Poster Spotlight 9: Lobular Carcinoma 

S. Oesterreich¹, S. Rivero-Hinojosa², V. N. Aushev², S. Satta², E. Kalashnikova², R. Green², A. Ramu², A. Rodriguez², M. C. Liu², J. Foldi¹, M. Balic¹, A. V. Lee¹; ¹UPMC Hillman Cancer Center, Magee-Womens Research Institute, University of Pittsburgh and UPMC, Pittsburgh, PA, ²Oncology, Natera, Inc., Austin, TX.

Introduction: Invasive lobular carcinoma (ILC) represents a distinct subtype of breast cancer with unique clinical and molecular features compared to invasive ductal carcinoma (IDC). A hallmark of ILC is the loss of E-cadherin due to CDH1 alterations, yet large-scale genomic profiling of this histology, which is predominantly hormone receptor-positive (HR+)/HER2-negative, remains limited. Here, we comprehensively characterized the somatic mutational landscape of HR+/HER2-negative ILC and explored associations with clinical outcomes using real-world data. Methods: Whole-exome sequencing (WES) was performed as part of the Signatera^TM (Natera, Inc.) workflow. WES-derived somatic variants from 1,451 patients with HR+/HER2- ILC in Natera’s large real-world database. Variants were annotated using the Ensembl Variant Effect Predictor (VEP), and only non-synonymous mutations with variant allele frequency (VAF) ≥ 5% were included in the prevalence analysis. Mutational signatures were assessed by comparing 96-trinucleotide mutation contexts with COSMIC single base substitution (SBS) signatures (v3.3). Event dates, including curative-intent surgery and distant relapse, together with treatment information, were sourced from claims data. Subtype was inferred based on treatment received. Associations between gene-level mutation and distant relapse-free survival (DRFS) were evaluated. Results: Among 1,451 HR+/HER2- ILC cases, CDH1 mutations were observed in 68%, with truncating variants accounting for 91% of these alterations. The most frequent CDH1 variants were p.Q23* (6.2%), p.I650Yfs13 (1.03%), p.Q129 (1.03%), and p.Q610* (1.03%). Comparative analysis with HR+/HER2- IDC revealed differential mutation frequencies in key genes, including PIK3CA, TBX3,TP53, and ERBB2 (47.3% vs 37.33%, 11.8% vs 4.0%,6.9% vs 25.2%, and 5.8% vs 2.2% P<0.001 for all), in addition to CDH1 (68% vs 2.6%, P<0.001). Despite its prevalence, CDH1-truncated tumors were not associated with worse clinical outcomes when compared to wild-type or non-truncating counterparts. In contrast, mutations in TP53, ERBB2, and RYR1 were significantly associated with inferior DRFS in ILC (HR(95% CI), p-val); 2.2 (1.2-3.9), P=0.006; 2.4(1.2-4.6), P=0.009; 2.4(1.2,4.9), P=0.018). Mutational signature analysis revealed enrichment of age-related, APOBEC-mediated, and mismatch repair (MMR)-associated processes. Tumors harboring the APOBEC signature exhibited the highest tumor mutational burden (TMB), with a mean TMB of 7.1 Muts/MB (vs. 1.31 Muts/MB, p-val < 0.001). Co-mutation analysis showed significant co-occurrence of CDH1 and PIK3CA mutations (P<0.01), while PIK3CA and AKT1 mutations were mutually exclusive (P<0.01). VAF analysis suggested that CDH1 mutations generally precede PIK3CA alterations, indicating CDH1 loss as an early event in ILC tumorigenesis. Conclusions: This study provides one of the largest genomic profiling studies of HR+/HER2- ILC, highlighting the central role of CDH1 loss and its likely function as an early initiating event. CDH1 mutations alone did not portend worse outcomes, mutations in TP53, ERBB2, and RYR1 were associated with significantly poorer DRFS and may serve as clinically relevant prognostic markers. The presence of APOBEC-driven mutational signatures and the co-occurrence of PIK3CA and CDH1 mutations further underscore the molecular heterogeneity within ILC. These insights enhance our understanding of ILC biology and may guide future therapeutic stratification and biomarker development in this histologic subtype.

A. Wedn¹, H. Waltermire², F. Chen³, C. Merkel⁴, D. Brown⁵, O. Shah⁶, J. Hooda⁷, R. Bhargava⁸, A. Lee⁹, S. Oesterreich⁹; ¹Molecular Pharmacology Graduate Program, University of Pittsburgh, School of Medicine, Pittsburgh, PA, ²Department of Medicine, University of Pittsburgh, School of Medicine, Pittsburgh, PA, ³Breast Oncology Center Translational Hub, Dana-Farber Cancer Institute, Boston, MA, ⁴4Cellular and Molecular Pathology Graduate Program, University of Pittsburgh, School of Medicine, Pittsburgh, PA, ⁵Institute for Precision Medicine, University of Pittsburgh, School of Medicine, Pittsburgh, PA, ⁶Pancreatic Cancer Center of Excellence, University of Nebraska Medical Center, Omaha, NE, ⁷Women’s Cancer Research Center, Magee Women’s Research Institute, UPMC Hillman Cancer Center, University of Pittsburgh, School of Medicine, Pittsburgh, PA, ⁸Department of Pathology, University of Pittsburgh, School of Medicine, Pittsburgh, PA, ⁹Department of Pharmacology and Chemical Biology, University of Pittsburgh, School of Medicine, Pittsburgh, PA.

Background: Invasive lobular carcinoma (ILC) accounts for 10-15 % of breast cancers and is linked to poorer long-term survival than invasive carcinoma of no special type (NST), yet current therapy remains identical to stage matched NST regimens. Seeking ILC-specific vulnerabilities and targets, we have recently identified the transcription factor AP-2β (TFAP2B) as one of the most hypomethylated genes in ILC relative to NST and normal breast tissue. Concordantly, TCGA, METABRIC and SCAN-B datasets show significant TFAP2B upregulation in ILC, with hypomethylation of the first-exon CpG island showing the strongest correlation with mRNA expression. Clinically, high TFAP2B expression is associated with worse survival in patients with ILC in the METABRIC dataset, further supporting its potential role in disease progression. Therefore, the current study aimed to elucidate the ILC-specific role of and dependency on TFAP2B and to assess its potential as a therapeutic target. Methodology: TFAP2B mRNA and protein expression as well as DNA methylation levels were assessed across breast-cancer cell lines, patient-derived organoids (PDOs), and tumor samples. Functional roles of TFAP2B were tested by siRNA or inducible shRNA knockdown in ER-positive and ER-negative ILC cell lines (including MDA-MB-134 and MDA-MB-453) and by inducible TFAP2B over-expression in NST models (MCF7, T47D, and MDA-MB-231), followed by assays for anchorage-dependent and independent growth, colony formation, soft agar growth, cell-cycle analyses and apoptotic cell death. To assess regulation by E-cadherin, TFAP2B expression was evaluated after CDH1 knockout or inducible CDH1 expression in NST and ILC models, respectively. Preliminary mechanistic insights were obtained through CUT-RUN, RNA-seq and reverse-phase protein arrays following TFAP2B knockdown or overexpression. Results and conclusions: TFAP2B exhibited selective upregulation with uniform nuclear localization in ILC cell lines, PDOs and tumors. Knocking down TFAP2B markedly impaired ILC proliferation in both anchorage-dependent and independent contexts, suppressed colony formation and soft agar growth, reduced S-phase entry and EdU incorporation, expanded the G0-G1 population, lowered the anti-apoptotic protein BCL2 and increased apoptotic cell death. Conversely, inducible TFAP2B over-expression in NST cells hindered proliferation and blocked the G1 to S phase transition. Altering E-cadherin levels did not affect TFAP2B levels, indicating independence from CDH1 status in the transformed cells. Preliminary CUT-RUN, and transcriptomic data point to TFAP2B regulation of cell-cycle, chromatin-remodeling and metabolic pathways. Further studies are ongoing to uncover downstream targets, pathways and interaction partners specific to TFAP2B in ILC. Collectively, the data establish TFAP2B as an ILC specific driver that promotes proliferation and survival and is associated with adverse clinical outcomes. Functional dependence of ILC but not NST cells on TFAP2B underscores its promise as a therapeutic target uniquely tailored to lobular breast cancer.

H. Maynard¹, D. Liu¹, A. S. Moghaddam², S. Ali¹, S. Solomon³, M. Balic¹, J. Foldi¹; ¹Division of Hematology/Oncology, University of Pittsburgh Medical Center Hillman Cancer Center, Pittsburgh, PA, ²Department of Medicine, University of Pittsburgh Medical Center, Harrisburg, PA, ³Medical Affairs, Guardant Health, Palo Alto, CA.

BackgroundNext generation sequencing (NGS) of circulating tumor DNA (ctDNA) provides non-invasive and real-time information on tumor biology including resistance mechanisms, which is especially important given the increasing availability of targeted treatment options for patients (pts) with metastatic breast cancer (mBC). This study aims to understand the real-world use of ctDNA-based assays by reviewing testing patterns, genomic findings, and potential treatment impact among pts treated at a large US Comprehensive Cancer Center Network, with a focus on differences in histology: invasive lobular (ILC) versus no special type (NST) breast carcinomas. MethodsFrom January 2016 – March 2025, 1,384 mBC pts had Guardant360 (Guardant Health) liquid biopsy testing obtained at the University of Pittsburgh Medical Center (UPMC) Hillman Cancer Center and its community locations. A total of 1,810 tests were sent. Guardant360 evaluated ctDNA for up to 83 genomic biomarkers, microsatellite instability (MSI) status, blood tumor mutation burden (bTMB) as well as epigenomic-based tumor fraction and other signals in the newest generation assay (Guardant360 Liquid; May 2025). Clinical data was obtained via the UPMC Hillman Cancer Center Tumor Registry. Clinical and genomic data were paired for analysis. Clinically relevant alterations were defined as those in PIK3CA, ESR1, ERBB2, BRCA1, BRCA2, and PALB2.Results A total of 915 pts were identified with histology information: 783 (85.6%) with NST and 132 (14.4%) with ILC. 838 pts had genomic alterations detected: 715 (91.3%) with NST and 123 (93.2%) with ILC. 77 pts did not have ctDNA detected. The 915 pts had a total of 1,165 Guardant360 tests sent; 226 pts had multiple serial NGS tests (1-9 per pt, median 2). A total of 5,392 unique gene mutations were detected with 586 unique genes mutated. Clinically relevant mutations were detected in 60.1% of pts (59.7% NST, 62.6% ILC). PIK3CA was the most common clinically relevant gene with mutations found in 36.8% of pts, and more common among pts with ILC (49.6%) compared to NST (34.6%) (p=0.002). ESR1 mutations were found in 27.4% of pts with similar rates in NST and ILC. BRCA1 mutations were found in 5.8% of pts (6% NST, 4.9% ILC). BRCA2 mutations were found in 9.7% of pts with similar rates between NST and ILC. In 208 pts (178 NST, 30 ILC) who had Guardant360 on Infinity platform, only one pt had an MSI-high tumor. Median bTMB was 8.54 mut/Mb among NST and 11.39 mut/Mb among ILC (p=0.25).ConclusionAmong a large cohort of pts with mBC, 91.6% had detectable alterations including clinically relevant mutations on NGS-based ctDNA analysis. Most frequent actionable alterations were in PIK3CA and ESR1, and PIK3CA mutations were more common in ILC tumors. Additional clinical data collection is ongoing. Full analysis including longitudinal changes in genomic biomarkers will be reported in the presentation.

H. Dopeso¹, L. Ferrando², L. Gusain¹, S. Farahani¹, S. Shen³, A. Mamtani⁴, G. Joergensen⁵, M. Repetto⁶, C. Schwartz¹, D. Ross¹, H. Wen¹, E. Brogi¹, H. Zhang¹, A. Lee⁷, S. Oesterreich⁷, B. Weigelt¹, F. Pareja¹; ¹Deparment of Pathology, Memorial Sloan-Kettering Cancer Center, New York, NY, ²Department of Internal Medicine, University of Genoa, Genoa, ITALY, ³Deparment of Medicine, Memorial Sloan-Kettering Cancer Center, New York, NY, ⁴Deparment of Surgery, Memorial Sloan-Kettering Cancer Center, New York, NY, ⁵Connecticut Institute for the Brain and Cognitive Sciences, University of Connecticut, Storrs, CT, ⁶Early Drug Development Service, Memorial Sloan-Kettering Cancer Center, New York, NY, ⁷Department of Pharmacology & Chemical Biology, UPMC Hillman Cancer Center, University of Pittsburgh School of Medicine, Pittsburgh, PA.

Background: RHOA is a small GTPase that regulates actin cytoskeleton dynamics, playing key roles in various solid tumors, such as in diffuse gastric cancer (DGC). DGC and invasive lobular carcinoma (ILC) of the breast share biologic and phenotypic features, including E-cadherin (CDH1) deficiency, and a characteristic single-cell invasive pattern due to impaired cell-cell adhesion. In breast cancer (BC), the prevalence, spectrum and functional impact of RHOA mutations remain to be determined. Here, we sought to characterize the RHOA mutational landscape in BC compared to gastric cancer (GC), assess its association with CDH1 mutations, and investigate its functional impact in BC cell models. Methods: We conducted a re-analysis of targeted sequencing data from 8,939 BCs (n=4,182 primary, n=4,757 metastatic) and from 1,378 GC (n=1,044 primary; n=334 metastatic). RHOA mutational frequency, spectra and co-occurrence with other genomic alterations were assessed. The functional impact of RHOA alterations was investigated in the MDA-MB-134-VI ILC cells with RHOA knockdown (RHOA-KD) and inducible CDH1 overexpression. Results: RHOA alterations were rare in BC overall (0.76%) but significantly enriched in ILC, occurring in 2.7% of cases (29/1,058), compared to 0.25% (15/5,940) in invasive ductal carcinoma of no special type (IDC-NST; P<0.001). In BC, the most common RHOA mutations were the dominant negative loss-of-function (LOF) G17 hotspot mutation (61%), followed by E40 (10%), L22 and R5 (8%, each), L57 (6%), and G2 and Y42 (3%, each) mutations. In contrast, in GC, gain-of function (GOF) Y42 mutations, associated with altered effector binding and signaling activity, were the most frequent ones (32%), followed by R5 (15%), G62 (14%), L57 (11%), and G17 (11%) mutations.In BC, RHOA mutations frequently co-occurred with CDH1 (54%) and PIK3CA (49%) mutations, significantly more often than in GC (CDH1, 21%, P<0.001; PIK3CA, 10%, P<0.001). Compared to RHOA-wild-type BCs, RHOA-mutant BCs were also significantly enriched for mutations in chromatin remodeling genes such as CBFB (23.7% vs. 4%, [MOU1] P<0.001) and KMT2C (22.7% vs. 8%, P=0.0003), as well as in transcription factors including FOXP1 (9.3% vs. 1.7%, P<0.0001) and TBX3 (16% vs. 5.8%, P=0.001). In addition, RHOA-mutant GCs had a significantly lower KRAS mutation rate (4.2% vs. 12.4%, P=0.014) compared to RHOA-wild-type GCs.We assessed the functional impact of co-occurring RHOA and CDH1 alterations using the MDA-MB-134-VI ILC BC cells with and without CDH1 overexpression. These analyses showed that RHOA knockdown increased cell migration, invasion, anoikis resistance and proliferation, with these effects being markedly amplified in CDH1-deficient cells. Conclusions: RHOA mutations, while rare in BC overall, are enriched in ILC and frequently co-occur with CDH1 and PIK3CA mutations. Our data further suggest lineage specific differences in the role of RHOA in oncogenesis between BC and GC, with RHOA GOF mutations prevailing in GC, and RHOA LOF mutations in BC. Through functional studies we demonstrate that RHOA inactivation synergizes with CDH1 deficiency to promote oncogenic phenotypes in BC. Taken together, these findings support a context dependent tumor suppressor role of RHOA in BC, particularly in ILC, through its cooperation with CDH1 loss.

