Citations

Background: Several breast cancer (BC) risk prediction models have been developed to provide personal risk assessments. Though individually validated, their performance has not been systematically evaluated across a wide range of populations or ages. Methods: We harmonized individual-level baseline questionnaire data and incident BC diagnoses from 21 cohorts from North America, Europe, and Australia participating in the Breast Cancer Risk Prediction Project. Five-year absolute risk of invasive BC was estimated for five established risk prediction models using classical risk factors only. Discrimination was evaluated by area under the curve (AUC). Calibration was assessed using average and risk-decile specific expected to observed (E/O) ratios. Performance metrics were meta-analyzed across cohorts and models. Metaregression tested associations between cohort characteristics and performance metrics. Results: This analysis included 1,595,977 women aged 20-75 years, enrolled in studies between 1976-2015, with 19,062 (1.2%) invasive BC cases ascertained within 5 years from exposure assessment. Age-adjusted AUCs were similar across models and cohorts (pooled AUCs by model: 0.57-0.58), while E/O ratios varied substantially (pooled E/O ratios by model: 0.83-1.25). Overestimation was common among predicted high-risk individuals (>3%). No appreciable differences in model performance by cohort age, birth year, race, and variable missingness emerged. Calibration improved after assigning race-specific incidence rates. Conclusion: Existing BC risk prediction models provided similar risk discrimination across multiple cohorts, although there was overestimation of risk for high-risk individuals. Performance variation across cohorts was not driven by specific characteristics, which supports development of a unified risk model for diverse populations that leverages appropriate incidence rates.

Abstract Background Breast cancer is etiologically heterogeneous, but which risk factors differ in their associations across tumor subtypes remains unclear. We conducted a large, pooled analysis to evaluate independent, dose-response associations between breast cancer risk factors and quantitative tumor features. Methods Analyses of 15,731 invasive breast cancers from 24 studies evaluated associations (p-trend) between reproductive and hormonal factors, body mass index (BMI), alcohol, smoking, and family history in relation to quantitative immunohistochemistry measures on tissue microarrays (ER, PR, HER2, KI67, TP53) and tumor grade. Analyses in a subset of 10 population-based studies estimated subtype-specific odds ratios (ORs) comparing cases to controls. A Bayesian False Discovery Probability (BFDP) <0.2 was used to identify associations with strong statistical evidence. Results Nulliparity and later age at menopause were associated with higher ER-positivity (p-trend=0.021 and 0.001, respectively), with corresponding OR[ER+] (95% CI) = 1.49 (1.16-1.90) for nulliparous vs. parous and 1.07 (1.03-1.11) per 5 years. Current combined menopausal hormone therapy (MHT) use was associated with lower grade (p-trend<0.001), with OR [grade1] = 3.37 (2.69-4.21) for current vs. never users. Higher BMI was associated with lower ER-positivity and higher grade in premenopausal women (p-trend<0.001 and <0.001), with OR[ER+] = 0.80 (0.74-0.87) and OR[grade1] = 0.75 (0.63-0.88) per 5 units, and with higher PR-positivity and higher grade in postmenopausal women (p-trend<0.001 and <0.001), with OR[PR+] = 1.08 (1.03-1.14) and OR[grade 3] = 1.10 (1.04-1.17) per 5 units. Conclusion This pooled analysis of 15,731 cases showed that nulliparity, age at menopause, MHT, and BMI have independent, dose-response associations with ER, PR, and grade, clarifying patterns of etiologic heterogeneity. Associations with HER2, KI67 and TP53, or other risk factors did not meet our threshold for strong evidence.

