Learning Curve of Robotic End-to-Side Microanastomoses

Abstract

BACKGROUND AND OBJECTIVES: Robotics are becoming increasingly widespread within various neurosurgical subspecialties, but data pertaining to their feasibility in vascular neurosurgery are limited. We present our novel attempt to evaluate the learning curve of a robotic platform for microvascular anastomoses. METHODS: One hundred and sixty one sutures were performed and assessed. Fourteen anastomoses (10 robotic [MUSA-2 Microsurgical system; Microsure] and 4 hand-sewn) were performed by the senior author on 1.5-mm caliber tubes and recorded with the Kinevo 900 (Zeiss) operative microscope. We separately compared interrupted sutures (from needle insertion until third knot) and running sutures (from needle insertion until loop pull-down). Average suture timing across all groups was compared using an unpaired Student's t test. Exponential smoothing (α = 0.2) was then applied to the robotic data sets for validation and a second set of t tests were performed. RESULTS: We compared 107 robotic sutures with 54 hand-sewn sutures. There was a significant difference between the average time/stitch for the robotic running sutures (n = 55) and the hand-sewn running sutures (n = 31) (31.2 seconds vs 48.3 seconds, respectively; P-value = .00052). Exponential smoothing (α = 0.2) reinforced these results (37.6 seconds vs 48.3 seconds; P-value = .014625). Average robotic running times surpassed hand-sewn by the second anastomosis (38.8 seconds vs 48.3 seconds) and continued to steadily decrease with subsequent stitches. The average of the robotic interrupted sutures (n = 52) was significantly longer than the hand-sewn (n = 23) (171.3 seconds vs 70 seconds; P = .000024). Exponential smoothing (α = 0.2) yielded similar results (196.7 seconds vs 70 seconds; P = .00001). However, average robotic interrupted times significantly decreased from the first to the final anastomosis (286 seconds vs 105.2 seconds; P = .003674). CONCLUSION: Our results indicate the learning curve for robotic microanastomoses is short and encouraging. The use of robotics warrants further study for potential use in cerebrovascular bypass procedures.