Y. Sung¹, J. Sim¹, J. Gim², T. Jeon³, J. Lee⁴, J. Oh³, S. Jung⁵, K. Park⁶, A. Kim³; ¹Pathology, Korea University Anam Hospital, Seoul, KOREA, REPUBLIC OF, ²Medical Science, Soonchunhyang University, Asan, KOREA, REPUBLIC OF, ³Pathology, Korea University Guro Hospital, Seoul, KOREA, REPUBLIC OF, ⁴Pathology, Korea University College of Medicine, Seoul, KOREA, REPUBLIC OF, ⁵Division of Breast and Endocrine Surgery, Korea University Anam Hospital, Seoul, KOREA, REPUBLIC OF, ⁶Division of Oncology/Hematology, Korea University Anam Hospital, Seoul, KOREA, REPUBLIC OF.

Background: Invasive lobular carcinoma (ILC) is typically characterized by lobular morphology, CDH1 inactivation and reduced E-cadherin expression. However, discordance between morphologic and immunohistochemical (IHC) findings often complicates diagnosis. To address this issue, we performed whole transcriptome sequencing (WTS) on selected breast cancer cases exhibiting variable E-cadherin IHC patterns and morphology, including cases with and without CDH1 mutations identified by prior targeted sequencing. Methods: Whole transcriptome sequencing (WTS) was performed on 11 FFPE tissues (8 tumors and 3 normal breast tissue) using poly(A)-selected RNA libraries and 100 bp paired-end reads on the Illumina NovaSeq 6000 platform (Macrogen, Seoul, Korea). All 8 tumor cases had previously undergone targeted DNA sequencing using the TruSight Oncology 500 assay (TSO500; illumina). Based on morphologic, IHC, and molecular features, samples were stratified into three groups: Group 1 (n=4) consisted of tumors lacking lobular morphology, despite harboring CDH1 mutations identified by prior targeted sequencing. Group 2 (n=4) comprised tumor with classic lobular morphology and complete loss of E-cadherin expression, but no detectable CDH1 mutations on targeted sequencing. Group 3 (n=3) consisted of normal breast tissue samples serving as controls. Results: In Group 1, WTS confirmed the presence of CDH1 mutations previously identified by targeted sequencing (E880K in three cases, L630V in one case). Additionally, a novel transcript-derived variant, p.D756Y, was detected in one case. In Group 2, all four cases harbored CDH1 mutations that had not been detected by prior targeted sequencing. These included a stop-gained variant (Q610*), a frameshift mutation (D476fs), and several missense mutations (F69Y, S70P, I584V, I707T). Additionally, an intronic alteration near a canonical splice site (c.163+19330C>G) was identified, suggesting potential splicing disruption. No pathogenic CDH1 mutations were detected in any of the three normal breast tissue samples in Group 3. Conclusion: WTS confirmed the presence of CDH1 mutations in breast tumors lacking lobular morphology, suggesting that molecular alterations can precede or occur independently of classic morphologic features. This finding suggests a diagnostic gray zone, where tumors with CDH1 mutations may exhibit non-lobular morphology and thus risk misclassification. Furthermore, WTS detected pathogenic CDH1 variants in tumors with classic lobular morphology, despite the absence of detectable mutations by targeted sequencing. This finding underscores the added sensitivity of transcriptome-based analysis. Together, these findings reveal the molecular and diagnostic complexity of lobular carcinoma and support an expanded role for RNA-level profiling in its classification.

G. Nader-Marta¹, S. Kumar Deshmukh², K. Fanucci¹, P. Tarantino¹, S. Schnitt³, A. Lee⁴, S. Oesterreich⁵, S. Wu², J. Xiu², P. Advani⁶, M. Lustberg⁷, N. Lin¹, S. Tolaney¹, E. Mayer¹, O. Metzger¹, G. Sledge Jr.², R. Jeselsohn¹; ¹Breast Medical Oncology, Dana-Farber Cancer Institute, Boston, MA, ²Medical Affairs, Caris Life Sciences, Tempe, AZ, ³Department of Pathology, Dana-Farber Cancer Institute, Boston, MA, ⁴Medical Oncology, UPMC Hillman Cancer Center and University of Pittsburgh, Pittsburgh, PA, ⁵Breast Medical Oncology, UPMC Hillman Cancer Center and University of Pittsburgh, Pittsburgh, PA, ⁶Oncology (Medical) Mayo Clinic Comprehensive Cancer Center, Mayo Clinic, Jacksonville, FL, ⁷Yale School of Medicine, Yale University, New Haven, CT.

Background: Invasive lobular carcinoma (ILC) is the second most common breast cancer (BC) subtype, accounting for up to 15% of cases. Although several biologic differences between ILC and invasive carcinoma of no special type (NST) have been described, most data are derived from primary tumors. In contrast, the biology of metastatic ILC remains poorly defined. This study aimed to investigate the unique genomic, transcriptomic, and immune landscapes of metastatic ILC using a multi-omic approach. Methods: This real-world study included patients with confirmed NST or pure ILC who had a biopsy of a metastatic lesion or a breast lesion in the setting of known metastatic disease. Tumors underwent whole exome sequencing (NGS 592, NextSeq; WES, NovaSeq), whole transcriptome sequencing (WTS, NovaSeq) (Caris Life Sciences, Phoenix, AZ). Demographic and baseline molecular features were compared across histologic subtypes. Tumor mutational burden (TMB; high ≥10 mut/MB) was assessed by NGS. Immune cell populations were estimated by RNAseq deconvolution (quanTIseq). Pathway enrichment analysis was determined by GSEA. Overall survival (OS) was defined from the date of biopsy to death from any cause using Kaplan-Meier estimates. Statistical tests included chi-square and Mann-Whitney U with multiple testing correction (q<0.05). Results: Among 2,651 metastatic BC samples (ILC n = 608; NST n = 2,043), patients with ILC were older (median 67 vs 63 yrs) and more likely white (81.5% ILC vs 73.1% NST), all p<0.05. By IHC, ILC were more frequently HR+/HER2- (83.6% vs 55.9%), less commonly triple-negative (11.9% vs 29.8%) or HER2+ (4.6% vs 14.3%), all p<0.05. PAM50 subtypes differed between ILC and NST: luminal A (Lum A) (31.1% vs 13.7%) and luminal B (Lum B) (44.9% vs 43.8%) intrinsic subtypes were enriched in ILC, while NST was more frequently basal (1.2% vs 22.4%) and HER2-enriched (19.8% vs 22.2%), all p<0.05. Within luminal BC, Lum B was more common than Lum A in both histologies, but Lum A was proportionally more frequent in ILC (ILC: 40.3% Lum A vs 59.7% Lum B; NST: 23.4% Lum A vs 76.6% Lum B), p<0.0001. Comparative multi-omic analysis focused on the 922 HR+/HER2- tumors (ILC n = 275; NST n = 647). Compared to NST, ILC had higher frequency of CDH1 (87.3% vs 6.1%), PIK3CA (52.7% vs 38.5%), FOXA1 (9.0% vs 3.4%), ERBB2 (7.1% vs 2.0%), ARID1A, and NF1 (both q < 0.05) mutations and TMB high (17.7% vs 9.9%), but lower frequency of TP53 (15.7% vs 33.3%), ESR1 (10.9% vs 18.1%) and GATA3 (2.3% vs 14.9%) mutations along with FGFR1 amplifications (15.3% vs 6.7%), all q<0.05. ILC had higher expression of androgen receptor (91.2% vs 83.5%, q<0.05) and higher MAPK activation score (-0.81 vs -1.19, q<0.05), consistent with a luminal, hormone-driven phenotype. NST tumors demonstrated enrichment of pathways involved in the cell cycle and metabolic activity, including E2F targets, G2M checkpoint, MYC targets, mTORC1 signaling, glycolysis, and unfolded protein response (NES: 1.5-2.7, FDR<0.25). ILC tumors had higher infiltration of B cells, dendritic cells, neutrophils, NK cells and M2 macrophages, but lower M1 macrophages, all q<0.05. ILC had differential expression of immune checkpoint genes (upregulation: CD274 (PD-L1), PDCD1 (PD-1), TNFSF14, CEACAM1, CD160, fold change (FC): 1.2-1.5; down: HAVCR2 (TIM-3): FC: 1.2), all q<0.05. Median OS (mOS) was similar between luminal ILC and NST (33.8 vs 35.4 months (mo); HR 1.0, 95% CI 0.91-1.3). No mOS difference was seen between Lum A ILC and NST (47.4 vs 50.5 mo; HR 1.0, 0.84-1.43), while Lum B ILC had worse mOS than NST (27.4 vs 35.3 mo; HR 1.3, 1.1-1.6). Conclusions: This study represents the largest cohort to date with multi-omic characterization of metastatic ILC, revealing distinct genomic, transcriptomic, and immune features compared to NST. These findings highlight the unique biology of ILC and may inform future subtype-specific therapeutic strategies.

A. Zatsepina¹, F. Mohsin¹, P. Yang², L. Zhao², M. Silver³, A. Corben⁴, Y. Xu²; ¹Internal Medicine, Maimonides Medical Center, Brooklyn, NY, ²Hematology and Medical Oncology, Maimonides Medical Center, Brooklyn, NY, ³Biostatistics, Maimonides Medical Center, Brooklyn, NY, ⁴Pathology, Maimonides Medical Center, Brooklyn, NY.

Background: Lobular carcinoma in situ (LCIS) is a noninvasive lesion associated with a 7- to 11-fold increased risk of developing invasive breast cancer. Its incidence in biopsy specimens ranges from 0.5% to 3.6%, typically found incidentally during evaluation for another abnormality. LCIS has been reported in 5% of patients diagnosed with stage I or II invasive breast cancer and is present in approximately 50% of invasive lobular carcinoma (ILC) cases and 2% of invasive ductal carcinoma (IDC) cases. However, no recent studies have evaluated the prevalence of LCIS among patients with invasive breast cancer or its association with cancer subtypes, demographic factors, and clinicopathological features. Methods: This retrospective study included all newly diagnosed, non-metastatic breast cancer patients who underwent surgical resection at Maimonides Medical Center in 2023. The majority of pathology diagnoses were rendered by a single breast pathologist. Collected data included patient demographics, hormone receptor status, presence of multifocal disease, LCIS, and other high-risk pathological features. Fisher’s exact test was used to assess differences between LCIS-positive and LCIS-negative groups. Results: A total of 152 patients were included (150 female, 2 male). The ethnic distribution was 38.2% Caucasian, 26.3% Asian, 31.6% African American, and 3.3% Hispanic. There were 140 cases of IDC and 13 cases of ILC. The overall prevalence of LCIS in surgical specimens was 26.97% (41/152). Among LCIS-positive cases, 73% were associated with IDC and 29% with ILC. LCIS was present in 12 of 13 ILC cases (92.3%) and in 30 of 140 IDC cases (21.4%). Atypical ductal hyperplasia (ADH) or atypical lobular hyperplasia (ALH) was identified in 46.3% of LCIS-positive cases and 26% of LCIS-negative cases. The LCIS detection rate in biopsy specimens was 30.3% (46/152). LCIS was more frequently associated with estrogen receptor (ER)-positive disease (92.7%) compared to the LCIS-negative group (74%). However, 3 patients (7.3%) in the LCIS-positive group had triple-negative disease, and 8 patients (19.5%) had ER-positive, human epidermal growth factor receptor 2 (HER2)-positive tumors. Multifocal disease was also more common among LCIS-positive cases (27% vs. 9%). Conclusions: The prevalence of LCIS in invasive breast cancer surgical specimens was 26.97%, and the detection rate in biopsy specimens was 30.3%. LCIS was found in both IDC (21.4%) and ILC (92.3%) cases and was predominantly associated with ER-positive tumors (92.7%), although a minority of triple-negative cases were observed. These findings suggest a higher LCIS prevalence in both biopsy and surgical specimens than previously reported, raising questions about the potential biological significance of LCIS in breast cancer development. The presence of LCIS in ER-negative disease prompts consideration of whether endocrine-based chemoprevention should be offered to reduce the risk of future ER-positive malignancies.

T. Pai¹, C. Bullock², N. Wiest³, E. Parent², P. Advani³; ¹Department of Hematology and Medical Oncology, Emory University, Atlanta, GA, ²Department of Radiology, Mayo Clinic, Jacksonville, FL, ³Division of Hematology/Oncology, Mayo Clinic, Jacksonville, FL.

