Heart failure & cardiomyopathiesDiagnostic or prognostic study
External validation of an explainable electrocardiogram-only deep learning algorithm for the prediction of response after cardiac resynchronization therapy
- external validation study
Authors' conclusion
In this external validation study, an explainable ECG-only algorithm for the prediction of nonresponse after CRT device implantation generalized well to a lower-risk population from a different hospital. Adding indices of mechanical dyssynchrony and right ventricular function might be of additional value when evaluating volumetric response.
Study context
- Population
- Patients with cardiac resynchronization therapy (CRT) from the University of Virginiaexternal validation cohort, model developed in 1306 patients with CRT from the Netherlands · Development cohort (Netherlands) 1,306 · External validation cohort (University of Virginia) 161
- Participants
- 161 patients
- Comparison
- The original model vs American Heart Association criteria for left bundle branch block for the prediction of nonresponse
Further findings
| Outcome | FactorECG algorithm | American Heart Association criteria for left bundle branch block |
|---|---|---|
| Prediction of volumetric nonresponse (C-statistic) | ||
| C-statistic | Not reported in the abstract | 0.51 (95% CI 0.41-0.60) |
- Prediction of volumetric nonresponse (C-statistic): Original model C-statistic 0.67 (95% confidence interval [CI] 0.59-0.76)
- Prediction of nonresponse: refitted FactorECG-based model vs model also integrating indices of mechanical dyssynchrony: Refitted FactorECG-based model C-statistic 0.74 [95% CI 0.66-0.82]
- Prediction of nonresponse: refitted FactorECG-based model vs model also integrating indices of mechanical dyssynchrony: Model also integrating indices of mechanical dyssynchrony C-statistic 0.70 [95% CI 0.62-0.77]
Prediction of volumetric nonresponse (C-statistic)
- The original model significantly outperformed American Heart Association criteria for left bundle branch block
Prediction of nonresponse: refitted FactorECG-based model vs model also integrating indices of mechanical dyssynchrony
- Performed similarly
Prediction of response: combination of both models
- C-statistic
- 0.79 [95% CI 0.71-0.85]
Original publication
External validation of an explainable electrocardiogram-only deep learning algorithm for the prediction of response after cardiac resynchronization therapy. Heart Rhythm. Published 10 January 2026. doi:10.1016/j.hrthm.2026.01.014
Summarised from the published abstract; every number is checked against it