7518 Background: Monoclonal gammopathy of undetermined significance (MGUS) is common, yet only a minority progress to malignant plasma-cell disorders. Mayo risk stratification uses baseline factors and is commonly implemented/validated in a Cox proportional hazards (CoxPH) framework, which may miss prognostic longitudinal patterns. Methods: We assembled 15,158 MGUS patients with longitudinal laboratory data and progression outcomes (smoldering multiple myeloma, multiple myeloma, AL amyloidosis, or Waldenström macroglobulinemia). We compared a Mayo-factor Cox proportional hazards (CoxPH) baseline versus STRaTS, a transformer-based survival model, to predict progression risk at 3-, 5-, and 10-year horizons. STRaTS leveraged longitudinal trajectories of routine labs including hemoglobin, platelets, differential counts (neutrophils/lymphocytes/monocytes), immunoglobulins (IgG/IgA/IgM), free light chains (kappa/lambda), M-spike, and creatinine. To minimize bias, we used laboratory results collected from 1–3 years after diagnosis and started follow-up at year 3; patients progressing before year 3 or lacking follow-up beyond year 3 were excluded. A total of 7,305 patients met eligibility and were split into train/validation/test cohorts of 4,383/1,461/1,461. Discrimination was assessed by concordance index (C-index). Clinical operating performance was evaluated on a held-out test cohort by comparing sensitivity and specificity at the same proportion of patients selected for intensified surveillance (same workload). Results: Among MGUS patients with complete baseline Mayo factors(N=3,388), Kaplan–Meier–estimated 3-/5-/10-year cumulative progression risk was 0.1/0.4/3.4% (Low), 3.0/4.3/7.9% (Low-Intermediate), 4.4/6.4/14.1% (High-Intermediate), and 8.9/12.5/16.9% (High). On a held-out test cohort, STRaTS demonstrated superior discrimination versus the Mayo-factor CoxPH (C-index 0.75 vs 0.538; p<0.0001). Both models were trained and evaluated on identical train/test splits. At the same surveillance workload, STRaTS achieved higher sensitivity with comparable specificity; when 40.3% of patients were selected for closer surveillance, 3-year sensitivity was 81.0% for STRaTS versus 47.6% for CoxPH (Table). Conclusions: Among MGUS patients who remain progression-free at 3 years, transformer-based survival modeling of longitudinal laboratory trajectories provides more accurate individualized conditional progression risk estimates and improved workload-matched identification of progressors compared with the conventional Mayo/Cox approach. Sensitivity/specificity (%). Operating Point Flagged % STRaTS Sens / Spec (%) CoxPH Sens / Spec (%) Int+High, 3y 40.3 81.0 / 58.5 47.6 / 58.9 Int+High, 5y 40.3 81.2 / 60.4 46.9 / 58.8 Int+High, 10y 40.3 25.0 / 56.3 23.8 / 58.7 Two-Group High (≥2 factors), 5y 5.6 37.4 / 93.4 35.4 / 93.1

Image-derived artificial intelligence (AI) risk models have shown promise in short-term risk assessment for improving breast cancer (BC) screening. No image-derived long-term AI risk model for primary prevention has been developed and externally validated. Individuals aged 31 to 94 years were recruited between 2010 and 2020 in two population-based case cohorts (Olmsted County, Minnesota, US, and KARMA, Sweden) and one hospital-based case-control study (EMBED, Atlanta, US). Median follow-up in the case cohorts was 10 years, with BCs diagnosed through June 2022. Additional validation was performed in EMBED with 3-year follow-up. An AI risk model was developed in Sweden, and we report independent validation in Olmsted/KARMA/EMBED. Absolute 10-year risks were estimated at study entry. Time-dependent discriminatory performance [AUC( t )] and expected-to-observed events (E/O) were estimated. Comparisons were performed with clinical risk tools and the image-based Mirai tool. Across Olmsted/KARMA, 8696 individuals (mean age, 54.4±10.6) and 1633 individuals with incident BC (mean age, 57.0±10.6) were included. Average 10-year risks were 3.83 and 3.14%, and E/O ratios were 0.99 and 0.99. Ten-year AUC( t ) for invasive BCs were 0.72 in Olmsted and 0.72 in KARMA. Similar discriminatory results were observed in EMBED. Our AI risk tool performed significantly better than Mirai in Olmsted/KARMA/EMBED. In KARMA, in the top 10% of high-risk individuals, the AI risk tool predicted 33% of BCs compared with 23%/20%/24% predicted by Tyrer-Cuzick-v8/BCSC-v3/Mirai (all P < 0.01). The 10-year image-derived AI risk model showed significantly higher performance than clinical risk models and Mirai in diverse populations, supporting its clinical potential for primary prevention.

