Learning Curve in Robotic-Assisted Microsurgery: A Systematic Review of Skill Acquisition

Background Robotic-assisted microsurgery offers enhanced precision, tremor filtration, and improved ergonomics, but adoption is limited by undefined learning curves. This systematic review evaluates the learning curve of microvascular anastomosis performed on dedicated robotic platforms. Methods A systematic review was conducted across MEDLINE, EMBASE, Web of Science, and Cochrane databases (inception to June 2025). Evaluated outcomes included time metrics (anastomosis time, time per stitch), microsurgical skills and anastomosis quality for the robot-assisted (RA) and hand-sewn (HS) techniques. Results After screening 6,743 records, 19 studies were included. In preclinical studies (n=11), mean RA anastomosis time decreased by 18.54 [4.01–47.00] minutes versus 8.43 [0.31–30.00] minutes for HS, although mean final RA time remained longer at the final attempt (20.94 [8.99–36.00] vs 14.78 [4.58–30.00]). RA time per stitch decreased by a mean of 2.96 [0.83–5.62] minutes compared to 1.58 [-0.31–4.84] minutes in the HS group. Preclinical microsurgical skill scores improved by 0.99 [0.50–1.63] in the RA group and by 0.76 [-0.10–1.30] in the HS group. In clinical studies (n=8), RA anastomosis time was reduced by 18.03 [6.30–37.27] minutes over anastomosis attempts. Preclinical Anastomosis Lapse Index scores were comparable between techniques. Reported immediate patency in clinical settings ranged from 95.66% to 100%. Conclusion Performance on dedicated robotic platforms improved with practice, but the current evidence did not demonstrate overall superiority over HS technique. In preclinical studies, final RA anastomosis and per-stitch times remained longer, while early anastomosis quality and patency were broadly comparable. Considering added time and platform-associated costs, the clinical settings in which robotic assistance may provide a meaningful benefit remain to be defined. Heterogeneity in models, operators, platforms, and outcome reporting limits definitive characterization of the learning curve. Future studies should standardize learning curve methodology, report complete sequential trajectories, and define clinically relevant proficiency thresholds before integration into credentialing frameworks.