Coronary disease & ACSDiagnostic or prognostic study
Myocardial Scar Assessment Using Artificial Intelligence-Powered Contrast-Free MRI
- Diagnostic accuracy study
- prospective
- multicentre
- blinded
Authors' conclusion
Accurate myocardial scar quantification can be achieved without contrast agents in high-quality and clear VNE images. VNE may triage the need for LGE and obviate its use in more than two-thirds of referrals for chronic myocardial infarction scar assessment without compromising diagnostic accuracy.
Study context
- Population
- Patients referred for cardiovascular magnetic resonance (CMR) myocardial scar assessment at 2 sites (Leeds and Fuwai)
- Comparison
- AI virtual native enhancement vs late gadolinium enhancement (LGE)
- Outcome
- Diagnostic performance of VNE in myocardial scar detection
Performance
As reported in the abstract
Diagnosis of MI
| Test or model | Accuracy, confident images (n = 107) | Accuracy, all images (n = 136) |
|---|---|---|
| Virtual native enhancement (VNE) | 94.4% | 87.5% |
Further findings
| Outcome | mean difference (%) (95% CI) | |
|---|---|---|
| Myocardial scar quantification, VNE vs LGE · correlation R = 0.90 | 3.2 (−10.4% to +16.8%) | |
| −20−1001020Difference, % (linear scale) |
| Outcome | VNE | LGE | Estimate |
|---|---|---|---|
| Clinical readers (4): VNE adequate without LGE | |||
| Patients VNE adequate | 69.7% | Not reported in the abstract | |
| Average diagnostic accuracy in these cases | 93.7% | Not reported in the abstract | |
| Diagnostic accuracy in these cases | Not reported in the abstract | 93.9% |
Agreement on infarcted myocardial territories
- VNE-LGE agreement
- 90.0%
Original publication
Myocardial Scar Assessment Using Artificial Intelligence-Powered Contrast-Free MRI. J Am Coll Cardiol. Published 15 July 2026. doi:10.1016/j.jacc.2026.05.029
Summarised from the published abstract; every number is checked against it