Machine Learning Model for Atherosclerosis Evaluation and Cardiovascular Risk Prediction Based on Coronary CT Angiography-Analysis From the CREATION Registry
Performance · MACE prediction (AUC)Result — go to the resultAUC 0.903 versus 0.830Authors' conclusion
The machine learning model incorporating CCTA plaque quantification, characterization, and stenosis assessment significantly enhanced the predictive capacity for major adverse cardiac events.
Study context
- Population
- Patients with suspected CAD undergoing coronary CT angiography at Fuwai Hospital, 2016–2019training cohort n=5901
- Participants
- 8,431 patients
- Comparison
- XGBoost coronary computed tomography angiography (CCTA) model vs clinical risk factors and CAC score model
- Outcome
- MACE:
- All-cause death
- Acute MI
- Coronary revascularization
- Stroke
- Follow-up
- Median 3.68 years
Methodological detailsregistry, machine learning, training and testing cohorts, single-centre
- Registry
- Machine learning
- Training and testing cohorts
- Single-centre
Performance
MACE prediction (AUC)
AUC
0.903 versus 0.830
P < 0.001
As reported in the abstract
Further findings
| Outcome | Adjusted HR (95% CI) | P value | |
|---|---|---|---|
| Lipid plaque burden and MACE | 2.524 (2.157–2.996) | P < 0.001 | |
| 0.20.5125Adjusted hazard ratio (log scale) |
Subgroups (sex, onset age, symptoms)
- Consistent predictive performance across sex, onset-age and symptom subgroups
Incremental value over 1- to 5-year follow-up; secondary composite (death, MI or stroke)
- Incremental value consistent across time points; findings unchanged for secondary composite end point
Parts of this Brief are withheld: the automatic checks could not confirm them against the abstract.
Original publication
Machine Learning Model for Atherosclerosis Evaluation and Cardiovascular Risk Prediction Based on Coronary CT Angiography-Analysis From the CREATION Registry. Circ Cardiovasc Imaging. Published 21 November 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →