Predicting Outcomes in Patients With Tricuspid Regurgitation Undergoing Transcatheter Edge-to-Edge Repair Using an Artificial Intelligence-Derived Risk Score: The EuroTR Risk Score

Comprehensive Summary

Hausleiter et al. developed the EuroTR score, an artificial intelligence-derived risk model to predict 1-year mortality in patients with severe tricuspid regurgitation undergoing transcatheter edge-to-edge repair (T-TEER). The model was derived from the EuroTR registry, including 1,225 patients in the derivation cohort and 601 in the validation cohort (N = 1,826). Using 18 clinical, laboratory, echocardiographic, and hemodynamic variables, an extreme gradient boosting (XGBoost) algorithm was trained and compared with established risk scores such as EuroScore and TRI-SCORE. The overall 1-year survival rate was 82.1% (95% CI: 80.1%-84.2%). The EuroTR score stratified patients into low- and high-risk groups for mortality after T-TEER (HR: 4.26; 95% CI: 2.71-6.67; p < 0.001) and demonstrated superior discrimination in the validation cohort (Harrell’s C-index 0.741; 95% CI: 0.699-0.783). Increasing EuroTR scores were also associated with a higher risk of the composite endpoint of mortality, heart failure hospitalization, or persistent NYHA class ≥III symptoms.

Outcomes and Implications

Accurate risk stratification is essential when selecting patients with severe tricuspid regurgitation for transcatheter repair. This study demonstrates that an AI-driven model trained on multicenter registry data can outperform traditional surgical risk scores for predicting outcomes after T-TEER. The EuroTR score may help guide patient selection, support shared decision-making within multidisciplinary heart teams, and improve the design of future clinical trials for transcatheter tricuspid therapies. However, further validation across broader patient populations is important before routine clinical adoption.