Integrating pericoronary adipose tissue radiomics features with CCTA-derived functional parameters significantly improved the prediction of major adverse cardiovascular events in patients with coronary atherosclerosis, increasing the AUC from 0.737 to 0.803.
Cohort (n=171)
No
Does a combined model integrating PCAT radiomics features and CCTA-derived functional parameters improve the prediction of MACE in patients with coronary atherosclerosis compared to CCTA-derived functional parameters alone?
Integrating perivascular coronary adipose tissue radiomics with CCTA-derived functional parameters significantly enhances the predictive accuracy for major adverse cardiovascular events in patients with coronary atherosclerosis.
Absolute Event Rate: 0.803% vs 0.737%
p-value: p=0.002
Objective To explore the predictive value of a combined model integrating perivascular coronary adipose tissue (PCAT) radiomics features and coronary computed tomography angiography (CCTA)-derived functional parameters for major adverse cardiovascular events (MACE) in patients with Coronary Atherosclerosis (CAS). Methods This retrospective study enrolled 171 CAS patients who underwent CCTA at Datong Third People's Hospital between November 2020 and September 2022 and stratified them into a MACE-positive group ( n = 72) and a MACE-negative group ( n = 99) based on the occurrence of MACE. Using support vector machine (SVM) and Gaussian process regression (GPR) algorithms, we constructed four MACE prediction models: two models relying on CCTA-derived functional parameters (stenosis severity and CT-FFR), and two combined models integrating these parameters with the radiomics score (Rad-score). Model performance was evaluated using the area under the receiver operating characteristic curve (AUC), F1-score, Delong test, calibration curves, and decision curve analysis (DCA). Results Among the CCTA-derived functional parameters, CT-derived fractional flow reserve (CT-FFR) and coronary stenosis severity emerged as independent predictors of MACE in patients with CAS (both P 0.05). Models integrating CCTA-derived functional parameters with the radiomics score (Rad-score) demonstrated superior predictive performance compared with models relying solely on CCTA-derived functional parameters. Specifically, the mean AUC for SVM and GPR models based exclusively on CCTA-derived functional parameters were 0.742 and 0.737, respectively. In contrast, the mean AUCs for the corresponding combined SVM and GPR models both increased to 0.803. Notably, the combined GPR model achieved the highest mean F1-score (0.686). The DeLong test confirmed that the combined models significantly outperformed the CCTA-only models in both the training and testing sets (all P 0.05). Calibration curves revealed the best goodness-of-fit for the combined GPR model, and DCA indicated that this model provided the greatest net clinical benefit across a broad range of decision thresholds. Conclusion PCAT radiomics features can enhance the predictive performance of models based on CCTA-derived functional parameters for MACE in CAS patients. Notably, the combined GPR model exhibits optimal predictive accuracy and clinical utility.
Wang et al. (2026) conducted a cohort in Coronary Atherosclerosis (CAS) (n=171). Combined model integrating PCAT radiomics features and CCTA-derived functional parameters vs. Model relying solely on CCTA-derived functional parameters was evaluated on Area under the receiver operating characteristic curve (AUC) for predicting MACE (95% CI 0.722-0.887, p=0.002). Integrating pericoronary adipose tissue radiomics features with CCTA-derived functional parameters significantly improved the prediction of major adverse cardiovascular events in patients with coronary atherosclerosis, increasing the AUC from 0.737 to 0.803.