1. Introduction

Minimally invasive proctectomy has been shown to provide important advantages in the short term when compared to open proctectomy for the treatment of several malignant diseases with at least the same long-term survival.[1,2] However, although minimally invasive proctectomy (MIP) has been accepted worldwide for early and middle rectal cancer, its application in advanced rectal cancer has not been widespread.[3,4] With the development of MIP, more and more patients choose minimally invasive.[5] Locally advanced rectal cancer, once considered a contraindication of MIP, is also feasible and has been proven to be effective and safe.[6–8]

Although many articles have been published on the comparison of early and middle rectal cancer.[9–11] There are few studies on locally advanced rectal cancer.[10] The treatment plan and results of the 2 stages of rectal cancer are significantly different.[12,13] Therefore, the results of minimally invasive early and middle rectal cancer proctectomy cannot be directly applied to locally advanced rectal cancer. Therefore, we think it is necessary to carry out in-depth research.

2. Methods 2.1. Protocol and guidance

The study was performed according to Preferred Reporting Items for Systematic Reviews and the meta-analysis[14] and the quality evaluation of this article was scored using the Newcastle-Ottawa Scale (NOS) score. The protocol for this review has been registered on PROSPERO (CRD42023407029).

2.2. Search strategy

This study involved literature published in the Embase, PubMed, Cochrane Library, Medline, and Web of Science up to August 03, 2023. We defined the eligibility criteria according to the population(P), intervention(I), comparator(C), outcome, and study design approach(O). P: the patients with locally advanced colon cancer. I: undergoing MIP. C: open was performed as a comparator. O: one or more of the following outcomes: perioperative period, postoperative indices, and oncologic outcomes. The search terms included (minimally invasive OR laparoscopic [Mesh] OR robot [Mesh]) AND (Open [Mesh]) AND (advanced colon cancer [Mesh]). The search strategy was not limited by language or year. It was not requested by the ethics or institutional review committee due to the study being designed as a systematic review and meta-analysis.

2.3. Inclusion and exclusion criteria

We have included the literature by the following criteria. Comparative data were available on the treatment of locally advanced colon cancer through minimally invasive and conventional open proctectomy (COP). Outcome indexes should include at least one of the following, perioperative period, postoperative, and oncologic outcomes. Any study which did not confirm the above inclusion criteria was excluded.

2.4. Data extraction and outcome measures

Two researchers (L.D. and Y.L.) independently reviewed the retrieved literature by the inclusion and exclusion criteria. The third researcher (Z.Y.C) was asked to participate in the discussion to decide whether to include when disagreements were encountered. The extracted data included the first author, publication, country, study type, group, age, follow-up, tumor height, and tumor size (if mentioned) (Table 1).

