Toward robust mammography-based models for breast cancer risk
Menée à partir de clichés mammographiques réalisés auprès de 70 972 patientes et à partir de données de registres médicaux américains, suédois et taïwanais, cette étude évalue la performance d'un algorithme d'apprentissage profond, utilisant des données mammographiques, pour identifier les patientes présentant un risque élevé de développer un cancer du sein dans les cinq ans
Mammograms are a common but imperfect way of assessing breast cancer risk. Current U.S. breast cancer screening guidelines all use a component of cancer risk assessment to inform clinical course. Yala et al. developed a machine learning model called “Mirai” to predict breast cancer risk based on traditional mammograms. The authors’ risk model performed better than Tyrer-Cuzick and previous deep learning models at identifying both 5-year breast cancer risk and high-risk patients across multiple international cohorts. Mirai also performed similarly across race and ethnicity categories, suggesting the potential for improvement in patient care across the board.Improved breast cancer risk models enable targeted screening strategies that achieve earlier detection and less screening harm than existing guidelines. To bring deep learning risk models to clinical practice, we need to further refine their accuracy, validate them across diverse populations, and demonstrate their potential to improve clinical workflows. We developed Mirai, a mammography-based deep learning model designed to predict risk at multiple timepoints, leverage potentially missing risk factor information, and produce predictions that are consistent across mammography machines. Mirai was trained on a large dataset from Massachusetts General Hospital (MGH) in the United States and tested on held-out test sets from MGH, Karolinska University Hospital in Sweden, and Chang Gung Memorial Hospital (CGMH) in Taiwan, obtaining C-indices of 0.76 (95% confidence interval, 0.74 to 0.80), 0.81 (0.79 to 0.82), and 0.79 (0.79 to 0.83), respectively. Mirai obtained significantly higher 5-year ROC AUCs than the Tyrer-Cuzick model (P < 0.001) and prior deep learning models Hybrid DL (P < 0.001) and Image-Only DL (P < 0.001), trained on the same dataset. Mirai more accurately identified high-risk patients than prior methods across all datasets. On the MGH test set, 41.5% (34.4 to 48.5) of patients who would develop cancer within 5 years were identified as high risk, compared with 36.1% (29.1 to 42.9) by Hybrid DL (P = 0.02) and 22.9% (15.9 to 29.6) by the Tyrer-Cuzick model (P < 0.001).
Science Translational Medicine , résumé, 2020