Application and effectiveness evaluation of AI-assisted image analysis in clinical clerkship of vascular surgery

Background The study evaluated an integrated teaching model combining LBL, PBL, CBL, and AI assisted image analysis. Methods A total of 160 vascular surgery trainees, including interns and standardized training residents, were randomized into two groups. Post-intervention outcomes were compared using baseline-adjusted analyses. The 80 observation students received AI-assisted integrated teaching, and 80 controls received traditional LBL teaching. We compared OSCE scores, diagnostic accuracy, and satisfaction. The primary outcome was the post-intervention OSCE score. Results After baseline adjustment, baseline characteristics were comparable between groups. The intervention was associated with higher OSCE scores (86.74 ± 6.19 vs. 71.39 ± 7.06), higher diagnostic accuracy (87.59 ± 4.33 vs. 73.68 ± 5.88), shorter image interpretation time, and improved satisfaction, with a large effect size (Cohen's d = 2.32). Conclusion The AI-assisted integrated teaching model combining LBL, PBL and CBL improves imaging analysis teaching quality, which provides a promising strategy for intelligent medical education, although broader validation is still needed.