Background: Assessing invasive lobular carcinoma (ILC) with standard-of-care (SOC) imaging, such as ¹⁸F-fluorodeoxyglucose (FDG) PET-computed tomography (CT), is challenging. In May 2025, National Comprehensive Cancer Center Network guidelines added FES PET for recurrent/metastatic ILC staging. However, perceptions of the value added by FES PET to SOC imaging may vary, as only small ILC cohorts have been published. We evaluated FES PET’s clinical utility in the largest-studied cohort, to our knowledge, of ILC pts who have undergone FES PET. Methods: We identified all ILC pts who had an FES PET [PET-magnetic resonance (MR) or PET-CT)] between April 1, 2021, and May 1, 2025, at Mayo Clinic. Clinical and radiological data, including indications and management (mgmt) changes related to pts’ first FES PET, were obtained from clinical records. FES and SOC FDG PET were compared. Data were summarized with descriptive statistics.     Results:102 female pts were included: 88% White and 4.9% Black. Median age at initial breast cancer (BC) diagnosis was 57 (range, 36-79). At initial ILC diagnosis, most tumors were strongly estrogen receptor (ER) positive. Highest tumor grade was 1 in 25% of pts and 2 in 57%, with grade not documented in 19%; none had documented grade 3 ILC. Initial clinical prognostic stage was as follows: 26% stage 1, 21% stage 2, 10% stage 3, and 16% stage 4. At the time of FES PET, 37% had biopsy-confirmed metastatic ILC at the time of FES PET. The most common indications for FES PET were (re)staging ILC (49%), detecting suspected metastatic/recurrent BC (30%), and evaluating rising tumor markers like CA 15-3 or CA 27.29 (15%). 16% of pts were upstaged due to the FES PET result: 14 pts to stage 4 and 2 pts to stage 3.FES PET led to a mgmt change in 38 pts (37%), for whom the most common FES PET indications were (re)staging ILC (21 pts), detecting suspected recurrent/metastatic BC (11 pts), and detecting ER expression when biopsy was not feasible/nondiagnostic (7 pts). 19 scans (50%) led to a systemic therapy (tx) decision, 10 (26%) to a biopsy, 4 (11%) to additional imaging like MRI, 3 (8%) to locoregional tx, and 2 (5%) to a non-oncologic specialty referral to evaluate a suspected new metastasis. 11 pts had stage 4 BC pre-FES PET; their FES PET was ordered primarily for restaging (5 pts) or detecting ER expression (5 pts). Notably, of the 19 pts with a systemic tx-related change, 4 began a triple-negative BC regimen due to low FES uptake on the scan. Of the 15 pts who had FES PET to evaluate a rising tumor marker, a direct change in systemic tx occurred in 2 pts. 63/102 pts (62%) had FDG PET-CT prior to FES PET. Of these, 57% had discordant FDG and FES PET results. 39% had a higher number of avid metastatic lesions on FES PET as compared to FDG PET, primarily in the bone (8 pts) and lymph nodes (4 pts). 25% had bone, nodal, or soft tissue metastatic lesions that were FDG but not FES avid. 13 pts had FES PET-MRI; 9 of these had a prior FDG PET-CT. 2 PET-MRI pts had FES-avid metastatic lesions in unusual sites (brain, peritoneum) that were not FDG avid. Among the 38 pts with a mgmt change due to their FES PET result, 63% had prior FDG PET; 32% of the 38 pts had FES-avid metastatic lesions that were not seen on FDG PET, and 8% had FDG-avid but not FES-avid lesions. Conclusions: FES PET was useful for evaluating not only stage 4, but also early-stage ILC, with 1/3 of pts with an FES PET-driven mgmt change having early-stage BC. FES PET was most helpful for ILC (re)staging, detection of suspected recurrent/metastatic BC, and detection of ER expression when biopsy was not feasible/nondiagnostic. FES PET added clinically meaningful information beyond SOC imaging, with detection of FES-avid but not FDG-avid metastatic lesions in 1/3 of pts with an FES PET-driven mgmt change. Notably, FES PET-MRI demonstrated potential superiority over FDG PET in identifying atypical ILC metastases.

J. Avila¹, X. Xue², A. Gyamfi³, J. D. Anampa³; ¹Department of Hematology and Medical Oncology, Montefiore Medical Center, Bronx, NY, ²Department of Epidemiology, Albert Einstein College of Medicine, Bronx, NY, ³Department of Oncology, Montefiore Einstein Cancer Center, Bronx, NY.

Background:Breast cancer is the most common cancer and leading cause of cancer-related death in women. Invasive lobular carcinoma (ILC) is the second most common histological type of breast cancer after Invasive breast carcinoma of no special type (invasive ductal carcinoma [IDC]). ILC accounts for 10-15% of all breast cancer cases. Prior studies have suggested different patterns of short-term (5 years) survival between ILC and IDC. However, comprehensive data on both outcomes is limited. The objective of our study was to identify the factors associated with early and late mortality in patients with ILC. Methods:We conducted a retrospective analysis using the National Cancer Database (NCDB) from 2006-2020. Inclusion criteria were female gender, age ≥ 18 years, lobular histology, breast cancer stage I-III and removal of breast cancer tumor and axillary lymph nodes. Our primary endpoint, overall survival (OS), was defined as time in months from the date of diagnosis until death. The short-term OS analysis evaluated survival from diagnosis until 5 years from diagnosis (censored data at 5 years). The long-term OS analysis evaluated those patients who were alive after 5 years from diagnosis. We used the Pearson X² test to evaluate the relationship between categorical variables and the t test for continuous variables. Cox hazard regression models were used to identify factors associated with short-term ( 5 years). Results:The analysis of short-term OS included 115,159 patients, most of them were Non-Hispanic White (81%) with a median age of 63 years. Most tumors were T1 (57%), N0 (71%), Hormone Receptor positive (ER+ 100%, PR+ 87%), HER2 negative (81%). 52,503 (46%) patients underwent mastectomy. The 5-year OS rate was 93%. Old age (Hazard Ratio [HR] 2.16; p <0.001), large tumor size (HR 1.97 p <0.001), axillary node involvement (HR 5.15; p <0.001), high grade (HR 1.59; p <0.001), and lymphovascular invasion (LVI) (HR 1.19; p <0.001) were associated with worse short-term OS; whereas mastectomy (HR 0.84, p <0.001), chemotherapy (HR 0.77; p <0.001), radiation (HR 0.6; p <0.001), PR positivity (HR 0.74; p <0.001) and endocrine therapy (ET) (HR 0.5; p <0.001) were associated with improved short-term OS. Among patients who had OncotypeDx testing (N= 3,800), compared to low-OncotypeDx recurrence score, those with high OncotypeDx recurrence score had worse short-term OS (HR 1.68; p <0.001), whereas the short-term OS was similar for those with intermediate OncotypeDx recurrence score (HR 0.67; p = 0.3). The analysis of long-term OS included 67,927 patients with a median age of 62 years. The 10-year OS rate for this cohort was 86%. Old age (HR 1.89; p <0.001), large tumor size (1.58; p <0.001), axillary nodal involvement (HR 4.38; p <0.001), high tumor grade (HR 1.37; p <0.001), and LVI (HR 1.12; p = 0.01) were associated with worse long-term OS; whereas mastectomy (HR 0.94; p = 0.04), chemotherapy (HR 0.81; p <0.001), radiation (HR 0.84; p <0.001), PR positivity and ET (HR 0.78; p <0.001) were associated with improved long-term OS. Among patients who had OncotypeDx testing (N= 18,339), compared to low OncotypeDx recurrence score, those with high OncotypeDx recurrence score had worse long-term OS (HR 1.8; p <0.001); whereas the long-term OS was similar for those with intermediate OncotypeDx recurrence score (HR 1.17; p = 0.08) Conclusion: Pathological characteristics such as size and nodal status are prognostic for short-term and long-term survival outcomes in ILC. Adjuvant chemotherapy and endocrine therapy improve short-term and long-term OS in ILC. High OncotypeDx recurrence score can identify patients with worse short-term and long-term OS. Our results highlight the prognostic value of pathological characteristics, genomic profiling, and adjuvant therapies in ILC beyond 5 years.

