Cardio-oncology & otherDiagnostic or prognostic study
Artificial intelligence based multimodal fusion of electronic medical record data and ECG signals for pericarditis prediction: pericarditis calculator (p-Cal) as a screening aid
- development of an artificial intelligence (AI)-based multimodal fusion model (browser-based pericarditis risk calculator)
- decision-level (late) fusion of ECG image and clinical tabular data
- hold-out test set
- external validation in Medical Information Mart for Intensive Care (MIMIC)
Authors' conclusion
A late fusion AI model integrating tabular clinical electronic medical record data and ECG signal has robust predictive capability for pericarditis as a screening tool.
Study context
- Population
- Patients with confirmed pericarditis and normal controlspatients used for training and internal validation, 902 with confirmed pericarditis (13.8%) · hold-out test patients 1,302
- Participants
- 6,508 patients
- Comparison
- Fusion Model vs ECG and clinical data only models for pericarditis prediction
Performance
As reported in the abstract
Discriminatory performance for pericarditis (AUC) · hold-out test set
Fusion Model
0.89
AUC
ECG
0.85
AUC
clinical data only
0.74
AUC
Further findings
| Outcome | Fusion Model | ECG | clinical data only |
|---|---|---|---|
| External validation AUC (MIMIC) | AUC 0.81 | Not reported in the abstract | Not reported in the abstract |
Parts of this Brief are withheld: the automatic checks could not confirm them against the abstract.
Original publication
Artificial intelligence based multimodal fusion of electronic medical record data and ECG signals for pericarditis prediction: pericarditis calculator (p-Cal) as a screening aid. Heart. Published 20 August 2026. doi:10.1136/heartjnl-2025-327090
Summarised from the published abstract; every number is checked against it