Laparoscopic Natural Orifice Specimen Extraction Surgery versus Conventional Surgery in Colorectal Cancer: A Meta-Analysis of Randomized Controlled Trials

Objective This study was to quantitatively synthesize data in randomized controlled trials (RCTs) of laparoscopic resection comparing natural orifice specimen extraction (NOSE) versus conventional laparoscopy (CL) in colorectal cancer. Methods We identified eligible RCTs by searching seven electronic databases (PubMed, Cochrane Library, Embase, Web of Science, CNKI, CQVIP, Wanfang, and Sinomed). Mean differences (MDs) between groups with 95% confidence intervals (CIs) were used for continuous outcomes. Event rate ratios (RRs) were also calculated with their 95% CIs. Results 1,569 citations were identified from electronic database as of June 2020, and finally, 21 RCTs involving 2,112 patients met the study eligibility criteria and were included. Compared to the CL group, NOSE had longer operation time (MD: 8.14 min, 95% CI: 3.02 to 13.25, and p < 0.01), less estimated blood loss (-10.64 ml, 95% CI: -14.92 to -6.36, and p < 0.01), less hospital stay after surgery (-2.21 days, 95% CI: -3.36 to -1.06, and p < 0.01), shorter time of gas passage after surgery (-0.58 days, 95% CI: -0.82 to -0.34, and p < 0.01), better pain score (-1.06, 95% CI: -3.74 to -0.37, and p < 0.01), and improved cosmetic scores (1.93, 95% CI: 0.77 to 3.10, p < 0.01). Rate ratios of total complications, infection, and incision infection all favored NOSE surgery, with RRs (95% CIs) of 0.81 (0.71 to 0.93), 0.34 (0.21 to 0.54), and 0.24 (0.12 to 0.51), respectively. Conclusion This report appeared the first comprehensive meta-analysis of RCTs to synthesize data of laparoscopic resection with NOSE versus conventional laparoscopy. NOSE surgery seemed favorable with shorter hospital stay, less pain score, a shorter time to recover along with better cosmetic scores, and less postoperative complications.


Introduction
Colorectal cancer (CRC) remains one of primary causes of cancer-related morbidity and mortality worldwide [1]. As one of the treatment options, laparoscopic surgery has been accepted for decades widely [2]. In recent years, natural orifice specimen extraction surgery (NOSES) is gradually practicing in CRC's treatment and hence causes widespread interests among surgeons [3]. It is reported that NOSE surgery would reduce access trauma in laparoscopic colorectal surgery, with alleviated postoperative pain, faster patient recovery, and a favorable long-term outcome regarding cosmesis and incisional hernia rate [4]. However, a NOSE surgery guideline with adequate evidence has not been formulated to date yet. There were also negative arguments that NOSE surgery may be a risk factor of bacterial contamination of the peritoneal cavity [5]. Nevertheless, relevant studies on NOSE are increasing year by year while few metaanalyses, especially of randomized controlled trials (RCTs), have been carried out. As a result, this topic is still at the level of insufficient evidence [4,6]. Given these, we carried out this meta-analysis study of RCTs in a hope to summarize laparoscopic resection data comparing NOSE versus conventional laparoscopy in colorectal cancer. a diagnosis of colorectal cancer as study disease and compared the laparoscopic resection with NOSE versus conventional laparoscopic surgery, and (3) the report language was Chinese or English. If more than one article reported data from the same study, the most recent and complete articles were included. However, those studies without any valid information on resection outcomes were removed.

Data Extraction and Quality Assessment.
In this metaanalysis between laparoscopic resection with NOSE surgery (NOSE group) and conventional laparoscopy (CL group), the following data were extracted from each eligible individual study: (1) the name of first author; (2) year of publication; (3) study groups and number of patients; (4) baseline characteristics such as age and sex; and (5) resection outcomes including operation time, estimated blood loss, gas passage after surgery, various complications, and duration of hospital stay.
Two investigators utilized a uniform structured extraction sheet to extract data from included RCTs. If any disagreement was noted, a third investigator was asked to     scale was composed of 6 items, including the following: (i) randomization (yes scored 2 points and no scored 0), (ii) blinding (yes scored 2 and no scored 0), (iii) description of withdrawals and dropouts (yes scored 1 point and no scored 0 points), (iv) inclusion/exc1usion criteria (yes scored 1 point and no scored 0 points), (v) adverse effects (yes scored 1 and no scored 0), and (vi) statistical analysis (yes scored 1 and no scored 0). 4 Gastroenterology Research and Practice reach a final agreement. The potential risk of study bias was assessed according to the preferred reporting items for systematic reviews and meta-analysis recommendations [7]. The level of evidence was evaluated by using the Oxford Levels of Evidence [8,9]. Study quality was assessed by using the modified Jadad scale, which involves six items to evaluate the methodological quality of RCTs [10][11][12]. Its score range was 0 to 8, with a higher score showing better report quality. In this study, a score of 1 to 3 indicated low quality and 4 to 8 for high quality.

Gastroenterology Research and Practice
2.4. Statistical Analysis. We used R 3.4.4 (R Foundation for Statistical Computing, Vienna, Austria; http://www.Rproject.org/) and the Meta package [13] for this metaanalysis. For continuous outcome data, mean differences (MDs) along with their 95% confidence intervals (CIs) were used as their main effect measures. When the mean and standard deviation were not provided directly, we estimated them from the median, range, and size of the study samples [14]. For binary event data, the rate ratios (RRs) were calculated with 95% CIs. Heterogeneity was defined as an I 2 value of more than 50% [15] or p value of less than 0.10 from Cochrane Q test [16]. These two statistics evaluate the percentage of variability attributable to study heterogeneity instead of by chance. Therefore, when an outcome measure showed negligible heterogeneity, we used a fixed-effect model for its data pooling instead of random-effects model. The funnel plots were visually inspected for the measures of most included RCTs being conducted to statistically evaluate publication bias [15]. For any statistical test, significance was defined as a two-tailed p value of 0.05 or less.

Search Results and Study
Characteristics. Initially, 1,569 citations were identified from electronic database as of June      Gastroenterology Research and Practice    Of them, 88 studies were excluded due to their inappropriate study population or thesis type. Finally, a total of 21 RCTs  involving 2,112 patients met the study eligibility criteria and were included (Figure 1).
Only patients from the NOSE group or the CL group according to laparoscopic resection methods were included in our meta-analysis. Four studies [38][39][40][41] published as thesis and not in peer-reviewed journals were excluded. One study [42] with a printing error but was repaired and one study [43] in Russian were excluded. For four studies with more than two arms, we removed the open surgery group from two studies [19,27] and the laparoscopic surgery plus a traditional nursing group [30] or combined two NOSEStype arms into one [32]. The main study characteristics are shown in Table 1.

Study Quality and Publication Bias.
The results of quality assessment by the modified Jadad scale were as follows: two articles scored 6, five scored 5, twelve scored 4, one scored 3, and one scored 2. In summary, 19 out of 21 studies earned a score of 4 or more. All of the 21 articles were on RCT design and met 1b level of evidence. These generally suggested their high study quality ( Table 2).

Recurrence and Overall Survival.
Disease recurrent data were reported in five studies and overall survival in two studies (Table 6). No significant differences for both survival-related outcomes were found between the two groups: RR of 1.08, 95% CI 0.64 to 1.83, and p = 0:7791 for event recurrence rate and 1.08, 95% CI 0.92 to 1.27, and p = 0:3514 for overall survival rate ( Figure 5).

