Generalizability of an AI-based mammogram risk score (MRS) for breast cancer among diverse populations of women
Clinical Snapshot
PICO Framework
| P — Population | Women undergoing mammographic screening from two North American cohorts (>226,000 women) including non-Hispanic white, non-Hispanic Black, East Asian, South Asian, and Indigenous women |
| I — Intervention | AI-derived mammogram risk score (MRS) based on texture features |
| C — Comparator | Risk assessment across different racial and ethnic subgroups |
| O — Outcomes | Breast cancer incidence, hazard ratios for cancer development, calibration across populations |
Bottom Line
This large prognostic cohort study demonstrates that an AI-derived mammogram risk score shows consistent breast cancer prediction across diverse racial and ethnic groups, with hazard ratios of 2.24-2.32 per standard deviation. Unlike conventional risk models that underperform in non-European populations, this texture-based approach showed excellent calibration across all ethnic subgroups in over 226,000 women. The findings suggest MRS could address health equity concerns in breast cancer screening by providing equally accurate risk stratification regardless of ethnicity. However, the study lacks detail on follow-up duration, comprehensive confounder adjustment, and direct comparison with existing models. While promising for clinical implementation, particularly in diverse healthcare systems, further validation studies with longer follow-up and head-to-head comparisons with current risk assessment tools are needed before widespread adoption.
Key Findings
P Value: Statistically significant (p-values not specified)
Effect Size: Hazard ratios per standard deviation: 2.24 to 2.32
Primary Outcome: Breast cancer incidence across racial/ethnic groups
Nnt Or Sensitivity: Hazard ratio 2.24-2.32 per SD indicates strong prognostic discrimination
Confidence Interval: 95% CI: 2.03-2.46 and 2.25-2.39
Clinical Application
Requires AI infrastructure and standardized mammogram processing, but builds on existing screening programs Highly relevant for Australia's diverse population including Indigenous Australians and immigrant communities. Could enhance BreastScreen Australia program by providing more equitable risk stratification across ethnic groups Women undergoing mammographic screening, particularly valuable for diverse racial and ethnic populations where conventional models underperform
Abstract
Conventional prediction models incorporating genetic and clinical factors including breast density underperform in non-European populations. We investigated an artificial intelligence-derived mammogram risk score (MRS), a summary of texture features that captures intrinsic breast-tissue characteristics-the substrate for cancer development. This study leveraged data from two North American screening cohorts totaling >226,000 women, including non-Hispanic white, non-Hispanic Black, East Asian, South Asian, and Indigenous women. MRS distributions showed nonsignificant shifts (and similar SDs) between cohorts and across race and ethnic subgroups. MRS increased with age and was significantly associated with breast cancer risk, with hazard ratios per SD ranging from 2.24 [95% confidence interval (CI), 2.03 to 2.46] to 2.32 (95% CI, 2.25 to 2.39) after age adjustment. Associations remained significant within all subgroups. Calibration was excellent across the racial and ethnic groups and across full-field digital mammograms and tomosynthesis. These findings establish MRS as a strong predictor that is independent of race or ethnicity, demonstrating its potential for broader clinical utility.
References
- 1.Jiang, S., Bennett, D. L., Kraft, P., & Colditz, G. A. (2026). Generalizability of an AI-based mammogram risk score (MRS) for breast cancer among diverse populations of women. Science Advances. https://doi.org/10.1101/2021.11.16.21266374
Related Research
The American journal of pathology
An Integrated Machine Learning and Genomic Framework for Precise Detection of Gastric Cancer
3 Aug 2026
Current opinion in chemical biology
Prediction of protein-protein interactions and co-complex models with deep learning
3 Aug 2026
European journal of radiology
A CT-based deep learning model to differentiate between benign and malignant adrenal lesions
3 Aug 2026
This content is for educational purposes for healthcare professionals only and does not constitute clinical advice. Clinical decisions should be based on individual patient assessment, current guidelines, and appropriate specialist consultation. Editorial Standards · Privacy Policy · Terms of Service