AI-Enhanced Monitoring Solutions for Anesthetized Patients – Anesth-IA

General anesthesia, although routine with 12 million cases annually in France and 300 million worldwide, presents challenges. 15-25% of patients undergo post-operative complications, increasing post-operative mortality, neurocognitive disorders, and burdening healthcare systems. Intra-operative hypotension is a significant risk, but focusing only on blood pressure doesn't suffice to prevent complications. A comprehensive view must encompass the interplay of blood pressure, brain function during anesthesia, and anesthetic dosage. Current models, however, fall short in capturing this synergy.
The Anesth-IA initiative seeks to fill this gap, developing a decision-support tool for tailored anesthesia doses, aiming to reduce post-operative complications. It will recommend optimal strategies for sedation and blood pressure modulation during anesthesia, based on non-invasive physiological signals, and is designed for integration into operating rooms.
Our team already compiled a database housing 50 TB of physiological data from operating rooms. We are now addressing the challenge of real-time data streaming, facilitating communication between operating rooms and a central server for tailored predictions. The Anesth-IA tool employs techniques like photoplethysmography for blood pressure and EEG for brain monitoring. Using pharmacological models, it aims to guide clinicians in personalizing anesthesia and blood pressure to each patient's needs
Anesth-IA integrates mathematical modeling of multimodal intra-operative data, aligned with the AAPG2024 call's H.14 research topics. Our advanced methodology aims at data management, from collection to application. The project is segmented into 3 workpackages: WP1 focuses on real-time data gathering, WP2 on modeling vessel-brain interaction during anesthesia, and WP3 on setting up the decision-support tool in the operating room.