Key result
CTA-derived myocardial radiomics improve clinical risk models for predicting future MI.
Why the study?
The prognostic value of myocardial radiomic phenotypes beyond coronary artery findings on cardiac CTA was unknown for predicting long-term myocardial infarction.
Does the incorporation of AI-derived myocardial radiomic phenotypes from cardiac CTA improve the prediction of fatal or nonfatal myocardial infarction compared to standard clinical models?
Observational (n=1,736)
Does the incorporation of AI-derived myocardial radiomic phenotypes from cardiac CTA improve the prediction of fatal or nonfatal myocardial infarction compared to standard clinical models?
Answers are grounded in this paper and cite their sources.
AI-derived myocardial radiomic phenotypes from cardiac CTA provide incremental prognostic value for predicting long-term risk of myocardial infarction beyond traditional clinical and imaging risk factors.
Background Cardiac computed tomography angiography (CTA) is widely used to assess coronary artery disease, but the prognostic value of myocardial radiomic phenotypes beyond coronary artery findings remains unknown. Objectives To investigate whether myocardial radiomic phenotypes derived from cardiac CTA provide incremental prognostic value for fatal or nonfatal myocardial infarction (MI) during long-term follow-up. Methods Post hoc analysis of CTA was performed within Scottish Computed Tomography of the Heart (SCOT-HEART) trial. A multivariable linear regression model was built to assess associations between myocardial radiomic phenotypes and cardiovascular risk factors, including hypertension, diabetes, and calcium score. Univariable and multivariable Cox proportional hazards models were used to determine the associations between myocardial radiomic phenotypes and MI. The prognostic model incorporating myocardial radiomic phenotypes was compared to clinical models based on cardiovascular risk factors, calcium score, and presence of obstructive disease to determine the improvement in predictive performance. Results Over a median follow-up of 8.6 years, 82 of 1736 participants (4.7%) experienced a fatal or nonfatal MI. Myocardial radiomic phenotypes were associated with age, sex and hypertension, and diabetes (P<0.01 for all), with older age, male sex, and cardiometabolic risk linked to increased myocardial volume and density. Among myocardial radiomic features associated with cardiovascular risk factors, univariable analysis identified 22 features associated with MI. Incorporation of myocardial radiomic phenotypes into a clinical model (including cardiovascular risk score, CAC score and presence of obstructive disease) improved discrimination for future MI events (C-index from 0.693 to 0.722, Δ=0.029, P<0.001) and improved model fit (likelihood ratio test, χ2=13.5, P=0.009). Conclusion Myocardial radiomic phenotypes based on CTA may offer incremental prognostic value for long-term myocardial infarction prediction, beyond conventional established cardiovascular risk factors and imaging features. Trial Registration: ClinicalTrials.gov Identifier: NCT01149590
No takes yet. Share an insight, caveat, or question.
Chen et al. (2026) conducted an observational in Coronary artery disease (n=1,736). Myocardial radiomic phenotypes derived from cardiac CTA vs. Clinical models based on cardiovascular risk factors, calcium score, and presence of obstructive disease was evaluated on Fatal or nonfatal myocardial infarction (MI) (ΔC-index 0.029, p=<0.001). Incorporating CTA-derived myocardial radiomic phenotypes into a clinical risk model improved discrimination for future fatal or nonfatal myocardial infarction.
Synapse has enriched 10 related papers. Consider them for comparative context: