Couch, Fergus J.

Abstract Polygenic risk scores (PRSs), which quantify inherited susceptibility to complex traits and diseases, have emerged as valuable tools for risk stratification and precision medicine. Despite their promise, PRS developed on European cohorts often demonstrate substantially reduced predictive accuracy in non-European populations, due to differences in genetic architecture. The disproportionate representation of European ancestry cohorts in genome-wide association studies (GWAS) leads to inequitable deployment of PRS technologies across diverse populations. Here, we introduce PRANA (Polygenic Risk Adaptation via Neural-network Architecture), a deep learning framework that adapts an existing PRS developed on one population to other ancestries. Unlike methods that require large-scale GWAS in the target population, PRANA leverages pre-trained PRS models derived from European cohorts and adapts them using modestly sized cohorts from the target population. We evaluated PRANA on seven complex traits in South Asian, East Asian and Ashkenazi Jewish populations, as well as in selected smaller East Asian subpopulations where the scarcity of training data poses a particular challenge. PRANA mostly improved predictive performance of the baseline PRS models by 5%-20% in terms of effect size (β) and Nagelkerke’s R², and, in most cases, outperformed existing cross-ancestry multi-PRS approaches. These results highlight PRANA as a scalable and practical strategy to reduce disparities in genomic risk prediction and advance the equitable application of PRS in diverse populations.

ABSTRACT Background Alcohol intake is a modifiable risk factor for breast cancer, yet patterns of alcohol intake in women with breast cancer remain underexplored. This study evaluates demographic and clinical predictors of alcohol intake in patients with breast cancer. Methods Patients at Mayo Clinic Rochester diagnosed with breast cancer (Stage 0–3) between July 2014 to March 2022 were surveyed about their average alcohol intake over the 10 years prior to their diagnosis. Alcohol intake was categorized as < 1, 1–4, 5–14, and ≥ 15 drinks per week. Demographic, behavioral, and clinical data were analyzed using univariate Fisher exact tests and multivariate multinomial logistic regression. Results Of 2030 participants (mean age at diagnosis: 59 years; 99.3% female; 95.9% White), 29% consumed < 1 drink/week, 48% reported 1–4 drinks/week, 20% consumed 5–14 drinks/week, and 3% reported ≥ 15 drinks/week. Older age at diagnosis (≥ 60 years), non‐White race, financial difficulty, and never smoking were associated with less alcohol intake. Patients who reported higher levels of mild–moderate intensity exercise reported more alcohol intake. Tumor stage, tumor receptors, and BRCA mutation status were not associated with alcohol intake. PROMIS‐10 physical health scores correlated with alcohol intake in univariate but not multivariate models. Conclusion Alcohol intake is more common prior to a breast cancer diagnosis for patients who are younger, White, financially secure, exercising, and smoking. These findings may inform alcohol intake‐related breast cancer prevention efforts and highlight patient subgroups who could benefit from targeted counseling at the time of diagnosis.

Abstract Background Black women are more likely to develop and die from early‐onset breast cancer. Hereditary breast cancer (HBC) is strongly associated with early age of onset. Accordingly, HBC characteristics in young Black women were evaluated. Methods Black women diagnosed with invasive breast cancer at age 50 years or younger between 2005 and 2018 were recruited through the state cancer registries in Florida and Tennessee. Participants were asked to complete a questionnaire, sign medical records releases, and donate a saliva sample for free genetic testing to identify HBC germline pathogenic or likely pathogenic variant (GPV). Women with sporadic versus hereditary breast cancer were compared, based on age distribution, clinical characteristics, and family cancer history. Results Among 686 participants with confirmed genetic testing, 15.3% had GPVs in breast and/or ovarian cancer predisposition genes, including BRCA1 (7.7%), BRCA2 (4.5%), PALB2 (0.9%), and ATM (0.7%), with 1.5% in other genes. Compared to the sporadic group, triple‐negative breast cancers were overrepresented among BRCA1 carriers (23.6% vs. 75.5%; p < .001), and family history of breast cancer was significantly more frequent among BRCA1 , BRCA2 , and PALB2 carriers (all p < .05). The majority of breast cancers occurred by age 40 years in BRCA1 carriers (52.8%), compared to other genes, where age was more evenly distributed. Conclusion Among unselected young Black women with breast cancer, our findings demonstrate that the highest frequency of HBC GPVs are in BRCA1 , BRCA2, and PALB2, with identification of distinct histopathologic and demographic characteristics across HBC genes, underscoring the critical need for genetic testing in all populations.

