Learning curve for robotic thyroidectomy using BABA: CUSUM analysis of a single surgeon's experience.

Background: This study assessed the safety and oncologic outcomes of robotic thyroidectomy via the bilateral axillary breast approach (BABA RT) for conventional open procedures. The learning curves of BABA RT were further evaluated.

Methods: An exact 1:1 matching analysis was performed to compare the technical safety and oncologic outcomes between robotic thyroidectomy and conventional open surgery. Learning curves were assessed using cumulative summation analysis.

Results: There was no significant difference in general characteristics, short time outcomes (including transient hypoparathyroidism, transient postoperative hoarseness, hematoma/seroma, mean postoperative hospital stay, and other complications), the number of retrieved central lymph nodes, and recurrence rates between robotic BABA and conventional groups. The mean number of retrieved lateral LNs in the robotic group was significantly less than those in the conventional group. The learning curve for working space making, robotic lobectomy, and total thyroidectomy are approximately 15, 30, and 20 cases, respectively. No differences except for operation time were found between the learning group and the proficient group.

Conclusions: Robotic thyroidectomy and neck dissection via BABA are feasible in terms of surgical completeness, surgical safety, and oncological safety. Our results provide a criterion for judging whether the surgeon has entered the stable stage of robotic thyroidectomy via BABA in terms of the operative time.