Artificial-Intelligence-based Surgical Phase Recognition in Robot-Assisted Radical Prostatectomy and Cross-Surgeon...

Area of Science:

• Robotics and Artificial Intelligence in Medicine
• Surgical Technology
• Medical Informatics

Background:

• Artificial intelligence (AI) applications in surgery are currently limited.
• An AI system was developed to automatically recognize surgical phases during robot-assisted radical prostatectomy (RARP).
• Cross-surgeon validation was performed to confirm the AI system's accuracy.

Purpose of the Study:

• To develop and validate an AI system for automated surgical phase recognition in RARP.
• To assess the AI system's accuracy and generalizability across different surgeons.
• To explore the potential of AI in evaluating surgical quality and enhancing surgical education.

Main Methods:

• Analysis of clinical data from 102 RARP patients.
• Development of an AI model using Temporal Convolutional Networks for the Operating Room (TeCNO).
• Classification of surgical operations into nine phases with surgeon-annotated video data for AI training and validation.

Main Results:

• The AI model was trained and validated on a large dataset of surgical videos (over 1.1 million frames).
• High precision of 0.94 was achieved when analyzing videos from the primary surgeon.
• A precision of 0.83 was obtained when analyzing videos from multiple other surgeons, indicating generalizability.

Conclusions:

• The developed AI system demonstrates high accuracy and generalizability across different surgeons in RARP.
• The AI has the potential to serve as a tool for objective surgical quality assessment.
• This AI can provide valuable feedback to surgeons and improve surgical education effectiveness.