Related Experiment Video
Updated: Apr 5, 2026
08:05
Detection and Isolation of Cancer in Prostate Biopsies Using Stimulated Raman Histology and Artificial Intelligence
Published on: June 10, 2025
1.4K
Artificial Intelligence-Based Recognition of the Prostatic Capsule During Nerve-Sparing Robot-Assisted Radical
Kazuki Honda1, Suguru Oka1, Kazuhide Makiyama2
1Department of Urology, Toranomon Hospital, Tokyo, Japan.
Summary
An AI deep-learning model assists in recognizing the prostatic capsule during nerve-sparing robot-assisted radical prostatectomy. This technology may enhance surgical safety and improve patient outcomes by aiding in precise layer identification.
Area of Science:
- Surgical Technology
- Artificial Intelligence in Medicine
- Urology
Background:
- Intraoperative identification of anatomical structures is crucial for surgical decision-making and patient outcomes.
- Nerve-sparing robot-assisted radical prostatectomy aims to preserve postoperative function, but relies heavily on surgeon experience for identifying critical layers.
- The prostatic capsule's precise recognition is vital for preserving neurovascular structures during radical prostatectomy.
Purpose of the Study:
- To develop and evaluate a convolutional neural network (CNN)-based deep-learning model for assisting in prostatic capsule recognition.
- To improve the accuracy and consistency of identifying the prostatic capsule during robot-assisted radical prostatectomy.
- To enhance the nerve-sparing technique by providing objective, AI-driven guidance.
Main Methods:
- A deep-learning model was trained using videos from 13 nerve-sparing robot-assisted radical prostatectomy procedures.
- The model's performance was evaluated quantitatively using video frames from 5 independent cases (Dice score: 0.621, Intersection over Union: 0.451).
- Qualitative assessment by 7 experienced urologists evaluated the model's concordance and effectiveness in still images and video clips.
Main Results:
- The AI model demonstrated moderate performance in prostatic capsule recognition with a median Dice score of 0.621 and IoU of 0.451.
- Qualitative evaluations showed a mean concordance score of 3.6/5 and effectiveness score of 3.1/5 for still images.
- Surgeons reported a mean stress level of 3.6/5 and effectiveness score of 2.9/5 when using the AI model during video assessment.
Conclusions:
- The developed AI model shows potential in assisting with prostatic capsule recognition during robot-assisted radical prostatectomy.
- This AI-driven assistance may contribute to safer nerve-sparing techniques.
- The use of AI in identifying critical surgical layers could lead to improved postoperative functional outcomes.