Abstract Context Prior reports of plasma estradiol concentrations and their reference ranges in postmenopausal women identified marked variation, in part due to assay method and body mass index (BMI). Objective To develop an international reference range for plasma estradiol in postmenopausal women, support accurate assessment of concentrations during systemic and local (eg, vaginal) estradiol therapy, and improve risk prediction for diseases, including osteoporosis, breast, and endometrial cancers. Methodology Estradiol concentrations were measured using primarily liquid chromatography-tandem mass spectrometry (LC-MS/MS) assays in 5 international cohorts of community-dwelling, postmenopausal women of primarily European ancestry aged 38-100 years who were not using exogenous estrogens (n = 7206). Measurements were harmonized to the Centers for Disease Control and Prevention (CDC) reference measurement procedure (RMP) by re-assaying a subset of samples for each cohort and generating recalibration equations. Reference intervals were determined for all women, non-obese women, and specific BMI categories. Results Estradiol concentrations from separate LC-MS/MS assays correlated closely with the CDC RMP, and harmonization minimized inter-assay variability. Estradiol concentrations correlated with BMI but not with chronological age. The reference range (2.5th-97.5th percentile) was 1.1-18.2 pg/mL (4.0-66.8 pmol/L) with median of 4.9 pg/mL (18.0 pmol/L) for all postmenopausal women and 1.1-12.5 pg/mL (4.0-45.9 pmol/L) with median of 4.1 pg/mL (15.1 pmol/L) for non-obese women (BMI &lt; 30). Reference ranges increased progressively with higher BMI categories. Conclusion Cross-calibration of estradiol measurements with the CDC RP enabled the development of harmonized plasma estradiol reference ranges for postmenopausal women who were not using exogenous estrogens, suitable for future studies to establish clinical utility. Increasing obesity in postmenopausal women was associated with increasing estradiol concentration and BMI category-specific reference ranges.

Abstract Background. Genome-wide association studies (GWAS) have identified 35 genetic susceptibility single-nucleotide polymorphisms (SNPs) for multiple myeloma (MM) in individuals of European ancestry (EA) and shown strong genetic correlation between MM and its precursor, monoclonal gammopathy of undetermined significance (MGUS). We evaluate the contribution of the 35 MM variants to MGUS susceptibility overall and by prognostic subgroups. Methods. The study included 20,756 participants (14,486 controls, 1,883 with MGUS, and 2,163 with MM) from the Mayo Clinic. Logistic regression assuming an additive model estimated odds ratios (ORs) and 95% confidence intervals (CIs) for individual SNPs and for the MM-PRS, adjusted for age, sex, study, and principal components. The PRS was a weighted sum of 35 SNPs with effect estimates from the largest MM GWAS and was modeled continuously (per SD) and by quintiles (Q1-Q5). SNPnexus annotated variants that replicated in MGUS (P &lt; 0.05 and OR &gt; 1.01) versus those that did not (P &gt; 0.05 and OR &lt; 1.01). Results. The 35-SNP MM-PRS was strongly associated with MM risk and modestly with MGUS. Compared with the middle quintile (Q3), MM odds rose from OR=0.55 (CI=0.46-0.65, P=2.2e-11) in Q1 to OR = 1.82 (CI=1.59-2.09, P=6.2e-18) in Q5. MGUS showed a similar, attenuated pattern (Q1 OR = 0.74, CI=0.64-0.84, P=6.4e-6 / Q5 OR=1.36, CI=1.21-1.53, P=3.4e-7). Each SD increase in PRS corresponded to OR=1.52 (CI=1.45-1.59, P=1.3e-66) for MM and OR = 1.22 (CI=1.18-1.27, P=1.0e-23) for MGUS. Higher PRS values were linked to larger M-protein ((0.1-1.5 g/dL: OR=1.40, CI=1.18-1.65, P=5.8e-12) vs &lt;0.1 g/dL: OR = 1.19, CI=1.13-1.24, P=5.8e-12)) and abnormal free light chain (FLC) ratio ((OR = 1.33, CI=1.21-1.46, P=1.0e-9) vs normal ratio (OR=1.19, CI=1.14-1.25, P=1.4e-13)). Ten risk loci replicated in MGUS, mapping to genes involved in plasma-cell function, immune regulation, and DNA repair and enriched for Rho GTPase signaling, NF-κB-mediated apoptosis, and RNA polymerase II transcription, implicating early plasma-cell activation and transcriptional control. The 12 non-replicating loci, including PHC3, ATG5, and NFIC, mapped to genes involved in chromatin remodeling, autophagy, and SUMOylation, suggesting roles in stress response and genomic maintenance. Conclusions. The MM-PRS captures shared heritability between MM and MGUS and correlates with MGUS subtype and severity. Replicating variants highlight immune and cell-cycle pathways relevant to MGUS onset, whereas non-replicating loci cluster in DNA-repair and stress-response processes, underscoring their potential role in progression. Citation Format: Alyssa Ione Clay-Gilmour, Angelica Macauda, Cristine Allmer, Danelle Moonen, Aaron D. Norman, Nicholas Boddicker, Janet E. Olson, Elizabeth E. Brown, Vincent S. Rajkumar, Esteban Braggio, David Murray, Susan Slager, Shaji Kunnathu Kumar, Celine Vachon. Implications of multiple myeloma polygenic risk scores (PRS) for MGUS [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 3603.

