Are we developing the right intraoperative AI assistance? Surgeons' perspectives and desired functions

Area of Science:

• Surgical Technology
• Artificial Intelligence in Medicine
• Robotic Surgery

Background:

• Clinical implementation of artificial intelligence (AI) in robotic surgery is limited.
• Understanding surgeon perceptions of AI's value is crucial for progress.
• This study assesses surgeon knowledge, attitudes, and use of AI intraoperative assistance.

Purpose of the Study:

• To evaluate surgeons' perceptions of AI-driven intraoperative assistance.
• To gauge surgeons' knowledge and attitudes towards surgical AI tools.
• To identify perceived usefulness of different AI functions in robotic surgery.

Main Methods:

• A web-based survey of 53 surgeons across 5 continents was conducted.
• Surgeons rated demographics, attitudes, knowledge, and use of 5 AI intraoperative guidance components.
• Data were analyzed using non-parametric tests, with statistical significance at p < 0.05.

Main Results:

• 83% of surgeons perceived their AI knowledge as average or lower; 79% had never used AI intraoperatively.
• 75.5% were confident in clinically validated AI tools; 86.8% believed AI could improve surgical performance.
• Anatomy recognition and risk detection were rated most useful; step recognition and decision-making guidance were rated lowest.

Conclusions:

• Surgeons perceive high usefulness for AI in areas like anatomical guidance.
• Despite positive perceptions, adoption of intraoperative AI in routine surgical practice is limited.
• A significant gap exists between the perceived potential of AI and its current clinical implementation.