A complexity evaluation system for mitral valve repair based on preoperative echocardiographic and machine learning

Area of Science:

• Cardiovascular Surgery
• Artificial Intelligence in Medicine
• Echocardiography

Background:

• Mitral valve repair complexity assessment is crucial for surgical success.
• Preoperative echocardiographic data offers valuable insights into valve anatomy and function.
• Machine learning presents a promising avenue for developing advanced predictive models.

Purpose of the Study:

• To create a novel system for evaluating mitral valve repair complexity.
• To utilize preoperative echocardiographic data and machine learning algorithms for this purpose.

Main Methods:

• A cohort of 231 patients undergoing mitral valve repair between March 2021 and March 2023 was analyzed.
• Clinical and echocardiographic data were collected and analyzed.
• Various machine learning algorithms were employed to develop the complexity evaluation system, with outcomes including immediate repair failure and recurrent regurgitation.

Main Results:

• The linear support vector classification model demonstrated superior predictive performance in both training and testing datasets.
• Key risk factors identified for mitral valve repair failure included patient age, A2 prolapse, leaflet height abnormalities, and mitral regurgitation severity.
• The overall success rate for mitral valve repair in the study cohort was 90.9%.

Conclusions:

• A linear support vector classification model shows potential for evaluating mitral valve repair complexity.
• Factors such as age, A2 lesions, leaflet height, and mitral regurgitation grades are associated with mitral repair failure, informing preoperative risk assessment.