Table 1 - The main characteristics of included studies. Author Publication Country Study period Study design Group Cases Age BMI(Body mass index) (kg/m2) ASA Follow-up (months) Confounders adjustment NOS score (max: 9) Zhang et al (2017)[15] Surgical Laparoscopy, Endoscopy and Percutaneous Techniques China 2007–2013 Retrospective MIP (LPN) 47 71.9 ± 10.1 21.2 ± 2.2 97.9 47 (22–106) No 6 COP 65 69.4 ± 10.5 20.9 ± 2.1 93.8 59 (22–106) Yang et al (2020)[16] Int J Colorectal Dis China 2010–2016 Retrospective MIP (LPN) 30 57.7 (12.6) 23.2 (2.5) 66.7 41.5 (3–98) No 7 COP 34 61.2 (13.3) 23.8 (2.7) 67.6 41.5 (3–98) Yamaguchi et al (2018)[17] Surg Endosc Japan 2002–2014 Retrospective MIP (RPN) 78 63 (36–78) 22.7 (16.7–29.7) 94.9 54 (13.6–135) Yes (propensity score matching) 8 COP 78 61 (30–79) 22.7 (16.2–37.1) 93.5 54 (13.6–135) Yamaguchi et al (2017)[18] Dis Colon Rectum Japan 2010–2011 Retrospective MIP (LPN) 118 59 (27–83) 90 (76.3) 98.3 43 Yes (propensity score matching) 8 COP 118 61 (40–84) 95 (80.5) 97.5 41.6 Yamaguchi et al (2016)[19] Surg Endosc Japan 2010–2014 Retrospective MIP (RPN) 85 63 (36–78) 22.8 (16.7–29.7) 92.9 No 8 COP 88 63 (26–84) 23.3 (16.2–30.2) 94.4 Watanabe et al (2021)[20] Surg Endosc Japan 2005–2016 Retrospective MIP (LPN) 93 64 (58.5–70.0) 23.0 (20.5–24.8) 55.6 (45.8–66) Yes (propensity score matching) 8 COP 93 64 (59.5–71.5) 23.0 (20.5–24.8) 55.6 (45.8–66) Sujatha-Bhaskar et al (2017)[5] Ann Surg USA 2010–2014 Retrospective MIP (RPN) 2009 57 ± 12 No 9 COP 3399 58 ± 11 Seshadri et al (2012)[21] Surg Endosc India 2004–2010 Prospective MIP (LPN) 72 48 (22–73) 21 (15–33) 99 Yes (propensity score matching) 8 COP 72 48 (19–71) 22 (14–38) 96 Sasi et al (2021)[22] Colorectal Dis India 2014–2017 Retrospective MIP (LPN) 293 47.49 98.3 Yes (propensity score matching) 7 COP 293 47.04 98.3 Nonaka et al (2017)[23] Anticancer Res Japan 2008–2014 Retrospective MIP (LPN) 27 64.9 No 6 COP 17 62.4 Nishikawa et al (2019)[24] Dis Colon Rectum Japan 2007–2015 Retrospective MIP (LPN) 31 70.2 ± 10.8 2.8 Yes (propensity score matching) 7 COP 50 67.4 ± 13.0 2.8 Lim et al (2023)[10] Surg Today Japan 2005–2020 Prospective MIP (RPN) 64 63 (SD 10.4) 22.4 (SD 3.3) 60 (48–70) Yes (propensity score matching) 6 COP 65 61 (SD 9.3) 23.3 (SD 3.1) 116 (78–142) Lee et al (2022)[7] J Laparoendosc Adv Surg Tech A Korea 2007–2017 Prospective MIP 30 56.83 23 No 7 COP 19 60.37 24 Law et al (2009)[25] Ann Surg Oncol China(Hong Kong) 2000–2006 Prospective MIP (LPN) 111 68.6 90.1 34 No 8 COP 310 67 85.8 34 Kim et al (2012)[26] Surg Endosc Korea 2001–2009 Retrospective MIP (LPN) 38 69.5 ± 8.8 23.3 ± 3.0 57.9 40 Yes (propensity score matching) 8 COP 16 68.1 ± 12.8 20.6 ± 2.4 62.5 35 Kusano et al (2014)[6] Jpn J Clin Oncol Japan 2002–2012 Retrospective MIP (LPN) 19 58 (32–82) 73.7%(<25) 35 No 6 COP 14 55 (39–73) 64.3%(<25) 40 Hida et al (2018)[27] Annals of surgery Japan 2010–2011 Retrospective MIP (LPN) 482 63.4 ± 13.1 22.5 3.6 95 39.6 Yes (propensity score matching) 9 COP 482 63.4 ± 10.9 22.4 3.4 94.4 39.6 Goto et al (2021)[28] Int J Colorectal Dis Japan 2008–2014 Retrospective MIP (LPN) 237 67 (60–73) 22.0 (20.2–24.0) 93.7 54 Yes (propensity score matching) 7 COP 237 67 (60–73) 22.5 (20.4–24.3) 94.1 54 Fleshman et al (2015)[29] Jama USA 2008–2013 Retrospective MIP (LPN) 242 57.7 (11.5) 26.4 (4.0) No 9 COP 239 57.2 (12.1) 26.8 (4.2) de’Angelis et al (2017)[30] Surgical Endoscopy USA 2005–2015 Retrospective MIP (LPN) 62 70 (37–89) 26 (18.5–33) 56.5 34.37 (23.09) No 8 COP 75 68 (28–86) 25 (16–34) 62.7 43.81 (32.54) Tang et al (2023)[31] Tech Coloproctol China 2012–2019 Retrospective MIP (LPN) 100 58 (30–85) 23.4 (17.5–33.0) 98 37 Yes (propensity score matching) 7 COP 100 57 (18–80) 23.8 (16.9–34.1) 100 46

Matching: (1) age; (2) BMI; (3) follow-up.

COP = conventional open proctectomy, LPN = laparoscopic invasive proctectomy, MIP = minimally invasive proctectomy, NA = data not available, NOS score = Newcastle-Ottawa Scale score, RPN = robotic invasive proctectomy.

2.5. Statistical analysis

Statistical analysis was performed by Stata v.12.0 (Stata Corp LLC, College Station, TX). For this meta-analysis, if the heterogeneity test was I2 > 50%, P < .1, we used the random effect model; if the heterogeneity test was I2 < 50%, P > .1, we used the fixed utility model. The combined r values and 95% confidence intervals (CIs) of each study were calculated, and the forest map displayed the characteristics of each study result. The quality of the included literature was evaluated using the Newcastle–Ottawa scale (NOS). The Begg and Egger tests were used to test the publication bias. The P < .05 was indicated as statistically significant.

3. Results 3.1. Eligible studies and study characteristics

We initially searched 1402 records. 686 literature that was published repeatedly and cross-published were deleted. After reading the title and abstract, 584 articles were excluded. After the remaining 132 pieces of literature were searched for full text, reading, and quality assessment, 21 pieces of literature[5–7,10,15–31] (10,132 participants: MIP: 4268 vs COP: 5864) were eventually included (Fig. 1: Guidelines Flow Diagram). The detailed information on this literature is listed in Table 1.

Figure 1.:

Flowchart for records selection process of the meta-analysis. (According to PRISMA template: Moher D, Liberati A, Tetzlaff J, Altman DG, The PRISMA Group (2009). Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med 6(7): e1000097. doi:10.1371/journal. Pmed 1000097).

3.2. Perioperative outcomes

Data on operation time (OP) were reported in 4 studies.[16,23,26,29] Compared with COP, patients who underwent MIP had less OP (standard mean difference [SMD] 0.41; CI 0.25 to 0.57; I2 = 76.3%, P = .000). Owing to high heterogeneity (I2 = 76.3%), sensitivity analysis was carried out by Stata 12.0. After removing the studies by Nonaka et al[23] as the sample that was “left out,” the pooled results did not change substantially but the heterogeneity was significantly reduced (SMD 0.48; CI 0.32 to 0.65; I2 = 0%, P = .000) (Fig. 2A). We included 5 studies[16,21,23,26,29] about estimated blood loss (EBL). Compared with COP, patients who underwent MIP had less EBL (SMD −0.67; CI −0.80 to −0.55; I2 = 95%, P = .000). Owing to sensitivity analysis and subgroup analysis cannot reduce heterogeneity, we choose the random-effects model (MD −1.23; CI −1.90 to −0.56; I2 = 95%, P < .0001) (Fig. 2B). We included 3 studies[26,29,30] about length of stay. Compared with COP, patients who underwent MIP had no statistical difference (SMD 0.03; CI −0.13 to 0.19; I2 = 0%, P = .745) (Fig. 2C). We included 3 studies[16,26,30] about the median time to a semi-liquid diet. Compared with COP, patients who underwent MIP had a less median time to semi-liquid diet (SMD −0.43; CI −0.70 to -0.15; I2 = 0%, P = .002) (Fig. 2D). Data on time to the first flatus were reported in 3 studies.[15,26,30] Compared with COP, patients who underwent MIP had less time to the first flatus (SMD −0.76; CI −1.01 to −0.50; I2 = 57.6%, P < .0001). Owing to high heterogeneity (I2 = 57.6%), sensitivity analysis was carried out by Stata 12.0. After removing the studies by De’Angelis et al[30] as the sample that was “left out,” the pooled results did not change substantially but the heterogeneity was significantly reduced (SMD −0.97; CI −1.30 to −0.63; I2 = 7%, P < .0001) (Fig. 2E).

Figure 2.:

Meta-analysis of minimally invasive proctectomy versus conventional open proctectomy for locally advanced colon cancer in perioperative outcomes: A: operation time, B: estimated blood loss C: length of stay, D: time to semi-liquid diet, E: time to the first flatus, F: intraoperative blood transfusion, G: retrieved lymph nodes, H: lymphatic invasion, I: venous invasion.

Data on intraoperative blood transfusion were reported in 7 studies.[6,18,20,26–28,31] Compared with COP, patients who underwent MIP had less intraoperative blood transfusion (relative risk [RR] 0.33; CI 0.24 to 0.46; I2 = 0%, P < .0001) (Fig. 2F). Data on the number of retrieved lymph nodes were reported in 4 studies.[10,16,22,24] Compared with COP, patients who underwent MIP had no statistical difference (SMD 0.06; CI −0.07 to 0.19; I2 = 68.1%, P = .386). Owing to high heterogeneity (I2 = 68.1%), sensitivity analysis was carried out by Stata 12.0. After removing the studies by Lim et al[10] as the sample that was “left out,” the pooled results did not change substantially but the heterogeneity was significantly reduced (SMD −0.01; CI −0.16 to 0.13; I2 = 29.9%, P = .240) (Fig. 2G).

We included 2 studies[6,10] about lymphatic invasion. Compared with COP, patients who underwent MIP had no statistical difference (SMD 0.84; CI 0.32 to 2.22; I2 = 0%, P = .724) (Fig. 2H). Data on venous invasion were reported in 2 studies.[10,24] Compared with COP, patients who underwent MIP had no statistical difference (RR 0.81; CI 0.60 to 1.09; I2 = 73.3%, P = .170). Owing to sensitivity analysis and subgroup analysis cannot reduce heterogeneity, we choose the random-effects model (RR 0.81; CI 0.60 to 1.09; I2 = 73.3%, P = .170) (Fig. 2I).

3.3. Postoperative outcomes

Data on overall complications were reported in 4 studies.[10,22,25,29] Compared with COP, patients who underwent MIP had lower overall complications (RR 0.85; CI 0.73 to 0.98; I2 = 22.4%, P = .023) (Fig. 3A). We included 7 studies[13,