Abstract PS1-13-03: Achieving Disparity Reduction in Precision Oncology: A Population-Based Simulation Study

Abstract African-American women with breast cancer receive biomarker testing at lower rates than White women, despite similar overall frequencies of actionable genetic alterations—a gap that contributes to about a 40% higher mortality rate. AI-powered precision oncology platforms, particularly when paired with culturally tailored interventions, hold promise for reducing this inequity, though their race-specific impact has yet to be fully quantified. We modeled AI-enabled clinical decision support combined with equity-focused strategies in a population-based simulation integrating TCGA molecular data with SEER demographics (N=10,000). Baseline testing rates reflected recent literature (2020-2024): BRCA (AA 21.4% vs W 47.9%), PIK3CA (28.6% vs 50.5%), ESR1 (16.6% vs 29.0%). Intervention effects were derived from systematic reviews of patient navigation, cultural competency training, community health workers, and tailored educational materials. Realistic ceilings were applied (85% maximum testing, 70% uptake, 90% sustainability). Economic modeling included program costs ($580/patient). Culturally tailored AI interventions improved testing among African American women. Baseline testing rates showed significant post-intervention improvement: BRCA 21.4% to 32.7% (+53%), PIK3CA 28.6% to 58.6% (+105%), ESR1 16.6% to 42.5% (+156%). Disparity gaps narrowed by 14-50%, with African American women achieving 2-4x greater improvements than White women. The model projected prevention of 4,763 non-concordant treatments and 714 deaths, with African American women representing 21% of lives saved despite being 10.5% of the population. All interventions were highly cost-effective (p<0.0001). Among Black patients, modeled cost per QALY gained was $642- 156x more favorable than the $100,000/QALY benchmark, with results robust to +50% parameter variation. This equity-focused, AI-enabled framework demonstrates the potential to close up to half of existing biomarker testing gaps in breast cancer. By doing so, this approach may prevent thousands of non-concordant treatments and hundreds of deaths while achieving extraordinary cost-effectiveness far beyond accepted thresholds. Importantly, these findings reflect analyses only recently completed and provide new, timely evidence of how equity-focused AI interventions can address one of the most urgent challenges in breast cancer care, racial disparity in biomarker testing and patient outcomes. Citation Format: J. Zeineh, R. DeAngel, E. Grullon, T. J. Lawton, K. J. Bloom. Achieving Disparity Reduction in Precision Oncology: A Population-Based Simulation Study [abstract]. In: Proceedings of the San Antonio Breast Cancer Symposium 2025; 2025 Dec 9-12; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2026;32(4 Suppl):Abstract nr PS1-13-03.