In diesem Semester wird die im WiSe 2022/23 erfolgreich gestartete Vortragsreihe Medical Information Sciences fortgesetzt. Renommierte Wissenschaftlerinnen und Wissenschaftler unterschiedlicher Fachdisziplinen und Forschungsstandorte geben jeden Donnerstag ab 16:00 Uhr Einblicke in aktuelle Fragestellungen und Anwendungsgebiete des breiten Forschungsfeldes Medical Information Sciences.

Artificial intelligence (AI) is increasingly transforming gastrointestinal diagnostics and therapy. I will provide a comprehensive overview of state-of-the-art AI-based methods in endoscopy, with a particular focus on the early detection and treatment of Barrett’s-associated carcinoma.

Using clinically relevant examples, I will demonstrate how deep learning models are applied to detection, classification, and semantic segmentation of endoscopic image data. A key emphasis is placed on semi-supervised learning approaches, which enable the training of robust models despite the limited availability of annotated medical data. In this context, error-correcting supervision and Mean Teacher–based frameworks are presented as effective strategies to improve model consistency, robustness, and clinical reliability.

In addition, I will address the role of the human-in-the-loop, challenges related to trust in AI systems, and the problem of overconfident predictions, including methods for out-of-distribution detection in medical imaging. Recent clinical studies show that AI can significantly improve diagnostic accuracy when used in combination with expert judgment or as a standalone system.

Finally, AI-supported assistance systems for endoscopic submucosal dissection (ESD) are presented, including real-time overlays, phase recognition, and safety alerts, highlighting the successful translation of AI research into clinical practice. The lecture thus bridges methodological foundations, clinical evidence, and practical deployment of intelligent systems in gastrointestinal medicine.