Poster Spotlight 5: Liquid Biomarkers in Breast Cancer—Driving Precision Medicine

A. J. Medford¹, C. Scalise², S. Dhulekar¹, Z. Smolar¹, A. Niemierko¹, E. Kalashnikova², A. Rodriguez², L. M. Spring¹, J. Kim¹, A. Comander¹, A. Rosenstock¹, J. M. Peppercorn¹, J. A. Shin¹, L. Schnipper³, H. Benjamin⁴, S. J. Isakoff¹, L. W. Ellisen¹, D. Juric¹, B. Moy¹, M. C. Liu², A. Aleshin², A. Bardia⁵; ¹Oncology, Mass General Cancer Center/Harvard Medical School, Boston, MA, ²Oncology, Natera, San Carlos, CA, ³Oncology, Beth Israel Deaconess Medical Center, Boston, MA, ⁴Oncology, Mass General Cancer Center – Danvers, Danvers, MA, ⁵Oncology, UCLA Jonsson Comprehensive Cancer Center, Los Angeles, CA.

Background: Detectable ctDNA after definitive therapy for early breast cancer (eBC), termed molecular residual disease, is associated with a higher risk of cancer recurrence. The LEADER Phase II study (NCT03285412) aimed to evaluate the efficacy of adding the CDK4/6 inhibitor, ribociclib, to adjuvant endocrine therapy (aET) for patients (pts) with ER+/HER2- eBC with ctDNA positivity after definitive surgery. Methods: Pts with biopsy-proven ER+(≥10%)/HER2-, pathological stage IIA or higher eBC were enrolled in a ctDNA surveillance protocol as pre-screening for the LEADER study. All pts completed surgical resection prior to ctDNA testing and were receiving aET, which was performed using a tumor-informed assay (Signatera^TM, Natera, Inc). Pts with detectable ctDNA during surveillance were then enrolled in LEADER and treated with 400mg ribociclib daily (continuous dosing) plus an aromatase inhibitor for 12 cycles. Imaging studies at the time of ctDNA detection were not required and left to the discretion of the treating physician. On-treatment plasma samples were collected serially (3, 6, and 12 months) during follow-up visits. The primary endpoint was the proportion of pts who had ctDNA clearance after completion of 12 cycles of adjuvant ribociclib in combination with endocrine therapy, and secondary endpoints included the proportion of pts who had a molecular response after 3 cycles on therapy, time to ctDNA clearance and recurrence-free survival. Results: A total of 162 pts registered and 152 had complete sample sets for assay design between 9/3/20 and 2/12/24 (prior to FDA approval of ribociclib for HR+/HER2- eBC). ctDNA testing was performed on 140 pts over a median of 11.4 (range: 0-35) months with a median of 2 (range: 1-4) tests per pt. In this cohort, 89% (125/140) had persistently ctDNA-negative results and experienced favorable distant recurrence-free survival (DRFS), measured from first screening ctDNA test, with a median clinical follow-up of 31.1 (range: 0-52.9) months, yielding a negative predictive value of 98.1% (103/105), assessed based on each individual’s status at last follow-up. A cumulative ctDNA detection rate of 11% (15/140) was observed during surveillance, with 87% (13/15) testing ctDNA-positive on their first surveillance time point. Of the 15 positive cases, 33% (5/15) were ineligible due to concomitant radiographic evidence of recurrence; 80% (4/5) of these were ctDNA-positive on their first test. The median ctDNA level of the 10 pts starting ribociclib was 7.7 (range: 0.12-163.8) mean tumor molecules (MTM)/mL. One pt was excluded due to early treatment discontinuation. Of the remaining 9 pts, 66% (6/9) and 33% (3/9) showed a ctDNA decrease (>25% baseline) and complete clearance after 3 months on ribociclib, respectively. Pts with favorable ctDNA dynamics (N=6; decrease/cleared) had lower pre-treatment ctDNA levels (median 4 MTM/mL) compared to pts with unfavorable dynamics (N=3; increase ctDNA) (median: 73.8 MTM/mL). DRFS measured from time of enrollment was 16.5 months in pts with favorable dynamics and 5.3 months for unfavorable dynamics; particularly for pts who achieved ctDNA clearance (N=3), the time to recurrence was longer (18.6 months) vs. those without clearance (7.2 months). Conclusions: ctDNA positivity during surveillance accurately identified pts at high risk of recurrence, while persistent ctDNA negativity was associated with numerically lower recurrence. Pts starting ribociclib with lower molecular levels of disease were more likely to exhibit ctDNA-decrease/clearance and delayed metastasis. These findings support the utility of ctDNA based molecular surveillance to evaluate therapeutic interception and monitoring treatment response in ER+/HER2− eBC.