Background Tumor-infiltrating lymphocytes (TILs) are established determinants of clinical outcomes in breast cancer. 1–6 While stromal TILs (sTILs) are commonly evaluated due to easier scoring and higher interobserver reproducibility, intraepithelial TILs may provide additional insights into anti-tumor immunity. 7 We aimed to comprehensively characterize immunotype-associated cellular landscapes in triple-negative breast cancer (TNBC) and identify features associated with clinical outcomes. Methods 212 TNBC tumors were classified into three immunotypes based on immune infiltrate location and density: immune desert (ID), immune excluded (IE), and immune inflamed (IN). 8 Single-cell spatial transcriptomics was performed on 75 tumors to define the cellular composition and functional states of the tumor microenvironment across immunotypes. Findings were validated and integrated with bulk RNA sequencing data and clinical outcome datasets from FinXX, CALGB-40603, I-SPY 2, and Real-World Clinico-Genomic Data (RWCGD) cohorts. Results Patients with IN tumors had significantly improved outcomes compared with ID tumors. Despite high sTILs, IE tumors showed poor outcomes similar to ID tumors. Single-cell spatial analysis revealed that ID and IE tumors exhibited reduced major histocompatibility complex I/II expression and fewer tumor-resident plasmacytoid dendritic cells (pDCs) compared with IN tumors, which were enriched for interferon-alpha (IFNα) and interferon-gamma (IFNγ) responses. High IFN response scores were associated with favorable outcomes across multiple therapy types in independent datasets. Deconvolution of RWCGD bulk RNA-seq data confirmed that pDC abundance correlated with improved outcomes specifically in hormone receptor–negative subtypes. Conclusions Our study highlights pDCs and IFN signaling as hallmarks of effective anti-tumor immunity in TNBC. Immunotype-based profiling underscores the limited prognostic value of sTILs alone in immune-excluded tumors and supports pDCs and IFN pathways as potential biomarkers for prognosis and therapeutic development in TNBC.

e12697 Background: Postpartum-associated breast cancer (PABC) has been defined clinically as breast cancer (BC) diagnosed within the first 5 years of the most recent birth. Though generally rare, incidence rates are increasing in conjunction with delayed childbearing and older maternal age at first birth. Currently, there are limited data regarding serial patient-reported outcomes in this patient population compared to non-PABC patients. Methods: Patients with PABC from the Mayo Clinic Breast Disease Registry who completed baseline, 2, and 4-year surveys were matched 1:1 to BC patients whose cancers occurred more than 5 years after their most recent birth (or in the setting of nulliparity) based on age, clinical stage, treatment received, and tumor subtype. Survey results from the Patient Health Questionnaire-2 (PHQ-2), Patient-Reported Outcomes Measurement Information System-10 (PROMIS-10), and Impact of Event Scale-Revised (IES-R) collected at approximately two and four years after diagnosis were compared between PABC and non-PABC patients. Univariable analyses used Wilcoxon rank sum tests, whereas multivariable analyses employed linear regression models adjusting for potential confounding effects. Results: A total of 222 individuals were included (111 with PABC). A higher proportion of individuals in the PABC cohort were married compared to the non-PABC cohort (80% vs. 42%, p < 0.001); otherwise, the groups were well-balanced across multiple categories, including age, gender, race/ethnicity, menopausal status, and germline mutation status. In univariable analyses, the Year 2 composite PHQ-2 score was significantly worse among individuals with non-PABC (p = 0.04), and the non-PABC cohort reported greater difficulty with sleep (p = 0.03) as measured by IES-R at Year 4. Both associations attenuated in a multivariable analysis controlling for marital status. In marital status-adjusted analyses of Year 4 survey results, PABC participants reported greater difficulty in carrying out social activities but were less likely to report emotional distress as evaluated by PROMIS-10 (compared with non-PABC participants). No other variables were significant univariably or multivariably (p > 0.05). Conclusions: In general, quality of life is similar among survivors of PABC compared to those diagnosed with breast cancer outside the postpartum period (or when nulliparous). However, our findings reveal that individuals with PABC are less likely to experience emotional distress 2 and 4 years after the BC diagnosis, which is unexpected, given that postpartum patients typically report higher levels of distress compared to their non-postpartum counterparts. Further investigation into these findings is warranted.

10513 Background: Prior reports of high frequency of BRCA1 and BRCA2 ( BRCA1/2 ) PVs among South Asian women from Bangladesh, India, Nepal, and Pakistan with breast or ovarian cancer require validation in large prospective unselected cohorts. Methods: Individual-level anonymized patient data from two prospective cohorts were pooled to evaluate the prevalence of germline PVs among South Asian women: the South Asian Regional Alliance for Hereditary Cancers (SARAH Consortium) in Bangladesh, Nepal, and Pakistan, and the previously reported Nurse-led Genetic Counseling and Awareness (NuGenA) expansion cohort in India. Both studies enrolled consecutive cases of histologically confirmed newly diagnosed breast or epithelial ovarian cancer irrespective of age at diagnosis or family history. In both studies, participants completed questionnaires on personal and family history (first-, second- or third-degree) and provided blood for germline genetic testing after written consent. PVs in cancer predisposition genes including ATM, BARD1, BRCA1, BRCA2, BRIP1, CHEK2, MLH1, MSH2, MSH6, PALB2, RAD51C, RAD51D, and TP53 were identified as part of the studies utilizing a multigene panel and the results were returned to ordering clinicians. The data cutoff date for this analysis was November 30, 2025. Results: The analysis included 1,711 South Asian women with breast (n=1,210) or ovarian (n=501) cancer. A family history of breast or ovarian cancer was noted in 177 (10.3%). Among breast cancers, 94% were invasive, 84% were ductal, 35% were triple-negative and over two-thirds presented with nodal involvement. Germline PVs across 13 predisposition genes were observed in 17.7% of breast and 35.9% of ovarian cancers. Notably, BRCA1/2 PVs accounted for most of these PV (13.6% of breast and 28.3% of ovarian cancer cases) with PVs in all genes except BRCA1/2 under 1%. The frequency of BRCA1/2 PVs was 24.4% in triple-negative breast cancer and 8.0% in ER+/HER2-. BRCA1/2 PV carriers were diagnosed with breast cancer at a younger age than non-carriers (Median age: 42 vs 46 years, P<0.001), with no age difference observed in ovarian cancer (50 vs. 51 years, P=0.79). Notably, among the 306 BRCA1/2 PV carriers identified, 229 (74.2%) did not report a family history of breast or ovarian cancer. Conclusions: We confirm a high prevalence of BRCA1/2 PVs among unselected South Asian women with either breast or ovarian cancer. The observations of younger age at cancer diagnosis, and low rates of family history support the need for population-specific germline genetic testing and risk management strategies in South Asia.

ABSTRACT Reversion mutations (REVs) restore homologous recombination repair (HRR) and confer resistance to PARP inhibitors (PARPi) in HR-deficient cancers. Yet, their prevalence, mechanisms, and biological constraints remain undefined. We analyzed genomic profiling of 609,464 tissue and liquid biopsy samples across multiple cancer types to delineate the pan-cancer landscape of REVs. REVs were identified in eight HRR genes, most frequently BRCA2 and BRCA1 and notably never in ATM or CHEK2 . REVs exclusively impacted truncating pathogenic variants, predominantly through large in-frame and exon-level deletions, associated with repetitive sequences. Conserved functional domains are relatively depleted of REVs. Structural modeling and functional studies support that exon-level deletions preserve critical domain architecture and confer PARPi resistance. The study establishes HRR reversion as a structurally permissive, yet evolutionarily constrained, resistance mechanism with implications on response, monitoring, and therapeutic strategy. One Sentence Summary Homologous recombination reversion evolves through structurally tolerated, microhomology-driven alterations that restore DNA repair under therapeutic selection across cancers.

