Accurate risk assessment is essential for the success of population screening programs and early detection efforts in breast cancer. Mirai is a new deep learning model based on full resolution mammograms.

Mirai is 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 found to be significantly more accurate than the Tyrer-Cuzick model, a current clinical standard.

The primary aim of this study is to prospectively quantify the clinical benefit (i.e. MRI/CEM cancer detection rate) of Mirai-based guidelines and to compare them to the current standard of care.

1. Conduct a prospective study where patients who are identified as high risk by Mirai guidelines are invited to receive supplemental MRI within 12 months.
2. Compare cancer outcomes between patients only identified as high risk by Mirai and patients identified as high risk by existing guidelines The secondary aim is to study the impact of new guidelines by race and ethnicity, to ensure equitable improvements in cancer screening.

Inclusion Criteria:

• Women who were identified as high risk on the retrospective study (dating from 2017-2025) using MIRAI will be recruited and consented for the prospective study

• Women over 40 years of age identified as high risk according to traditional guidelines will also be potentially eligible for this study

• Following consent and enrollment in the study, a participant will subsequently receive the following:

  1. These patients will be invited to receive a supplemental MRI examination currently considered the most sensitive test for breast cancer detection.
  2. Any positive diagnosis on MRI will be followed by biopsy to confirm 'truth" of diagnosis.

• To be selected, a given record must include the following:

  1. A report of a routine screening mammogram or diagnostic mammogram, and availability of the DICOM images from that report with the PACS system.
  2. Reports of all follow up screening and diagnostic studies documented on PACS.
  3. Some may have interventional procedures (as long as all of these are done at one of Umass sites) and documentation of these biopsy results in the hospitals EHR.

Exclusion Criteria:

• Under age 40. Women under 40 years are not routinely xrayed with a mammogram.
• Xray breast cancer screening imaging study that has artifacts, corruption, or other image quality degradation.
• Pregnant patients because they do not routinely receive screening mammogram
• Adult male patients with breast cancer