Jelle Ruurda

UMC Utrecht

In short, Jelle Ruurda's research entails the following: 

INTelligent computeR-Aided Surgical gUidance for Robot-assisted surGEry – the INTRA-SURGE-project for the future of surgery

Every year, 3000 patients are diagnosed with esophageal cancer in the Netherlands. Approximately 25% of them is a candidate for surgical resection of the esophagus. Despite efforts to reduce complications, including performing minimally invasive surgery and dedicated training programs, morbidity and mortality remain substantial. During robot-assisted minimally invasive esophagectomy (RAMIE) procedures, several vital anatomical structures are located within the narrow operating space. The INTRA SURGE-project proposes to advance surgery by developing intelligent surgical guidance based on artificial intelligence to realize computer-aided anatomy recognition to support surgeons intraoperatively, prevent unnecessary tissue injury and increase patient safety.

Publications

• Scaling up self-supervised learning for improved surgical foundation models
  • Medical Image Analysis, February 1, 2026
  • Click here to view the publication.
• Do segmentation metrics reflect clinical reality? A surgeon-centered evaluation in robot-assisted minimally invasive esophagectomy
  • Surgical Endoscopy, October 10, 2025
  • Click here to view the publication.
• Benchmarking and Enhancing Surgical Phase Recognition Models for Robotic-Assisted Esophagectomy
  • Proc. SPIE 13408, Medical Imaging 2025: Image-Guided Procedures, Robotic Interventions, and Modeling, April 7, 2025
  • Click here to view the publication
• Exploring the Effect of Dataset Diversity in Self-supervised Learning for Surgical Computer Vision
  
  • ata Engineering in Medical Imaging. DEMI 2024, April 25, 2024
  • Click here to view the publication.

Brief summary of progress / results February 2026

INTRA SURGE develops artificial intelligence to support surgeons during complex robot-assisted esophageal surgery. By combining advanced machine learning with large surgical video datasets, the project has made considerable progress in real-time recognition of anatomy and surgical phases, even with limited manual annotations. Lightweight and efficient AI models were developed that are suitable for use in the operating room. In addition, stereoscopic surgical video recording was successfully implemented, enabling future patient-specific surgical guidance. These results have been widely published and presented internationally, and form a strong foundation for upcoming clinical validation.

Brief summary of progress / results February 2025

The INTRA SURGE project focuses on enhancing real-time surgical guidance during robot-assisted minimally invasive esophagectomy (RAMIE). By leveraging deep learning, we develop AI-driven models for anatomy recognition, surgical phase identification, and 3D registration. The INTRA SURGE database, containing annotated videos and stereoscopic surgical footage, supports this work. Our preliminary results indicate that self-supervised learning improves model performance, enabling better recognition of anatomical structures and surgical phases. Future steps include refining these AI models for real-time surgical assistance and validating them in both laboratory and clinical settings.