LAUSR

Natural-orifice specimen-extraction surgery (NOSES) for colorectal disease is not a new surgical procedure, although its lack of penetration into common colorectal surgical practice may make it a relatively new procedure. Stewart et al. [1] were among the first to report the extraction of a colectomy specimen through the vagina in 1991 and, shortly thereafter, Franklin et al. [2] published the first report of partial colectomy with naturalorifice specimen extraction (NOSE) via the anus. Since then, there have been several publications on the extraction of both malignant and benign colonic diseases from the caecum to the distal rectum through natural-orifice extraction sites. Unlike procedures such as laparoscopic cholecystectomy, which was introduced in 1985 and rapidly diffused into the surgical community over a 2to 3-year period, the dispersion of NOSES has not been equivocal [3]. As there are several reasons why some new surgical innovations may be taken up more quickly than others, the adoption curves of new procedures can take many forms. However, the tipping point that describes the onset of the peak rate of diffusion of the new technology usually occurs after the first 10%–20% of users have adopted it [3]. From a global perspective, the adoptions curve for NOSES has not obtained a peak rate of diffusion and appears to have arrested in the developmental and explorative phase of innovation [3]. The recent publication by Guan et al. [4] from the International Alliance of NOSES provides a succinct description of the classifications and indications for NOSES procedures. Strategies in reducing the risk of bacterial contamination of the peritoneal cavity and oncological safety when extracting the specimen are discussed. However, in order to properly evaluate NOSES, we need to ask the following questions: (i) Does it make sense? (ii) Are there any significant short-term benefits? (iii) Are there any adverse complications? (iv) Are there any long-term oncological implications? Surgery is generally very slow to scrutinize the rapid progression of new surgical innovations until Level-1 evidence such as randomized control trials (RCTs) have shown them to be effective. However, it is challenging to evaluate a new surgical procedure in an RCT due to many potential practical problems: recruiting patients may be difficult, as they may refuse to be randomized; measuring appropriate outcomes may require years of follow-up; there may be differences in the surgical skills of the techniques being analysed; therefore, analysis should take account of how experienced each surgeon is in performing the new operation. Ideally, randomization should begin as soon as it is feasible, as this would enable the researchers to monitor the learning curve. The latter was emphasized in the recent ROLARR randomized clinical trial, which compared the effect of robotic vs laparoscopic surgery for rectal cancer. The median laparoscopic cases performed in the laparoscopic arm was 91 vs 50 in the robotic arm and, despite a recruitment of 471 patients, there was no difference in the primary endpoints of conversion to open laparotomy and positive rate of circumferential resection margin. A subsequent publication exploring and adjusting for potential learning effects showed that the initial ROLARR analysis was confounded by the learning effect and that the estimated odds ratio of conversion in the robotic arm was significantly lower after 70 cases were performed when compared with the laparoscopic arm with a median of 91 cases [5].