To evaluate the efficacy of AI-enhanced training in microsurgery compared to traditional methods, specifically focusing on technical performance, learning efficiency, and skill retention.
Key Findings:
13 studies included from 2,056 records, involving 3–50 participants, mostly single-centre with varied designs, indicating a need for broader research.
AI models used included Mask R-CNN, YOLOv2, ResNet-50, focusing on instrument tracking (30.8%) and motion analysis (23.1%), highlighting the diversity of applications.
Median accuracy of AI systems was 83.8% (IQR 78.4–88.2%), suggesting reliable performance.
AI improved technical skills by reducing errors and enhancing learning curves through real-time feedback, indicating significant potential for training.
Interpretation:
AI enhances microsurgical training by providing objective metrics and personalized feedback, showing potential technical advantages in simulations.
Limitations:
High risk of bias in studies, particularly in methodology and reporting.
Very low evidence certainty, necessitating cautious interpretation of results.
Poor external validation and heterogeneous, low-quality evidence limiting generalizability.
Conclusion:
Future research should focus on multi-centre RCTs, standardized outcomes, external validation, and ethical considerations for clinical translation, emphasizing the need for rigorous ethical frameworks.
These 10 states make it more practical for physicians to participate in hospital ownership by aligning statutory structure, corporate practice of medicine rules, and population trends.
