all of whom had advanced-stage cancer at diagnosis. Overall survival from the time of ovarian cancer diagnosis was 94.2 months. Most patients (65%) opted for active surveillance every 6 months, with magnetic resonance imaging (MRI), ultrasound, and mammography. Imaging findings classified as BI-RADS 3 were identified in 9 patients (16%), and BI-RADS 4/5 in 8 patients (14%). A total of 6 patients (10%) were ultimately diagnosed with breast cancer, half of which were luminal subtype tumors. Among these 6 cases, half were clinical stage IA, two were stage IIA, and one was stage IIIA breast cancer. Conclusion: In this cohort of OC patients with gBRCA1/2mut, the majority underwent imaging-based surveillance, with a relevant proportion presenting with suspicious findings and subsequent breast cancer diagnoses. These findings highlight the importance of individualized counseling and the integration of breast cancer prevention into the long-term management of ovarian cancer survivors with gBRCA1/2 mutations, particularly in the context of improving survival outcomes.

C. Alem, A. Cabello, b. N. Duarte, S. R. teixeira, T. Gaspar, A. viaro, D. I. Silva, M. L. Sousa, T. Cabello, S. Ramalho, L. R. Silva, C. Cabello; Obstetric and Gynecology, State University of campinas – UNICAMP, campinas, BRAZIL.

The Self-Declared White Ethnicity is Associated with the Presence of Pathogenic Variants in Brazilian Public Health Patients Undergoing Extended Multigene Panel Testing Christine Elisabete Rubio Alem ^1,, Ana Elisa Ribeiro da Silva Cabello¹, Bárbara Narciso Duarte¹; Sandra Regina Campos Teixeira; ¹ Leonardo Roberto da Silva ¹, Susana Oliveira Botelho Ramalho ¹, Thiago Cabello ¹ , Márcio Lopes de Souza ¹, Thiago Gaspar, ¹ Daniel Imbassahy SBC Silva ¹, Ana Lidia Viaro ¹, Cesar Cabello ¹1.State University of Campinas (UNICAMP)Corresponding author: cesar cabelloEmail: cabello@unicamp.brAddress: Alexander Fleming, 101- Cidade Universitaria, City: Campinas State; São Paulo, ZIP Code: 13083-881 Introduction: Breast cancer is the most common cancer among women in Brazil, with high mortality rates, especially in younger women. Hereditary factors, including mutations in BRCA1 and BRCA2, are responsible for a significant number of cases. The use of multigene panels for genetic testing has increased, though data on ethnically admixed populations remain limited. Objective: To evaluate the prevalence of pathogenic variants in a Brazilian public health population and their association with epidemiological variables. Methodology: This study involved women with a personal or family history of breast cancer. Blood samples were collected for genetic sequencing of 144 genes associated with hereditary syndromes. The study took place at the High-Risk sector of the Hospital da Mulher Prof. Dr. José Aristodemo Pinotti – CAISM, UNICAMP, between November 2021 and October 2022. Data were analyzed using next-generation sequencing (NGS) and statistical tests (Chi-Square, Fisher’s Exact, Kruskal-Wallis, Bowker, McNemar, logistic regression) with significance set at p < 0.05. Results: Of the 364 participants, 85.9% had genetic variants, with 29.7% having pathogenic or likely pathogenic variants. White women (OR: 2.1; CI 1.31-3.36) and those with a personal history of breast cancer (OR: 1.90; CI 1.16-3.10) had significantly higher risks for these variants. The most commonly affected genes were BRCA1, BRCA2, TP53, PALB2, and MUTYH. Conclusion: The study revealed a high prevalence of pathogenic variants, highlighting the role of race and personal history in genetic predisposition to breast cancer.

M. Sporn¹, C. Marsh¹, A. Sharma¹, A. Marmer¹, B. Greenfield¹, V. Tatapudi², B. Lonze², M. Gemignani¹, F. Schnabel¹, C. DiMaggio²; ¹Department of Surgery, NYU Langone, New York, NY, ²Department of Transplant, NYU Langone, New York, NY.

Introduction:Within the general population, male breast cancer (BC) is exceedingly rare, with approximately one case for every 100 female cases. However, as a high-volume transplant center, we have anecdotally observed several cases of male BC following kidney transplant. Despite this, the gender-specific BC incidence in solid organ transplant remains a gap in the current literature. We therefore sought to examine the gender-specific incidence of BC among kidney transplant recipients. Methods:Retrospective analysis of data from 1987-2022 from the Scientific Registry of Transplant Recipients (SRTR) database was used to identify patients who developed BC following kidney transplant. Clinical, pathological, treatment, and follow-up data were collected. Using R statistical software program comparisons of continuous variables were based on t tests and chi squared tests were used to compare categorical variables. Measures of association were based on odds ratios from multivariable logistic regression for dichotomous outcomes. Survival analyses were conducted using Kaplan-Meier curves and proportional hazard modeling for time to event outcomes.Results:Within the study period a total of 214 patients with BC were identified, with 42 developing BC in the 1,267 post kidney transplant recipients. Among the 484 women post kidney transplant, 15 (3%) developed BC compared to 27 of 783 men (3%). The majority of BC was identified in men 27 of 42 (64%). The incidence of BC in women was 0.08 per 1,000 transplants vs. 0.09 per 1,000 transplants in men. The odds ratio for association of BC with male gender was OR=1.1. (95% CI 0.6, 2.2) with no significant difference noted.The median time to mortality among post-transplant patients with BC was 144 months (95% CI, 87, NA) compared to 198 months for patients without BC (95% CI 179, 252). This difference was not statistically significantly different on Cox Proportional Hazard modeling controlling for gender (HR=1.1, 95% CI 0.6, 2.1).Controlling for gender, patients with post-transplant tumors were nearly 3 times more likely to have died. (OR=2.9, 95% CI 2.8, 2.9). There wasn’t a statistically significant association of BC with reported post-transplant mortality when controlling gender (OR = 1.0, 95% CI 0.5, 2.0).Conclusion:We observed a three-fold increase in mortality for patients who had BC after transplant surgery, regardless of gender. Additionally, we interestingly found an unexpected shift in the gender distribution of BC cases among patients who underwent kidney transplant. Unlike the typical pattern seen in the general population, male transplant recipients exhibited a slightly higher risk of developing breast cancer compared to their female counterparts. This striking reversal of the usual gender disparity in BC represents a highly significant finding. Further investigation into the factors affecting the internal hormonal milieu in these patients may shed light on the physiologic basis of this heightened risk. Understanding the risk of BC in well-defined subsets of post-transplant populations may help guide screening approaches with hopes of lowering morbidity and mortality.

J. T. Santos¹, D. B. Macedo¹, P. B. Porto², I. B. Rocha², R. S. Guindalini¹; ¹Oncogenetics, Oncology D’Or, São Paulo, BRAZIL, ²Endocrinology, All Clinic, Rio de Janeiro, BRAZIL.

Introduction: Hormone replacement therapy (HRT) is a widely recommended strategy for relieving menopausal symptoms, particularly for healthy women under 60 years of age or within 10 years of menopause. However, HRT has been excessively used in Brazil, including less studied compounds and delivery methods, such as subcutaneous implants of gestrinone. Although HRT provides symptomatic benefits, the increased risk of breast cancer has already been confirmed with certain types of hormone therapy. While numerous risk factors—such as age at menarche, menopause, parity, and obesity—are well established, 5-10% of breast cancer cases are linked to hereditary genetic predisposition. In this context, tools such as CanRisk, when integrated with genetic counseling (GC) and germline testing, enable personalized risk assessment and inform clinical decision-making. Objective: To investigate how the integration of GC and CanRisk (v3.0.0) influences BC risk stratification and modifies screening and prevention recommendations in women using or planning to initiate HRT. Methods: This exploratory, quantitative, retrospective study analyzed electronic medical records of women referred for BC risk assessment at a private oncogenetics service in Brazil between September 2022 and May 2025. The primary reason for referral was to discuss HRT initiation or continuation. Testing indications followed the 2025 NCCN guidelines. Risk management recommendations were based on CanRisk estimates: breast MRI for lifetime BC risk >20%; chemoprevention for 5-year BC risk >1,7% or 10-year BC risk >5%; and risk-reducing surgery for carriers of high-penetrance pathogenic variants (PVs). Results: A total of 144 women were included (mean age: 51 years). Of these, 100 (69,4%) met NCCN criteria for genetic testing, and 95 (94,1%) proceeded with testing. Among the 44 women (30,6%) who did not meet criteria, 40 (90,9%) opted for testing after a shared decision-making process, totaling 135 tested individuals. The integration of CanRisk and GC resulted in a change in clinical management for 29 women (21,5%). Regarding chemoprevention, 117 women (86,7%) were initially eligible, and after genetic testing, 9 (6,7%) lost this indication, while 4 (3,0%) gained it due to findings of PVs in BC susceptibility genes. For MRI surveillance, 15 patients (11,1%) were initially eligible; 5 (3,7%) lost and 8 (5,9%) gained the indication based on revised lifetime risk estimates. A total of 14 pathogenic variant (PV) carriers (10.4%) were identified. Among these, nine women (6.7%) received new surgical recommendations due to high-penetrance PVs: four women with BRCA1 mutations were advised to undergo bilateral risk-reducing mastectomy and bilateral salpingo-oophorectomy; two women with BRCA2 mutations were recommended bilateral mastectomy—one had already undergone oophorectomy, while the other was additionally advised salpingo-oophorectomy. One woman with a PALB2 mutation was advised bilateral mastectomy. Two women with RAD51C and RAD51D mutations were recommended bilateral salpingo-oophorectomy. Notably, one patient was diagnosed with ductal carcinoma in situ after undergoing a risk-reducing mastectomy. Discussion and Conclusion: Integrating CanRisk and GC significantly impacted risk stratification, with the new BC risk estimates resulting in reclassification in over one-fifth of the cohort (21,5%). Importantly, genetic testing should not be universally indicated for all women considering or using HRT; however, this context provides a valuable window of opportunity to identify patients who already meet criteria for genetic counseling and testing but had not yet been referred. This proactive approach supports precise cancer prevention, better clinical safety, and informed, shared decision-making for HRT use

B. Resendes, R. Sandoval, T. Correa, A. Castro, Z. Souza, I. Guttieres; Centro de Oncologia Clínica, Hospital Sírio Libanes, Brasília, BRAZIL.

Introduction: Family history (FH) of breast cancer (BC) is an independent risk factor for BC, regardless the presence of a germline mutation in a BC predisposition gene. Women with a FH of BC could be diagnosed at earlier stages due to increased awareness and screening. Nevertheless, other factors could be associated with diagnosis at later stages such as early onset, fear, and misinformation. Objective: To determine if women with a family history of breast cancer present earlier stages of disease. Methods: This single institution study was conducted at the Oncogenetics Department of Hospital Sirio-Libanês (Brasília DF, Brazil). Eligibility included patients with a personal history of BC who received genetic counseling between 2017 and 2025, and a genetic testing result without a pathogenic/probably pathogenic germline variant in a BC gene. Lack of FH data was an exclusion criteria. Data were collected retrospectivelly from medical charts. FH of BC was defined as a BC cancer diagnosis in 1st, 2nd, and/or 3rd degree relatives (defined as close relatives). For patients who received neoadjuvant therapy, BC stage was based on clinical staging and pathological or clinical staging were used for the remaining cases. Chi-square test was used for comparison of patients with and without a FH of BC. Results: Among 161 patients, 155 met study criteria. Median age of BC diagnosis was 49 years (23.0-87.0), 6 patients were diagnosed during pregnancy or breastfeeding, 6 had bilateral synchronous BC and 7 metastatic de novo disease. Out of 155 patients, 88 (56.7%) had at least one close relative with BC (FH+), including 37 (23.9%) with a family member affected before 50 years and 7 (4.5%) with bilateral BC. In total, 68 (43.8%) patients were diagnosed due to symptoms, 68 (43.8%) through screening, 3 incidentally found during mamoplasty and in 16 cases presentation was not specified. Among those screen-detected cancers (n=68), 22 (32.3%) were detected by ultrasound, 35 (51.5%) by mammogram, 7 (10.3%) by MRI, and in 4 cases the diagnostic modality was unavailable. There were no differences between groups (FH+ vs FH-) regarding age of BC diagnosis, menopausal status, BC presentation, and BC stage. Conclusion: In this BC Brazilian cohort from a private health setting without a cancer predisposition syndrome, a family history of BC did not impact BC stage at diagnosis. Cancer risk assessment based on risk models such as Tyrer Cuzick could inform the benefit of BC screening with MRI in this scenario.

Across all variables tested, no statistically significant association was found with family history status (FH+ vs FH-) at the conventional p < 0.05 threshold.

X. Kuang, Y. Lin, N. Shao; Breast Cancer, The First Affiliated Hospital, Sun Yat-sen University, Guangzhou, CHINA.

Background: Approximately 5-10% of breast cancers (BC) are attributed to inherited germline pathogenic variants, primarily in genes involved in homologous recombination repair (HRR), most notably BRCA1/2. This study aimed to identify clinical risk factors associated with pathogenic variants in patients with a high genetic risk of breast cancer and evaluate the utility of multigene panel testing. Methods: A total of 533 high-risk BC patients underwent next-generation sequencing (NGS)-based germline multigene panel testing covering 61 HRR-related genes. Eligibility was defined by the presence of at least one of the following: triple-negative breast cancer (TNBC), early-onset disease (≤50 years), a family history of malignancy, or multiple primary tumors (including bilateral BC). Targeted capture-based sequencing was performed using DNA extracted from peripheral blood leukocytes. Results: The median age at diagnosis was 40.6 years old. All patients had ductal carcinoma. Pathogenic HRR variants were detected in 120 patients (22.5%), with 123 total variants identified; one patient (HR-positive/HER2-negative) carried two variants and another (TNBC) carried three. Pathogenic BRCA1 and BRCA2 variants were observed in 36 (6.75%) and 37 (6.94%) patients, respectively. Non-BRCA HRR pathogenic variants were present in 47 patients (8.82%), with PALB2 being the most frequently mutated non-BRCA gene (n=19, 3.56%). TNBC patients (43/145, 29.66%; P=0.012) and those with ≥2 clinical risk factors (84/329, 25.53%; P=0.021) had a significantly higher prevalence of pathogenic variants. A trend toward increased mutation frequency was noted in younger age groups (≤30y: 27.54%, 31-40y: 23.61%, 41-50y: 20.96%, ≥51y: 17.19%), although the difference was not statistically significant (P=0.284). Among HR-positive/HER2-negative and HR-negative/HER2-positive patients, pathogenic variant prevalence was 20.65% and 18.75%, respectively. Conclusions: This study highlights the clinical relevance of germline HRR gene testing in high-risk BC patients. Multigene panel testing should be considered routine for patients with TNBC or those presenting with multiple risk factors to guide personalized management and familial risk assessment.

S. SINGH, S. Vidaurri, M. Noorani, S. Goyal, A. Dhasmana, S. Dhasmana, M. Yallapu, S. Fofana, S. C. Chauhan, D. Nguyen, S. Khan; School of Microbiology and Immunology, University of Texas Rio Grande Valley, McAllen, TX.

