• Journal of Internet Computing and Services
• /
• v.25 no.5
• /
• pp.31-40
• /
• 2024

Aortic valve stenosis is disease caused by the calcification of the aortic valve, which is located between the left ventricle of the heart and the aorta, preventing the backflow of blood. Transcatheter Aortic Valve Replacement (TAVR) has become the standard non-surgical procedure for treating aortic stenosis. However, patients who undergo TAVR are still face the risk of complication, which calls for a systematic solution for complication prediction and management. In this study, we designed a platform that manages patient complication risks using various prediction models, and we utilized explainable AI to ensure reliable prediction models and solutions. For the implementation of this platform, digital twin technology was employed, allowing for complication prediction and monitoring based on a patient's digital twin model constructed from Real-World Data (RWD). The digital twin platform has been developed with a microservice architecture using Kubernetes, enhancing the flexibility and availability of the platform. Additionally, we propose a feedback system for continuous improvement of prediction model performance and TAVR procedures to ensure ongoing development.