Abstract Background For many evidence criteria within v3.0 of the ACMG/AMP guidelines, methodologies have been developed to empower their use outside the stipulated evidence strengths. However, no such methodology has been established for case-control data (PS4). With the release of large-scale unselected case-control datasets and expansion of nationally-collected laboratory datasets enriched for pathogenic variant carriers, there is potential to combine datasets across ascertainment contexts in a more quantitative manner using novel likelihood ratio tools. Methods Using our published PS4-LR-Calculator, we calculated a combined log likelihood ratio (PS4-LLR) across five datasets (three unselected, and two enriched), and estimated enrichment of pathogenic variants in clinically-ascertained laboratory data using truncating variant prevalence. Results Data were combined for 10,817 missense variants from 325,345 female breast cancer patients and 671,006 controls of Western European ancestry for five breast cancer susceptibility genes ( BRCA1, BRCA2, PALB2, ATM, CHEK2 ). A combined LLR was produced for 4,690 missense variants; 927 variants received evidence towards pathogenicity (LLR≥ 1), and 3,242 received evidence towards benignity (LLR≤ -1). Conclusion This flexible, variant-level methodology combines nationally-collected ‘enriched’ datasets with unselected case-control cohorts, expanding the available information for case-control analysis, boosting power, enabling exploration of atypical penetrance and empowering variant classification.

Abstract Inherited susceptibility plays a critical role in prostate cancer (PCa) risk. Using large biobank and case-control datasets, we evaluated the contribution of both common variants and rare germline pathogenic variants (PVs) to overall PCa risk. The Prostate Cancer Exome Sequencing Consortium currently includes 427,388 male participants (51,452 PCa cases and 375,936 controls) with whole-exome sequencing data from ten biobanks and studies: UK Biobank (14,669 cases/195,600 controls), All of Us Research Program (7,577/75,226), African Ancestry Prostate Cancer Consortium (7,176/4,675), Mayo Clinic Biobank (6,031/15,084), Mass General Brigham Biobank (3,393/14,095), Geisinger’s MyCode Community Health Initiative (3,026/15,130), UCLA ATLAS Precision Medicine Biobank (2,850/16,904), Penn Medicine Biobank (2,598/16,255), Colorado Center for Personalized Medicine (2,269/13,399), and Malmo Diet and Cancer (1,863/9,568). Based on self-reported race/ethnicity and estimated genetic ancestry, the cases comprise approximately 79% European, 18% African, and 3% other ancestry populations. Single-variant association analyses tested all variants on chromosomes 1-22 and X with a minor allele count ≥ 5. In gene-based analyses, PVs were defined as rare variants (minor allele frequency [MAF] &lt; 1% in controls) that had either a Variant Effect Predictor (VEP) impact score of “high” or a pathogenic or likely pathogenic ClinVar classification. Associations were estimated using Firth logistic regression, adjusting for age and the top ten genetic principal components. Results from individual studies were combined using fixed-effect meta-analysis. In single-variant association analyses, 496 variants reached genome-wide significance (p&lt;5×10-8; MAF&gt;0.02%). Among these, 458 (92%) variants mapped to previously known risk regions, including three rare PVs in HOXB13 (rs138213197), CHEK2 (rs555607708), and FAM111A (rs533676902). Characterization of the remaining 38 variants is ongoing. Gene-based analyses identified significant associations (p&lt;2.4×10-6) for eight genes: HOXB13 (OR=3.7, 95% CI=3.3-4.2), BRCA2 (OR=2.0, 95% CI=1.7-2.3), CHEK2 (OR=1.6, 95% CI=1.5-1.8), ATM (OR=1.6, 95% CI=1.4-1.9), FAM111A (OR=1.4, 95% CI=1.3-1.5), BIK (OR=1.4, 95% CI=1.2-1.6), SAMHD1 (OR=2.1, 95% CI=1.6-2.7), and SMOC2 (OR=3.2, 95% CI=2.0-5.1). All genes except SMOC2 have been previously implicated in PCa susceptibility. Among cancer predisposition and DNA repair genes, nominal associations were also observed for XRCC2 (OR=1.6, 95% CIs=1.2-2.3) and BRCA1 (OR=1.2, 95% CI=1.0-1.4), whereas the association was not significant for PALB2 (OR=1.2, 95% CI=0.9-1.5). These findings reinforce the role of rare germline PVs, particularly in cancer predisposition and DNA repair genes, in PCa susceptibility. As additional studies are incorporated into the Consortium, we expect this work to provide a more comprehensive characterization of the genetic architecture of PCa. Citation Format: Yifan Zhang, Shuyan Cheng, Nicholas Boddicker, Matthew Lebo, Alexander S. Berry, Roni Haas, Ryan Hausler, Tokhir Dadaev, Heena Desai, Alex A. Rodriguez, Ravi K. Madduri, Andrew Hill, Xin Sheng, Susan M. Gundell, Mine Cicek, Penn Medicine Biobank, Olle Melander, Chris R. Gignoux, Isla P. Garraway, Bogdan Pasaniuc, Paul C. Boutros, Matt Oetjens, Adam S. Kibel, Robert J. Klein, Zsofia Kote-Jarai, Fergus J. Couch, Kara N. Maxwell, Burcu F. Darst, David V. Conti, Christopher A. Haiman, Fei Chen. Genetic risk of prostate cancer: Insights from the Prostate Cancer Sequencing Consortium [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(8_Suppl):Abstract nr LB390.

