Predicting Sudden Cardiac Death in Patients With Sarcoidosis Using a Multimodal Artificial Intelligence Model
Authors' conclusion
MAARS-CS offers superior predictive accuracy and consistency over LVEF in predicting SCD risk among patients with sarcoidosis.
Study context
- Population
- Patients with sarcoidosis
- Participants
- 317 patients
- Comparison
- MAARS-CS vs left ventricular ejection fraction (LVEF) ≤35% criterion and continuous LVEF
- Outcome
- Prediction of sudden cardiac death (SCD) risk (AUROC)
Methodological detailsretrospective cohort, 5-fold cross-validation
- Retrospective cohort
- 5-fold cross-validation
Performance
As reported in the abstract
SCD risk prediction
| Test or model | AUC |
|---|---|
| MAARS-CS | 0.86 (95% CI: 0.80-0.91) |
| LVEF ≤35% | 0.59 (95% CI: 0.53-0.66) |
| Continuous LVEF | 0.77 (95% CI: 0.67-0.85) |
AUC
MAARS-CS0.86 (95% CI: 0.80-0.91)
LVEF ≤35%0.59 (95% CI: 0.53-0.66)
Continuous LVEF0.77 (95% CI: 0.67-0.85)
- MAARS-CS vs LVEF ≤35%: P < 0.0001
- MAARS-CS vs Continuous LVEF: P = 0.019
Further findings
| Outcome | MAARS-CS | Continuous LVEF |
|---|---|---|
| Area under precision-recall curve and balanced accuracy | ||
| Area under precision-recall curve | 0.54 | 0.43 |
| Balanced accuracy | 0.83 | 0.74 |
Robustness
- MAARS-CS maintained robust performance across varying image qualities and MR sequences
Original publication
Predicting Sudden Cardiac Death in Patients With Sarcoidosis Using a Multimodal Artificial Intelligence Model. JACC Clin Electrophysiol. Published 4 December 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →