Deep learning LV circumferential strain improves CMR diagnosis of myocarditis with preserved EF

DL-LVCS on cardiac MRI achieved AUC 0.887, sensitivity 85.9%, specificity 79.2% for acute myocarditis with preserved EF, surpassing conventional feature tracking (AUC 0.806, sensitivity 67.2%).

• In the preserved EF myocarditis group, DL-LVCS was significantly impaired vs controls (−8.0%±1.4 vs −10.4%±1.8; P<0.001).
• The deep learning model (PPIM-Net + U-Net) reduces operator dependency and analysis time compared with conventional feature tracking.

Automated summary

RadPigeon summaries are original and for information only. They are not clinical advice.