Targeting Breast Cancer via Dual Antagonism of Estrogen Receptor α and GPERBackgroundEstrogen’s diverse biological effects are primarily driven by the classical receptors ERα and ERβ,which function as ligand-activated transcription factors. Among these, ERα plays a pivotal role inbreast cancer development, making it a key target for therapies such as tamoxifen. However,tamoxifen’s clinical success is often compromised due to both inherent and acquired resistance.Notably, tamoxifen inhibits ERα but paradoxically acts as an agonist for GPER/GPR30, estrogenreceptor that mediates non-classical estrogen signaling in normal and cancerous cells. Activationof GPER can promote gene expression and cell proliferation, contributing to tamoxifen resistancein hormone-sensitive tumors. Our study addresses this challenge by focusing on a dual-targetingstrategy that simultaneously inhibits both ERα and GPER, aiming to overcome resistance andimprove therapeutic outcomes in breast cancer treatment.MethodsGPER expression was evaluated in breast cancer cell lines and tumor tissues using immunoblotting,qRT-PCR, and immunohistochemistry. Cell proliferation was assessed via MTT and colonyformation assays. Migration and invasion capabilities were measured using wound healing andtranswell invasion assays. Intracellular calcium signaling was monitored using fluorescence-basedcalcium imaging following GPER stimulation. Metagenomic analysis of Hispanic breast cancertissues were performed using 16S rRNA sequencing to identify microbial populations correlatedwith GPER expression and tumor subtypes. Statistical analyses included correlation tests andsubtype stratification.ResultsGPER expression was detected in both breast cancer cell lines and tumor tissues. Our resultsdemonstrate that GPER is significantly overexpressed in ERα-positive breast tumors, with aunique correlation observed in a Hispanic breast cancer patient population (P < 0.001). GPERexpression was also observed in Triple-Negative subtypes. The correlation of GPER expressionwas analyzed across different breast cancer subtypes, classified as Luminal A (HR+/HER2−),Luminal B (HR+/HER2+), HER2-Enriched (HR−/HER2+), and Triple-Negative (HR−/HER2−).In vitro studies confirmed the presence of GPER protein and RNA in breast cancer cells of varioussubtypes. Furthermore, this study identifies Ormeloxifene (ORM), a selective estrogen receptormodulator (SERM), as a dual antagonist of ERα and GPER. GPER expression was associated withincreased cancer aggressiveness, migration, and invasion, all of which were effectively inhibitedby Ormeloxifene. ORM acts as an antagonist ligand for both ERα and GPER by suppressing genetranscription and inhibiting growth signaling pathways in breast cancer cells. Mechanistically,Ormeloxifene suppresses GPER-induced calcium signaling and downstream activation of EGFR,ERK, YAP, and TAZ pathways, thereby inhibiting gene transcription and tumor growth. Additionally, metagenomic analysis of Hispanic breast cancer tissues revealed specific microbialpopulations associated with GPER expression and tumor subtypes.ConclusionsGPER is significantly overexpressed in ERα-positive and some Triple-Negative breast cancers,especially within the Hispanic population. The exclusive antagonistic activity exerted byOrmeloxifene on both ERα and GPER represents an innovative pharmacological approach fortargeting breast carcinomas expressing one or both receptors at diagnosis or during progression.Simultaneous inhibition of ERα and GPER may offer greater therapeutic benefits compared toselective estrogen receptor antagonists alone. Additionally, the association of GPER expressionwith specific microbial populations suggests new directions for personalized breast cancertreatment.

K. Rupani¹, E. Moshier², C. Jin³, D. Jaber⁴, N. Casasanta⁵, C. Wu⁶, A. Tiersten*⁷, R. Patel*⁷; ¹Division of Hematology and Medical Oncology, Department of Medicine, Icahn School of Medicine at Mount Sinai, New York, NY, ²Division of Biostatistics, Department of Population Health Science and Policy, Icahn School of Medicine at Mount Sinai, New York, NY, ³Department of Pathology, Donald and Barbara Zucker School of Medicine at Hofstra/Northwell, New Hyde Park, NY, ⁴Department of Medicine, Northwestern University Feinberg School of Medicine, Chicago, IL, ⁵Division of Hematology and Medical Oncology, Department of Medicine, Yale-New Haven Hospital, New Haven, CT, ⁶Department of Anesthesiology, Perioperative and Pain Medicine, Icahn School of Medicine at Mount Sinai, New York, NY, ⁷*Denotes co-final authors, Division of Hematology and Medical Oncology, Department of Medicine, Icahn School of Medicine at Mount Sinai, New York, NY.

Background: Patients diagnosed with non-invasive breast lesions such as DCIS, LCIS, ADH, and ALH are at increased risk of developing invasive breast cancer (BC). The TAM-01 study found that low-dose tamoxifen (LDT) for chemoprevention resulted in an approximate 50% risk reduction in BC events compared to placebo. A recent study at our institution (PMID: 39837695) showed that the uptake of LDT was 19%. The objective of our present study was to evaluate the recurrence patterns for patients who initiated LDT. Methods: We performed a retrospective review of patients with DCIS, LCIS, ADH, and/or ALH who started LDT from January 2019 to December 2021. We evaluated rates of LDT completion and recurrence characteristics. A time-dependent Cox regression model was used to evaluate the association between early discontinuation of LDT and risk of recurrence, accounting for changes in LDT use over time. Results: 88 eligible patients were identified and 7 were excluded due to lack of follow up. Of the remaining 81 patients, 67% (N=54) completed 3 years of LDT and 30% (N=24) stopped early due to adverse events (AEs); the remaining 3% stopped due to recurrence (N=1), to attempt pregnancy (N=1), and diagnosis of another malignancy (N=1). Overall, the mean age of our population was 54 years (SD 12) and 39% (N=32) were postmenopausal. 39% (N=32) of patients had DCIS, and 61% (N=49) had either LCIS, ADH, and/or ALH. At a median follow up time of 4.74 years, the overall recurrence rate was 10% (N=8). Among the 27 patients who discontinued LDT early, 19% (N=5) recurred: 4 after discontinuing LDT due to AEs and 1 recurred while on LDT. Among 54 patients who completed LDT, 6% (N=3) recurred: 1 during treatment and 2 after 3 years of LDT (Table 1). Based on the time-dependent Cox model, the estimated 5-year probability of recurrence was 8% for patients who completed LDT and 15% for those who discontinued early. Early discontinuation was associated with nearly a two-fold increased risk of recurrence, although this was not statistically significant due to the low event rate [HR: 1.94; 95% CI: 0.42-9.07; p=0.40]. All recurrences were in patients who had LCIS, ADH, and/or ALH. Most recurrences (75%) were in premenopausal patients, and the 3 patients who recurred after completing 3 years of LDT were all premenopausal. Except for 1 patient who developed HR-, HER2+ IDC and DCIS while on LDT, all other recurrences, irrespective of completion of 3 years of LDT, were atypia or LCIS. Conclusions: In patients initiated on LDT for chemoprevention, at a median follow up of 5 years, the estimated probability of recurrence was 8% in those who completed 3 years of LDT and 15% in those who stopped early. Most recurrences were seen in premenopausal patients consistent with a subgroup analysis of the TAM-01 study suggesting decreased efficacy of LDT in premenopausal women.

A. Athman Omar¹, R. Mukadam², A. Kalebi³; ¹Oncology, KUTRRH, Nairobi, KENYA, ²Pathology, Dr. Kalebi Lab – DKL, Nairobi, KENYA, ³Pathology, DKL, Nairobi, KENYA.

Background: Breast cancer (BC) is the most common cancer in Kenya, with patients being diagnosed a decade earlier than in the USA. Patients harbouring BRCA1/2 mutations have aggressive disease, but paradoxically respond well to tailored treatments such as poly ADP-ribose polymerase (PARP) inhibitors. However, BRCA testing is not accessible for most patients in Africa; thus, the prevalence largely remains unknown, with very scant published literature from Sub-Saharan Africa showing a wide variation from 2% up to 17%. We hereby report the prevalence of BRCA1/2 mutations for patients screened in the first six months in a recently opened independent referral lab in Kenya. Methods: The BC patients with high-risk clinical features such as young age, triple-negative breast cancer (TNBC), bilateral BC, and familial history of breast cancer who had been referred for germline (gBRCA 1/2) or somatic (sBRCA 1/2) testing were screened. After consent, patients’ samples, either peripheral blood or Formalin-Fixed Paraffin-Embedded, were tested for BRCA 1/2 mutations using the Oncomine BRCA Research Assay. Gene variant ID, variant class, and variant allele frequency, among others, were reported. Only patients with a confirmed BC diagnosis and de-identified data were included in this analysis. Results: In the six months, 42 patients were screened for BRCA 1/2 mutation. Of which 14 females had confirmed BC diagnosis, and their median age is 47 years (range: 29 – 70). The majority (10 out of 14) had triple-negative breast cancer, one with bilateral breast cancer, while another had metastasis to the ovary. The family history was not provided for any of these patients. The prevalence of BRCA 1/2 mutations was 21% among the patients who underwent germline testing (n=12). Two had pathogenic variants in BRCA1, while one in BRCA2. The most common mutation was indels. Moreover, three patients had a BRCA 2 variant of uncertain significance (VUS). The most common mutation was missense. Conclusion: In this study, the prevalence of BRCA 1/2 mutations among BC patients in Kenya is 20%. To our knowledge, this is among the first works on BRCA testing in BC reported from Kenya, and increased availability and accessibility to oncogenetic testing will not only lay a foundation for future research work but also facilitate better management of BC patients with PARP inhibitors.

M. Zafar¹, M. Krishnakumar², A. Reddy³; ¹Department of Internal Medicine, Arkansas College of Osteopathic Medicine, Mercy Hospital Fort Smith, Fort Smith, AR, ²Department of Internal Medicine, SRM Medical College Hospital and Research Center, Chennai, INDIA, ³Department of Hematology Oncology, Mercy Clinic, Fort Smith, AR.

Background: Breast cancer is the most common non-cutaneous malignancy among women, with around 10% of cases involving germline mutations in DNA repair genes. Germline mutations in BRCA1/2 genes significantly increase breast cancer risk. Other relevant genes to increase the risk of breast cancer include TP53, PTEN, PALB2, CHEK2, and ATM. Multigene panel testing can sequence many genes, sometimes identifying non-disease-related genes and variants of unknown significance. Methods: We reviewed patients with newly diagnosed breast cancer diagnosis from 2022 to 2024 in our institution. Germline genetic testing was offered to all newly diagnosed breast cancer patients at the time of diagnosis irrespective of their family history. Testing was offered to patients with invasive breast cancer and ductal carcinoma in situ. The goal of this study was to identify the prevalence of genes not attributed to increasing the risk of breast cancer and the implications of those results. Results: A total of 478 patients (Invasive cancer and DCIS) were seen in our oncology clinic from January 2022 to December 2024. The median age at diagnosis was 66 years (range 28-91). Invasive ductal cancer (IDC) was the most common subtype (66%), followed by DCIS (18%). Most patients had estrogen-positive cancer (81%), and approximately 14% had HER2-positive disease (62% HER2 low). We found 27 patients with pathogenic germline mutations which are not related to breast cancer risk. The most common mutation was MUYTH (7 patients), followed by FH gene (3 patients). MSH3, LZTR1, and HOXB13 genes were each found in 2 patients. APC, BAP1, BLM, BRIP1, NTHL1, SDHA, SDHB, MSH6, MITF, and FLCN genes were each identified in one patient. All patients with MUYTH mutations had hormone-positive (HER 2 negative) disease contrary to the limited observation of triple-negative breast cancer related to the MUYTH gene in the available literature. Patients with positive mutations were offered genetic counseling and referred to a high-risk genetics’ clinic. FANCM was observed in one patient which has rarely been associated with an increased risk of triple-negative breast cancer. Conclusion: We present our single institution’s experience of germline mutations in breast cancer patients. Previously, we observed CHEK2 being the most common breast cancer-related germline mutation in our patient population. This analysis focuses on identifying germline mutations without an increased risk of breast cancer. MUTYH was the most common pathogenic mutation identified. Further studies are needed to determine the actual risk percentage of breast cancer associated with the MUTYH gene. Universal testing for all breast cancer patients is feasible with the lower cost of multi-gene panels. However, identifying other non-disease-related genes, as observed in our study, has implications. Patients must be informed about the consequences of a positive result, including genetic counseling, additional screening tests, and testing of immediate family members.

R. Campbell¹, L. Houghton², D. Walker³; ¹School of Nursing- Community Health Systems, University of CA, San Francisco, San Francisco, CA, ²Mailman School of Public Health, Columbia University, New York, NY, ³Medical Humanities and Ethics, Columbia University, New York, NY.

Introduction/Background: Early-onset breast cancer remains a critical public health concern. Despite advances in medical research, young women continue to account for approximately 15% of breast cancer (BC)-related deaths. Among these, Black young women have double the mortality rate at 12.1 per 100,000 compared to young White women. Black young women face significant disparities in both access to comprehensive BC preventive care, and the personalization of BC treatments. There is a crucial need for improved primary prevention strategies to reduce BC incidence and ensure access to personalized treatment upon diagnosis. Biomarker-based preventative BC screening, and monitoring post-diagnosis offers a promising avenue to combat these disparities. These approaches have the potential to provide individualized risk assessments and individualized BC treatment. The purpose of this study was to explore the perspectives of Black women under the age of 50 living with BC on the acceptability of biomarker-based primary BC prevention screening, as well as biomarker’s role post BC diagnosis. Study Aims: (1) Examine the perspectives of Black women living with BC on the acceptability of preventative BC biomarkers, and use of biomarker monitoring after BC diagnosis; (2) Understand the breast health and BC risk information available to young Black women prior to their diagnosis; (3) Explore the pathways leading to BC diagnosis of these women; and (4) Assess the extent to which biomarker monitoring and genomic testing was received after their BC diagnosis. Methods: Nineteen Black women living with breast cancer were invited to participate in virtual focus groups (FG) via the Zoom platform in November 2025. Each FG session was preceded by an educational PowerPoint presentation detailing the role of primary preventive BC biomarker screening, and biomarker and genomic screening following a BC diagnosis. Using a critical lens informed by Intersectionality Theory and the Embodied Health Movement Theory, Reflective Thematic Analysis was employed to analyze FG transcripts, extracting both articulated and emergent themes. Results: Participants generally expressed openness to the idea of preventive BC biomarker screening, recognizing its potential benefits. However, concerns arose regarding the financial burden of screening, insurance coverage, and fear of how test results might impact access to job benefits. Two participants reported a comprehensive breast assessment due to a known family genetic BC mutation. While 17 participants reported receiving minimal breast health education, no comprehensive BC risk assessments prior to diagnosis, and identified their breast change(s) themselves, which lead to BC diagnosis. Regarding the use of biomarkers in post-diagnosis treatment, responses varied widely. Access to biomarker monitoring was influenced by factors such as the time since diagnosis, disease aggressiveness, and physician preference. Several emergent themes surfaced during the discussions, including the need for self-advocacy both before and after diagnosis, the racialized experience of BC, and the inadequate marketing of BC prevention. Conclusion and Future Research: Biomarkers for BC prevention and use post- diagnosis are novel but promise more personalized BC risk assessments, individualized BC treatment, and improved disease monitoring. Future research should prioritize randomized controlled trials evaluating preventive BC biomarkers, and biomarker monitoring post-diagnosis among diverse young women.

M. E. Gonzalez¹, M. Mullins², C. Barriga², M. Bravo², D. Carvajal², J. Camacho²; ¹Surgery, Clinica Alemana, Santiago, CHILE, ²Oncology, Clinica Alemana, Santiago, CHILE.

IntroductionApproximately 5-10% of breast cancer cases are linked to genetic predisposition, primarily involving pathogenic germline variants in BRCA1 and BRCA2, which confer a lifetime risk of 80-90% and 60-85%, respectively. Other genes like CHEK2, BRIP1, and PALB2 are associated with moderate risks (20-40%). While pathogenic variants can predispose to triple-negative breast cancer, most cases occur without these mutations. High-risk patients may opt for genetic screening and prophylactic mastectomy, which reduces breast cancer risk and improves disease-free survival, though it does not eliminate risk or increase overall survival. ObjectiveThis study aims to analyze genetic mutations and variants of uncertain significance in breast cancer cases treated at Clínica Alemana, Santiago, over the past ten years, and compare them with global data. It also examines how these findings influence clinical management, including prophylactic mastectomy. Methods and materialsRetrospective descriptive study of genetic mutations in 200 women aged ≤40 with breast cancer over 10 years at Clínica Alemana, Santiago, Chile. ResultsIn this study, 47.4% of patients received genetic counseling and all underwent testing. Pathogenic mutations were found in 18 patients, mostly in BRCA1/2, while 28 had variants of uncertain significance, primarily in CHEK2. DiscussionGenetics are implicated in about 10% of breast cancer cases, consistent with our finding of 9.4% positive genetic tests. Variants of uncertain significance (VUS) occur in 2-7% of BRCA1/2-only tests and up to 10-41% in multigene panels; our study found 14.6% VUS after multigene testing. BRCA1/2 pathogenic variants were present in 4.2% of cases, aligning with reported rates of 3-4%. Prophylactic mastectomy rates vary between 40-80% in carriers of high-penetrance mutations, and are lower for moderate-penetrance genes. In our cohort, 63% of BRCA1/2 carriers and 10% with moderate-risk gene mutations underwent prophylactic mastectomy.

N. Liao¹, J. Wu², G. Zhang¹, J. Weitzel³, C. Balch⁴; ¹Department of Breast Cancer, Guangdong Provincial People’s Hospital, Guangdong Academy of Medical Sciences, Guangzhou, CHINA, ²Clinical Research Devision, Kanghui Biotech, Beijing, CHINA, ³Department, The University of Kansas Comprehensive Cancer Center, Leawood, KS, ⁴Department of Surgical Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX.

Purpose: Germline mutations were evaluated in 530 Chinese breast cancer (BC)patients with different HER2 expression status plus 98 patients with atypical ductalhyperplasia or a breast cancer family history.Methods: DNA extracted from blood samples was analyzed with a next generationsequencing based multigene panel, with reporting of likely pathogenic and pathogenicvariants (PV) of 102 cancer associated genes, and correlation betweenclinicopathologic characteristics and known BC-associated genes. Results: The 71 identified PVs were categorized into five distinct pathway clusters:BRCA/FANC, DDR, HRR, FANC, and Other. The distribution of PVs enriched withinthese clusters differed significantly between BC patients and unaffected individuals(p=0.031). This difference was primarily driven by the BRCA/FANC, FANC, and DDRclusters, with enrichment percentages of 57.7% vs. 16.7% (BRCA/FANC), 7.0% vs.25.0% (FANC), and 15.5% vs. 41.7% (DDR) in BC vs. unaffected groups, respectively.Notably, the three BC groups stratified by HER2 expression level exhibited distinct PVdistributions. Both the HER2-low and HER2-zero BC groups showed significantlydifferent distributions compared to unaffected individuals (p=0.0132 and p=0.0081,respectively). The BRCA/FANC cluster was the predominant pathway enriched in bothHER2-low (59.6%) and HER2-zero (81.8%) groups. Furthermore, the PV distributionin the HER2-zero group was significantly different from that in the HER2-high group(p=0.0028). In contrast, the distribution in the HER2-high group resembled thatobserved in non-BC individuals. These findings indicate that BC patients with differentHER2 expression statuses harbor distinct germline PV signatures, which correlate withdifferential clinical outcomes following neoadjuvant systemic therapy.Discussion: Chinese breast cancer patients with different HER2 expression statusdemonstrated different pathogenic germline variant, which correlated with differentclinical outcome after neoadjuvant therapy. Multi-gene genetic test for pathogenicgermline variant should be offered to breast cancer patients as well as high-riskindividuals at risk for future breast cancer who would benefit from prevention and earlydetection programs.

L. Shahid¹, A. J. Patev², M. A. Danso³, R. A. Burke⁴, N. Balanchivadze³; ¹Internal Medicine, Macon & Joan Brock Virginia Health Sciences at Old Dominion University, Norfolk, VA, ²Research and Infrastructure Service Enterprise, Macon & Joan Brock Virginia Health Sciences at Old Dominion University, Norfolk, VA, ³Hematology/Oncology, Virginia Oncology Associates, Norfolk, VA, ⁴Surgical Oncology, Macon & Joan Brock Virginia Health Sciences at Old Dominion University, Norfolk, VA.

Title: Real-world adherence to germline genetic testing guidelines in breast cancer: patterns and disparities in a community oncology setting Authors: Lyeba Shahid, DO; Alison J. Patev, PhD; Michael Danso, MD; Rachel Burke, MD; Nina Balanchivadze, MD Objectives Germline genetic testing is a critical component of personalized care in breast cancer, informing surveillance, treatment, and risk-reduction strategies for patients and their families. National Comprehensive Cancer Network (NCCN) recommend testing patients with early breast cancer and specific high-risk features. However, adherence to these guidelines in real-world community settings, and the role of social determinants such as insurance and geography, remains incompletely understood. Methods: We retrospectively reviewed charts of patients who were diagnosed with breast cancer at Sentara Norfolk General Hospital between June 1, 2019 to June 1, 2024. The sample included 89 patients, between 36 to 65 years old (M = 55.53), with 88 female patients and 1 male patient. Patients were deemed eligible for genetic screening in accordance NCCN Guidelines Version 1.2025, which included: 1) diagnosis of breast cancer at < 50 years old, or 2) triple-negative breast cancer, multiple primary breast cancers, or male breast cancer, or 3) a family history of one or more close blood relative with breast cancer at or before age 50, pancreatic cancer, male breast cancer, ovarian cancer, prostate cancer with high-risk features, or three or greater breast and/or prostate cancer diagnoses on the same side of the family [1]. Demographic variables included race, marital status, insurance, and social vulnerability, which was derived from zip code. Results: Genetic testing was ordered for 75% of patients (n=67). Of the total sample, 65 out of 89 patients had testing completed. Fisher’s exact tests explored relationships between social determinants and testing. For eligible patients, no significant relationships were found. However, examination of subgroup testing completion rates showed some group differences. Nearly 92% of White eligible patients received testing, compared to 77% of Black eligible patients. Married patients completed testing at a rate of almost 87%, while 70% of single people completed testing. Additionally, ninety-two percent of eligible patients with a qualifying family history were tested, compared to 83% based on age and 80% based on cancer type. Discussion: Despite high overall adherence to NCCN guidelines for germline genetic testing in this community-based breast cancer cohort, disparities emerged across several social dimensions. While statistical significance was not reached, subgroup analyses revealed meaningful trends. Black patients were less likely to complete testing compared to White patients, highlighting potential racial disparities in access, communication, or trust. Similarly, single patients had lower testing completion rates than married patients, suggesting that social support may influence engagement with genetic services. Overall, these findings suggest that even in a high-performing clinical setting, social determinants of health may influence the delivery of precision oncology, and targeted interventions are needed to ensure equitable access to genetic testing for all eligible patients. References:– [1] National Comprehensive Cancer Network. Genetic/Familial High-Risk Assessment: Breast, Ovarian, and Pancreatic. Version 1.2025. https://www.nccn.org/professionals/physician_gls/pdf/genetics_bop.pdf.

Z. Mitri¹, K. Preston², A. Harrison¹, A. Smith¹, T. Nghiem¹, R. Cheifetz¹; ¹Medical Oncology, BC Cancer Vancouver, Vancouver, BC, CANADA, ²Medicine, University of British Columbia, Vancouver, BC, CANADA.

Background. Individuals with germline hereditary mutations in breast cancer (BCa) predisposition genes are at significantly higher risk of developing BCa during their lifetime compared to the general population. In British Columbia, patients identified as carrying a deleterious mutation are followed by the Hereditary Cancer Program (HCP) High Risk Clinic (HRC) for management of their increased BCa risk. This study evaluates the clinical and pathologic features, and outcomes for BCa diagnosed in hereditary mutation carriers under the HRC.Methods. Patients followed in HRC with a confirmed deleterious hereditary mutation, and at least one BCa diagnosis after HCP enrollment were included in this study. Patient and tumor characteristics, detection method, treatment, and survival outcomes were collected. BCa diagnoses were categorized as prevalent (detected on first screening after HCP enrollment); incident (detected on routine imaging beyond first screening). Prevalent or incident BCa detected within 6 months of a normal screening and outside of the usual screening schedule were characterized as interval cancers. Results. Clinico-Pathologic Characteristics. Between 1997-2023, 166 patients were diagnosed with BCa out of 2011 patients (166/2011, 8.2%) seen in HRC in that period. Median age of BCa diagnosis was 45 (IQR 39-57). The most common deleterious mutations were BRCA1 (44.6%) and BRCA2 (40.4%). Other mutations included CHEK2 (4.2%), PALB2 (4.2%), TP53 (3.6%), ATM (2.4%), and NF1 (0.6%). In our cohort, 39 (23.5%) BCa were classified as prevalent, 127 (76.5%) as incident. Twenty-four cases (14.5%) met the criteria for interval BCa. MRI was the most common detection modality (50.6%), followed by mammography (26.5%), clinical exam (10.8%), and ultrasound (2.4%). Of note, 16 (9.6%) cases were diagnosed at time of prophylactic surgery and were not identified on prior imaging. Among all BCa diagnoses, 37 cases (22.3%) were in-situ disease only, and 129 (77.7%) had invasive disease on final pathology. Within invasive tumors, 66.7% had stage 1, 26.4% stage 2, 6.2% stage 3, and one patient had de-novo stage 4 disease at diagnosis. Incident cancers were more likely to be diagnosed at stage 1 compared to prevalent cancers (71% vs 48%, p=0.015). Invasive BCa subtype distribution was 51.9% hormone receptor positive HER2-negative (HR+HER2-), 32.6% triple negative (TNBC), and 15.5% HER2-positive. Of note, BRCA1 tumors were more likely TNBC (52.2%, p<0.001) and BRCA2 tumors were more likely HR+HER2- (60.3%, p<0.001). Treatment. Most patients (153/166, 92.2%) underwent bilateral mastectomies as the surgical treatment of choice. Among patients with invasive BCa, 79 (47.6%) received chemotherapy either in the neo-adjuvant or adjuvant setting. There was no statistically significant association between incident or prevalent BCa and receipt of chemotherapy. There is a marginally statistically significant association between interval cancer and receipt of chemotherapy (p =0.067). Approximately 20% of all patients received adjuvant radiation therapy, including 26/153 (17%) who had bilateral mastectomies; 114/166 (68.7%) underwent prophylactic risk-reducing bilateral salpingo-oophorectomy. Outcomes. At time of last follow up, there were 8 deaths in our cohort, 5 due to metastatic BCa, and 3 due to ovarian cancer. An additional 10 patients were diagnosed with recurrent BCa and undergoing treatment. Three and 5-year overall survival rates in our cohort were 98% and 95.2%.Conclusions. Patients diagnosed with BCa under the HCP program have excellent outcomes. BRCA mutations are the most common identified alteration, and BRCA1 is enriched for TNBC. Further work is needed to identify clinical and biological features of interval cancers in this population.

R. Basho¹, E. Heredia², H. McArthur³, Z. Fang², V. Narumi⁴, A. LeVee⁵, T. Jonsson⁶, F. Dadmanesh⁷, S. Bose⁷, S. Patre², T. Keller², J. Paine², B. Tran⁴, B. Simental⁴, O. Castellanos⁴, J. Lee⁸, S. Shiao⁹, D. Agus¹, N. Matasci²; ¹Medicine, Ellison Medical Institute, Los Angeles, CA, ²AI and Advanced Molecular Medicine, Ellison Medical Institute, Los Angeles, CA, ³Internal Medicine, UT Southwestern, Dallas, TX, ⁴Clinical Research, Ellison Medical Institute, Los Angeles, CA, ⁵Medicine, University of California Los Angeles, Santa Monica, CA, ⁶Program Management, Ellison Medical Institute, Los Angeles, CA, ⁷Pathology, Cedars-Sinai Medical Center, Los Angeles, CA, ⁸Molecular Carcinogenesis, Ellison Medical Institute, Los Angeles, CA, ⁹Radiation Oncology, Cedars-Sinai Medical Center, Los Angeles, CA.

BackgroundNeoadjuvant taxane-based chemotherapy in combination with anti-HER2 therapy is standard in patients with locally advanced HER2-overexpressing (HER2+) breast cancer. Pathologic complete response (pCR) after combination therapy is associated with improved long-term outcomes and occurs less frequently in patients whose disease is also estrogen and/or progesterone receptor positive (HR+) compared to patients with estrogen and progesterone receptor negative (HR-) disease. In patients with HER2+, HR+ disease, improved patient selection for standard neoadjuvant therapy is needed to optimize outcomes and minimize toxicity in patients unlikely to achieve pCR. Here we sought to utilize digital pathology and artificial intelligence/machine learning (AI/ML) algorithms on baseline biopsy tissue of newly diagnosed patients with operable HER2+ breast cancer to predict likelihood of pCR after standard neoadjuvant therapy, with a particular focus on HR status-specific differences.Methods Digitized H&E-stained baseline biopsy slides from patients with HER2+ operable breast cancer treated neoadjuvantly with taxane-based chemotherapy plus anti-HER2 therapy with pathologic response data available were identified from 2 publicly available datasets and an institutional cohort. Utilizing CanvOI, a pre-trained computational pathology foundation model, we developed an AI/ML model for predicting pCR from whole-slide images (WSIs). For processing, WSIs were divided into 380×380 pixel tiles. CanvOI’s tile encoder converted tile image data into numeric vectors, and subsequently, CanvOI’s slide encoder converted tile vectors into an aggregated slide vector. The slide vectors were used as predictors of pCR for HR+ and HR- (ER and PR ≤ 10%) cohorts independently, using logistic regression as a vector-data classification method. Model performance and generalizability were evaluated using 4×4 cross-validation, consisting of 4-fold cross-validation repeated 4 times with randomly mixed data. In each iteration, 75% of the data was used for model training and 25% for evaluation. Multivariable logistic regression analysis was conducted in each cohort to evaluate the model and pathological features associated with pCR. To address bias due to the naturally occurring imbalance of pCR rates within the HR+ and HR- cohorts, we implemented sample trimming in the dominant class within each cohort.ResultsA total of 231 patients had adequate baseline tissue for analysis, N=119 HR+ and N=112 HR-. The rates of pCR for these patients was 40.4% for HR+ disease and 64.1% for HR- disease. For prediction of pCR status, the model achieved a mean Area Under the Curve (AUC) of 0.69 (95% CI: 0.66-0.73) and Matthews Correlation Coefficient (MCC) of 0.304 in HER2+/HR+ patients, and a mean AUC of 0.65 (95% CI: 0.61-0.69) and MCC of 0.213 in HER2+/HR- patients. After multivariable analysis evaluating the ML model, tumor grade (1-2 vs 3) and HER2 immunohistochemistry status (2+ vs 3+) for patients in whom the pathological data was available (N=70 HR+ and N=73 HR-), the ML model remained a significant predictor of pCR in both HR+ and HR- cohorts. Applying data balancing techniques improved model performance, with a mean AUC of 0.70 (95% CI: 0.67-0.73) and MCC of 0.318 in HER2+/HR+ patients, and a mean AUC of 0.67 (95% CI: 0.63-0.71) and MCC of 0.181 in HER2+/HR-.Conclusions These findings demonstrate the potential of a digital pathology-based AI/ML model to capture information about tumor sensitivity to therapeutic agents that are not provided by current clinical parameters. In patients with HER2+, HR+ operable breast cancer, where rates of pCR remain low, such a tool has the potential to guide the evaluation of novel therapeutic strategies in appropriately selected patients. Larger datasets are needed to build upon these findings.

