{"entity":{"id":"mirai","kind":"technology","name":"Mirai (MIT breast cancer risk from mammograms)","aka":[],"tldr":"Reads a mammogram to estimate five-year breast cancer risk, consistently across races and devices.","summary":"Mirai is a deep-learning model from MIT that estimates five-year breast cancer risk from a standard mammogram, using device-conditional adversarial training so that its predictions do not shift between mammography machines. The Science Translational Medicine 2021 paper reported validation across seven hospitals in several countries, with consistent performance across races and devices, and the model is used in risk-adapted screening trials to decide who might need supplemental imaging or shorter intervals. It is intended for screening programmes and researchers rather than for diagnosis. Its prospective impact on cancer detection and outcomes is still under study, so it complements rather than replaces established risk models. For a newcomer: Mirai looks at a routine mammogram and estimates how likely breast cancer is over the next five years.","status":"emerging","asOf":"2026-09-08","links":[{"label":"Sci Transl Med 2021","url":"https://doi.org/10.1126/scitranslmed.aba4373"}],"tags":["risk-model","radiology"],"related":[],"cancers":["breast-hr-positive","tnbc"],"sections":["ai-computation","early-detection"],"technologies":["radiology-ai-screening","mammography"],"targets":[],"drugs":[],"companies":[],"institutions":["mgh"],"pathways":[],"terms":[],"trials":[],"people":[],"bottlenecks":[],"keyPapers":["paper-yala-sci-transl-med"],"journals":[],"dependsOn":[],"notes":[],"principle":"Mirai applies deep learning to mammograms with device-conditional adversarial training.","strengths":["Cross-site consistency"],"limitations":["Prospective impact still under study"],"since":2021},"route":"/technologies/mirai/","neighbours":{"cancer":[{"id":"breast-hr-positive","kind":"cancer","name":"HR-positive / HER2-negative breast cancer","route":"/cancers/breast-hr-positive/"},{"id":"tnbc","kind":"cancer","name":"Triple-negative breast cancer (TNBC)","route":"/cancers/tnbc/"}],"section":[{"id":"ai-computation","kind":"section","name":"AI & Computation","route":"/fronts/ai-computation/"},{"id":"early-detection","kind":"section","name":"Early Detection & Screening","route":"/fronts/early-detection/"}],"technology":[{"id":"radiology-ai-screening","kind":"technology","name":"AI in radiology","route":"/technologies/radiology-ai-screening/"},{"id":"mammography","kind":"technology","name":"Mammography & tomosynthesis","route":"/technologies/mammography/"}],"institution":[{"id":"mgh","kind":"institution","name":"Massachusetts General Hospital Cancer Center","route":"/institutions/mgh/"}],"paper":[{"id":"paper-yala-sci-transl-med","kind":"paper","name":"Toward robust mammography-based models for breast cancer risk","route":"/key-papers/paper-yala-sci-transl-med/"}],"roadmap":[{"id":"ai-oncology-roadmap","kind":"roadmap","name":"AI in oncology roadmap: pattern readers → foundation models → agents in the workflow","route":"/roadmaps/ai-oncology-roadmap/"},{"id":"ai-oncology-clinic","kind":"roadmap","name":"AI in the oncology clinic: from narrow cleared tools to multimodal decision support","route":"/roadmaps/ai-oncology-clinic/"}]}}