Predicting Late Gadolinium Enhancement of Acute Myocardial Infarction in Contrast-Free Cardiac Cine MRI Using Deep Generative Learning
Authors' conclusion
CGE achieved superior image quality to LGE and accurate scar delineation in patients with acute MI of both internal and external data sets.
Study context
- Population
- Patients with acute myocardial infarction (MI) from 3 centres; plus normal subjects from a public cine cardiac magnetic resonance (CMR) data setacute MI, 3 centres
- Participants
- 430 patients
- Comparison
- Cine-generated enhancement (CGE) from contrast-free cine CMR vs late gadolinium enhancement (LGE) CMR
- Outcome
- Image quality and MI quantification (scar size and transmurality) of CGE vs LGE
Methodological detailsmulticentre, internal and external testing, Training: 1525 pairs (289 patients), centre I; internal testing: 293 slices (52 patients); external testing: 40 patients (401 slices) from 2 other centres
- Multicentre
- Internal and external testing
- Training: 1525 pairs (289 patients), centre I; internal testing: 293 slices (52 patients); external testing: 40 patients (401 slices) from 2 other centres
Performance
As reported in the abstract
| Test or model | Scar size Pearson correlation (internal/external) | Scar size ICC (internal/external) | Transmurality Pearson correlation (internal/external) | Transmurality ICC (internal/external) | Sensitivity for scar (all data sets) | Specificity for scar (all data sets) |
|---|---|---|---|---|---|---|
| CGE | 0.79/0.80 | 0.79/0.77 | 0.76/0.64 | 0.76/0.63 | 91.27% | 95.83% |
CGE
Scar size Pearson correlation (internal/external)0.79/0.80
Scar size ICC (internal/external)0.79/0.77
Transmurality Pearson correlation (internal/external)0.76/0.64
Transmurality ICC (internal/external)0.76/0.63
Sensitivity for scar (all data sets)91.27%
Specificity for scar (all data sets)95.83%
- Significant correlation between CGE and LGE for scar size and transmurality
- P < 0.001
Further findings
Robustness in normal subjects (public cine CMR data set)
- Realistic enhancement images without false positive subjects
- 20 normal subjects
Image quality, CGE vs LGE
- CGE image quality superior to LGE in internal and external data sets
Original publication
Predicting Late Gadolinium Enhancement of Acute Myocardial Infarction in Contrast-Free Cardiac Cine MRI Using Deep Generative Learning. Circ Cardiovasc Imaging. Published 10 September 2024. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →