Computed Tomography Radiomics Model Predicts Procedure Success of Coronary Chronic Total Occlusions
Authors' conclusion
CT radiomics-based model outperformed CT-derived Multicenter CTO Registry of Japan score for predicting PCI success.
Study context
- Population
- Patients with coronary chronic total occlusion (CTO) undergoing percutaneous coronary intervention (PCI)training 202, internal validation 98, external test 75
- Comparison
- Computed tomography (CT) radiomics model vs CT-derived J-CTO score for predicting PCI success
- Outcome
- Prediction of PCI (revascularization) success
Methodological detailsretrospective, 2 tertiary hospitals, external test set
- Retrospective
- 2 tertiary hospitals
- External test set
Performance
As reported in the abstract
Prediction of PCI success (external test set, 83 CTO lesions)
| Test or model | AUC | Accurately identified CTO lesions with procedure success |
|---|---|---|
| CT radiomics model | 0.920 | 89.16% (74/83) |
| CT-derived J-CTO score | 0.752 |
AUC
CT radiomics model0.920
CT-derived J-CTO score0.752
Accurately identified CTO lesions with procedure success
CT radiomics model89.16% (74/83)
CT-derived J-CTO score— not stated in the abstract
- CT radiomics model vs CT-derived J-CTO score: P = 0.008 — 0.920 vs 0.752
Further findings
Occlusion length, PCI success vs PCI failure (external test set)
P = 0.007
Tortuous course, PCI success vs PCI failure (external test set)
P = 0.004
| Outcome | PCI success | PCI failure | P value |
|---|---|---|---|
| Radiomics score, PCI success vs PCI failure (external test set) | 0.10 | 0.55 | P < 0.001 |
Original publication
Computed Tomography Radiomics Model Predicts Procedure Success of Coronary Chronic Total Occlusions. Circ Cardiovasc Imaging. Published 21 February 2023. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →