KLEVER- Confocal laser endomicroscopy for real-time visual detection of resection borders

Oncological surgery in the gastrointestinal and upper aerodigestive tract requires highly precise tumor resections to ensure both oncological safety and preservation of patient quality of life. However, current techniques such as white-light endoscopy only allow a rough estimation of tumor margins, which can result in either incomplete resections or unnecessarily radical procedures with significant functional impairment.

In critical anatomical regions, this may lead to severe consequences such as permanent stomas. Confocal laser endomicroscopy (CLE) has emerged as a promising non-invasive imaging modality, enabling real-time, in vivo visualization of epithelial structures at near-cellular resolution and offering the potential for an “optical biopsy” without tissue removal.

This project aims to develop algorithms for the automated analysis of CLE images to distinguish malignant from benign tissue based on histopathological annotations.

The AI-generated results will be visualized in real time and integrated into conventional white-light endoscopy through an intuitive, user-centered interface. By projecting tumor boundaries directly into the surgeon’s field of view, the system aims to support more accurate intraoperative decision-making and enable precise, organ-preserving resections.