Y. Lin¹, Y. Yu¹, Q. Wang¹, K. Huang², S. Tang³, J. Yuan⁴, C. song¹; ¹Department of Breast Surgery, Clinical Oncology School of Fujian Medical University, Fujian Cancer Hospital, Fuzhou, CHINA, ²Department of Breast Surgery, The Second Affiliated Hospital of Fujian Medical University, Quanzhou, CHINA, ³Department of Breast Surgery, Yunnan Cancer Hospital, Kunming, CHINA, ⁴Department of Medical Imaging, Henan Cancer Hospital, Zhengzhou, CHINA.

Background: Triple-negative breast cancer (TNBC) is an aggressive subtype of breast cancer (BC) and neoadjuvant therapy (NAT) has become an essential treatment strategy. Recent studies have showed tertiary lymphoid structures (TLSs) is associated with better treatment response to NAT in BC. However, there is still a lack of non-invasive biomarkers to predict the presence of TLSs in TNBC. This study aimed to develop a novel non-invasive machine learning model for early identification and prediction of the presence of TLSs and treatment response to NAT in TNBC, which were essential for timely adjustments in treatment strategies for TNBC. Methods: In this study, 698 patients from multicenter were divided into a training cohort (n = 137), the TLSs validation cohort (n = 63) and the NAT response validation cohorts (n = 561). A total of 4788 radiomic features per patient were extracted from the intratumoral and peritumoral regions of DCE-MRI. After extracting the optimal features, five machine learning models were developed to predict the presence of TLSs, including K-Nearest Neighbors (KNN), Extreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine (LightGBM), Support Vector Machine (SVM) and Multilayer Perceptron (MLP) and were subsequently applied to predict the treatment response to NAT in NAT response validation cohorts. The performance of the models was assessed by accuracy, sensitivity, specificity, negative predictive value (NPV), positive predictive value (PPV), and area under curve values (AUC), and prognostic analysis were performed to evaluate its predictive value. Then,the rTLS predictive model was interpreted by SHapley Additive Explanations (SHAP). Finally, the correlation between key radiomic features extracted from H&E-stained histological slide images using CellProfiler and key pathomic features was further analyzed to reveal the tumor heterogeneity of TNBC from a pathological perspective.Results: The XGBoost model, which outperformed all other predictive models, was selected as the radiomics-based TLS (rTLS) predictive model. It achieved AUCs of 0.922 in the training cohort, 0.852 in the TLS validation cohort, 0.724 in the Chinese-TNBC cohort, 0.919 in the DUKE cohort, and 0.883 in the I-SPY2 cohort. The rTLS predictive model demonstrated robust predictive performance, including across various patient subgroups defined by age, menopausal status, Ki67 level, cT stage, cN stage, and MammaPrint gene status. In YNCH cohort and DUKE cohort, prognostic analysis showed that low rTLS predictive score was significantly correlated with better disease-free survival in TNBC receiving NAT, and Cox regression analysis also confirmed the rTLS predictive score was a strong independent prognostic factor. H&E-stained tumor slides of patients with high predictive score revealed the presence of TLSs, and these patients achieved pathological remission after NAT, resulting in a favorable outcome. Pathomic features further explained the pathological heterogeneity of TNBC with different responses to NAT. Conclusions: The rTLS predictive model, which accurately predicted the presence of TLSs and treatment response to NAT in TNBC, held promise for future clinical application in formulating personalized and effective treatment strategies for TNBC, ultimately improving prognosis.

J. Kim¹, C. HyeongJin², H. Lee³, E. Cho³, J. Lee⁴, S. Kim⁴, Y. Park¹, H. Kim²; ¹Division of Hematology-Oncology, Department of Medicine, Samsung Medical Center, Seoul, KOREA, REPUBLIC OF, ²School of Pharmacy, Sungkyunkwan University, Suwan, KOREA, REPUBLIC OF, ³Department of Pathology, Samsung Medical Center, Seoul, KOREA, REPUBLIC OF, ⁴Department of Surgery, Samsung Medical Center, Seoul, KOREA, REPUBLIC OF.

BACKGROUND: Focal copy number amplification is known to play a crucial role in oncogenesis of breast cancer (BC). Especially, the amplification of interchromosomal oncogenic translocation such as t(8;11), t(8,17) and t(11;17) is known to be the main feature of BCs. However, there is a lack of clinical evidence that such interchromosomal focal amplifications promotes tumor progression. Here, we investigated the prevalence of interchromosomal focal amplifications in primary and metastatic BCs collected at a single hospital. METHODS: 65 primary tumors were collected from BC surgical specimens obtained after curative surgery following NAC (neoadjuvant chemotherapy), and 10 metastatic BC tumors from post-adjuvant biopsies. Whole genome sequencing (mean coverage 62X, range 51~74X for tumors; mean coverage 33X, range 30~40X for patient-matching normal bloods) was generated and analyzed to profile focal amplifications and genomic characteristics. BCs were categorized into three prognostic groups: primary with dismal prognosis (TP-D; total N = 34; HER2+ N = 12), primary with good prognosis (TP-ND, total N = 31; HER2+ N = 6), and metastatic (TM; total N = 10; HER2+ N = 6). Genomic characteristics including focal amplifications were compared between these three prognostic groups. RESULTS: Interchromosomal focal amplifications co-carrying ERBB2/CDK12/C17orf37 were highly prevalent in HER2+ TM (5/6: 83.3%) and TP-D (10/12: 83.3%) patients compared to HER2+ TP-ND (1/6: 16.7%) (P = 0.040 in TM vs TP-ND; P = 0.013 in TP-D vs TP-ND; one-sided Fisher’s exact tests), indicating that those amplifications may be early events during aggressive tumor progression in HER2+ BCs. Interestingly, we also found that interchromosomal oncogenic focal amplifications cargo more oncogenes as the patients carrying those amplifications have worse prognosis (TP-ND: mean 6.5 oncogenes; TP-D: mean 15.3 oncogenes; TM: mean 21.4 oncogenes present in interchromosomal oncogenic focal amplification), which was statistically significant (P = 0.0043 in TM vs. TP-ND; P = 0.034 in TP-D vs TP-ND; wilcox ranksum test). CONCLUSION: The high prevalence of ERBB2/CDK12/C17orf37 co-amplified interchromosomal focal amplifications in HER2+ poor prognosis groups (TM and TP-D) indicate a role for these amplicons in tumor progression. Additionally, the correlation between an increasing number of oncogenes in interchromosomal focal amplifications and poor prognosis suggests that these amplifications may evolve by accumulating oncogenic elements during tumor progression, thereby promoting tumor aggressiveness.

J. Park¹, F. Howard², P. Wysocki³, A. Przywara³, K. Zeng¹, J. Cappadona¹, B. Machura¹, J. Szpor⁴, C. Bifulco⁵, B. Piening⁵, V. Reid⁶, J. Witowski¹, F. Esteva⁷, K. J. Geras¹; ¹AI, Ataraxis AI, New York, NY, ²Department of Medicine, University of Chicago, Chicago, IL, ³Department of Oncology, Jagiellonian University Medical College, Kraków, POLAND, ⁴Department of Pathomorphology, Jagiellonian University Medical College, Kraków, POLAND, ⁵Earle A. Chiles Research Institute, Providence Cancer Center, Portland, OR, ⁶Hall Perrine Cancer Center, Mercy Medical Center, Cedar Rapids, IA, ⁷Division of Hematology/Oncology, Northwell Health Cancer Institute, Lenox Hill Hospital, New York, NY.

Background: Neoadjuvant therapy (NAT) is standard of care for many early-stage breast cancer patients. Traditional clinical characteristics such as ER and HER2 receptor status provide useful prognostic information, but lack sufficient precision when used in isolation. Basic histological features, such as tumor infiltrating lymphocytes, are also associated with NAT response. With the advent of computational pathology and machine learning, it is now possible to extract complex spatial and morphological predictive patterns from digitized slides that are not apparent to the human eye. Integrating histopathologic features with clinical variables holds promise for building robust predictive models that can guide personalized treatment planning, reduce overtreatment, and improve survival outcomes in breast cancer patients undergoing neoadjuvant therapy. Ataraxis Breast Neo (ATX-N) is a multi-modal artificial intelligence test that integrates morphological features extracted from standard H&E-stained slides with baseline clinical data (ER/PR/HER2 status, staging, age, IDC/ILC presence). Here we validate the accuracy of ATX-N in diverse patient populations. Methods: Whole slide images of H&E-stained biopsy specimens and corresponding clinical variables were collected from 563 patients who have undergone neoadjuvant therapy across UChicago (n=257), Providence Health (n=65), Jagiellonian University (n=115), and public IMPRESS dataset (n=126) cohorts. Across the cohorts, there are 186 TNBC patients, 201 HER2+ patients and 164 HR+HER2- patients. The remaining 12 patients had HER2 equivocal status or the information was missing. ATX-N scores were generated using a locked model. Performance of ATX-N was assessed for pathologic complete response (pCR) AUC. Reported AUCs are pooled using random effects models. Results: The overall pCR rate was 40.1%. ATX-N achieved a pooled AUC of 0.721 (95% CI: 0.646-0.796) across four independent cohorts, with cohort-level AUCs ranging from 0.644 to 0.804 (Table 1). Specifically, ATX-N achieved an AUC of 0.644 (CI: 0.577-0.716) on UChicago cohort, 0.734 (CI: 0.598-0.846) on Providence cohort, 0.804 (CI: 0.717-0.881) on Jagiellonian cohort, and 0.714 (CI: 0.624-0.797) on IMPRESS cohort. Notably, the confidence intervals for all cohorts did not overlap with 0.5, indicating that the model’s discriminative performance was better than random classification. In stratified analyses across molecular subtypes, pooled AUCs for TNBC, HER2+, and HR+HER2- patients are 0.675 (CI: 0.597-0.752), 0.633 (CI: 0.488-0.778), and 0.689 (CI: 0.578-0.799), respectively. When stratifying all patients based on ATX-N predictions, the bottom quarter had 15.6% pCR rate whereas the top quarter had 63.8% pCR rate. Conclusions: ATX-N demonstrates promising performance in prediction of pCR in breast cancer across diverse datasets. Larger studies are needed to confirm its efficacy and utility.

W. Zhu, A. Gutierrez, L. Hsu, D. Tripathy, J. Litton, B. Arun, C. Barcenas,, A. Singareeka Raghavendra; Breast Medical Oncology, MD Anderson Cancer Center, Houston, TX.

Background: Vast critical precision oncology data, including results for actionable findings like germline BRCA (gBRCA) mutations, remain hidden in unstructured electronic health record (EHR) notes. We sought to develop and evaluate a systematic, zero-shot large language model (LLM) framework to accurately classify gBRCA mutation status in patients using real-world EHR clinical note narratives. Methods: In this retrospective study at a large single academic cancer center, we utilized two distinct cohorts randomly sampled from a population of metastatic breast cancer patients whose genetic data were confined to unstructured notes. The development cohort (n=439) was used for model optimization, and an independent validation cohort (n=510) was used for objective performance evaluation. Our framework involved selecting an optimal base model by evaluating three LLMs with standardized prompts to ensure an unbiased comparison. The selected model that demonstrated the best performance then underwent further optimization of its input data and iterative prompt engineering to enhance performance on a granular 11-class schema. Key performance metrics included classification accuracy and F1-score (the harmonic mean of precision and recall). Results: Our final optimized framework achieved an accuracy of 86.1% and a weighted F1-score of 0.88 on the 11-class task in our development cohort. Crucially, the model demonstrated robust performance and generalization on our validation cohort with an accuracy of 82.4% and weighted F1 score of 0.81. A qualitative error analysis of the test cohort set clarified that many discrepancies were attributable to heterogenous clinical annotation criteria between the two cohorts. Significant performance gains on the development cohort set were attributed to prompt engineering, which improved accuracy by 7.5%: from 78.6% to 86.1%. Conclusion: Zero-shot LLMs, when applied within a systematic framework characterized by its multi-phase optimization of inputs and iterative engineering of prompts that enforce step-by-step reasoning, can accurately and reliably extract critical genetic data from unstructured EHR notes. This approach provides a reproducible and scalable blueprint for developing trustworthy clinical artificial intelligence methodology to accelerate research and precision oncology.

Z. Shah¹, S. S. Afridi², M. Ombada¹, A. M. Roy³, A. LeVee⁴, S. Premji⁵, V. Gupta¹, N. M. Lopetegui³, D. M. Quiroga⁶, R. L. Sacks⁷, S. Shaikh⁸, Y. Abdou⁹, H. Yu¹⁰, A. Madabhushi¹¹, R. Parikh⁷, L. N. Chaudhary¹², E. Levine¹, M. Lambertini¹³, K. Kalinsky⁷, S. Kabraji¹, S. Gandhi⁷; ¹Medicine, Roswell Park Cancer Center, Buffalo, NY, ²Medicine, SUNY Upstate Medical University, Buffalo, NY, ³Medicine, OSU Comprehensive Cancer Center, Colombus, OH, ⁴Medicine, UCLA Medical Cente, Los Anjeles, CA, ⁵Medicine, Sarah Cannon Cancer Institute, Nashville, TN, ⁶Medicine, Roswell Park Cancer Center, Colombus, OH, ⁷Medicine, Emory Winship Cancer Institute, Atlanta, GA, ⁸Medicine, UT Health San Antonio, Mays Cancer Center , San Antonio, TX, ⁹Medicine, UNC Lineberger Comprehensive Cancer Center, Chapel Hill, NC, ¹⁰Biostatistics & Bioinformatics, Roswell Park Cancer Center, Buffalo, NY, ¹¹Radiology & Imaging Sciences, Biomedical Informatics, Emory, Atlanta, GA, ¹²Medicine, Medical College of Wisconsin, Milwaukee, WI, ¹³Medicine, University of Genoa, Genova, ITALY.

Background: Artificial intelligence (AI) and large language models (LLMs) are increasingly explored as tools to support clinical decision-making in oncology. However, evidence validating their performance in complex breast cancer clinical scenarios (BCCS) remains limited. Given breast cancer’s diverse subtypes, evolving standards of care, and the need for nuanced, personalized treatment, we compared three LLMs for treatment decision-making to assess their capabilities and determine their readiness for integration into real-world breast oncology clinics. Methods: Ten breast cancer cases mimicking real-world scenarios were posed to three LLMs: ChatGPT-4o (GPT), DeepSeek-R1 (DS), and OpenEvidence (OE). 16 breast medical oncologists (BMOs; assistant to full professors) from 9 academic centers graded responses using a 5-point Likert scale (1 = poor, 5 = excellent) for clinical accuracy, clarity, relevance, and usability. Questions(Q) 1-9 assessed treatment decision-making cases; Q10 tested multimodal image interpretation (MMI) skills. Repeated measures ANOVA evaluated model differences, followed by Tukey’s post hoc comparisons. Results: On BCCS Q1-9 spanning all major breast cancer subtypes and treatment settings, including early-stage, metastatic, neoadjuvant, and adjuvant, OE achieved the highest mean score (3.91 ± 0.48; 2.57-4.43), significantly outperforming both GPT (3.19 ± 0.67; 2.36-3.93) and DS (2.93 ± 0.54; 1.50-3.86) in overall performance (p < 0.0001), with large effect sizes (Cohen’s d = 1.12 vs. GPT; d = 2.84 vs. DS). Repeated measures ANOVA identified significant differences among models in 5/9 (56%) Q (p < 0.05), as shown in the Table. Pairwise comparisons showed OE outperformed GPT in 4/5 (80%) and DS in 5/5 (100%) significant BCCS. GPT modestly outperformed DS in 2/5 (40%) (Cohen’s d = 0.41). GPT showed the highest inter-reviewer variability (SD = 0.67 vs. 0.48-0.54), indicating less agreement among BMOs on its responses. OE provided well-supported treatment recommendations and recurrence risk assessments with citations, though it lacked MMI. DS relied on optical character recognition in Q10 (inflammatory breast cancer image; 2.93 ± 1.33), limiting its utility in image-based BCCS. GPT, the only vision-enabled model, scored highest (4.29 ± 1.14), highlighting its strong potential for MMI integration in breast cancer diagnostic workflows. Conclusions: This is the first comparison of GPT, DS, and OE in BCCS. OE generated the most guideline-concordant treatment choices across BCCS, showing strong potential as a clinical decision support tool, though its verbosity may require streamlining. GPT showed moderate performance, while DS lagged in clinical relevance and accuracy. These findings highlight the promise of LLMs in breast oncology and the need for further refinement to ensure reliability and real-world applicability.

A. Khalil¹, J. Juybari², J. Hamilton³; ¹CompuMAINE Lab, Chemical and Biomedical Engineering, University of Maine, Orono, ME, ²CompuMAINE Lab, Electrical and Computer Engineering, University of Maine, Orono, ME, ³CompuMAINE Lab, Graduate School of Biomedical Sciences and Engineering, University of Maine, Orono, ME.

Introduction: Mammography-based risk prediction models show promise, with top AI models achieving AUCs near 0.80¹, potentially outperforming traditional clinical models (TC8, BCSC, BRCAPRO, Gail) whose AUCs are ~0.60-0.65². However, AI models face limitations in widespread use due to issues with generalizability, calibration, and interpretability. Building on the understanding that tumor microenvironment changes promote breast tumorigenesis^3,4 and tissue organization influences tumor invasion⁵, our patented interpretable computational biophysics approach^6,7 separates mammographic dense tissue into “active” (structurally reorganizing) and “passive” subtypes^8,9. We found that active dense tissue associates with early tumor onset¹⁰, that our method is agnostic to processing vs. presentation mammograms¹¹, and our measurements synergize with volumetric breast density (VBD) and fibro-glandular volume (FGV) for a 3-yr risk AUC of 0.82 [under review]. We present a retrospective case-control analysis demonstrating how a novel morphological metric, called anisotropic mass, augments the predictive power of our active dense tissue measurement. Methods: Using the OPTIMAM database^12,13, bilateral mammographic screenings were acquired on Hologic’s Selenia units ~3 years prior to diagnosis for n=197 cases (mean age = 59.4), or the last screening on record for n=3712 controls (mean age = 59.8). We calculated areas and anisotropic mass of fatty, active dense, and passive dense tissues. Anisotropy measures directional preference by quantifying departure from uniform directional gradients¹⁴ and anisotropic mass is the sum of anisotropy across an area. VBD and FGV were obtained from Volpara® (Lunit Inc.) software. An interaction generalized linear model (GLM) predicted cancer vs. control status using VBD, FGV, patient age, amounts of fatty, passive dense, and active dense tissue, our mammographic percent density, and active dense tissue’s anisotropic mass. AUC and 95% confidence intervals were calculated from bootstrapped ROC curves.The images and data used in this publication are derived from the OPTIMAM imaging database (https://medphys.royalsurrey.nhs.uk/omidb/about-omi-db/)^12,13. We acknowledge the OPTIMAM project team and staff at the Royal Surrey NHS Foundation Trust who developed the OPTIMAM database, and Cancer Research UK who funded the creation and maintenance of the database. Results: Our model outputs an AUC of 0.85, with (0.84-0.86 95% CI), a very significant improvement (p-value < 10^-7) over our prior model (AUC = 0.82), validating anisotropy as a complementary measure to active dense tissue. Conclusions: Our biophysics-based approach offers a significant advantage in breast cancer risk prediction. It generates interpretable measurements that match or exceed deep learning models’ performance while overcoming their drawbacks in generalizability, calibration, and interpretability. References:1. IBDW 2025, https://breastdensityworkshop.org/wp-content/uploads/2025/06/IBDW2025_program_online_6-2-2025.pdf.2. McCarthy et al. J Natl Cancer Inst, 2020. 112(5): p. 489-497.3. Bissel et al. Nat Med, 2011. 17(3): p. 320-9.4. Tanner et al. Proc Natl Acad Sci USA, 2012. 109(6): p. 1973-8.5. Nazari et al. Breast Cancer, 2018. 25(3): p. 259-267.6. Khalil & Batchelder U.S. Patent 10,467,755 B2, 2019.7. Khalil & Batchelder EU Patent 2,988,659 B1, 2022.8. Marin et al. Med Phys, 2017. 44(4): p. 1324-36.9. Gerasimova-Chechkina et al. Front. in Physiol., 2021. 12(595).10. Batchelder et al. medRxiv 2024: p. 2024.02.17.24302978.11. Khalil & Hamilton IBDW2025 poster #70 2025.12. Halling-Brown et al. Radiol Artif Intel, 2021. 3(1): p. e200103.13. OPTIMAM (OMI-DB) project: https://medphys.royalsurrey.nhs.uk/omidb/about-omi-db/.14. Hamilton et al. Front. in Oncol., 2022. 12: p. 991850.

M. Filipits¹, D. Hlauschek², J. Zhang³, M. Balic⁴, R. Kates⁵, R. Greil⁶, F. Fitzal⁷, N. Toro-Bauer⁸, G. Rinnerthaler⁹, H. Pinckaers¹⁰, K. Sotlar¹¹, Z. Bago-Horvath¹², W. Hulla¹³, P. Regitnig¹⁴, A. Piehler¹⁵, C. E. Geyer⁴, H. Kreipe¹⁶, J. Griffin¹⁷, N. Harbeck¹⁸, N. Wolmark¹⁹, M. Gnant²⁰; ¹Center for Cancer Research, Medical University of Vienna and ABCSG, Vienna, AUSTRIA, ²Statistics, ABCSG, Vienna, AUSTRIA, ³Biostatistics, Artera, Inc, Mountain View, CA, ⁴Department of Medicine, NSABP and the University of Pittsburgh, Pittsburgh, PA, ⁵Statistics, WSG, Moenchengladbach, GERMANY, ⁶Center for Clinical Cancer and Immunology Trials, Salzburg Cancer Research Institute, Salzburg, AUSTRIA, ⁷Department of Surgery, Hanusch Hospital Vienna, Vienna, AUSTRIA, ⁸Department of Surgery, Hospital Wiener Neustadt, Wiener Neustadt, AUSTRIA, ⁹Division of Clinical Oncology, Department of Internal Medicine, Medical University of Graz, Graz, AUSTRIA, ¹⁰AI, Artera, Inc, Mountain View, CA, ¹¹Institute of Pathology, Paracelsus Medical University Salzburg, Salzburg, AUSTRIA, ¹²Department of Pathology, Medical University of Vienna, Vienna, AUSTRIA, ¹³Institute of Clinical and Molecular Pathology, Hospital Wiener Neustadt, Wiener Neustadt, AUSTRIA, ¹⁴Diagnostic- and Research Institute of Pathology, Medical University of Graz, Graz, AUSTRIA, ¹⁵Biostatistics, Artera Inc., Mountain View, CA, ¹⁶Pathology, WSG, Hannover, GERMANY, ¹⁷Clinical Development, Artera, Inc, Mountain View, CA, ¹⁸LMU Breast Center, WSG and LMU Munich, Munich, GERMANY, ¹⁹Department of Surgery, NSABP and the University of Pittsburgh, Pittsburgh, PA, ²⁰Comprehensive Cancer Center, Medical University of Vienna and ABCSG, Vienna, AUSTRIA.

Background: Genomic assays are currently used to determine prognosis and predict treatment outcome in patients with estrogen receptor-positive (ER+), HER2-negative early breast cancer (EBC). However, their use is expensive, time-consuming and not available in many clinical facilities. We previously developed a multimodal artificial intelligence (MMAI) model using clinical and histopathological data from more than 12,000 patients from six phase III trials, offering a faster, non-genomic alternative for personalized risk stratification. We present the first external independent validation of this MMAI model using the ABCSG trial 8 – a large, prospective phase III trial of a low- to moderate-risk cohort of ER+ postmenopausal breast cancer patients receiving endocrine therapy (ET) only – to assess its prognostic performance for distant metastasis (DM). Methods: ER+ HER2- patients from the prospective-randomized trial ABCSG 8 with digitized baseline H&E-stained sections, complete clinical data (age, tumor size and nodal status), and long-term follow up were included in this pivotal validation study. The locked MMAI model includes a continuous score and predefined risk groups. Prognostic performance was assessed using univariable and multivariable Cox proportional hazard models, with hazard ratios (HR) and corresponding 95% confidence intervals (CI). Results: MMAI raw scores were generated for 2,109 patients. The median follow-up time was 9.5 years. Despite the overall lower risk profile (G1 and G2 only) of the ET-only population, the MMAI classified 77% patients as low risk, 9% as intermediate risk, and 14% as high risk. The estimated 10-year DM-free rates were 94.7% (93.3% – 95.8%) for low-risk patients, 89.1% (82.5% – 93.3%) for intermediate-risk patients, and 77.3% (71.3% – 82.2%) for high-risk patients. In univariable analysis (UVA), both the continuous MMAI raw score and MMAI risk groups were significantly associated with risk of DM: MMAI raw score (HR [95% CI] = 2.22 [1.91 – 2.58], p<0.001) and MMAI risk group (intermediate vs. low risk HR [95% CI] = 2.19 [1.34 – 3.57], p=0.002; high vs. low risk HR [95% CI] = 4.45 [3.19 – 6.19], p<0.001). In multivariable analysis (MVA) adjusting for age, tumor size, nodal status, both continuous MMAI raw score (p<0.001) and MMAI risk group (p<0.001) remained independently associated with DM. The MMAI raw score remained significantly associated with the risk of DM across clinically meaningful subgroups, including lymph node status (N0 and N1), tumor grade (Grade 1 and 2), histology (invasive ductal carcinoma and invasive lobular carcinoma), and trial-defined Ki67 levels (≤5%, 6-29%, ≥30%). Additionally, the image-only component of the MMAI was independently associated with risk of DM (p<0.001 in both UVA and MVA). Conclusion: This study validates the MMAI model, originally developed for HR+ HER2- EBC, as a strong prognostic tool for DM in lower-risk postmenopausal women with ER+ HER2- EBC. This MMAI technology promises to improve personalized risk stratification and adjuvant treatment decisions in ER+ HER2- EBC as a cost-effective, non-tissue consuming, and faster alternative to genomic assays.

J. Moon¹, J. Lee¹, B. Kang¹, W. Kim², J. Kim³, J. Baek², H. Park¹, H. Kim², G. Baek⁴; ¹Department of Surgery, School of Medicine, Kyungpook National University, Kyungpook National University Chilgok Hospital, Daegu, KOREA, REPUBLIC OF, ²Department of Radiology, School of Medicine, Kyungpook National University, Kyungpook National University Chilgok Hospital, Daegu, KOREA, REPUBLIC OF, ³School of Computer Science and Engineering, Kyungpook National University, Daegu, KOREA, REPUBLIC OF, ⁴Research and Development, BeamWorks Inc., Daegu, KOREA, REPUBLIC OF.

Using a Real-Time Artificial Intelligence Ultrasound System with Computer-Aided Detection and Diagnosis to Distinguish Ductal Carcinoma In Situ and Invasive Ductal Carcinoma Background: Real-time artificial intelligence-based computer-aided detection/diagnosis (AI-CAD) for breast ultrasound enables immediate assessment of suspicious lesions with probability of malignancy (POM) outputs, potentially aiding in the differentiation between ductal carcinoma in situ (DCIS) and invasive ductal carcinoma (IDC). Identifying ultrasonographic features with significant POM differences between DCIS and IDC may refine clinical triage. Methods: We retrospectively analyzed 34 cases (DCIS=12, IDC=22) assessed by a real-time AI solution (CadAI-B for Breast cancer) during breast ultrasound at a tertiary center. POM outputs and BI-RADS categorization were compared across ultrasonographic features including shape, margin, echo pattern, and orientation. Features demonstrating significant POM differences between DCIS and IDC were identified, and comparative analyses were performed to evaluate CadAI-B’s differential diagnostic utility. Results: Overall, IDC demonstrated a higher mean POM (0.365 ± 0.306) compared to DCIS (0.126 ± 0.196, p<0.001). Ultrasonographic features with significant POM differences included irregular shape (IDC: 0.344 vs DCIS: 0.135, p=0.002), indistinct margins (IDC: 0.318 vs DCIS: 0.012, p=0.014), microlobulated margins (IDC: 0.419 vs DCIS: 0.132, p=0.031), hypoechoic pattern (IDC: 0.345 vs DCIS: 0.135, p=0.018), and parallel orientation (IDC: 0.362 vs DCIS: 0.044, p=0.003). These features were associated with larger POM differentials between IDC and DCIS, suggesting enhanced discrimination capacity. Under CadAI-B, IDC cases were more frequently categorized into higher BI-RADS categories, with IDC cases showing a substantial shift towards BI-RADS 4B, 4C, and 5 with high POM outputs. Notably, DCIS cases retained lower POM even when categorized into BI-RADS 4A and 4B. Conclusions: CadAI-B may help distinguish IDC from DCIS during real-time breast ultrasound examinations by reflecting differences in probability of malignancy. This may assist clinical decision-making during scanning by providing consistent, objective data when evaluating suspicious breast lesions. Further studies are needed to confirm its diagnostic value in broader clinical settings. Keywords: AI-CAD, Breast Ultrasound, DCIS, IDC, Probability of Malignancy

L. Schwartzberg¹, H. S. Rugo², A. Forsythe³, D. Flora⁴, S. Glück⁵, S. Grieve⁶, R. Campden⁶, R. Liu⁶, J. M. Rege³, P. A. Kaufman⁷; ¹Medical Oncology and Hematology, Renown Health-William N. Pennington Cancer Institute, Reno, NV, ²Duarte Cancer Center, City of Hope, Duarte, CA, ³Executive, Oncoscope-AI, Miami, FL, ⁴Yung Family Cancer Center, St. Elizabeth Healthcare, Edgewood, KY, ⁵Consultant, Oncology Immunotherapy Biotech and AI, Miami Beach, FL, ⁶Data management, Eviviz, Vancouver, BC, CANADA, ⁷Oncologist, The University of Vermont Medical Center, Burlington, VT.

Background: Oncologists face growing difficulty in staying current with rapidly evolving data from congresses, journals, guidelines, and regulatory updates. Creating a well-organized, annotated clinical trial (CT) data library in an easily digestible format is time- and labor-intensive. To address this, we developed and validated a living-DST for breast cancer (BC), powered by an agentic AI system that supports daily human conducted / AI augmented SLR integrated with guidelines, regulatory approvals and ongoing CT results. Methods: An agentic AI system using GPT-4.1 & o3 (OpenAI), and Claude Sonnet-4 (Anthropic) was designed to emulate expert-led Cochrane-compliant SLR processes. The system follows an annotation manual, decomposes tasks into subtasks, and self-debugs and validates. The annotation manual was constructed by human scientists manually annotating 29,236 clinical trial abstracts across BC, lung, and prostate cancer from 2019-current. Each study was annotated with 4 review variables (population, intervention/comparator, reported outcomes, study design) and 32 extraction variables: i.e. TNM staging, histology, biomarkers, risk factors, treatment line, intervention, study design, size, follow-up, outcomes (overall and progression/disease-free survival, response, QoL), subgroup data, and toxicity. Structured data were integrated with national guideline-based treatment pathways, including new evidence beyond guidelines, forming a real-time, evidence-linked DST. Accuracy was assessed vs. 1,997 human annotations as validation set. DST evidence precision was evaluated against the AI chatbots: ChatGPT, Perplexity, and Consensus using 6 criteria for 8 cases on treatment options in BC with different treatment path, subpopulation, and biomarker variables. Accuracy for each criterion was defined as the percentage of questions evaluated as positive over all eight questions. Results: Accuracy across review variables ranged from 95.1-97.2%. For extraction variables, accuracy exceeded 90% for all variables (91.5%-99.1%) with 50% of variables above 95%. Agentic AI annotated 1,997 publications in 7.25 hours vs. an estimated 727.4 hours required by human experts—yielding 99% of time savings. When compared to other AI tools, our DST outperformed by providing more complete and accurate evidence, retrieving recently published, potentially practice-changing pivotal trials (23% of all evidence), reporting all available clinical evidence, and linking to original citations and FDA labels where available (Table). Conclusion: By integrating human-conducted, AI-augmented daily SLR integrated with guidelines and regulatory approvals, our Living-DST delivers real-time, clinically actionable decision support. This system offers oncologists precise, patient-specific insights and may improve cancer treatment outcomes.

J. Casali da Rocha¹, A. C. Ricciardi², G. B. Pinheiro¹, D. S. Pegos¹, R. A. Romero³; ¹Oncogenetics, A.C.Camargo Cancer Center, São Paulo, BRAZIL, ²Department of Computer Science, Institute of Mathematics and Computer Sciences (ICMC), University of São Paulo, São Carlos – SP, BRAZIL, ³Department of Computer Science, Institute of Mathematics and Computer Sciences (ICMC), University of São Paulo, São Carlos, BRAZIL.

Background: Clinical criteria, such as Chompret 2015, guide genetic testing for Li-Fraumeni Syndrome (LFS) and TP53 mutations. However, these guidelines may overlook hereditary breast cancer risk in genetically diverse populations. In Brazil, the founder mutation TP53 p.R337H is highly prevalent, presenting with atypical tumor patterns and age of onset, often leading to underdiagnosis by conventional criteria. This study aims to assess whether machine learning (ML) and explainable artificial intelligence (XAI) can enhance the prediction of hereditary breast cancer subtypes, particularly TP53-related, in a real-world Brazilian cohort.Methods: We curated a dataset of 576 probands (515 with complete cancer history data) subjected to genetic testing and with structured clinical and family information.Three multiclass machine learning models were trained to predict three groups: TP53-negative, TP53 R337H, and other TP53 mutations. Performance was evaluated using ROC-AUC, and SHAP values were employed to interpret variable importance. Among the tested algorithms, CatBoost achieved the highest ROC-AUC values, although Random Forest was selected for interpretability. Kaplan-Meier analysis assessed age at diagnosis across groups.Results: Among 515 probands, 70.1% had a personal cancer history. Among those with cancer, 32.7% carried pathogenic mutations, with TP53 variants being the most frequent: 7.8% for R337H and 6.1% for other TP53 mutations, totaling 13.8%. Other recurrent mutations included BRCA2 (3.0%), BRCA1 (2.5%), CHEK2 (1.4%), ATM (1.1%), and NF1 (0.6%). The remaining 67.3% had no identifiable pathogenic variant. The model achieved robust discrimination (AUC: 0.73 for TP53 R337H, 0.84 for other TP53, 0.85 for TP53-negative), outperforming the Chompret clinical criteria (AUC = 0.67). Unlike Chompret, the ML model distinguished between R337H and other TP53 mutations, offering a more tailored and data-driven approach to hereditary cancer risk stratification in genetically diverse populations. SHAP analysis identified key predictors for TP53 p.R337H, including early-onset maternal cancer, proband’s breast cancer, and early-onset brain tumors. Notably, TP53 p.R337H cases showed a significantly earlier median age at diagnosis (43 years) compared to TP53-negative cases, yet later than classical TP53 mutations. The model also accurately revealed that several TP53-negative individuals carried pathogenic variants in BRCA1, BRCA2, and ATM, with early-onset breast cancer mirroring TP53 carriers, highlighting the limitations of TP53-centric screening.Conclusion: Our findings reveal that Chompret criteria may miss substantial hereditary breast cancer risk associated with LFS, especially in populations with founder mutations like TP53 p.R337H. Explainable ML models integrating comprehensive clinical and familial data provide accurate and interpretable risk stratification, demonstrating superior performance in identifying diverse hereditary cancer patterns. These insights support the use of multigene panel testing and population-tailored triage strategies for improved genetic counseling and surveillance in hereditary breast cancer care, particularly in regions with high genetic heterogeneity.

T. Cantore¹, D. Hoang¹, L. R. Pal¹, A. Stemmer¹, S. Dhruba¹, T. Chang¹, S. Sammut², S. Lipkowitz³, R. S. Padma¹, C. Caldas⁴, N. Ulhas Nair¹, E. Ruppin¹; ¹Cancer Data Science Laboratory (CDSL), Center for Cancer Research (CCR), National Cancer Institute, National Institute of Health (NIH), Bethesda, MD, ²Division of Breast Cancer Research, Breast Cancer Now Toby Robins Research Centre, The Institute of Cancer Research, London, UK., London, UNITED KINGDOM, ³Women’s Malignancies Branch, Center for Cancer Research (CCR), National Cancer Institute (NCI), National Institute of Health (NIH), Bethesda, MD, ⁴Department of Clinical Biochemistry and Institute of Metabolic Science, School of Clinical Medicine, University of Cambridge, Cambridge, UK, Cambridge, UNITED KINGDOM.

BACKGROUND:Neoadjuvant therapy is central to breast cancer treatment, yet response rates vary widely, and robust biomarkers to predict pathological complete response (pCR) are lacking. While expression-based signatures such as Oncotype DX and MammaPrint routinely guide adjuvant treatment decisions, no tools are approved for the neoadjuvant setting. Although some predictive markers have been explored for HR-positive and HER2-positive disease, triple-negative breast cancer (TNBC) lacks reliable predictors despite its aggressive nature and unmet clinical need. Recent single-cell transcriptomic studies reveal that multiple molecular subtypes can coexist within individual tumors, challenging the conventional one-subtype-per-patient paradigm. We hypothesized that modeling tumors as compositions of intra-tumoral subtypes would improve neoadjuvant therapy response prediction. METHODS: We developed BRIDGE (Breast Intratumoral Deconvolution of Gene Expression), the first computational framework that deconvolves bulk breast cancer transcriptomes to quantify the relative abundance of malignant subtype populations within each tumor and predict neo-adjuvant therapy response. BRIDGE involves three key steps. (I) Subtype Deconvolution: We constructed a reference matrix of ‘pure’ molecular subtype profiles from an integrated single-cell RNA-seq compendium and applied support vector regression to infer subtype abundances per tumor. (II) Response prediction from bulk expression: We trained logistic regression models to predict pCR taking as input subtype compositions using 33 datasets from 17 independent neoadjuvant cohorts. These included 11 HR-positive, 10 TNBC chemotherapy, and 12 anti-HER2 datasets. For each treatment, models were trained on the three largest datasets and validated on the remaining 24. (III) Response prediction from histopathology: we integrated BRIDGE with Path2Omics, a deep learning framework that infers gene expression from histopathology slides, enabling transcriptome-free prediction in six datasets with available image data. RESULTS:BRIDGE accurately estimated subtype composition, achieving a mean Spearman correlation of 0.80 on simulated pseudo-bulk data. For HR-positive chemotherapy response prediction, BRIDGE achieved a mean ROC-AUC of 0.78 and an odds ratio (OR) of 7.8 – outperforming existing biomarkers. For anti-HER2 therapy, BRIDGE reached a mean ROC-AUC of 0.77 and OR of 7.75, substantially outperforming the use of HER2-enriched subtype classification alone (OR = 3.1). For TNBC chemotherapy-treated cohorts, BRIDGE achieved a mean ROC-AUC of 0.71 and an OR of 4.7, addressing a major clinical gap due to the lack of established predictors. When applied to AI-inferred transcriptomes from H&E slides, BRIDGE maintained high accuracy (AUC > 0.85 for chemotherapy and > 0.70 for anti-HER2), with response scores highly concordant with those derived from true transcriptomic data (correlation > 0.7). CONCLUSION: BRIDGE is the first robust and interpretable framework for deconvolving malignant subtypes to predict neoadjuvant therapy response across diverse breast cancer subtypes. By modeling intra-tumoral heterogeneity and leveraging histopathology images, BRIDGE bridges the gap between expression-based profiling and routine pathology. Its low-cost, image-based approach supports personalized, subtype-informed treatment strategies and helps democratize precision oncology in the neoadjuvant setting.

H. Lin¹, Y. Wang², J. Dai², L. Pusztai²; ¹Internal Medicine, Danbury Hospital, Danbury, CT, ²Yale Cancer Center, Yale School of Medicine, New Haven, CT.

Identifying recurrent gene amplifications in breast cancer for antibody-drug-conjugate development Hao-Kuen Lin ¹, Yueyue Wang ², Jiawei Dai ², Lajos Pusztai ^{2* 1}Danbury Hospital, Danbury, CT ²Yale School of Medicine, Yale Cancer Center, New Haven, CT Background: We hypothesized genes with frequent copy number amplifications in breast cancer resulting in overexpression of membrane proteins could serve as potential targets for novel antibody drug conjugates (ADC). Methods: DNA copy number aberrations (CNA) and mRNA expression data was retrieved from cBioPortal (https://www.cbioportal.org/) for 3,970 patients with breast cancer. We defined gene amplification as CNA value >2. Chi-square or Fisher’s Exact test was used to test for significant difference in amplification frequency between 3 clinical subtypes ER+ (n=2474), HER2+ (n=712), and TNBC (n=632). Wilcoxon signed-rank test was used to test if gene mRNA expression is higher in cancers with amplification compared to wild type. We shortlisted amplified genes as candidates if amplification has led to overexpression and the gene encoded a membrane protein based on information in the Protein Data Bank (https://www.rcsb.org/) and Human Protein Atlas (https://www.proteinatlas.org/). We retrieved normal tissue mRNA expression data from GTEx (https://www.gtexportal.org/home/) covering 21 different tissue types and compared candidate ADC target expression in cancer versus normal tissues using DESeq2. We prioritized ADC candidates by the following criteria (i) CNA frequency > 15% across all breast cancers or within a subtype (FDR<.0001) (ii) higher expression in amplified cancers (FDR<0.001), (iii) mean Log₂ Fold expression > 1.5 in cancer compared to each normal tissue one by ones, and (iv) known membrane protein that is primarily located in the cell membrane and ability for internalization. Results: 963 unique genes were recurrently amplified and showed higher gene expression in amplified cancers across all breast cancers (722 genes) and/or significantly amplified only in ER+ (206 4 genes), HER2+ (216 200 genes), and TNBC (114 37 genes), and. how many were amplified in all subtypes?. Of the amplified and overexpressed genes, ⋯N=?…154 genes were membrane proteins. were membrane proteins. Among these membrane membrane proteins, 9 genes had higher expression in breast cancer than in normal tissues and primary located on cell membrane with ability for internalization. They (including HER2) are highlighted in Table 1 met all 4 criteria as potential novel ADC targets⋯N=?… had higher expression in breast cancer than in normal tissues. Nine genes (including HER2) highlighted in bold in Table 1 met all 4 criteria as potential novel ADC targets. Conclusion: We have identified several novel potential targets for ADC development in breast cancer.

M. Jahanzeb¹, K. Haines², T. Buchholz³, R. Butler⁴, W. J. Gradishar⁵, S. A. Hurvitz⁶, T. A. King⁷, R. L. Mahtani⁸, T. Mamounas⁹, H. McArthur¹⁰, M. Morrow¹¹, A. P. O’Dea¹², J. O’Shaughnessy¹³, M. D. Pegram¹⁴, H. S. Rugo¹⁵, C. Shah¹⁶, F. Vicini¹⁷, N. Wolmark¹⁸; ¹Medical Oncology, Florida Atlantic University Schmidt College of Medicine, Boca Raton, FL, ²Medical Oncololgy, OncAdvisor, Delray Beach, FL, ³Radiation Oncology, Scripps MD Anderson Cancer Center, San Diego, CA, ⁴Diagnostic Radiology, Yale University School of Medicine, New Haven, CT, ⁵Medical Oncology, Northwestern University Feinberg School of Medicine, Chicago, IL, ⁶Medical Oncology, Fred Hutch Cancer Center, Seattle, WA, ⁷Surgical Oncology, Emory Winship Cancer Institute, Atlanta, GA, ⁸Medical Oncology, Miami Cancer Institute, Miami, FL, ⁹Surgical Oncology, AdventHealth Cancer Institute, Orlando, FL, ¹⁰Medical Oncology, UT Southwestern Medical Center, Dallas, TX, ¹¹Surgical Oncology, Memorial Sloan Kettering Cancer Center, New York, NY, ¹²Medical Oncology, University of Kansas Medical Center, Westwood, KS, ¹³Medical Oncology, Baylor University Texas Oncology, Dallas, TX, ¹⁴Medical Oncology, Stanford University School of Medicine, Standford, CA, ¹⁵Medical Oncology, UCSF Helen Diller Family Comprehensive Cancer Center, San Francisco, CA, ¹⁶Radiation Oncology, AHN Cancer Institute, Pittsburgh, PA, ¹⁷Radiation Oncology, MHP Radiation Oncology Institute, Farmington Hills, MI, ¹⁸Surgical Oncology, University of Pittsburgh Hillman Cancer Center, Pittsburgh, PA.

Background: Second opinions and multidisciplinary review (MDR) can inform and impact the management of cancer patients, yet these resources are not universally accessible. We previously presented over 400 anonymized complex cancer scenarios (cases) via video conference to MDR panel experts—including specialists in medical oncology, surgical oncology, radiation oncology, radiology, and a moderator—across 5 major tumor types and tracked their treatment recommendations. In this study, we focus on breast cancer cases to compare how recommendations from 3 leading AI models compare to those of MDR panel experts. Methods: We selected 50 complex breast cancer scenarios previously adjudicated by MDR panels from a larger database of over 400 anonymized cancer cases collected between 2020 and 2021. These cases were analyzed by 3 different foundational AI models (OpenAI’s ChatGPT 4.5, Anthropic’s Claude Opus 4 and Google’s Gemini Ultra) using PrecisCa’s proprietary method for prompts. The AI recommendations from each system were scored on a scale of 1-5 (5 being the highest) for completeness, reasoning, clarity, menu of options, recency, and relevance as compared to the MDR panel recommendations. The maximum possible competence score was 30 points per scenario and an aggregate score of 1500. Final AI recommendations were also compared with current National Comprehensive Cancer Network (NCCN) guidelines for serious omissions. Comparison in reverse (additional AI options that the experts missed) was not performed due to interval changes in treatment recommendations over the past 4 years. Results: Patient characteristics and aggregate concordance and competence scores are listed in Table 1 below. While there was some variation in the concordance and competence rates of the 3 AI models, there was excellent concordance with the expert opinion for all of them. Discordant cases were reviewed and primarily involved minor differences that would not have altered the patients’ management significantly. Relatively speaking within the 6 categories listed above, AI systems excelled in clarity and reasoning and less so in relevance and menu of options. Conclusions: This study demonstrates a high degree of concordance between 3 leading AI models and expert MDR panels for common yet complex breast cancer clinical scenarios. These findings suggest that AI tools have matured sufficiently to serve as valuable decision support aids in oncology practice, particularly where expert review may be limited or unavailable. Careful human oversight remains essential to ensure safe and personalized cancer care.