Discussion
To our knowledge, this report appeared the first comprehensive meta-analysis to synthesize RCT data regarding NOSE versus traditional laparoscopic colorectal cancer surgery. The large-sized meta-analysis of 21 RCTs demonstrated that laparoscopic resection with NOSE surgery reduced intraoperative estimated blood loss, relieved postoperative pain, accelerated postoperative recovery, and decreased the incidence of postoperative complications as well.
The terminology regarding NOSE surgery means that the surgical specimen resection is conducted intra-abdominally, and then, the specimen is taken out by opening a hollow organ such as anus, vagina, or mouth to communicate with the outside of the body [44]. Laparoscopic surgery combined with NOSE avoids incisions on the abdominal wall and reduces pain and wound complications, along with a shorter recovery time, etc. [45]. Besides, there was no auxiliary incision on the abdominal wall, and only a few small puncturing scars remained, indicating an excellent minimally invasive effect [46]. Given these reasons above, it was expected that NOSE surgery showed a better prognosis in terms of intraoperative data, postoperative recovery, and complications. NOSE surgery had less estimated blood loss (approximately 11 ml), and it may be due to no auxiliary incision, reducing the amount of wound bleeding. In the meantime, these results suggested that patients in NOSE group had less postoperative pain, faster recovery than the CL group, which also might be due to no auxiliary incision. The incidence of postoperative complications is an important indicator to evaluate the feasibility of NOSES. The total postoperative complication results suggested a significantly lower risk of complications (RR = 0:62, 95% CI 0.48 to 0.82, and p = 0:0006), especially in the incision infection. Therefore, in recent years, great advances in NOSES lead to a new tendency in CRC's surgical therapy in China and even other countries around the world. Given these, "Expert consensus of natural orifice specimen extraction surgery in colorectal neoplasm (2019)" and "International consensus on NOSES for colorectal cancer (2019)" were published along with individual reports [44,46].
On the other hand, however, NOSE surgery had a slightly longer mean operation time (8 minutes) as compared to the CL group. The reasons behind it may include (1) the operation space inside the natural cavity is narrow so that the anastomosis is more time-consuming and (2) surgical proficiency of the surgeon with a possible learning curve. Beginners require a learning process to perform this new type of surgery. As for disease recurrence and overall survival rate, there was no significant difference noted between the NOSE group and the CL group, suggesting that there was likely no significant difference in the long-term efficacy. For postoperative complications, new studies with adequate sample size may be also needed to differentiate them later in the future. Even so, laparoscopic NOSES was, to some extent, a safe extraction method for colorectal diseases.
There were several limitations in this report. First, the meta-analysis was based on secondary study-level data, and the evaluation indicators varied greatly among different RCTs. Low quality of RCTs (2 out of 21 RCTs scored less than 4 by the modified Jadad scale) might influence the pooled results. Unlike one meta-analysis report recently published with only one RCT included [47], we only included RCTs (n = 21). Second, few studies reported the disease recurrence and overall survival data and the like. For them, it was difficult to adequately measure the long-term efficacy of NOSE surgery. Third, of the 21 included RCTs, one was reported in Belgium, one was in Hong Kong, China, and the others were all reported in mainland China. The enrolled studies were not widely distributed all over the world, which would limit the study finding to extrapolate further. Last, different operation skills and study population might induce potential bias among the included RCTs. Therefore, a large-sized wellcontrolled RCT is warranted to further verify the advantages and disadvantages of NOSES after following a uniform surgery guideline.

Conclusion
This report appeared the first comprehensive meta-analysis to quantitatively synthesize data from RCTs of laparoscopic resection with NOSE versus conventional laparoscopy. Compared with CL, NOSE surgery demonstrated multiple advantages in terms of shorter hospital stay after surgery, less pain, faster recovery from surgery, better cosmetic results, and most importantly, fewer postoperative complications. Even so, well-controlled RCTs of the NOSES following a uniform surgery guideline are warranted in the future.

CRC:
Colorectal cancer CI: Confidence interval CL: Conventional laparoscopy ERAS: Enhanced recovery after surgery MD: Mean difference NA: Not applicable NOSE: Natural orifice specimen extraction NOSES: Natural orifice specimen extraction surgery NOTES: Natural orifice transluminal endoscopic surgery