Comparison of learning curves and related postoperative indicators between endoscopic and robotic thyroidectomy: a systematic review and meta-analysis

Background:Endoscopic thyroidectomy (ET) and robotic thyroidectomy (RT) yield similar perioperative outcomes. This study investigated how the learning curve (LC) affects perioperative outcomes between ET and RT, identifying factors that influence the LC.Materials and methods:Two researchers individually searched PubMed, EMBASE, Web of Science, and Cochrane Library for relevant studies published until February 2024. The Newcastle-Ottawa Scale assessed study quality. A random-effects model was used to compute the odds ratio and weighted mean difference (WMD). Poisson regression comparison of the number of surgeries (NLC) was required for ET and RT to reach the stable stage of the LC. Heterogeneity was measured using Cochran's Q. Publication bias was tested using funnel plots, and sensitivity analysis assessed findings robustness. Subgroup analysis was done by operation type and patient characteristics.Results:This meta-analysis involved 33 studies. The drainage volume of ET was higher than that of RT (WMD=-17.56 [30.22, -4.49]). After reaching the NLC, the operation time of ET and RT was shortened (ET: WMD=28.15 [18.04-38.26]; RT: WMD=38.53 [29.20-47.86]). Other perioperative outcomes also improved to varying degrees. Notably, RT showed more refined central lymph node resection (5.67 vs. 4.71), less intraoperative bleeding (16.56 ml vs. 42.30 ml), and incidence of transient recurrent laryngeal nerve injury (24.59 vs. 26.77). The NLC of RT was smaller than that of ET (incidence-rate ratios [IRR]=0.64 [0.57-0.72]). CUSUM analysis (ET: IRR=0.84 [0.72-0.99]; RT: IRR=0.55 [0.44-0.69]) or a smaller number of respondents (ET: IRR=0.26 [0.15-0.46]; RT: IRR=0.51 [0.41-0.63]) was associated with smaller NLC. In RT, transoral approach (IRR=2.73 [1.96-4.50]; IRR=2.48 [1.61-3.84]) and retroauricular approach (RAA) (IRR=2.13 [1.26-3.60]; IRR=1.78 [1.04-3.05]) had smaller NLC compared to bilateral axillo-breast and transaxillary approach (TAA). In ET, the NLC of RAA was smaller than that of TAA (IRR=1.61 [1.04-2.51]), breast approach (IRR=1.67 [1.06-2.64]), and subclavian approach (IRR=1.80 [1.03-3.14]).Conclusions:Rich surgical experience can improve surgical results of ET and RT. After reaching the NLC, the perioperative outcomes of RT are better than those of ET. Study subjects, surgical approaches, and analysis methods can affect NLC.