Abstract Artificial intelligence (AI)-scores estimated from digital mammograms predict future breast cancer (BC) risk. MIRAI is a deep learning BC risk model that provides a continuous 5-year risk of overall BC from four screening full field digital mammograms. Common germline genetic variation in the form of a polygenic risk score (BC-PRS) is associated with increased BC risk and may improve MIRAI’s 5-year risk prediction. Our goal was to develop an updated MIRAI 5-year risk model that incorporates the BC-PRS (MIRAI+PRS) and to evaluate its discriminatory accuracy and calibration for both overall and invasive BC compared to MIRAI alone. We developed the MIRAI+PRS model by multiplying each woman’s 5-year MIRAI risk estimate by their relative risk based on their BC-PRS relative to the population mean. We evaluated the models within the Mayo Clinic Biobank mammography cohort, comprised of 12,307 women without a prior history of BC; 176 invasive and 250 overall BC were diagnosed within 5 years. MIRAI was estimated on screening mammograms closest to enrollment but at least 6 months prior to BC. Discriminatory accuracy, assessed by C-statistic, was high and similar for MIRAI+PRS vs. MIRAI models, for overall BC and invasive BC (Table). Calibration assessed by observed to expected (O/E) ratios was also similar for MIRAI+PRS compared to MIRAI predictions for BC outcomes (Table), although there was improvement in decile-specific O/E ratios across the lowest risk deciles (&lt;1.67%) for MIRAI+PRS. Calibration for invasive cancer was poor for MIRAI with or without PRS. In summary, the MIRAI+PRS risk model did not result in significant difference of discriminatory accuracy or overall calibration compared to the MIRAI model, but there was evidence for improved calibration for women with 5-year risk below 1.67%. For invasive BC, the model had poor calibration regardless of whether PRS was included, underscoring the importance of training AI models for BC outcomes that are associated with a clinical intervention. Citation Format: Christopher G. Scott, Peter Kraft, Imon Banerjee, Ramon Correa Medero, Aaron D. Norman, Fergus J. Couch, Karla Kerlikowske, Stacey J. Winham, Celine M. Vachon. Development and evaluation of a MIRAI 5-year risk model with a breast cancer polygenic risk score [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 2779.