AI in pelvic lymph node dissection improves surgeons’ ability to identify key pelvic anatomical structures and enhances sensitivity and specificity across specialties, according to new findings.
Pelvic Lymph Node Dissection
The lateral pelvis is a critical anatomical region encountered during colorectal, gynaecological and urological surgery. However, its complex anatomy and natural variation can make pelvic lymph node dissection technically demanding, increasing the importance of accurate identification of surrounding structures. The researchers therefore developed an AI model to determine whether it could improve surgeons’ recognition of key pelvic anatomy during the procedure.
AI-Assisted Anatomical Recognition
The study involved 36 surgeons from colorectal, gynaecological and urological specialties with varying levels of experience, who reviewed 640 video snippets of 0.5 s extracted from pelvic lymph node dissection procedures. An AI model was developed using 23,259 annotated images and 653 unannotated images derived from 293 surgical videos.
The model was trained to identify key anatomical structures relevant to pelvic lymph node dissection, including the ureter, obturator nerve, external iliac artery and external iliac vein. It was designed to support anatomical recognition during surgery across different specialties and experience levels.
Across threefold cross-validation, Dice similarity coefficients were 0.6483 for the ureter, 0.8654 for the obturator nerve, 0.8619 for the external iliac artery and 0.8736 for the external iliac vein.
Impact on Surgeon Performance
AI assistance significantly improved surgeons’ sensitivity and specificity in identifying pelvic anatomical structures during assessment (p<0.001). Improvements were observed across colorectal, gynaecological and urological specialties, regardless of experience level. The findings suggest that AI support enhanced surgeons’ ability to recognise the anatomical structures evaluated during the study, demonstrating consistent benefits across participants with differing levels of surgical experience.
Clinical Implications
The authors suggest that the model could provide support for surgeons during pelvic lymph node dissection regardless of specialty or experience. They note, however, that additional studies using continuous intraoperative workflows will be necessary to determine how the technology performs in routine clinical practice.
Reference
Kitaguchi D et al. Enhancing anatomical recognition by surgeons during pelvic lymph node dissection using artificial intelligence. npj Digit. Med. 2026;DOI:10.1038/s41746-026-02936-4.
Featured on: Fahng Adobe Stock
- Author:
