Learning curves and outcomes of robotic colorectal surgery: A single-surgeon experience within a structured dual-console training program

Robotic colorectal surgery has gained broad acceptance, but defining the learning curve and safety profile during structured implementation remains essential. This single-surgeon study aimed to analyze procedural proficiency and oncological outcomes during the introduction of robotic colorectal surgery within a dual-console training framework. All robotic colorectal resections performed between September 2021, and December 2024 were retrospectively analyzed. Procedures included robotic anterior resection (R-AR), robotic low anterior resection (R-LAR), and robotic right colectomy (R-RC). Cumulative sum (CUSUM) and risk-adjusted CUSUM (RA-CUSUM) analyses were used to evaluate operative performance and safety. Demographic, perioperative, and histopathological parameters were analyzed descriptively. A total of 102 procedures were performed by the surgeon as the primary console operator (R-AR = 46, R-LAR = 32, R-RC = 24). Mean operative times were 163 ± 44 min (R-AR), 228 ± 56 min (R-LAR), and 154 ± 25 min (R-RC), respectively. Distinct learning curve turning points were observed for two procedures, with proficiency reached after approximately 28 R-AR cases and 16 R-RC cases. In contrast, R-LAR did not show a clear turning point but demonstrated a prolonged plateau between cases 6 and 27. Conversion rate was 1%, major complications (≥ Clavien–Dindo IIIb) occurred in 4.9%, and there were no intraoperative adverse events. The R0 resection rate exceeded 97% (R-AR: 100%, R-LAR: 96.9%, R-RC: 95.8%), and mean lymph-node yield was 28.2 ± 13. CUSUM and RA-CUSUM curves confirmed stable performance and consistent oncological quality throughout the learning phase. Robotic colorectal surgery can be safely implemented in academic centers within a structured dual-console training environment. Procedural proficiency was achieved after 28 rectal resections and 16 right colectomies.