Artificial Intelligence in Free Flap Monitoring: A Systematic Review and Meta-Analysis

Background: Postoperative free flap monitoring is critical for early detection of vascular compromise and successful flap salvage. Artificial intelligence (AI)?based monitoring systems have recently emerged as potential adjuncts to conventional clinical assessment. Methods: A systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. PubMed/MEDLINE, Scopus, and Google Scholar were searched for studies evaluating AI-based postoperative free flap monitoring. Original clinical studies involving human subjects were included. Extracted data included study characteristics, AI model type, input modality, monitoring outcomes, and reported diagnostic performance metrics. Due to heterogeneity in study design and outcome reporting, narrative synthesis was performed, with limited quantitative analysis of reported accuracy and area under the curve (AUC). Results: Eight studies met the inclusion criteria. AI models primarily employed deep learning and machine learning techniques applied to clinical photographs or hyperspectral imaging. Reported diagnostic accuracy ranged from 95.3% to 98.7%, with AUC values between 0.96 and 0.99. Sensitivity and specificity were inconsistently reported across studies. Overall, AI-based systems demonstrated earlier and more reliable detection of flap perfusion compromise compared with conventional clinical monitoring. Conclusion: AI-based free flap monitoring demonstrates high diagnostic performance and promising clinical utility. However, heterogeneity in study methodology and outcome reporting limits formal meta-analysis. Standardized reporting and multicentre prospective validation studies are required before widespread clinical adoption.