Clinical likelihood models calibrated against observed obstructive coronary artery disease on computed tomography angiography
Authors' conclusion
CCTA-calibrated clinical likelihood models improve calibration and discrimination of observed obstructive CAD at CCTA.
Study context
- Population
- Symptomatic outpatients with suspected obstructive coronary artery disease (CAD)training cohort and several validation cohorts · training cohort 38,269 · validation cohorts 28,340
- Index test or model
- RF-CLCCTA and CACS-CLCCTA models vs Basic PTP model, against observed obstructive CAD at CCTA (>50% diameter stenosis)
- Target condition
- observed obstructive CAD at CCTA
Methodological detailsadvanced machine learning algorithm, training cohort, pooled validation cohorts
- Advanced machine learning algorithm
- Training cohort
- Pooled validation cohorts
Performance
As reported in the abstract
Basic PTP
0.71
area under the receiver operating curve · 95% CI 0.70–0.72
RF-CLCCTA
0.74
area under the receiver operating curve · 95% CI 0.73–0.75
CACS-CLCCTA
0.87
area under the receiver operating curve · 95% CI 0.86–0.87
Further findings
Calibration
- Basic PTP underestimated the prevalence of observed obstructive CAD at CCTA; RF-CLCCTA and CACS-CLCCTA models showed superior calibration
Observed obstructive CAD at CCTA in the pooled validation cohort
- Patients with observed obstructive CAD at CCTA
- 6443 (22.7%)
Original publication
Clinical likelihood models calibrated against observed obstructive coronary artery disease on computed tomography angiography. Eur Heart J Cardiovasc Imaging. Published 7 February 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →