AI-Driven Navigation Enhances Surgical Precision in Urological Surgery

Strategic Insights -

Researchers at the Institute of Science have developed an artificial intelligence (AI) model that enables real-time fluorescence-like navigation (FLN) for robot-assisted radical cystectomy (RARC), potentially enhancing surgical outcomes and training efficacy. The AI model, built on a convolutional neural network (CNN), achieved high accuracy in identifying and tracking the ureter during surgery, reducing the risk of misrecognition by surgeons. In a study involving 17 RARC cases, the CNN-based prediction model demonstrated an average Dice score of 0.71, IoU of 0.55, recall of 0.90, precision of 0.60, and accuracy of 0.99. The AI system scored an average of 4.74 on a scale of 0 to 5 for identification performance, indicating 80% recognition accuracy.

Key Takeaways:

  • The AI model achieved an average Dice score of 0.71, IoU of 0.55, recall of 0.90, precision of 0.60, and accuracy of 0.99 in identifying and tracking the ureter during RARC.
  • The AI system scored an average of 4.74 on a scale of 0 to 5 for identification performance, indicating 80% recognition accuracy.
  • The AI model demonstrated high accuracy in real-time fluorescence-like navigation, reducing the risk of ureter misrecognition by surgeons.
  • The study involved 17 RARC cases, with 41 test images used for quantitative evaluation and 18 surgeons participating in qualitative assessment.
  • The AI model has the potential to enhance surgical accuracy, safety, and training efficacy in urological surgery.
  • The research was conducted by a team of researchers from the Institute of Science, including Wei Chen, Shohei Fukuda, Soichiro Yoshida, and others.

Statistics:

  • Average Dice score: 0.71
  • IoU: 0.55
  • Recall: 0.90
  • Precision: 0.60
  • Accuracy: 0.99
  • Identification performance: 4.74/5 (average score)
  • Recognition accuracy: 80% (based on identification performance)
  • Number of RARC cases: 17
  • Number of test images: 41
  • Number of participating surgeons: 18

Sources:

  • Institute of Science Reports Findings in Radical Cystectomy (Pioneering AI-guided fluorescence-like navigation in urological surgery: real-time ureter segmentation during robot-assisted radical cystectomy using convolutional neural network).
  • Journal of Engineering. May 12, 2025; p 1245.
  • SpringerNature. Journal of Robotic Surgery. 2025;19(1):188.