Prevention & lipidsCohort study
Artificial intelligence–based quantification of breast arterial calcifications to predict cardiovascular morbidity and mortality
- Retrospective cohort study
- multi-institutional
- two healthcare systems (internal and external cohorts)
Authors' conclusion
Automatically quantified BAC is an independent predictor of MACE and mortality, adding prognostic value to the PREVENT score.
Study context
- Population
- Women with screening mammograms, racially diverse, 2 healthcare systems; Breast arterial calcification (BAC) quantified by AI (transformer-based neural network)
- Participants
- 123,762 women
- Exposure
- Mild, moderate and severe BAC vs zero BAC on screening mammography
- Outcome
- MACE
Main finding
As reported in the abstract
Any MACE (PREVENT-adjusted)
Each group vs zero BACHazard ratio (95% CI)
Mild BACinternal cohort
1.32(1.10–1.59)
Mild BACexternal cohort
1.28(1.17–1.39)
Moderate BACinternal cohort
1.75(1.23–2.50)
Moderate BACexternal cohort
1.79(1.55–2.06)
Severe BACinternal cohort
3.29(2.15–5.05)
Severe BACexternal cohort
2.80(2.36–3.32)
- Each 1 mm² increase in BAC: additional 2%-3% risk for major adverse cardiovascular events (MACE)
- P < 0.001
Observational data: associations, not effects.
Further findings
BAC prevalence
- BAC detected, internal cohort
- 16.1%
- BAC detected, external cohort
- 20.6%
Original publication
Artificial intelligence–based quantification of breast arterial calcifications to predict cardiovascular morbidity and mortality. Eur Heart J. Published 9 March 2026. doi:10.1093/eurheartj/ehag128
Summarised from the published abstract; every number is checked against it