Culturally tailored AI interventions increased biomarker testing in African-American women by 53-156%, narrowing disparities by 14-50% and preventing 714 deaths.
Does AI-enabled clinical decision support combined with equity-focused strategies improve biomarker testing rates and reduce disparities in African-American women with breast cancer?
An equity-focused, AI-enabled framework demonstrates the potential to close up to half of existing biomarker testing gaps in breast cancer, preventing non-concordant treatments and deaths while being highly cost-effective.
Absolute Event Rate: 0% vs 0%
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 (p0. 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.
Zeineh et al. (Tue,) reported a other. Culturally tailored AI interventions increased biomarker testing in African-American women by 53-156%, narrowing disparities by 14-50% and preventing 714 deaths.