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.

Abstract Background: Hypogammaglobulinemia at the time of CLL diagnosis predicts shorter time to first therapy (TTFT). However, the depth of immunoparesis and its impact on TTFT in newly diagnosed MBL/CLL remains unexplored. Methods: We identified previously untreated MBL/CLL patients from Mayo Clinic CLL Database (2000-2024) who had serum immunoglobulin (Ig; IgG, IgM and IgA) and monoclonal protein assessed within 1 year of diagnosis. Immunoparesis was evaluated: a) qualitatively by classifying patients into preserved, partial and full Ig suppression, based on the number of suppressed Igs; and b) quantitatively by calculating the average relative difference (ARD), defined as the mean percentage below normal for all Igs. For both methods, only uninvolved Igs were considered for those with monoclonal protein. TTFT was estimated by Kaplan-Meier with competing risk of death. Cox models estimated hazard ratios (HR) and 95% confidence intervals (CI). Results: Among 1420 patients, median age at diagnosis was 65 (range 28-96); 954 (67%) were male; 493 (44.8%) had unmutated IGHV, and 86 (6.8%) had TP53 disruption. The median [range, mg/dL] serum IgG, IgA, and IgM were 889 [111-5290], 132 [1-4880], and 49 [4-9480]. Low IgG was present in 472 (33.2%), low IgA in 223 (16.4%), and low IgM in 530 (37.3%) patients. An M-spike was detected in 122 (8.6%) patients; distribution was: IgG (73), IgM (30), IgA (6), or multiple (13). By qualitative assessment, preserved Ig were seen in 664 (46.8%), partial Ig suppression in 604 (42.5%), and full suppression in 152 (10.7%) patients. By quantitative assessment, median ARD was 0.6 (-0.8 to 11.3); 280 (19.7%) patients had negative ARD, and 1140 (80.3%) had positive ARD. In those with negative ARD, 128 (9%) had partial Ig suppression and 152 (10.7%) had full Ig suppression; in those with positive ARD, 664 (46.8%) had preserved Ig and 476 (33.5%) had partial Ig suppression. Median follow-up was 14.7 years; 510 patients progressed requiring therapy. Median TTFT was 9.6 years. Median TTFT for patients with preserved Ig was 12.6 years, 8.0 years for partial Ig suppression (HR 1.4, 95%CI 1.1-1.7), and 1.8 years for full Ig suppression (HR 3.1, 95%CI 2.4-4.0). The median TTFT was 2.3 and 12.1 years for patients with negative vs positive ARD; negative ARD was associated with shorter TTFT (HR 2.4, 95%CI 2.0-2.9). After adjusting for sex and CLL-International Prognostic Index (CLL-IPI) score, full Ig suppression (HR 2.9, 95%CI 2.2-3.9) and partial Ig suppression (HR 1.6, 95% CI 1.3-2.0) were associated with shorter TTFT (model 1, c-stat 0.77); and negative ARD (HR 2.4, 95%CI 2.0-2.9) was associated with shorter TTFT (model 2, c-stat 0.76). Negative ARD identified an additional 9% of patients at higher risk for shorter TTFT beyond full Ig suppression. Conclusions: The depth of immunoparesis at diagnosis predicts TTFT and enhances risk stratification in newly diagnosed MBL/CLL. Citation Format: Yuan Yao, Kari G. Rabe, Eli Muchtar, Paul Hampel, Yucai Wang, Lindsey Roeker, Saad Kenderian, Amber Koehler, Catherine Wagner, Amy Behnken, Jose F. Leis, Mazie Tsang, Talal Hilal, Ricardo Daniel Parrondo, Susan M. Schwager, Min Shi, Curtis A. Hanson, Celine M. Vachon, Shaji Kunnathu Kumar, Esteban Braggio, Neil E. Kay, Susan Slager, Sameer A. Parikh. Depth of immunoparesis predicts time to first therapy in newly diagnosed monoclonal B-cell lymphocytosis and chronic lymphocytic leukemia [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 1191.

Abstract Background Artificial intelligence (AI)–based, mammography breast cancer risk prediction models show improved discriminatory accuracy relative to clinical risk models. However, data on their calibration are limited. This study compared model performance of 3 clinical breast cancer risk models: Gail, Tyrer–Cuzick v8, and Breast Cancer Surveillance Consortium v3 with the MIRAI AI-risk model. Methods Digital mammograms were ascertained from a screening mammography cohort of 12 308 women within the Mayo Clinic Biobank with 250 incident breast cancers (176 invasive) within 5 years. We predicted 5-year breast cancer risk, estimated discriminatory accuracy (C-index), and calibration (observed-to-expected ratio) of both overall and invasive breast cancer and compared estimates using bootstrapping approaches. Results MIRAI demonstrated similar or improved discriminatory accuracy of overall breast cancer (C-index = 0.71, 95% confidence interval [CI] = 0.68 to 0.74) and invasive breast cancer (C-index = 0.71, 95% CI = 0.67 to 0.75) compared with clinical models (overall breast cancer: C-index = 0.59-0.68; invasive breast cancer: C-index = 0.60-0.68). MIRAI’s calibration for risk of overall breast cancer (observed-to-expected ratio = 0.96, 95% CI = 0.85 to 1.08) was improved compared with Gail (observed-to-expected ratio = 1.22, 95% CI = 1.07 to 1.38) and Breast Cancer Surveillance Consortium (observed-to-expected ratio = 1.38, 95% CI = 1.22 to 1.56) but similar to TC with volumetric percent density and polygenic risk score (observed-to-expected ratio = 0.99, 95% CI = 0.87 to 1.13). However, for low-risk women (approximately 50%), MIRAI overestimated risk of overall breast cancer. MIRAI also overestimated risk of invasive breast cancer across the risk spectrum (observed-to-expected ratio = 0.68, 95% CI = 0.58 to 0.78), whereas clinical models had good calibration (observed-to-expected ratio = 0.86–0.99). Conclusion MIRAI demonstrated stronger discriminatory accuracy than clinical models for 5-year overall and invasive breast cancer risk prediction but overestimated risk for both breast cancer endpoints. AI-based risk models should consider discriminatory accuracy and calibration for invasive cancer before implementation.