T. Goncalves¹, D. Pulido-Arias¹, C. Perrino², M. Cleveland¹, E. Gerstner³, J. Kalpathy-Cramer⁴, K. Flaherty³, V. Zanfagnin⁵, A. Iafrate⁵, S. Sirintrapun⁵, A. Parwani⁶, G. Tozbikian⁶, J. Cardoso⁷, D. Sgroi⁵, C. Bridge¹, A. E. Kim³; ¹Athinoula A. Martinos Center for Biomedical Imaging, Mass General Hospital, Boston, MA, ²Department of Pathology, Lahey Hospital & Medical Center, Burlington, MA, ³Cancer Center, Mass General Hospital, Boston, MA, ⁴Division of Artificial Medical Intelligence, University of Colorado, Aurora, CO, ⁵Department of Pathology, Mass General Hospital, Boston, MA, ⁶James Cancer Center, Ohio State University, Columbus, OH, ⁷Faculty of Engineering, University of Porto, Porto, PORTUGAL.

Background: At present, clinical decision making and drug discovery are based on an incomplete picture of tumor biology. Deep learning (DL) has demonstrated the ability to infer oncogenic molecular signatures from data acquired during standard oncologic care (e.g., histopathology [H&E]). To advance the paradigm of DL from a tool that merely predicts outcomes into an engine that uncovers new therapeutic insights, we present MERLIN: Multimodal Evaluation of therapeutic Resistance through deep Learning INtegrating multimodal data. Methods: To develop MERLIN, we curated a dataset of 3111 H&E whole slide images (WSIs) and patient-matched clinico-genomic data from The Cancer Genome Atlas-Breast Cancer dataset and cBioPortal. This dataset was split into training (70%), validation (15%), and testing (15%) sets. Using attention-based multiple instance learning and several multimodal fusion techniques, we trained uni-modal (input: H&E) and multimodal (input: H&E and clinical data) models that quantify a patient’s tumor microenvironmental (TME) phenotypes. The ground truth for training these models was defined by each patient’s bulk RNA-sequencing data and Gene Set Enrichment Analysis protocols. Consistent with our prior work, image patching, feature extraction, and data augmentation strategies were applied during model training. Results: MERLIN achieved robust performance in deriving TME insights using only a patient’s H&E WSI and clinical metadata (Table 1). Using attention maps, MERLIN correctly identified morphology on H&E that governs tumor and immune function. A board-certified pathologist (C.M.P.) verified that high-attention areas, as determined by MERLIN, consistently identified spatial patterns of cell populations (e.g. cluster of lymphocytes) that matched with the phenotype of interest (e.g. T-cell cytotoxicity). Additional validation with tissue-matched immunofluorescence is ongoing. To demonstrate MERLIN’s ability to guide clinical decision making, we curated a multi-institutional dataset of 232 H&E WSIs of core biopsies from stage 2 & 3 triple negative breast cancer patients. Consistent with a known link between tumor-infiltrating lymphocyte activity and checkpoint inhibitor response, we identified correlations between the patient’s likelihood of achieving a pathologic complete response to neoadjuvant Keynote-522 and increased expression of T-cell cytotoxicity (p=0.017) as determined by MERLIN’s analysis of a patient’s H&E WSI. Conclusion: MERLIN has illustrated promise in obtaining TME-derived insights and identifying therapeutically relevant spatial patterns of cell populations from H&E and clinical metadata. We are now applying MERLIN to large datasets of clinically acquired data to identify new insights for cancer therapeutics without known a priori mechanisms of resistance.

G. Das¹, S. Wang¹, C. Xue¹, M. Vasquez¹, W. Yang², L. Xiaoxian², H. Zhao¹, W. Dong¹, S. T. Wong¹; ¹Systems Medicine and Bioengineering, Houston Methodist Research Institute, Houston, TX, ²Pathology and Laboratory Medicine, Emory University School of Medicine, Atlanta, GA.

Decoding resistance to trastuzumab deruxtecan in metastatic breast cancer: S100P signaling and immune-niche barriersGlori Das^1,2, Shiruo Wang¹, Catherine Xue¹, Matthew Vasquez^1,3, Wei Yang⁴, Xiaoxian Li⁴, Hong Zhao^1,3,5,*, Wenjuan Dong^1,*, Stephen T.C. Wong^1,2,3,5,*1 Department of Systems Medicine and Bioengineering, Houston Methodist Neal Cancer Center, Houston Methodist Hospital, Weill Cornell Medical College, Houston, TX 77030² Department of Biomedical Engineering, Texas A&M University College of Medicine, College Station, TX, 77843³ Advanced Cellular and Tissue Microscopy Shared Resource, Houston Methodist Research Institute and Houston Methodist Neal Cancer Center, Houston, TX 77030⁴ Department of Pathology and Laboratory Medicine, Emory University School of Medicine, Atlanta, GA 30322⁵ Houston Methodist Neal Cancer Center, Houston Methodist Hospital, Houston, TX 77030*Corresponding authors Background: Trastuzumab deruxtecan (T-DXd) has shown strong clinical efficacy in HER2-high and HER2-low metastatic breast cancer (MBC). However, many patients develop intrinsic or acquired resistance, particularly within sanctuary sites like the brain and bone. These niche-specific mechanisms remain poorly understood, limiting durable treatment responses and the design of effective combination therapies.Methods: We analyzed 109 MBC patients treated with T-DXd, stratifying response using clinical and radiographic criteria. Pre-treatment biopsies from 31 patients underwent spatial immune profiling using GeoMX Digital Spatial Profiler. Eight brain and bone metastases were profiled using a 570-plex antibody panel with spatial transcriptomics. Single-cell spatial transcriptomics (CosMX) was performed on 6 brain metastases. Functional validation included CRISPR-Cas9–mediated knockout and lentiviral overexpression of S100P, RAGE inhibition with azeliragon, and in vivo xenograft models.Results: T-DXd response was not associated with HER2 (p=0.128), ER (p=0.832), PR (p=0.617), or Ki-67 (p=0.212). Prior T-DM1 exposure predicted improved response (OR=5.27, p=0.048). Spatial profiling identified enhanced immune activation (CD3, CD8α, FABP4) in responders, especially in bone lesions. In contrast, brain metastases displayed conserved immunosuppressive profiles regardless of response. S100P was the most upregulated protein in resistant tumors (6.8-fold, p=0.0018), with marked spatial heterogeneity. S100P knockout sensitized HER2+ cells to T-DXd, while overexpression induced resistance. RAGE inhibition enhanced T-DXd cytotoxicity (~49-fold reduction in EC50). Single-cell analysis of resistant brain metastases revealed astrocytic upregulation of CAT and FN1, and macrophage expression of DR5 and CD163, consistent with a neuroprotective, immune-silent niche impeding ADC activity.Conclusions: Resistance to T-DXd is driven by tumor-intrinsic S100P-RAGE signaling and reinforced by niche-specific microenvironments—particularly in the brain—characterized by immunosuppressive astrocyte and macrophage programs. These findings nominate S100P, RAGE, and brain niche-derived stromal pathways as actionable targets to overcome resistance and extend the efficacy of T-DXd in MBC.

S. Ramezani, H. T. Le-Petross, M. Naser, R. V. Young, T. K. Morris, M. Kai, M. C. Stauder, B. Lim, L. Anthony, MDACC-IBC, C. D. Fuller, W. Woodward, C. R. Goodman; University of Texas MD Anderson Cancer Center, Houston, TX

Background: The diagnosis of inflammatory breast cancer (IBC) depends upon subjective clinical criteria, often leading to misdiagnosis and delayed treatment. While multiparametric MRI (mpMRI) is sensitive to characteristic IBC features such as global skin thickening, chest wall edema, and non-mass enhancement, its interpretation remains subjective and prone to inter-observer variability. An objective, automated clinical decision support system has the potential to improve patient outcomes enabling earlier and more accurate identification of IBC. Methods: DenseNet121-based deep learning models were developed for the automated classification of IBC using mpMRI sequences from a prospective institutional patient registry. Four separate datasets were analyzed based on sequence availability: T1-weighted without contrast (T1; N=791), diffusion weighted imaging (DWI; N=611), dynamic contrast-enhanced (DCE) imaging (N=484), and T2-weighted with contrast (T2c; N=267). Models for each sequence were evaluated using both midline-cropped single-breast (ipsilateral) images and bilateral full-breast images. For each sequence, stage-aware stratified splitting allocated 80% of patients for training/validation and 20% for testing. Image preprocessing included resizing to 144³ voxels, intensity normalization, and augmentation with random rotations and flips. Binary classification models were trained using 5-fold cross-validation and performance was evaluated using area under the receiver operating characteristic curve (AUC-ROC) on ensemble predictions. Results: Deep learning classification performance demonstrated strong overall discriminative ability across all sequences. DWI achieved the highest ensemble performance (AUC=0.96, mean fold AUC=0.92±0.02), followed by T2c (ensemble AUC=0.96, mean fold AUC=0.89±0.04), T1 (ensemble AUC=0.89, mean fold AUC=0.84±0.04), and DCE (ensemble AUC=0.80, mean fold AUC=0.84±0.02). Stage-specific analysis revealed high accuracy for early-stage disease (cT1-cT3: 92-97%) and classic IBC (cT4d: 94-95% accuracy), but reduced performance in distinguishing cT4b from cT4d cases (40-42% accuracy). Cross-validation demonstrated robust model stability with minimal variance across folds. The ensemble consistently outperformed individual fold averages, with DWI and T2c achieving clinically significant discrimination for overall IBC detection. A bilateral imaging approach did not significantly improve performance compared to analysis of the involved breast alone. Conclusion: Automated deep learning classification of IBC using mpMRI achieved excellent diagnostic performance, with DWI demonstrating superior discriminative ability (AUC=0.96). While the models consistently distinguished classic IBC (cT4d) from early-stage disease, classification of IBC from locally advanced non-IBC with extensive skin involvement and/or edema (cT4b) remains a challenge, reflecting known clinical diagnostic difficulties. Future work will incorporate combined-sequence analysis, probabilistic deep learning for uncertainty quantification, and an expanded cT4b cohort to improve distinction between IBC and non-inflammatory locally advanced breast cancers. These results demonstrate the potential of a deep learning-based clinical decision support system to improve diagnostic accuracy and help guide personalized treatment decisions for patients with IBC.

A. Fu, K. Wang, H. Jin, W. Gu; R&D, NVIGEN, Inc, Santa Clara, CA.

Background: Circulating tumor cells (CTCs) are valuable biomarkers for early detection, prognosis, and therapeutic monitoring in breast cancer. However, their rarity and the need for multiplexed marker detection present significant technical challenges. We introduce a novel platform combining quantum dot-based fluorescent magnetic nanoparticles (MyQuVigen) with AI-enhanced imaging to achieve sensitive, multiplexed CTC isolation and detection using standard fluorescence microscopy. Methods: NVIGEN’s MyQuVigen fluorescent magnetic nanoparticles are conjugated with antibodies targeting epithelial tumor markers (e.g., EpCAM, HER2, PD-L1, VEGF, EGFR), enabling efficient magnetic capture of CTCs from whole blood. The nanoparticles incorporate semiconductor quantum dots as fluorescent reporters, offering key advantages: broad excitation with narrow, tunable emissions for true multiplexing, high photostability, low autofluorescence background, and enhanced signal-to-noise ratio. An AI-powered image correction algorithm was developed to process images captured from standard color CCD cameras, improving contrast and signal clarity without requiring specialized multispectral equipment. Results: Using spiked breast cancer cell lines, the platform demonstrated >90% CTC recovery efficiency and successful multiplexed visualization of 3-5 protein markers per cell. AI image correction improved signal-to-noise ratio by 2-10×, enabling clear detection of low-abundance markers with simplified imaging systems. The workflow is compatible with widely available fluorescence microscopes, making it suitable for both research and translational applications. Conclusions: This integrated approach leveraging quantum dot fluorescence, magnetic nanoparticle-based CTC capture, and AI-enhanced imaging offers a powerful, scalable solution for sensitive and multiplexed CTC analysis in breast cancer. The technology’s simplicity, sensitivity, and adaptability make it well-positioned for broader use in biomarker-driven oncology research and diagnostics.

J. Tan¹, P. Garodia², R. Williams³, L. Cook³, M. Hasanova⁴, A. Kirby², B. Lamb⁵, G. Hunnings², G. Langton⁶, R. Pearson⁷, S. Adomah⁸, O. Ayodele⁹, A. Ghose⁴, A. Maniam³; ¹University of Manchester, London, United Kingdom, ²University College London, London, United Kingdom, ³Isle of Wight NHS Trust, Newport, United Kingdom, ⁴OncoFlow AI, London, United Kingdom, ⁵North East London Cancer Alliance, London, United Kingdom, ⁶The Clatterbridge Cancer Centre NHS Foundation Trust, Wirral, United Kingdom, ⁷Lord Lister Health Centre, London, United Kingdom, ⁸The Royal Marsden NHS Foundation Trust, London, United Kingdom, ⁹University of Leicester, Leicester, United Kingdom

Background Cancer Multi-Disciplinary Team Meetings (MDTMs) in the UK ensure comprehensive, patient-centred decision-making. They operate at an industrial scale across the National Health Service England (NHSE), with an estimated annual cost of approximately £154.3. Each patient may be discussed upto an average of four times, with a base cost of £428 per patient, increasing to £485 when accounting for opportunity costs. While MDTMs aim to improve outcomes, the substantial financial burden raises questions about efficiency and value. Standard NHS Consultant Programmed Activity (PA) is 4 hours, worth almost £100. The Isle of Wight (IOW) Breast Cancer MDT comprises 2 Oncologists/Equivalent, 3 Surgeons, 1 each of Radiologist, Pathologist, Radiographer, Clinical Nurse Specialist (CNS) and MDT Coordinator, with weekly MDTMs involving 1560 to 2080 cases annually. The OncoflowTM Artificial Intelligence (AI) powered Cancer MDTM CoPilot platform is a class 1 UKCA (UK Conformity Assessed) MHRA (Medicines and Healthcare products Regulatory Agency) registered medical device (RN 32434) that uses proprietary Large Language Model (LLM) based Data Extraction and Treatment Matching. A simulation trial of this intervention vs standard manual practice was conducted to inform a cost benefit and effective analysis. Methods The Simulation study involved 2 prospective Breast Cancer MDTMs with 10 cases each, matched to similar disease stage (early or metastatic) and complexity (simple/edge/complex). Phase 1 was standard practice and Phase 2 was the OncoFlow intervention. Manual preparation for this 10 case set involved 120 minutes or 1.5 hours of time from 10 IOW MDT members as aforementioned, with role-specific hourly rates applied. OncoFlow was assumed to eliminate labour input, incurring only an annual licence fee of £25,000. A cost-analysis model compared the resource use and associated costs of standard versus OncoFlow AI assisted breast cancer MDTM preparation. Results Standard manual preparation cost per 10 breast cancer cases taking a total 1.5 hours time was £378.00, equating to £37.80 per case. This included £262.50 for 7 consultants’ time (£25/hour) and £115.50 for 3 other MDT staff (£25 for radiographer, £22 for MDTM coordinator, £30 for CNS). In contrast, OncoFlow’s cost per case was calculated for two scenarios based on MDTM volume, i.e., £16.03 for 1,560 cases/year and £12.02 for 2,080 cases/year, based solely on the annual licence fee. This resulted in savings of £21.77 and £25.78 per case, respectively. Annualised savings were projected from £33,961.20 (1,560 cases) to £53,622.40 (2,080 cases) for 52 weekly breast cancer MDTMs. OncoFlow also reduced preparation time from 13.5 hours per case to negligible levels (~0.01 hours), yielding an Incremental Cost Effectiveness Ratio (ICER) of £1.91 per hour saved. This implies that for every hour of clinician time saved through automation, the cost difference was £1.91, favouring OncoFlow as a cost-saving intervention. Conclusion The OncoflowTM AI Cancer MDTM CoPilot significantly reduces the cost burden associated with manual MDTM preparation in breast cancer pathways. The system delivers cost savings of up to £25.78 per case. With an ICER of £1.91 per hour saved, OncoFlow offers a highly cost-effective solution for modernising cancer MDTM workflows, operating at scale. Broader implementation across tumour types and MDTMs could amplify these cost benefits, contributing to more sustainable, efficient NICE (National Institute for Health and Care Excellence) commissioned cancer care delivery across the NHS.

Z. Qu¹, X. Wang², S. Pei³, Y. Fang³; ¹Department of Biostatistics, Epidemiology, and Informatics, Perelman School of Medicine, University of Pennsylvania, PHILADELPHIA, PA, ²College of Information Sciences and Technology, The Pennsylvania State University, State College, PA, ³Department of Breast Surgical Oncology, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, CHINA.

Background: Large Language Models (LLMs) have shown promising accuracy in generating guideline-based cancer treatment recommendations. However, their alignment with real-world breast cancer management decisions remains poorly characterized. This study presents a novel benchmark evaluating LLM-generated recommendations against both clinical guidelines and actual treatments recorded in electronic health records (EHR), while also assessing the models’ capacity for patient-centered communication. Methods: We extracted structured and unstructured EHR data from 500 newly diagnosed non-metastatic breast cancer patients (2023-2025) at Penn Medicine, including imaging data. Each case was processed through a multi-turn, agent-based LLM prompting framework using a range of models, including general-purpose LLMs (GPT-4o, Claude 3.7, DeepSeek R1, Grok), as well as medically tuned models (MedGemma, HealthBench). Each model received basic case data and dynamically queried for additional clinical information before recommending a treatment plan. Recommendations were compared to (1) NCCN/ESMO guidelines and (2) the actual treatment received (ATR). Concordance rates, over-/under-treatment frequencies, and the “Reality Gap” (cases where recommendations matched guidelines but diverged from real-world practice) were analyzed. Propensity score-based causal inference was used to identify demographic and socioeconomic drivers of discordance. Additionally, a multidisciplinary panel rated LLM responses to common patient questions in terms of clarity, empathy, and shared decision-making support. Results: Across all models, guideline concordance was high (GPT-4o: 90.2%, Claude 3.7: 89.4%, Deepseek: 88.6%, MedGemma: 87.9%). However, concordance with real-world treatment was notably lower (GPT-4o: 67.8%, Claude 3.7: 65.1%, MedGemma: 62.3%), with LLMs generally recommending more intensive treatment than was actually administered. In 22.3% of cases, the LLM recommendation aligned with guidelines but diverged from the real-world decision, reflecting a significant “Reality Gap.” Multivariable analysis identified Medicaid insurance (OR 2.14, 95% CI 1.31-3.49), age ≥70 (OR 1.72, 95% CI 1.08-2.75), and documented treatment refusal as independent predictors of discordance. In the communication task (N=100 dialogues), GPT-4o showed relatively strong performance in accuracy (4.61/5) and decision support (4.55/5), though differences between models were modest. Conclusions: This study directly addresses a central question in AI-enabled oncology: how far are large language models (LLMs) from actual clinical decision-making? While LLMs are capable of generating medically sound and guideline-concordant recommendations, real-world factors—such as patient preferences, socioeconomic barriers, and health literacy—are often overlooked, despite playing a critical role in shaping care delivery. By jointly evaluating clinical validity and contextual fidelity, our framework quantifies the “reality gap” between LLM recommendations and actual treatment. Importantly, we do not advocate for LLMs to conform to all real-world deviations, nor should models be encouraged to offer suboptimal plans based on non-clinical attributes. Instead, our findings highlight the importance of using social context to enhance interpretation and patient-centered communication—without compromising evidence-based care. Our work provides a concrete model for evaluating and improving AI alignment with real-world clinical decisions in breast cancer care. Further analyses are underway to expand the scope and robustness of these findings.

K. Vishwanath¹, C. I. Crawford², A. G. Sorace², T. E. Yankeelov³, E. A. Lima¹; ¹Oden Institute for Computational Engineering & Sciences, The University of Texas at Austin, Austin, TX, ²Radiology, The University of Alabama at Birmingham, Birmingham, AL, ³Biomedical Engineering, The University of Texas at Austin, Austin, TX.

Introduction. Triple-negative breast cancer (TNBC) disproportionately affects obese individuals, with emerging evidence suggesting that obesity may decrease or paradoxically enhance the efficacy of certain immunotherapies. However, obesity-driven intratumoral hypoxia presents additional challenges for treatment, motivating the use of evofosfamide (a hypoxia-activated prodrug) as a therapeutic adjunct. With this dual challenge in mind, we set out to build a clinically oriented ‘virtual trial’ platform that can forecast tumor response and pinpoint dosing schedules that maximize immunotherapy + evofosfamide benefit while minimizing toxicity in obesity-linked TNBC. Methods. We have developed a mathematical framework that allows us to conduct ‘virtual trials’ on the computer, screening thousands of possible dosing protocols before any animal or patient is treated. The framework is built on a nonlinear system of ordinary differential equations that simultaneously describe intrinsic tumor growth, immune response, angiogenesis and vessel regression, and the effects of evosfamide and immunotherapy—whose interaction is modulated by the well-vascularized fraction of the tumor. Model parameters are inferred with Markov Chain Monte Carlo, an iterative random-sampling algorithm that scans the parameter space and converges on the combinations that most closely replicate the observed data. Model calibration relies on 34 days of serial data from C57BL/6J mice (N=24). Prior to data acquisition, the obese model of TNBC was established with 13 weeks of a high-fat diet (HFD) and then orthotopically implanted into the mammary fat pad with E0771 TNBC cells 12 days before baseline (day 0). For each mouse, we tracked tumor volume by caliper and the well-vascularized fraction by [¹⁸F]-fluoromisonidazole (FMISO) PET across four study arms—control, evofosfamide alone, immunotherapy on normoxic mice, and the combination on hypoxic mice. Mice were grouped seven days before baseline. By varying dosing and scheduling, we optimize treatment for three objectives: 1) reducing final tumor volume, 2) minimizing total tumor burden (tumor volume integrated over the experimental interval), and 3) limiting therapy toxicity. Results. Our mathematical framework accurately reproduces (i.e., calibrate) tumor growth dynamics under control and individual treatment conditions, achieving an average concordance correlation coefficient (CCC) of 0.97 ± 0.005. The model is also able to accurately predict the response of tumors to combination therapy, yielding a CCC of 0.93. Through differential evolution optimization—a mathematical technique that mimics natural selection by iteratively combining and selecting the best treatment parameter sets—we identify candidate treatment protocols that we hypothesize will yield a 67.5% reduction in average final tumor volume, a 68.2% reduction in tumor burden, and an 18.7% reduction in total therapy dosage compared to the standard of care. Conclusion. The findings underscore the efficacy of reducing hypoxia to enhance immunotherapy in HFD mice, which can be significantly strengthened through optimized scheduling of evofosfamide and immunotherapy combination therapies. The proposed mathematical framework demonstrates a robust ability to predict tumor response and optimize treatment strategies, highlighting the potential for translating mathematical systems to guide combination therapies and develop digital twins for personalized medicine. The optimized schedules point to the possibility of achieving significantly better tumor control with less total drug dose, potentially translating into reduced toxicity while preserving benefit. These optimized regimens should be regarded as hypothesis-generating; their safety and therapeutic benefit must be experimentally confirmed.

K. Park, C. A. Minami, L. N. Butler, J. Gittzus, K. McLean, A. Dunn, E. A. Mittendorf, T. A. King; Surgery, Dana-Faber Cancer Institute/ Brigham and Women’s Hospital, Boston, MA.

Introduction: Pre-charting for oncology patients (pts) requires manual review of documentation, often with varying formats. AI trained to summarize records for pre-charting may enhance efficiency in providing consistent information to clinicians, thereby enhancing the quality of care. We developed and clinically validated an AI agent to summarize records of pts with breast cancer. Methods: A generative AI agent utilizing GPT-4o and Retrieval-Augmented Generation (RAG) for processing radiology, pathology, and consultation reports was developed on a secure platform. Prompt engineering was employed to guide GPT-4o in creating oncology-specific summaries. For 50 consecutive breast cancer pts, history of present illness summaries (HPIs) generated by 4 breast surgery physician assistants (PAs) were compared to AI HPIs of identical records. Two breast surgeons evaluated the AI HPIs by rating the organization (clarity’) and contextual relevance (relevance’) of the information on a 5-point Likert scale (1=strongly disagree, 5=strongly agree). They clinically validated the AI HPIs by comparing to the PA HPIs using a structured evaluation framework focused on accuracy (factually correct/incorrect) and completeness of information (‘complete’ if AI HPIs aligned with the key information that oncologists would anticipate in a PA HPI) across predefined domains: patient age, imaging findings (mammogram, ultrasound, MRI, biopsy modality), pathology findings (histology, receptor status), medical history, and pending workup. Results: The overall clarity and relevance of the AI HPIs were high (average clarity 4.01, average relevance 3.87). AI accurately stated pt age in 94% of HPIs, with discrepancies attributed to birthday timing since initial workup. All pts had a mammogram, 86% had an US, and 36% MRI. Accuracy of imaging report summaries varied by modality, with MRI showing lowest accuracy (83.3%) and completeness rates (44.4%) (Table). Among the 85 biopsies, the biopsy modality was missing in 39.8%; AI HPI stated ‘core biopsy’ rather than ‘stereotactic guided’, ‘US guided’ or ‘MRI guided’ core biopsy. Pathology reporting demonstrated high accuracy among the 85 biopsies: histology 97.6%, hormone receptors and HER2 receptor 89.8%, and tumor grade 95%. Inaccurate AI responses resulted when grade or receptor stain percentage were reported in ranges. Two HPIs misrepresented past medical history. Among 12 patients requiring additional workup, 41.7% of AI HPIs appropriately identified pending studies. Conclusion: RAG-enabled GPT-4o demonstrated a high level of accuracy in breast cancer record summarization for pre-encounter preparation. Human review is necessary, but AI shows promise for clinical integration. Iterative prompt engineering, integration into clinical workflow, and evaluating the effectiveness of AI-assisted breast oncology medical record summarization is ongoing.

J. Dai¹, M. Posta², M. Marczyk³, T. Qing¹, B. Győrffy², L. Pusztai¹; ¹Yale Cancer Center, Yale School of Medicine, New Haven, CT, ²Department of Bioinformatics, Semmelweis University, Budapest, HUNGARY, ³Department of Data Science and Engineering, Silesian University of Technology, Gliwice, POLAND.

Background:How germline and somatic alterations together contribute to alterations in cancer relevant pathways in individual cancers remains poorly understood.Methods:We mapped all germline and somatic alterations of individual cancer samples from The Cancer Genome Atlas (TCGA) and the Breast Cancer Genome Guided Therapy Study (BEAUTY) onto Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomics (KEGG) pathways. To quantify pathway-level disturbances from germline and somatic origins, we calculated gene-level impact scores by integrating the functional importance of a gene in sustaining cell survival with the predicted impact of its variants. These scores were then used to compute a pathway disturbance score (CanSys score), that quantified the combined effect of gene-level alterations across all pathway members. Results:We found that almost all cancers have common and rare protein function altering germline alterations in cell cycle, telomere maintenance, and DNA repair hallmark pathways. At somatic mutation level, 12 pathways were altered in >75% of cancers corresponding to core cancer hallmarks, however different genes within the same pathway are affected in different individuals, illustrating phenotypic convergence. In addition, 404 pathways were affected in >25% but <50% of cancers which may contribute to the unique clinical course of each cancer. A freely available web tool (https://cansysplot.com) developed for this study enables visualization of GO/KEGG pathway disturbances from germline and somatic origins at the individual sample level.Conclusion:Our findings implicate that the individuals with cancer harbor protein function altering germline alterations in genes involved with cell cycle regulation, telomere maintenance, and DNA repair, whereas somatic mutations primarily affect pathways involved in cell adhesion, cell motility, metabolism, and a broad range of signal transduction processes. The unique combination of pathway alterations might explain the unique behavior of each cancer.

E. Kim¹, I. Do², S. Chae²; ¹General Surgery, Kangbuk Samsung Hospital, Seoul, KOREA, REPUBLIC OF, ²Pathology, Kangbuk Samsung Hospital, Seoul, KOREA, REPUBLIC OF.

Background Tumor-infiltrating lymphocytes (TILs) have been recognized as prognostic and predictive biomarkers in breast cancer. However, challenges such as interobserver variability and lack of standardized cut-off values have limited their clinical implementation. Recent advancements in artificial intelligence (AI)-based pathology have shown potential to enhance reproducibility and objectivity in TIL quantification. Concurrently, circulating tumor DNA (ctDNA) has emerged as a minimally invasive biomarker capable of detecting treatment response after neoadjuvant chemotherapy (NAC). Objectives This study aims to: (1) evaluate the predictive value of AI-quantified stromal TILs from H&E-stained slides; and (2) validate the clinical utility of ctDNA-based biomarkers for detection of responses in breast cancer patients after NAC. Methods TILs were assessed from pre-treatment tumor H&E slides using Lunit SCOPE IO, a validated AI-powered WSI analysis tool. Immune phenotypes were classified as inflamed, immune-excluded, or desert. PD-L1 expression was evaluated using PD-L1 IHC 22C3 pharmDx. TIL cut-off was set at 30%, Formalin-fixed paraffin-embedded samples (FFPEs) were obtained from biopsy tissue, and plasma samples were collected at baseline, pre-surgery (after NAC). Forty-seven breast cancer-related genes were analyzed using next-generation sequencing. The diagnostic performance of ctDNA was evaluated, and logistic regression analyses were conducted to assess the impact of clinical and molecular factors on ctDNA status. Parallel plasma samples were collected at predefined intervals for ctDNA extraction and sequencing using a customized Breast NGS panel targeting 157 hotspot mutations across 12 genes. Mutation validation was performed using ddPCR. Preliminary Results A total of 101 patients with early breast cancer were enrolled. The most frequently identified gene was TP53 (FFPE, 66.7%; ctDNA, 46.4%), followed by PIK3CA (FFPE, 36.2%;ctDNA, 17.4%). The triple-negative breast cancer (TNBC) subtype exhibited the strongest association with ctDNA detection (odds ratio [OR] 209.50, p = 0.005) in multivariate analysis. Patients with inflamed TIL phenotype at diagnosis and ctDNA clearance after NAC had higher pathological complete response (pCR) rate (38.5% vs. 11.1%, p = 0.238). Conclusion Our integrated approach using AI-based TIL quantification and ctDNA analysis using NGS and ddPCR provides a feasible and sensitive approach for NAC response in breast cancer patients. These findings support the clinical utility of combining digital pathology and liquid biopsy for treatment monitoring and risk stratification in conjunction with precision oncology.

S. Ranjbar¹, A. San Lucas², P. Chen¹, K. Shah², M. Suchko², B. Guerrouahen², C. Ercan³, E. Barrientos Toro³, K. Sweeney⁴, X. Pan¹, J. DeLa Cruz², B. Rodriguez¹, N. Reddy², C. Yam⁵, G. Mantha², M. Riben⁴, L. Huo⁴, Y. Yuan¹; ¹Institute for Data Science In Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX, ²Division of Pathology and Laboratory Medicine Clinical Trial Operations, The University of Texas MD Anderson Cancer Center, Houston, TX, ³Translational Molecular Pathology, The University of Texas MD Anderson Cancer Center, Houston, TX, ⁴Department of Pathology, The University of Texas MD Anderson Cancer Center, Houston, TX, ⁵Department of Breast Medical Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX.