Purpose: Recent United States Preventive Services Task Force (USPSTF) guidelines highlight a concerning disparity in breast cancer outcomes among racial groups and a lack of data about the effectiveness of breast cancer screening for certain subgroups of women, such as African American women. The purpose of this study is to compare the performance metrics of screening mammography among Caucasian, African American, and Asian women at a large academic medical center.
Materials and Methods: In this IRB-approved and HIPAA-compliant study, screening mammograms obtained from 2007 to 2019 at an academic medical center were retrospectively reviewed. All mammograms obtained from 2013 onward included digital breast tomosynthesis. Each mammogram was classified as a true-positive, true-negative, false-positive, or false-negative examination based on one-year follow-up data. Performance metrics were calculated according to standard definitions in the 5th Edition of the BI-RADS Atlas and compared among the Caucasian, African American, and Asian populations using logistic regression models.
Results: Over the study period, 491,190 mammograms were obtained in 103,782 women. A total of 441,161 mammograms were obtained in Caucasian women (mean age 59 +/- 12), 24,505 mammograms in African American women (mean age 57 +/- 11), and 25,524 mammograms in Asian women (mean age 55 +/- 11). The cancer detection rates were 5.0, 4.1, and 3.8 per 1,000 examinations in the Caucasian, African American, and Asian populations, respectively (p < 0.01). Positive predictive value 1 was also highest in the Caucasian population (7.8% versus 6.2% in African American population versus 6.3% in Asian population, p< 0.01). No significant differences in sensitivities were observed (p=0.08). Abnormal interpretation rate was highest in the African American population (6.7% versus 6.4% in Caucasian population versus 6.1% in Asian population, p=0.04), and specificity was lowest in the African American population (93.7% versus 94.0% in Caucasian population versus 94.2% in Asian population, p=0.04).
Conclusion: This comprehensive analysis of screening mammography performance metrics based on race reveals disparities in key performance metrics, including cancer detection rates, positive predictive values, abnormal interpretation rates, and specificity, among Caucasian, African American, and Asian women. These results underscore the need for further research and targeted interventions to understand and address disparities in mammography outcomes among different racial populations.
Clinical Relevance Statement: Understanding racial differences in screening mammography performance metrics is crucial for designing effective breast cancer screening strategies. These findings emphasize the importance of implementing targeted interventions to improve mammography outcomes, particularly among populations experiencing lower cancer detection rates and lower specificity.
