Acute cardiac care & resuscitationDiagnostic or prognostic study
DeepCRI: real-time EEG-based prognostication after cardiac arrest
- development and validation of a deep learning system
- multicenter
- independent internal and external validation cohorts
- bedside-integrated
- real-time
Authors' conclusion
By embedding DeepCRI into routine ICU EEG infrastructure, we demonstrate the technical feasibility and clinical promise of continuous, real-time AI-driven prognostication for comatose patients after cardiac arrest.
Study context
- Population
- Comatose patients after cardiac arrestdevelopment dataset 522, internal validation 219, external validation 167 · development dataset 522 · internal validation cohort 219 · external validation cohort 167
- Participants
- 219 vs 167
- Comparison
- internal validation cohort vs external validation cohort
- internal validation cohortn = 219
- external validation cohortn = 167
Performance
As reported in the abstract
Lock-in classification and performance, internal validation cohort · first 36 h after arrest
| Test or model | lock-in classifications | sensitivity, good outcome | specificity, good outcome | sensitivity, poor outcome | specificity, poor outcome | remained in gray zone |
|---|---|---|---|---|---|---|
| internal validation cohort | 179/219 patients (81.7%) | 94.7% (95% CI 90.0-97.6%) | 81.9% (95% CI 73.5-88.1%) | 49.5% (95% CI 40.2-58.9%) | 100.0% (95% CI 96.7-100.0%) | 40/219 patients (18.3%) |
Further findings
- Lock-in classification and performance, external validation cohort · first 36 h after arrest
- 100/167 patients (59.9%) · lock-in classifications · sensitivity, good outcome 67.2% (95% CI 54.7-77.7%) · specificity, good outcome 82.1% (95% CI 73.7-88.2%) · sensitivity, poor outcome 36.8% (95% CI 28.2-46.3%) · specificity, poor outcome 98.4% (95% CI 91.3-99.7%)
Post hoc analysis of false poor-outcome prediction
- residual EMG artifacts contributed to this false poor-outcome prediction
- post hoc
Original publication
DeepCRI: real-time EEG-based prognostication after cardiac arrest. Resuscitation. Published 16 April 2026. doi:10.1016/j.resuscitation.2026.111102
Summarised from the published abstract; every number is checked against it