A. DeMichele¹, L. Bayne¹, E. Walinsky¹, L. Berry², S. DeLuca¹, C. G. Smith³, A. Chevalier³, B. Ambasager³, P. Sanchez¹, J. Heaven¹, E. Chislock⁴, T. Pan⁴, J. Graves⁴, G. Belka⁴, J. Wang⁴, E. Taranto¹, A. Nayak⁵, M. Feldman⁶, A. Clark¹, L. Chodosh⁴; ¹Hematology/Oncology, University of Pennsylvania, Philadelphia, PA, ²Statistics, Berry Consultants, Austin, TX, ³NeoGenomics, NeoGenomics Laboratories, Inc., Fort Myers, FL, ⁴Cancer Biology, University of Pennsylvania, Philadelphia, PA, ⁵Pathology & Laboratory Medicine, University of Pennsylvania, Philadelphia, PA, ⁶Pathology & Laboratory Medicine, IU School of Medicine, Indianapolis, IN.

Background: Breast cancer (BC) recurrence may follow a dormant phase in which quiescent cells reside in niches such as bone marrow (BM). Indeed, dormant BM disseminated tumor cells (DTCs) are independently associated with BC recurrence/death. The CLEVER trial demonstrated feasibility, safety and a reduction of DTCs with inhibition of autophagy (hydroxychloroquine (HCQ), and/or mTOR signaling (everolimus [EVE]) in DTC-positive BC survivors (DeMichele, Nature Medicine, 2025). We now present data on concurrent circulating tumor DNA (ctDNA) testing and additional follow up. Methods: The CLEVER trial (NCT03032406) is a randomized, phase II trial in high-risk patients (pts) diagnosed within 5 years with either triple negative disease (dx), HER2+ or ER+ dx with either positive lymph nodes (LN) or residual dx (RD) after neaodjuvant chemotherapy, or, for ER+ dx, Oncotype RS>25. Subjects completed all treatment except endocrine therapy. DTCs were detected in BM aspirate (BMA) by immunohistochemistry (IHC) with pan-CK antibody AE1/AE3. DTC-positive pts were randomized to six 28-day cycles (C) of HCQ (600 mg BID), EVE (10 mg daily) or both (+/- 3-month (m0) observation period. DTC assessment and blood collection occurred prior to C1 (baseline), after C3, C6, C12 (if applicable) and 6-mo after end of treatment (EOT); additional blood samples were collected after 12, 18, 24, 30 and 36 months of follow up. Plasma ctDNA was assessed using the RaDaR assay (NeoGenomics), with whole-exome sequencing performed on primary tumor tissue to identify somatic mutations subsequently used in ctDNA assessment. Results: 51 DTC-positive pts were randomized to HCQ (n=15), EVE (n=15) or HCQ+EVE (n=21). At a median follow-up of 77 months, 2/51 (6%) have had a recurrence (1 distant only, 1 locoregional + distant). Overall 5-year RFS is 96%; 89.5% for ER+/HER2-, 100% for TNBC and 100% HER2+/any ER; RFS HR for ER+/HER2- vs. TNBC 3.64 (0.33, 40.5). ctDNA testing could be performed on plasma from 32/51 pts (63%) during treatment and follow-up; all were ctDNA-negative at baseline. Of the 29 who had at least one post-treatment DTC assessment, 2 were DTC-positive post-treatment and neither has become ctDNA-positive. Amongst all timepoints at which both DTCs and ctDNA were assessed, the concordance between DTCs and ctDNA was 85%. Of the 2 pts with distant recurrence, 1 pt with ER+/HER2- dx on EVE withdrew after C2 for toxicity, became ctDNA-positive 9.6 mo after first DTC-positive result and 4.6 mo before recurrence; 1 pt with ER+/HER2- dx on HCQ was DTC-positive post-C6, was DTC-negative after additional 6C EVE + HCQ, became ctDNA-positive 24 mo after first DTC-positive test, developed locoregional recurrence 3 mo later, became ctDNA-negative following treatment, then ctDNA-positive 31 mo later and developed distant recurrence after another 16 mo. Of note, an additional pt with TNBC prior to study entry subsequently developed a new ER+/PR+/HER2- breast cancer and concurrent liver metastases 36 mo after last negative ctDNA test. Conclusions: With longer follow-up, recurrence after treatment targeting DTCs remains low and ctDNA monitoring post-treatment identified both patients with recurrence in advance of overt metastasis. Longitudinal BMA, ctDNA and follow-up are ongoing; confirmatory trials, NCT04523857 and NCT04841148 are enrolling.

A. Horrmann¹, Y. Travadi¹, G. Schaap¹, K. Mallery¹, K. J. Kamalanathan¹, N. Bristow¹, C. Galeano-Garces¹, S. Y. Bae¹, A. Hesch¹, C. Rungkittikhun¹, B. R. Konety¹, J. M. Drake²; ¹Cancer Detection, Astrin Biosciences, Saint Paul, MN, ²Cancer Detection, Astrin Biosciences, St. Paul, MN.

