Abstract 2277: Validation and context-dependent effects of a prostate cancer polygenic risk score in the All of Us Research Platform

Background: Polygenic risk scores (PRS) are valuable tools for assessing genetic susceptibility to prostate and other cancers. The All of Us (AoU) Research Program, with its diverse and inclusive cohort, provides a unique opportunity to validate and further investigate context-dependent associations of cancer PRS, offering insights into their applicability in diverse populations. Methods: A previously established multi-ancestry PRS of 451 genetic variants associated with prostate cancer risk was calculated in AoU. Logistic regression models were employed to evaluate associations between the standardized continuous PRS and prostate cancer risk, adjusting for age and the top 10 principal components to account for population stratification. Comparisons were also made between individuals in the top PRS decile (90-100%) and those in the average range (40-60%). Analyses were stratified by genetic ancestry groups, age, BMI, and smoking status explored potential demographic and lifestyle factors affecting PRS predictive power. Results: PRS validation included 4,668 prostate cancer cases and 71,644 controls across six genetic ancestry groups defined by AoU. Odds ratios (ORs) per unit increase in PRS were 2.23 (95% CI = 2.14-2.33) in European, 1.83 (95% CI = 1.67-2.00) in African, 2.16 (95% CI = 1.86-2.51) in American, 2.67 (95% CI = 1.53-4.67) in East Asian, 2.46 (95% CI = 1.14-5.33) in Middle Eastern, and 1.67 (95% CI = 1.00-2.77) in South Asian populations. Individuals in the top PRS decile (90-100% of the PRS) had ORs of 3.97 (95% CI = 3.42-4.61) in European, 3.16 (95% CI = 2.41-4.14) in African, and 3.51 (95% CI = 2.31-5.36) in American populations compared to the average 40-60% PRS category. Age-stratified analyses showed diminishing PRS effect sizes with advancing age, with ORs declining from 2.44 (≤55 years, 95% CI = 2.10-2.83) to 2.05 (>70 years, 95% CI = 1.91-2.20) in European, and from 1.92 (≤55 years, 95% CI = 1.57-2.36) to 1.48 (>70 years, 95% CI = 1.19-1.85) in African populations. Moreover, BMI-stratified analyses showed slightly increased predictive performance in higher BMI categories, with ORs of 1.87 (95% CI = 1.22-2.88) in underweight, 2.06 (95% CI = 1.91-2.23) in normal weight, 2.12 (95% CI = 2.00-2.24) in overweight, and 2.26 (95% CI = 2.12-2.41) in obese individuals. Smoking stratified analyses demonstrated slightly higher PRS performance in never smokers, with ORs of 1.90 (95% CI = 1.68-2.16) in current smokers and 2.16 (95% CI = 2.03-2.29) in never smokers. Conclusion: This study validates the multi-ancestry PRS as an effective tool for stratifying prostate cancer risk across diverse populations. The findings emphasize the importance of ancestry-, age-, and lifestyle-specific interpretations in PRS applications. Analyses are underway for breast, colorectal, and lung cancers to evaluate the broader utility of PRS for cancer risk stratification and screening strategies within AoU. Citation Format: Shuyan Cheng, Austin Hammermeister Suger, Louisa B. Goss, Boya Guo, Sara Lindström, Burcu F. Darst. Validation and context-dependent effects of a prostate cancer polygenic risk score in the All of Us Research Platform [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 2277.