J. Van Cauwenberge¹, H. Nguyen¹, R. Prevos², A. Bernaerts³, S. Hendrickx⁴, C. Van de Velde⁴, K. Van Baelen¹, P. Neven⁵, H. Wildiers⁶, G. Floris⁷, A. Smeets⁸, I. Nevelsteen⁸, C. Van Ongeval², C. Desmedt¹; ¹Laboratory of Translational Breast Cancer Research, Department of Oncology, Catholic University of Leuven, Leuven, BELGIUM, ²Radiology, University Hospitals Leuven, Leuven, BELGIUM, ³Radiology, ZAS Hospitals Antwerp, Wilrijk, BELGIUM, ⁴Radiology, University of Ghent, Ghent, BELGIUM, ⁵Department of Gynecological Oncology, University Hospitals Leuven, Leuven, BELGIUM, ⁶Department of General Medical Oncology, University Hospitals Leuven, Leuven, BELGIUM, ⁷Pathology, University Hospitals Leuven, Leuven, BELGIUM, ⁸Department of Surgical Oncology, University Hospitals Leuven, Leuven, BELGIUM.

Introduction: Breast density (BD), a dynamic radiologist-assessed parameter, is impacted by factors such as – but not limited to – age, parity, hormone therapy, nutrition, family history and obesity. Higher BD is a known breast cancer risk factor, not only due to reduced screening sensitivity but also because of biological processess like increased collagen deposition, potentially promoting tumor growth. While some studies associate high BD with larger tumor size and greater nodal involvement, results remain inconclusive, and the prognostic impact is unclear. Invasive lobular carcinoma (ILC), characterized by distinct clinicopathological and biological features, is understudied in this context.Methods: We conducted a retrospective analysis of patients diagnosed with estrogen receptor (ER) positive, HER2 negative early pure ILC at University Hospitals Leuven (2000-2020). BD was scored using the BIRADS 4^th and 5^th edition (categories A-D). When feasible, mammograms were independently re-evaluated by radiologists, with the final score based on the majority. If re-evaluation was not possible, the clinical record score was used. Cases lacking a majority consensus were excluded. Survival data were collected until 05/2022. Multinomial logistic regression and Cox regression were employed to evaluate the associations of BD scoring (B/C vs. A, and D vs. A) with clinic-pathological characteristics and survival. The characteristics include age, parity, BMI, tumor grade, tumor size, focality, progesterone receptor, nodal status and treatments (radio- and chemotherapy).Results: We included 968 patients, BD scores were for 628 patients available from clinical records. Two radiologists re-evaluated 325 cases, showing a moderate concordance among themselves and the original score (Fleiss’ kappa = 0.488). For 101 cases with missing data, three radiologists re-assessed BD, showing moderate concordance (Fleiss’ kappa = 0.426). Final BD scores were available for 727 patients (Type A: 6.5%, B: 38.8%, C: 39.6%, D: 15.1%). Multivariable analyses showed significant associations between higher BD and both lower age and BMI. A non-significant trend suggested associations with lower tumor grade and more multifocality. Higher BD was independently associated with both better disease-free survival (DFS) and overall survival (OS) , but not with distant recurrence free survival (DRFS) (median follow-up of 9.94 years). Conclusion:In ER+ HER2- ILC, higher breast density is associated with younger age and lower BMI, and our data suggests a potential link with lower tumor grade and increased multifocality. Higher density was independently associated with better OS and DFS. These results highlight the need for further research into the biological mechanisms underlying these associations.