Related Experiment Video
Updated: Dec 23, 2025
05:27
Laparoscopic Anatomical Hepatectomy Using Takasaki's Approach and Indocyanine Green Fluorescence Navigation
Published on: May 16, 2025
641
Development of an artificial intelligence system using deep learning to indicate anatomical landmarks during
Tatsushi Tokuyasu1, Yukio Iwashita2, Yusuke Matsunobu3
1Faculty of Information Engineering, Department of Information and Systems Engineering, Fukuoka Institute of Technology, 3-30-1 Wajiro-higashi, Higashi-ku, Fukuoka-City, Fukuoka, 811-0295, Japan. tokuyasu@fit.ac.jp.
Surgical Endoscopy
|April 20, 2020
Summary
This study developed an AI system to identify key landmarks during laparoscopic cholecystectomy, aiming to reduce bile duct injuries. The system successfully highlighted critical anatomical structures, enhancing surgical safety.
Area of Science:
- Surgical Technology
- Artificial Intelligence in Medicine
- Medical Imaging Analysis
Background:
- Bile duct injury (BDI) is a significant complication of laparoscopic cholecystectomy (LC).
- Accurate identification of intraoperative landmarks is crucial for preventing BDI.
- Current methods rely on surgeon expertise to identify these landmarks.
Purpose of the Study:
- To develop an AI-powered system for real-time identification of critical landmarks during LC.
- To outline four key anatomical landmarks on endoscopic images to aid surgeons.
Main Methods:
- A deep learning object detection algorithm (YOLOv3) was employed.
- The system was trained on approximately 2000 endoscopic images from 76 LC videos.
- The model's accuracy was evaluated on 23 independent LC videos, identifying the cystic duct, common bile duct, liver segment edge, and Rouviere's sulcus.
Main Results:
- The system demonstrated quantitative and subjective performance in landmark identification.
- Average precision varied across landmarks, with higher scores for the common bile duct and liver segment edge.
- Expert surgeons confirmed landmark indications in 22 out of 23 videos, and the system increased surgical team awareness during a verification surgery.
Conclusions:
- The developed intraoperative landmark indication system successfully identified four critical landmarks during LC.
- This AI-driven approach has the potential to reduce the incidence of BDI and improve LC safety.
- The proposed system offers a novel method for preventing BDI in clinical practice.
