University of Twente Student Theses

Login

AI-based retrospective analysis of diathermic device use in fundoplication surgery

Ribbens, Vincent J. (2025) AI-based retrospective analysis of diathermic device use in fundoplication surgery.

Full text not available from this repository.

Full Text Status:Access to this publication is restricted
Abstract:Electrosurgical devices, such as Enseal, play a critical role in achieving hemostasis during surgical procedures. However, improper use can lead to complications, highlighting the need for objective skill assessment tools. This study presents a proof-of-concept framework to monitor Enseal usage in fundoplication surgery by analyzing device activations. A three-stage model was developed: an activation detection algorithm identified Enseal activations with an F1-score of 0.945; blood detection was optimized using a random forest classifier, achieving a Dice score of 0.495 after post-processing; and the YOLOv8 algorithm tracked the Enseal tool, achieving a mAP50 of 0.925. Combining these models, diathermic device-induced bleeding was detected with 78.2% accuracy, an AUC of 0.776, and 4.6% precision. Additionally, the tracked tool behavior yielded consistent data distributions, confirming the method's potential for skill assessment. The results demonstrate the feasibility of using laparoscopic video and current data to extract features for comprehensive evaluation of Enseal usage. With further refinement of bounding box accuracy and dataset expansion, this approach could support retrospective reporting systems to improve surgeons' proficiency with diathermic tools, contributing to safer and more effective surgical outcomes.
Item Type:Essay (Master)
Faculty:TNW: Science and Technology
Subject:44 medicine, 54 computer science
Programme:Technical Medicine MSc (60033)
Link to this item:https://purl.utwente.nl/essays/104910
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page