Valvular & structural heart disease
Deep Learning–Derived Right Ventricular Ejection Fraction Predicts Mortality in Patients Undergoing Transcatheter Tricuspid Valve Intervention
Authors' conclusion
Deep learning enabled an unbiased echocardiographic assessment of RV function after TTVI and identified a high-risk group with poor outcomes. These findings are exploratory and require external validation; if confirmed, deep learning-enhanced echocardiography may improve risk stratification and guide personalized follow-up strategies in patients undergoing TTVI.
Study context
- Population
- Patients undergoing transcatheter tricuspid valve intervention (TTVI) for severe tricuspid regurgitation between 2018 and 2023patients undergoing TTVI
- Participants
- 373 patients
- Comparison
- Postprocedural RV ejection fraction (RVEF) below 38% vs Postprocedural RVEF above 38%
- Outcome
- All-cause mortality (1-year survival)
- Follow-up
- 1 year
Main finding
All-cause mortality (1-year survival)
Hazard ratio
3.12
P < 0.001
By group
As reported in the abstract
Further findings
| Outcome | Postprocedural RVEF below 38% | Postprocedural RVEF above 38% | |
|---|---|---|---|
| RVEF change in high-risk group · baseline to postprocedural | |||
| Baseline RVEF | 41% (IQR: 38%-44%) | Not reported in the abstract | Postprocedural RVEF above 38%: not reported in the abstract |
| Postprocedural RVEF | 36% (IQR: 35%-37%) | Not reported in the abstract | Postprocedural RVEF above 38%: not reported in the abstract |
- RVEF change in high-risk group: P < 0.001
Deep learning-predicted RVEF before and after TTVI (cohort level) · baseline and 1 to 3 days after the procedure
- Median RVEFs similar before and after TTVI at the cohort level; individual trajectories diverged
Optimal prognostic threshold for postprocedural RVEF
- Optimal prognostic threshold for postprocedural RVEF
- 38%
- Maximally selected log-rank statistics
Original publication
Summarised from the published abstract; every number is checked against it