Design of Precision Medicine Web-service Platform Towards Health Care Digital Twin

Abstract

Recently, there has been a growing interest in researching and developing personalized medical AI services. The previous AI medical systems rarely provided model output compared to multiple datasets and AI models. Currently, only few medical AI systems offer integrated platforms for multidisciplinary precision medicine services. Most existing medical AI systems include AI prognosis with a singular discipline in focus, such as elderly healthcare. This paper proposes a novel digital twin-based integrated precision medicine web-services platform. Our proposed system architecture can be easily implemented in hospital organization interfaces because of the ensured platform independence. Based on the prognostic requirements, we design the service interface with a broad spectrum of patient medical parameter selection (survival time, vital signs, etc.) made available for each medical service. The data related to each patient can be effortlessly updated in real-time. The services will predict and evaluate the accuracy of the visualized output along with the patient clinical information. To verify the feasibility of the proposed architecture, we implemented it with different AI medical services, such as 5 year lung cancer survival prediction, survival analysis with lung tumor segmentation and rapid response analysis. We observed that the architecture showed excellent performance. The architecture for this comprehensive precision medicine web-service platform (Comp-Med) is highly efficient and flexible. It is easily extensible to the new features, services, and updates that may get accommodated in the future.