Employment of Artificial Intelligence for an Unbiased Evaluation Regarding the Recovery of Right Ventricular Function After Mitral Valve Transcatheter Edge-to-Edge Repair

A. Hesse (München)¹, V. Fortmeier (Bad Oeynhausen)², M. Lachmann (München)³, M. Tokodi (Budapest)⁴, E. Xhepa (München)⁵, E. Rippen (München)¹, J. Tervooren (München)¹, G. Harmsen (Johannesburg)⁶, S. Yuasa (Tokyo)⁷, T. Trenkwalder (München)⁵, M. Joner (München)⁵, V. Rudolph (Bad Oeynhausen)², K.-L. Laugwitz (München)^3
1Technische Universität München (TUM) Klinik und Poliklinik für Innere Medizin I München, Deutschland; ²Herz- und Diabeteszentrum NRW Allgemeine und Interventionelle Kardiologie/Angiologie Bad Oeynhausen, Deutschland; ³Klinikum rechts der Isar der Technischen Universität München Klinik und Poliklinik für Innere Medizin I München, Deutschland; ⁴Heart and Vascular Center, Semmelweis University Budapest, Ungarn; ⁵Deutsches Herzzentrum München Klinik für Herz- und Kreislauferkrankungen München, Deutschland; ⁶University of Johannesburg Department of Physics Johannesburg, Südafrika; ⁷Keio University School of Medicine Department of Cardiology Tokyo, Japan

Objective:

This study aims to use both conventional echocardiography and artificial intelligence to assess the recovery of RV function in patients undergoing MV TEER for severe MR.

 

Methods:

All patients from a bicentric registry undergoing MV TEER for severe MR were analyzed regarding their change in RV function from baseline to 3-month follow-up.

 

Results:

Among 851 patients (median age: 79.9 years [IQR: 74.5-83.5]; 56.9% male) undergoing MV TEER between 2017 and 2023, the 1-year survival rate was 86.8% (95% CI: 84.5–89.1%). Follow-up echocardiography was available for 475 patients (55.8%). MV TEER led to significant reductions in both left atrial (LA) volume and estimated systolic pulmonary artery pressure (sPAP) levels (median LA volume: from 123 mL [IQR: 92-169 mL] to 104 mL [IQR: 78-142 mL], p-value: < 0.001; median sPAP: from 46 mmHg [IQR: 35-58 mmHg] to 41 mmHg [IQR: 32-54 mmHg], p-value: 0.036). In contrast, right midventricular diameter and right atrial area did not show a reduction. Similarly, TAPSE remained unchanged at follow-up (median: from 17 mm [IQR: 14-21 mm] to 18 mm [IQR: 15-21 mm], p-value: 0.603). Deep learning-enhanced echocardiography confirmed this finding, showing no significant change in predicted RVEF after MV TEER (median: from 43.1% [IQR: 39.1-47.4%] to 43.2% [IQR: 39.2-47.2%], p-value: 0.475). Atrial fibrillation and arterial hypertension emerged as key factors preventing a successful reduction of pulmonary artery pressures and, consequently, RV afterload following MV TEER.

 

Conclusion:

This is the first study to employ deep learning to assess recovery of RV function following MV TEER. Our findings demonstrate while MV TEER improves left-sided hemodynamics, such as reducing LA volume and pulmonary artery pressures, it does not lead to significant recovery of RV function as confirmed by both conventional echocardiography and unbiased artificial intelligence. This finding underscores the importance of treating patients before the occurrence of irreversible right heart damage. Moreover, atrial fibrillation and arterial hypertension should be optimally treated to ameliorate pulmonary artery pressure levels and hence reduce RV afterload in damaged hearts.