Introduction: Women with dense breast tissue face a two-fold higher risk of breast cancer, yet mammography sensitivity is as low as 30% in this group, underscoring the need for more sensitive and accessible screening methods. Liquid-based biopsies for early breast cancer detection are emerging but currently remain out of reach for clinical use. Recent data on nucleotide assessment from plasma in breast cancer have been mixed with 87% sensitivity for late-stage disease (stage 3-4) but only 20% sensitivity for early-stage disease (stage 1-2). Proteomics is an exciting area for early cancer screening where improvements in sample preparation and equipment have enabled major advances in early cancer detection. This is particularly true for a deep proteome assay as the detection of proteins 8-9 orders of magnitude lower in abundance than common plasma proteins can be identified. Objective: This study evaluates the presence of breast cancer using label-free shotgun mass spectrometry-based proteomics on less than 1 ml of plasma in >1,100 women classified as either healthy or as newly diagnosed breast cancer patients with a focus on early-stage disease (stages 0-2). Methods: 363 banked blood samples from multiple sources were used to establish a training set consisting of 156 stage 0-4 treatment-naive breast cancer patients, 46 benign breast conditions, and 161 healthy controls. Samples were age and demographic matched between healthy, benign, and breast cancer cohorts. Samples were processed in a blinded manner using an automated and novel deep proteomic platform developed by Astrin Biosciences. This platform combines label-free quantitative mass spectrometry and a proprietary breast cancer-specific spectral library to extract features used to train a classifier to identify breast cancer patients. Blinded validation of this AI classifier was evaluated on the remaining 788 patient samples (healthy, n=374; benign, n=81; breast cancer, n=333) to confirm performance characteristics. Results: Preliminary analyses from our training set of deep proteomic profiling identified over 8,600 proteins in total and up to 7,500 proteins per patient (median 6,926 proteins) with a dynamic range spanning over 8 orders of magnitude. Differentially expressed proteins in early-stage breast cancer patients from healthy patient controls were also observed. Our AI classifier, using a leave-one-out cross-validation (LOOCV) approach on the interim cohort, separated healthy controls from breast cancer patients with an ROC of >95% and a specificity and sensitivity greater than 85% across all stages. Conclusion: We have developed a highly sensitive blood-based assay that utilizes deep proteomic profiling to identify distinctive cancer specific signatures in women who are undergoing screening for breast cancer. This work enables us to develop a protein-based classifier from plasma for early detection of breast cancer. Impact on screening strategies: This assay advances clinical diagnostics by identifying proteomic markers for early detection of breast cancer. This innovative approach enhances screening strategies for women with dense breasts who are at average or high-risk based on family history, specific genetic mutations such as BRCA1/2, race, or other factors. Acknowledgements: We thank all of the members at Astrin Biosciences who have contributed endless hours dedicated to ending cancer. We are also grateful to all the donors who generously contributed blood samples to this study.

C. Cheng¹, S. Sanchez Seyeddi², E. Brierley-Green², D. Vaghashia², M. Berkowitz², R. Callahan², B. DiCarlo², A. Master², J. Penn², M. S. Sedrak², A. Medford³, C. Gianni⁴, B. Pastò⁴, C. Reduzzi⁴, S. Tapiavala⁵, E. Podany⁵, A. A. Davis⁵, M. Cristofanilli⁴, A. Bardia², M. Lipsyc-Sharf²; ¹Department of Medicine, University of California, Los Angeles, CA, ²Department of Medicine, University of California, Los Angeles, Los Angeles, CA, ³Mass General Cancer Center, Massachusetts General Hospital, Boston, MA, ⁴Breast Center, Weill Cornell Medicine, New York, NY, ⁵Department of Medicine, Washington University in St. Louis, St. Louis, MO.

Background:Detection of adjuvant ctDNA in EBC is strongly associated with recurrence. Prior data (Lipsyc-Sharf et al, ASCO 2024, 2025) identified patterns of ctDNA testing for minimal residual disease (MRD) surveillance in the real-world setting. However, the impact of MRD testing on patient QoL is unknown. Here, we present initial results from I-SURV, a prospective survey study evaluating the effect of MRD monitoring using adjuvant ctDNA testing on QoL in EBC patients. Methods: EBC Patients undergoing adjuvant ctDNA testing in a single academic institution provided informed consent prior to enrollment in this IRB-approved study. All tests were ordered as part of clinical care based on patient/oncologist shared decision-making. Participants completed electronic surveys to assess 1) validated measures of anxiety/depression (HADS), fear of recurrence (FCR-4), and QoL (FACT-Bv4) and 2) their opinion on ctDNA utility. Patients undergoing their first ctDNA test at time of enrollment (cohort 1) were asked to complete paired surveys before and after receiving test results. Patients that received ctDNA testing results before enrollment (cohort 2) completed only post-result surveys. Survey scores were summarized. Results: Between March-June 2025, 71 patients were enrolled. 58/71 (81.7%) patients completed at least 1 survey and comprised the analytic cohort (43 in cohort 1; 15 in cohort 2) (see Table 1 for median baseline survey scores). Overall, 30/58 (51.7%) patients had stage II-III EBC, 47/58 (81.0%) had grade 2-3 EBC, and 30/58 (51.7%) patients received prior chemotherapy. 13/43 (30.2%) of patients in cohort 1 completed paired surveys; the median time between surveys was 42 days (range 29-79). 12/13 (92.3%) patients with paired surveys tested ctDNA-. In these 12 patients, there was improvement in scores evaluating anxiety via HADS (median -1[ -4 to -0.5]), depression via HADS (median -0.5 [-1 to +2]), fear of recurrence via FCR-4 (median -2 [-4.3 to -0.8]) and QoL via FACT-B (median +4 [0 to +12]). For the 1 patient tested ctDNA+, fear of recurrence via FCR-4 worsened (increased from low to high-fear category) as did QoL (FACT-B score decreased by 25 points). Among patients in cohort 2, all had prior ctDNA- result. Altogether, 54/58 (93.1%) participants valued the information provided by ctDNA testing and 52/58 (89.7%) wished to continue ctDNA surveillance. Updated data will be presented at the conference if selected. Conclusion: We studied the impact of adjuvant ctDNA testing on QoL in the real-world setting in EBC. Preliminary results suggest that most patients in this study value ctDNA surveillance and would like to continue. Patients who received a ctDNA- result had reduced anxiety and fear of recurrence. These data will guide the design of multicenter studies powered to investigate the impact of ctDNA surveillance on QoL among EBC patients.

C. R. Goodman¹, S. F. Shaitelman¹, S. Ramezani², K. Hoffman¹, M. P. Mitchell¹, N. Comeaux¹, A. Zia¹, C. Barcenas³, B. Lim¹, B. D. Smith¹, A. Lucci⁴, W. A. Woodward¹; ¹Department of Breast Radiation Oncology, UT MD Anderson Cancer Center, Houston, TX, ²Department of Radiation Oncology, UT MD Anderson Cancer Center, Houston, TX, ³Department of Breast Medical Oncology, UT MD Anderson Cancer Center, Houston, TX, ⁴Department of Breast Surgical Oncology, UT MD Anderson Cancer Center, Houston, TX.

Background: Detection of circulating tumor cells (CTCs) is a strong prognostic biomarker for clinical outcomes in non-metastatic breast cancer. The kinetics of CTC clearance with adjuvant radiotherapy (RT) and its utility as a predictive biomarker have not been well characterized. Here we present the results of a prospective observational cohort study evaluating the dynamics of CTC detection and response to adjuvant breast radiotherapy. Methods: Patients with clinical or pathologic T3-4 or node-positive non-metastatic breast cancer dispositioned to receive comprehensive regional nodal irradiation (RNI) were prospectively enrolled. CTCs were enumerated using the CellSearch® assay immediately prior to RT (Pre-RT), within 1 week of RT (End of RT), and at follow-up 3-6 months following RT (Post-RT). Rates of CTC detection at each time point and clearance with RT were estimated. Associations between CTC status with clinicopathologic and treatment variables were performed using McNemar’s Exact, Wilcoxon, Fisher’s Exact, or Mann Whitney U Tests, as appropriate. For all applicable analyses, statistical significance was defined as a P-value of <0.05 (two-tailed). Results: At 28 months post study activation, 249 patients completed Pre-RT assessment of CTCs. The majority of patients were hormone receptor positive (HR+Her2-, 66%; Her2+, 13%; TNBC: 21%) and were treated with (neoadjuvant, 61%; adjuvant, 17%; none, 23%), mastectomy (69%) and axillary lymph node dissection (61%). Prior to RT, 61 (25%) of patients had ≥1 detectable CTC (median=1 cell [IQR, 1-2]). Of the CTC-positive patients who completed subsequent blood draws, 69% (n=42 of 61) achieved CTC clearance by the End of RT (McNemar’s Exact, P<0.001), with a further increase to 90% (n=55 of 61) by first Post-RT follow-up (P=0.002). Of the CTC-positive patients, 22% (n=13 of 59), 17% (n=9 of 53) and 20% (n=16 of 79) demonstrated Her2+ CTCs at each timepoint, respectively (Cochran’s Q Test, P=0.69). Detection of Her2-expressing CTCs was more common in patients with Her2+ primary tumors (25% versus 12%, P=0.05). Pre-treatment CTC-positive status was significantly associated with advanced nodal disease (cN3, 23% vs 11% p=0.04) but not subtype, grade, or lymphovascular space invasion. Of the 47 patients who achieved pCR with surgery, 30% (n=14) had detectable CTCs prior to RT. Clearance of CTCs was not significantly associated with any tested clinicopathological or treatment-related variable, including pathologic complete response (pCR) and residual cancer burden score. Conclusion: In this large prospective analysis of patients with locally advanced breast cancer, 25% of patients had detectable CTCs prior to radiotherapy; 30% of patients with pCR had detectable CTCs. Following comprehensive RT, 90% of CTC-positive patients achieved clearance of CTCs by first follow-up. While delayed CTC clearance without RT cannot be assessed on this single arm study, these results support further study of longitudinal monitoring of circulating tumor material as a potential real-time biomarker for residual locoregional disease prior to RT as well as a potential surrogate for radiotherapeutic efficacy.

F. BIDARD¹, K. Jhaveri², S. Kim³, E. Tokunaga⁴, P. Aftimos⁵, C. Saura⁶, J. O’Shaughnessy⁷, N. Harbeck⁸, G. Curigliano⁹, A. Llombart-Cussac¹⁰, E. Lim¹¹, S. Im¹², H. Bando¹³, P. Neven¹⁴, S. T. Estrem¹⁵, B. Nguyen¹⁵, S. Cao¹⁵, M. Makena¹⁵, B. Desai¹⁵, L. A. Carey¹⁶; ¹Medical Oncology, Institut Curie, Paris, FRANCE, ²-, Memorial Sloan Kettering Cancer Center and Weill Cornell Medical College, New York, NY, ³Asan Medical Center, University of Ulsan College of Medicine, Seoul, KOREA, REPUBLIC OF, ⁴Department of Breast Oncology, National Hospital Organization, Fukuoka, JAPAN, ⁵-, Institut Jules Bordet, Brussels, BELGIUM, ⁶Vall d’Hebron Institute of Oncology (VHIO), Vall d’Hebron University Hospital, Barcelona, SPAIN, ⁷Texas Oncology, Baylor University Medical Center, Dallas, TX, ⁸Breast Center, Department of Obstetrics and Gynecology and Comprehensive Cancer Center, Ludwig Maximilians University Munich University Hospital, Munich, GERMANY, ⁹-, University of Milan and European Institute of Oncology (IRCCS), Milan, ITALY, ¹⁰Hospital Arnau de Vilanova, Universidad Católica de Valencia, Valencia, SPAIN, ¹¹Garvan Institute of Medical Research, St Vincent’s Hospital, University of New South Wales, New South Wales, AUSTRALIA, ¹²Department of Internal Medicine, Seoul National University Hospital, Seoul, KOREA, REPUBLIC OF, ¹³Breast and Endocrine Surgery, Institute of Medicine, University of Tsukuba, Tsukuba, JAPAN, ¹⁴-, University Hospitals Leuven, Louvain, BELGIUM, ¹⁵-, Eli Lilly and Company, Indianapolis, IN, ¹⁶-, University of North Carolina at Chapel Hill, Chapel Hill, NC.

Background: The EMBER-3 trial, which included patients with ER+/HER2- advanced breast cancer previously treated with endocrine therapy (ET), demonstrated a significant improvement in progression-free survival (PFS) with imlunestrant (imlu) over standard ET (SOC ET, fulvestrant or exemestane) in patients with ESR1 mutations (ESR1m). This improvement was also seen with imlu+abemaciclib (abema) over imlu alone in all patients regardless of ESR1m. ctDNA is a surrogate marker for early clinical response. Here, we present exploratory analyses of ctDNA dynamics from baseline to 4 weeks after treatment. Methods: A total of 874 patients were randomized across 3 treatment arms: SOC ET (n=330), imlu (n=331), and combination imlu+abema (n=213). Plasma samples from baseline and after 4 weeks of treatment (cycle 2 day 1 [C2D1]) were sequenced using the Guardant360 assay, covering 74 genes. ctDNA dynamics were assessed as a continuous measure (mean variant allele frequency [VAF] change). Median PFS (mPFS) was assessed in patients who achieved or did not achieve a ≥50% reduction in ctDNA from baseline to C2D1. Results: A total of 635 patients (73%) from the intention-to-treat population had evaluable ctDNA samples at baseline and C2D1: SOC ET (n=241), imlu (n=240), and imlu+abema (n=154). In patients with ESR1m (n=223), imlu led to a higher overall decline in ctDNA (-74% vs -40%) and a reduction of ESR1m (-98% vs -75%), compared with SOC ET. Among all patients randomized to imlu or imlu+abema (n=311), imlu+abema led to higher overall decline in ctDNA compared with imlu alone (-58% vs -37%), regardless of baseline ESR1m status. Across all treatment arms, patients who achieved a ≥50% reduction in mean VAF from baseline to C2D1 had a longer mPFS compared with those who did not (Table 1), with variability between treatment groups. Conclusions: Analyses of early ctDNA dynamics from EMBER-3 demonstrated greater early ctDNA reduction with imlunestrant compared with SOC ET in patients with ESR1m. The addition of abemaciclib to imlunestrant enhanced ctDNA decline in all patients, consistent with the primary study outcomes. Furthermore, while a ≥50% decline in ctDNA at C2D21 was associated with improved PFS, the allocated treatment remained a key determinant of patient outcomes.

S. Hilz¹, N. Turner², K. Kalinsky³, S. Loibl⁴, K. Jhaveri⁵, S. Im⁶, C. Saura⁷, P. Schmid⁸, S. Loi⁹, T. Stout¹⁰, C. Song¹¹, K. Hutchinson¹, D. Juric¹²; ¹Translational Medicine Oncology, Genentech, Inc., South San Francisco, CA, ²Royal Marsden Hospital, Institute of Cancer Research, London, UNITED KINGDOM, ³Winship Cancer Institute, Emory University, Atlanta, GA, ⁴Center for Hematology and Oncology Bethanien, Goethe University and The German Breast Group, Frankfurt, GERMANY, ⁵Breast and Early Drug Development Service, Department of Medicine, Memorial Sloan Kettering Cancer Center, New York, NY, ⁶Seoul National University College of Medicine, Cancer Research Institute, Seoul National University, Seoul, KOREA, REPUBLIC OF, ⁷Vall d’Hebron University Hospital, Vall d’Hebron Institute of Oncology, Barcelona, SPAIN, ⁸Centre for Experimental Cancer Medicine, Barts Cancer Institute, Queen Mary University of London, London, UNITED KINGDOM, ⁹Division of Cancer Research, Peter MacCallum Cancer Centre and The Sir Peter MacCallum Department of Oncology, University of Melbourne, Melbourne, AUSTRALIA, ¹⁰PDO Clinical Science, Genentech, Inc., South San Francisco, CA, ¹¹Product Development Oncology, Genentech, Inc., South San Francisco, CA, ¹²Mass General Cancer Center, Department of Medicine, Harvard Medical School, Boston, MA.

BACKGROUND Inavolisib (INAVO) + palbociclib (PALBO) + fulvestrant (FULV) has been approved for treatment of patients (pts) with PIK3CA-mutated, hormone receptor-positive, HER2-negative, endocrine-resistant advanced breast cancer with improved efficacy over placebo (PBO) + PALBO + FULV in INAVO120 (NCT04191499; progression-free survival [PFS] hazard ratio 0.43; 95% CI 0.32-0.59; p < 0.0001). Exploratory analyses were done to better understand the molecular features of response to this regimen using FoundationOne^ⓇLiquid CDx sequencing results from circulating tumor (ct)DNA collected during screening (baseline). METHODS Profiling was performed on baseline plasma-derived ctDNA from 277 pts enrolled in INAVO120. Circulating tumor fraction (TF; estimated plasma DNA fraction derived from tumor cells) was computed by Foundation Medicine, Inc. and reported as a percentage of total cell-free DNA (n = 267 pts with detectable TF). Only alterations annotated as known / likely to be oncogenic were analyzed unless noted. Q values indicate Benjamini-Hochberg false discovery-corrected p-values. Interaction analyses were used to explore the significance of the relationship between biomarkers and the differential benefit between arms. RESULTS Following PIK3CA mutation (mut), the most common alterations detected in enrolled pts were in TP53 (49.4%; 132 / 267), CDH1 (17.6%; 47 / 267), and ESR1 (14.5%; 39 / 267, 34 of which were short variants). Within each arm, pts with above median (> 3.8%) TF had shorter PFS (higher vs lower TF: INAVO arm hazard ratio 1.9; 95% CI 1.2-3.1; PBO arm hazard ratio 1.8; 95% CI 1.2-2.7). There was no interaction between TF and benefit from the INAVO regimen (p = 0.935), hazard ratio 0.5 (95% CI 0.3-0.8) and hazard ratio 0.5 (95% CI 0.3-0.8) for the above-median and at- / below-median pt subgroups, respectively. The benefit from the INAVO regimen was similar for pts with both ESR1 short variant mut (ESR1mut) detected (hazard ratio 0.2; 95% CI 0.1-0.7) and no ESR1mut detected (hazard ratio 0.5; 95% CI 0.4-0.7), interaction p = 0.705. PTEN deleterious alterations (PTENmut) were detected in 10.9% of samples (29 / 267), and benefit from the INAVO regimen in those with PTENmut was hazard ratio 0.6 (95% CI 0.2-1.4) and without PTENmut was hazard ratio 0.5 (95% CI 0.3-0.6), interaction p = 0.706. An AKT1mut was present in only one pt. The PFS benefit of the INAVO vs PBO arms was similar for pts with either multiple PIK3CAmut (20.3%; 54 / 266 hazard ratio 0.6; 95% CI 0.3-1.1) or only one PIK3CAmut (hazard ratio 0.5; 95% CI 0.3-0.7), interaction p = 0.677. An unbiased univariate association analysis of gene alteration status with PFS identified TP53mut as a negative prognostic factor in both arms (q = 6.89 × 10^-4 and q = 4.59 × 10^-3, INAVO and PBO arms, respectively); regardless, the INAVO arm remained superior to the PBO arm in pts with TP53mut ctDNA (PFS hazard ratio 0.5; 95% CI 0.4-0.8). In interaction analyses, no genes with an alteration prevalence ≥ 10% in the efficacy-evaluable population were associated with differential benefit between arms. CONCLUSIONS Together, these findings support the efficacy of this INAVO regimen across a variety of genomic subgroups, including those with low / high TF, a well-known prognostic feature, and regardless of ESR1mut status. TP53mut were associated with poor prognosis in both arms. Although sample sizes were small, subgroup analyses in pts with a PTENmut did not conclusively suggest lack of benefit of this INAVO regimen vs PBO. Overall, these post-hoc, exploratory findings warrant further investigation in larger datasets.

