Medical-AI trial: catching breast cancer opportunistically on chest CT with AI — compared with mammography and MRI (NCT07557654)

1The breast tissue already captured on chest CT

Breast-cancer screening usually means mammography, but uptake and eligibility have limits. Meanwhile many people undergo chest CT for lung conditions, health checks, or other symptoms, and the breast is captured in those images. If AI could pick up signs of breast cancer opportunistically from such CTs done for other reasons, it might catch people who slipped through the screening net, without added radiation or visits. This is the idea of opportunistic screening.

2Opportunistic screening from existing scans

This study evaluates opportunistic breast-cancer screening from chest CT, as an observational study.

Per the registry summary, it evaluates the feasibility and effectiveness of using non-contrast chest CT for breast-cancer screening and compares its diagnostic performance with conventional mammography and/or breast MRI. The primary outcome is the screening performance of chest CT for detecting breast cancer, including comparison with mammography and/or MRI. The size is large at 5,000, and it is recruiting.

3Using images already taken

Opportunistic screening does not add an examination. It uses information already lying in images taken for another purpose.

1. 1A CT is taken for another purposeChest CT for lung disease, a health check, or another symptom
2. 2The breast appears in itThe image includes breast tissue
3. 3AI picks up the signSigns of breast cancer may be detected without extra dose or an extra visit
4. 4Those missed by screening are caughtDiagnostic performance is compared against mammography and breast MRI

Mammography is the basis of breast screening, while uptake and eligibility are limited. CT captures the breast in three dimensions and may carry information mammography does not. Non-contrast CT was never designed for breast screening, however, so its performance and the effect of false positives and negatives require careful validation. The study covers 5,000 cases and is recruiting.