AI-Quantitative CT Coronary Plaque Features Indicate Higher Relative CV Risk in Women: CONFIRM2 Registry

Avatar
By Joseph Kim on

Key Takeaways:

  1. AI-based quantitative coronary plaque assessments significantly improve risk prediction for cardiovascular events in women compared to traditional clinical risk scores.
  2. Although women had lower absolute plaque burden, the same incremental increase in plaque conferred substantially higher relative risk for adverse cardiovascular outcomes compared to men, highlighting the importance of sex-specific risk evaluation and potential intensification of preventive therapy in women.

Artificial intelligence-based quantitative coronary computed tomography (AI-QCT) plaque features predict cardiovascular risk with greater accuracy in women compared to men, according to the international multicenter CONFIRM2 registry. Results presented at the American College of Cardiology’s Annual Scientific Session (ACC.25) and simultaneously published in Circulation: Cardiovascular Imaging demonstrate that women exhibit a significantly higher relative risk of major adverse cardiovascular events (MACE) despite having a lower absolute burden of coronary artery disease (CAD) compared with men.

The CONFIRM2 registry analyzed data from 3,551 symptomatic patients (49.5% women, mean age 59±12 years) undergoing coronary CT angiography for suspected CAD. Over a follow-up period averaging 4.8±2.2 years, women had significantly lower absolute MACE rates than men (3.2% vs. 6.1%, p<0.001). Despite this lower overall event rate, AI-QCT-derived plaque metrics including total plaque volume (TPV), noncalcified plaque (NCP), calcified plaque (CP), and percentage atheroma volume (PAV) conferred a notably higher relative risk in women compared to men, even after adjustment for age and cardiovascular risk factors. Specifically, every 50mm³ increase in TPV was associated with a 17.7% higher risk in women versus only 5.3% in men (p-interaction<0.001). Similarly, each 50mm³ increase in NCP conferred a 27.1% increase in risk among women compared to 11.6% in men (p-interaction=0.0015). Calcified plaque increases were also associated with greater risk in women (22.9% vs. 5.4%, p-interaction=0.0012).

Importantly, traditional risk stratification models, such as the Diamond Forrester score, exhibited limited predictive accuracy, particularly in women. Integration of AI-QCT-derived plaque features significantly enhanced the predictive accuracy for MACE, achieving an area under the curve (AUC) of up to 0.797 in women, significantly improving upon conventional cardiovascular risk factors alone.

Lead investigator Dr. Gudrun M. Feuchtner emphasized the clinical implications: “Using AI-QCT CAD feature-based risk stratification for MACE instead of relying on traditional risk scores as the potential to enhance the precision of cardiovascular risk stratification in women and should prompt more aggressive anti-atherosclerotic therapy and preventative interventions.”