"Abstract

Background

The Gail Model, a breast cancer risk calculator initially developed using data from primarily White women, has undergone recalibration using more diverse populations. However, concerns remain that risk is underestimated for certain racial groups.

Methods

This retrospective study evaluated the Gail Model’s performance across different racial groups. Electronic medical record data from 31,256 women aged 25–75 who underwent mammograms between July 2020 and June 2023 were collected. The cohort included 11,589 Black/African American, 11,008 White, 5,872 Hispanic/Latino, 1,429 Asian, and 1,358 individuals identifying as other.

Results

Significant differences in high-risk (HR) classification based on Gail scores were observed across racial groups. High-risk classification was highest among White women at 11.4%. Black compared with White women had an odds ratio (OR) 0.23 (95% confidence interval [CI] = 0.2–0.28) of HR classification. This difference persisted when controlling for all other variables. Breast cancer diagnosis rates were similar between Black (1.57%) and White women (1.6%).

Conclusions

These findings indicate a racial disparity in HR classification using the Gail Model. Black women are less likely to be identified as HR despite having comparable rates of cancer to White women. These results highlight the potential limitations of risk assessment algorithms incorporating race as a variable."