AI quantification of breast arterial calcifications (BACs) on mammograms enhances prediction of cardiovascular events beyond traditional risk models.
Key Details
- 1AI model (cmAngio, CureMetrix) used on 21,514 women aged 40+ from U.S. and Australia mammography data.
- 222.7% of women had BAC; prevalence rose to 61% for those over 70.
- 3Each 10-percentile BAC increase correlated to a 17% higher adjusted risk for major adverse cardiovascular events (MACE).
- 4Integration of BAC percentiles improved net risk reclassification by 5% and C-statistic from 0.67 to 0.71 (p = 0.04).
- 5Low and intermediate ASCVD risk women showed significantly increased MACE rates when BAC was above or below the median.
- 6Authors developed an online tool to aid clinical decision-making based on BAC percentiles.
Why It Matters
This study supports the use of mammogram-derived BAC scoring, automated by AI, to better stratify cardiovascular risk in women—potentially enabling earlier, more effective preventive care using routinely available imaging.
Source
AuntMinnie
Related News
•AuntMinnie
AI Models Show High Sensitivity but Moderate Specificity for Lung Nodule Classification on CT
AI shows high sensitivity but only moderate specificity in classifying lung cancer nodules on CT, indicating a role as a rule-out adjunct.
•HealthExec
Survey Reveals Top 6 Concerns About Healthcare AI for 2026
A new survey highlights six main concerns clinicians and patients have about healthcare AI in 2026, including bias, governance, deskilling, hallucinations, accountability, and source validation.
•Radiology Business
AI Triage Reduces Mammography Screening Workloads by 77%
AI used as a first reader in breast cancer screening can reduce radiologist workloads by 77%.