Background

In order to assess the various surgical modalities for local resection of rectal tumors, a systematic review of the current literature and a network meta-analysis (NMA) was designed and conducted.

Methods

The present study adhered to the PRISMA guidelines and the Cochrane Handbook for Systematic Reviews of Interventions principles. Scholar databases (Medline, Scopus, Web of Science) were systematically screened up to 23/12/2019. A Bayesian NMA, implementing a Markov chain Monte Carlo analysis, was introduced for the probability ranking of the available surgical methods. Odds ratio (OR) and weighted mean difference (WMD) of the categorical and continuous variables, respectively, were reported with the corresponding 95% confidence interval (95%CI).

Results

Overall, 16 studies and 2146 patients were introduced in our study. Transanal minimal invasive surgery (TAMIS) displayed the highest performance regarding the overall postoperative morbidity, the perioperative blood loss, the length of hospitalization, and the peritoneal violation rate. Transanal endoscopic microsurgery (TEM) was the most efficient modality for resecting an intact specimen. Although transanal local excision (TAE) had the highest ranking considering operative duration, it was associated with a significant risk for positive resection margins and tumor recurrence.

Conclusions

In conclusion, TEM and TAMIS display superior oncological results over TAE. Due to several limitations, validation of these results requires further RCTs of a higher methodological level.

中文翻译：

---

经直肠手术局部切除直肠肿瘤的技术比较：网络荟萃分析。

背景

为了评估直肠肿瘤局部切除的各种手术方式，对当前文献进行了系统综述，并进行了网络荟萃分析（NMA）。

方法

本研究遵守PRISMA指南和干预原则系统评价的Cochrane手册。截至2019年12月23日，系统地筛选了学者数据库（Medline，Scopus，Web of Science）。为实现可用手术方法的概率分级，引入了实施马尔可夫链蒙特卡洛分析的贝叶斯NMA。分别报告分类变量和连续变量的几率（OR）和加权平均差（WMD），并带有相应的95％置信区间（95％CI）。

结果

总体而言，我们的研究引入了16项研究和2146例患者。经肛门微创手术（TAMIS）在总体术后发病率，围手术期失血，住院时间和腹膜侵犯率方面表现最佳。经肛门内窥镜显微手术（TEM）是切除完整标本的最有效方法。尽管考虑到手术持续时间，经肛门局部切除术（TAE）的排名最高，但与阳性切缘和肿瘤复发的重大风险相关。

结论

总之，TEM和TAMIS的肿瘤学结果优于TAE。由于一些限制，对这些结果的验证需要更高方法水平的进一步RCT。