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

BACKGROUND: Risk stratification for tricuspid valve transcatheter edge-to-edge repair (T-TEER) is paramount in the decision-making process to appropriately select patients with severe tricuspid regurgitation. OBJECTIVES: The aim of this study was to develop and validate an artificial intelligence-driven risk score, the EuroTR (European Registry of Transcatheter Repair for Tricuspid Regurgitation) score, to predict 1-year mortality in patients undergoing T-TEER. METHODS: The EuroTR score was developed using data from the EuroTR registry, comprising 1,225 patients in the derivation cohort and 601 patients in the validation cohort. On the basis of 18 clinical, laboratory, echocardiographic, and hemodynamic parameters, an extreme gradient boosting algorithm was trained and independently validated against established risk models. RESULTS: Among the entire study cohort (N = 1,826), the overall 1-year survival rate was 82.1% (95% CI: 80.1%-84.2%), with no significant differences between the derivation and validation cohorts. The EuroTR score successfully stratified patients into low-risk and high-risk groups for 1-year mortality after T-TEER (HR: 4.26; 95% CI: 2.71-6.67; P < 0.001), and it significantly outperformed established risk scores such as the EuroScore and the TRI-SCORE in the validation cohort. Beyond mortality prediction (Harrell's C index [validation cohort] = 0.741; 95% CI: 0.699-0.783), increasing EuroTR score values were associated with a higher likelihood of a clinically relevant combined endpoint of 1-year mortality, need for heart failure hospitalization, or persistent dyspnea corresponding to NYHA functional class ≥III. The likelihood of poor outcomes increased from 30.6% in patients with the lowest EuroTR scores (EuroTR risk rank