Artificial intelligence-enabled electrocardiogram guidance for pulmonary valve replacement timing in repaired tetralogy of Fallot
Authors' conclusion
AI-ECG predicts post-PVR survival in rTOF patients with a PVR survival benefit in intermediate- and high-risk, but not low-risk, groups. AI-ECG may complement imaging biomarkers to determine rTOF PVR timing.
Study context
- Population
- Patients with repaired tetralogy of Fallot undergoing pulmonary valve replacement with analyzable ECGs ≤3 months pre-PVRPulmonary valve replacement (PVR) cohort, median follow-up 7.5 [IQR, 4.7–10.6] years · Boston Children's Hospital (BCH) 504 · Toronto General Hospital (TGH) 101
- Participants
- 605 patientsEvents: 3.6% mortality
- Predictors
- Pre-PVR AI-ECG risk probability
- Predicted outcome
- post-PVR mortality
Methodological details1:1 propensity score matching to non-PVR patients, two centres (BCH and TGH)
- 1:1 propensity score matching to non-PVR patients
- Two centres (BCH and TGH)
Model performance
As reported in the abstract
C-index
AI-ECG
0.77
Established imaging-based model benchmark
0.70
Further findings
| Outcome | HR (95% CI) | P value | |
|---|---|---|---|
| Survival, matched cohort (PVR vs non-PVR)PVR associated with increased survival overall | 0.28 (0.13–0.60) | P = 0.001 | |
| Survival by risk tertile, matched cohort (exploratory)PVRvs non-PVRIntermediate-risk group · exploratory | 0.10 (0.01–0.86) | P = 0.04 | |
| Survival by risk tertile, matched cohort (exploratory)PVRvs non-PVRHigh-risk group · exploratory | 0.3 (0.1–0.7) | P = 0.005 | |
| 0.050.21520Hazard ratio (log scale) |
| Outcome | Low- and intermediate-risk | High-risk |
|---|---|---|
| 15-year survival by AI-ECG risk tertile | 97-98% | 83% |
- Survival by risk tertile, matched cohort (exploratory) — No survival benefit associated with PVR in the low-risk group
- P = .8
- 15-year survival by AI-ECG risk tertile: low vs intermediate risk P = .6
- 15-year survival by AI-ECG risk tertile: increased mortality for the high-risk group P = .009
- Survival, matched cohort (PVR vs non-PVR): propensity score-matched 1:1
AI-ECG added to the benchmark model
- AI-ECG remained an independent predictor
- P < 0.001
- C-index
- 0.84
Original publication
Artificial intelligence-enabled electrocardiogram guidance for pulmonary valve replacement timing in repaired tetralogy of Fallot. Am Heart J. Published 29 August 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →