Effectiveness of artificial intelligence-based visualization for surgical anatomy education: A cluster...

Area of Science:

• Medical Education
• Surgical Training
• Artificial Intelligence in Medicine

Background:

• Traditional surgical education methods are insufficient for teaching intraoperative anatomy.
• Novice learners require more than passive observation for anatomical recognition.
• An AI-based visualization system was developed to highlight anatomy in real time.

Purpose of the Study:

• To evaluate the educational effectiveness of an AI-based visualization system in surgical education.
• To assess the impact of AI visualization on medical students' recognition of intraoperative anatomy.
• To compare AI-augmented teaching with conventional methods.

Main Methods:

• A cluster quasi-randomized controlled trial was conducted with fifth-year medical students.
• Students were divided into conventional teaching (cohort C) and AI-visualization (cohort A) groups.
• Performance was measured using image segmentation metrics (recall, precision, Dice coefficient) on pancreatic anatomy images.

Main Results:

• Both groups improved post-training, but cohort A showed significantly greater improvement in recall (P=.020).
• AI visualization enhanced sensitivity in identifying pancreatic regions.
• While the Dice coefficient showed a trend towards greater improvement in cohort A, it was not statistically significant (P=.097).

Conclusions:

• AI-based visualization can enhance medical students' recognition of intraoperative anatomy.
• This technology is particularly beneficial for complex structures like the pancreas.
• AI tools show promise for improving surgical anatomy education.