Related Experiment Video
Updated: Aug 6, 2026
Implementation of a Real-Time Psychosis Risk Detection and Alerting System Based on Electronic Health Records using CogStack
Published on: May 15, 2020
Development of a MACE risk prediction model based on CCTA-derived quantitative parameters: a proof-of-concept study
Tianyang Gao1, Mingyu Zou1, Wei Zhou1
1Department of Radiology, General Hospital of Northern Theater Command, Shenyang, China.
Background:
This study aimed to identify factors influencing major adverse cardiovascular events (MACE) in patients with coronary artery disease (CAD) using quantitative parameters derived from coronary computed tomography angiography (CCTA), and to develop a nomogram-based risk prediction model.
Methods:
Clinical data from 280 CAD patients (May 2020-December 2023) were retrospectively analyzed. Based on 1-year follow-up, patients were divided into MACE and non-MACE groups. Baseline characteristics and CCTA-derived parameters were compared, and independent predictors were identified via multivariable logistic regression. A nomogram was constructed and internally validated using Bootstrap resampling, followed by external validation in an independent cohort of 288 patients.
Results:
Significant intergroup differences were observed in age, cardiac function grade, smoking history, hypertension history, CT-FFR, stenosis degree, plaque length, total vessel volume, minimal lumen area (MLA), fibrous and fibrofatty plaque volumes, plaque burden (PB), coronary artery calcium score (CACS), perivascular fat attenuation index (FAI), and myocardial mass/volume ratio (MAS) (all P < 0.05). Multivariable analysis identified severe stenosis, longer plaque length, larger fibrous plaque volume, and elevated PB, CACS, FAI, and MAS as independent risk factors, while larger MLA was protective (P < 0.05). ROC analysis yielded AUCs of 0.735, 0.823, 0.747, 0.704, 0.691, 0.808, 0.810, and 0.689 for stenosis, plaque length, MLA, fibrous volume, PB, CACS, FAI, and MAS, respectively. The combined model achieved an AUC of 0.936, significantly outperforming each single indicator. Internal calibration curves closely matched the ideal line, demonstrating good predictive performance. External validation confirmed robust discrimination (AUC = 0.932) and excellent calibration (Hosmer-Lemeshow P = 0.382).
Conclusions:
The nomogram based on CCTA-derived quantitative parameters demonstrates high predictive accuracy for MACE in CAD patients and may serve as a reliable clinical screening tool. However, given the short follow-up duration and predominance of soft endpoints (e.g., rehospitalization), these results should be interpreted cautiously regarding hard outcomes, and longer-term prospective studies are warranted.