Exploring and analyzing learning curves in robotic-assisted thoracoscopic anatomical lung resections: A systematic review and meta-analysis

Abstract

Objectives: There has been a steady increase in the uptake of robotic-assisted thoracic surgery over recent years, with a necessary focus on training. The early learning curve has been extensively debated; however, a detailed understanding of how this extends as we gain experience has been poorly discussed. This study assesses the congruency and depth of the learning curve in robotic-assisted thoracic surgery of anatomical lung resection. Methods: All studies reporting a quantitative assessment of operator learning curve in robotic anatomical lung resection before March 1, 2024, were included. Two authors extracted data, and study quality was assessed according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. Meta-analysis was performed using random effects modeling. Results: Twenty-nine studies including 106 surgeons were identified. A triphasic learning curve was identified with "competency" achieved at 20 (interquartile range, 13) cases, and "proficiency" at 60 (interquartile range, 37.5) cases. Weighted mean difference operating time between novice (P1) and proficiency (P2) was 34.1 minutes and between "competency" (P2) and "proficiency" (P3) was 17.5 minutes. This decreased with newer generations of robotic technology (Da Vinci Xi, weighted mean difference P1 vs P3: 42.3 [22.0, 62.5] vs S/Si, 57.4 [45.6, 69.2] minutes). Conclusions: The learning curve in anatomical lung resection is triphasic, with a reproducible extension beyond the initial proficiency phase. As robotic lung resection becomes more widespread, we must better understand the translation of this learning pattern among trainee surgeons to maintain excellence in clinical outcomes while facilitating training of the next generation.