A Web-Based Telemonitoring Platform for Integrated Cardiometabolic Risk Analysis

The project, developed during a professional internship at Onit Sanità S.r.l., resulted in the engineering and validation of a web-based telemonitoring platform. The system architecture utilizes a server-client model (.NET with REST APIs) to facilitate the fusion of direct clinical measurements with TC-derived estimates. A significant advancement is the integration of a local AI module (LLM via Ollama), which provides clinical decision support while ensuring strict data privacy and sovereignty, as required by the medical context. Furthermore, the study addresses the technical challenges of deploying local AI, including computational constraints (GPU acceleration) and the "black box" problem regarding the interpretability of model outputs in clinical settings. Focusing on its fundamental relevance for early disease prevention and public health, in this thesis the web-based telemonitoring platform has been tested to explore the link between Visceral Adipose Tissue (VAT) and cardiometabolic risk. Thoracic Circumference (TC) is evaluated as a non-invasive anthropometric proxy for visceral adiposity, assessed along with a comprehensive battery of biochemical and hemodynamic measures. This represents a key methodological development in individual risk stratification. The resulting digital strategy enables proactive telemonitoring and establishes a framework for the future integration of automated alerts and deep learning-based segmentation to fully automatize the platform within this clinical scenario. By shifting the paradigm toward timely clinical intervention, consistent with the "Strike early, strike strong"[1] philosophy, this tool aims to significantly reduce cardiovascular risk and prevent adverse clinical events.