F. Andre¹, P. A. Fasching², A. Prat³, P. Neven⁴, Y. Lu⁵, A. Bardia⁶, G. Curigliano⁷, V. Kaklamani⁸, P. Razavi⁹, Y. Yap¹⁰, F. Su¹¹, E. Roux¹², K. Pantoja¹³, J. Wu¹⁴, A. Lteif¹⁵, D. Juric¹⁶; ¹Medical Oncology, Institut Gustave Roussy, Villejuif, FRANCE, ²Gynecology and Obstetrics, University Hospital Erlangen Comprehensive Cancer Center Erlangen, Erlangen, GERMANY, ³Medical Oncology, Hospital Clínic of Barcelona, Barcelona, SPAIN, ⁴Multidisciplinary Breast Centre, Universitair Ziekenhuis Leuven, Leuven, BELGIUM, ⁵Oncology, National Taiwan University Hospital, Taipei, TAIWAN, ⁶Medicine, Division of Hematology/Oncology, David Geffen School of Medicine at UCLA, Los Angeles, CA, ⁷New Drugs and Early Drug Development for Innovative Therapies, European Institute of Oncology, Milan, ITALY, ⁸Hematology-Medical Oncology, UT Health San Antonio, MD Anderson Cancer Center, San Antonio, TX, ⁹Medical Oncology-Breast, Memorial Sloan Kettering Cancer Center, New York, NY, ¹⁰Medical Oncology, National Cancer Centre, Singapore, SINGAPORE, ¹¹Precision Medicine, Novartis Pharmaceuticals Corporation, East Hanover, NJ, ¹²Global Medical Affairs, Novartis Pharma AG, Basel, SWITZERLAND, ¹³Senior Principal Biostatistician, Novartis Pharmaceuticals Corporation, Cambridge, MA, ¹⁴Biostatistician, Novartis Pharmaceuticals Corporation, East Hanover, NJ, ¹⁵VP Head of Solid Tumors, Novartis Pharmaceuticals Corporation, East Hanover, NJ, ¹⁶Cancer Therapeutics, Massachusetts General Hospital Cancer Center, Boston, MA.

Background: RIB demonstrated clinically and statistically significant PFS and overall survival (OS) benefit in all of its phase 3 trials in HR+/HER2− advanced breast cancer (ABC; ML-2, -3, -7). We studied clinical characteristics and biomarkers of pts who derived long-term treatment benefit. We present an exploratory analysis of the ML trials of all 1L pts with long-term response (LTR; PFS >4 y) to RIB and 1L postmenopausal pts with very long-term response (VLTR; PFS >5 y). Methods: In the ML trials, pre- (ML-7) and postmenopausal (ML-2, -3) pts with HR+/HER2− ABC received RIB + ET or placebo + ET. Only pts who received 1L RIB were included in this pooled analysis. Pts with early relapse (treatment-free interval [TFI] ≤12 months on [neo]adjuvant ET) and those in ML-7 treated with tamoxifen were excluded. Overall median follow-up (mfu) was 6.1 y (ML-2, 6.7 y; ML-3, 5.9 y; ML-7, 4.5 y). Given the median PFS of ≈2 y reported with CDK4/6i in 1L ABC, more than doubling that time (PFS >4 y) was considered LTR. VLTR was defined as PFS >5 y, calculated only for postmenopausal pts (ML-2, -3) as mfu in ML-7 was <5 y. Biomarker analyses were performed using baseline ctDNA and tumor samples. Nominal P values (unadjusted) were calculated to compare characteristics of pts with LTR and non-LTR. Results: Of 666 pts treated with 1L RIB + ET included in the analysis, 109 were on treatment at data cutoff. A total of 153 pts (23.0%) had LTR; 164 pts (24.6%) were censored before LTR cutoff. Median PFS (mPFS) in pts with LTR was 6.8 y (95% CI: 6.4 y-NE [not estimable]); median OS (mOS) was NE (95% CI: 7 y-NE). Age, body mass index, and menopausal status were balanced between pts with LTR and non-LTR (Table). Similar proportions of premenopausal (34/150 [22.7%]) and postmenopausal pts (119/516 [23.1%]) had LTR. Pts with LTR were less likely than those with non-LTR to have liver metastasis (16.3% vs 25.5%) and less likely to have ≥3 metastatic sites (30.1% vs 43.0%). Similar proportions of pts with LTR vs non-LTR had bone-only disease (24.2% vs 19.5%). ctDNA analysis used a panel of 558 genes; gene expression analysis used 800 genes. Pts with LTR had a lower mean ctDNA fraction and were less likely to have CCND1 or TP53 alteration detected by ctDNA. Additionally, they had lower gene expression of CCNE1 and a higher proportion of luminal A disease (PAM50 based). Of 516 postmenopausal pts, 88 (17.1%) had VLTR; further analysis will be provided in the presentation. Conclusions: In the ML studies, 1 of 4 pts treated with 1L RIB were able to derive LTR (PFS >4 y). Pts with LTR had mPFS of 6.8 y, and mOS was NE. Although LTR was more evident in pts with better prognostic factors, noticeably, some pts with unfavorable prognostic factors were still able to achieve LTR. This exploratory analysis suggests long-term benefit with 1L RIB in keeping a good number of pts with HR+/HER2− ABC free of disease progression for more than 4 y.

S. Morganti¹, C. Song², N. Zhou¹, K. Santos¹, P. Jain², L. Walsh², R. Li², J. Rhoades², K. Gilligan², G. Kirkner¹, C. Stever¹, A. Patel¹, M. E. Hughes³, N. Priedigkeit¹, I. E. Krop³, G. Curigliano⁴, E. P. Winer³, S. M. Tolaney¹, N. Tayob¹, H. Heiling¹, K. Xiong², N. U. Lin¹, V. A. Adalsteinsson², H. A. Parsons¹; ¹Dana Farber Cancer Institute, Boston, MA, ²Broad Institute of MIT and Harvard, Cambridge, MA, ³Yale Cancer Center, New Haven, CT, ⁴University of Milano, Milano, ITALY.

Background: HER2 targeted therapy (tx) has transformed the trajectory of HER2+ metastatic breast cancer (MBC), with a subgroup of pts remaining on first line (1L) tx for many years (yrs). Detection of minimal residual disease (MRD) via ctDNA to guide tx is promising but data supporting its use in HER2+ MBC are virtually absent. Methods: We identified 70 pts with HER2+ MBC treated with 1L HER2 targeted tx consented to the EMBRACE cohort study with archival tissue, buffy coat and plasma at ≥1 key timepoints while on 1L tx (baseline, first-on-tx, year [Y] 1 landmark [L], Y2L, Y3L), or at time of disease progression (PD). We employed MAESTRO, an ultrasensitive whole-genome, tumor-informed, mutation enrichment sequencing assay, in a pooled testing format (MAESTRO-Pool). Exceptional responders were pts without PD or death for ≥3 yrs from 1L start (real-world progression-free survival [rwPFS] ≥3 yrs). Conventional responders were pts with PD or death within 3 yrs of 1L start (rwPFS<3 yrs). Local- or brain-only PD that did not lead to tx switch were not considered events. The primary objective was the association between Y1L MRD status (plasma samples collected 1 yr from 1L start +/- 3 months) and exceptional response. Sixty-three pts (90%) (39 exceptional and 24 conventional responders) had sufficient samples and successful assay design; MAESTRO-Pool was run on 147 samples. Results: More exceptional than conventional responders had de novo MBC (69.2% vs 41.7%, p=0.03). High (3+ by IHC) HER2 expression (87.2% vs 75.0%), estrogen receptor (ER)+ status (38.5% vs 33.3%) and visceral disease (61.5% vs 62.5%) were similar between exceptional and conventional responders. Most pts in both groups had 1L chemotherapy plus trastuzumab/pertuzumab (82.1% vs 70.8%) and maintenance endocrine tx if ER+ (86.7% vs 62.5%). With 284.4 person-yrs of follow up, 6/39 exceptional responders had late PD. In a landmark analysis at Y3, the 2-yr rwPFS was 97.4% (95%CI 82.8%-99.6%) (5 yrs from 1L start); 5-yr overall survival (OS) was 91.5% (95%CI 68.7%-97.9%) (8 yrs from 1L start). For conventional responders, total follow up was 110.5 person-yrs, median rwPFS was 1.65 yrs (95%CI 1.1-2.51), median OS was 4.58 yrs (95%CI 3.78-6.95). A median of 1,370 (range 136-4,635) tumor-specific mutations were tracked per patient. MRD was detected in 50 (34%) samples (median tumor fraction (TFx) 1352.8 ppm; range 4-206,399 ppm), including 21 (42%) samples with TFx <500 ppm, the limit of detection (LOD) of tumor-agnostic MBC assays, and 14 (28%) with TFx <100 ppm, the LOD of 1^st generation tumor-informed MRD tests. After excluding baseline and PD samples, MRD was detected in 7 (7.4%) and 28 (80%) samples from exceptional and conventional responders, respectively. All exceptional responders who remained progression-free were always MRD- (n=30) or cleared MRD by Y1L (n=3). Indeed, Y1L MRD status was associated with exceptional vs conventional response (0/26 [0%] vs 10/13 [76.9%], p<0.001). Four of the 6 pts who had late PD had a Y3L sample, which was MRD+ in 3 cases (lead time range 2.77-13.47 yrs) and MRD- in one case of breast-only PD. For 20/24 conventional responders, the last sample collected before PD was MRD+. Of the 4 MRD- cases, one pt had 2 primary breast cancers, one had brain-only PD. Conclusions: Here we report, to our knowledge, the first data showing a significant association between MRD detection and duration of response to 1L tx for HER2+ MBC. MRD status at Y1L was highly predictive of exceptional response. Only 2 conventional responders had distant PD without a prior MRD+ sample. All exceptional responders with late distant PD were MRD+ at Y3L. Fourteen (28%) MRD+ samples had TFx<100ppm and would have been misclassified by 1^st generation MRD tests, supporting the value of ultrasensitive MRD tracking to predict outcomes and its potential to guide tx de-escalation in prospective clinical trials.