Patient selection and ethics
We retrospectively evaluated prospectively collected data of 47 consecutive patients (34 men, 13 women; mean age, 67.4 years; mean body mass index, 23.4 kg/m2) with rectal cancer who underwent TaTME with SST between November 2014 and October 2020 (Supplementary Figure 1). The inclusion criteria were age 20–85 years, histologically proven adenocarcinoma, and tumors in the rectum. Patients were excluded if they had T4b tumors, bowel obstruction, perforation caused by the primary tumor, or an American Society of Anesthesiologists physical status classification of ≥IV. A flow diagram of patients in this study is shown in Supplemental Figure 1. Patients with non-adenocarcinoma or T4b tumors and patients who underwent intersphincteric resection (ISR), Hartmann, or abdominoperineal resection (APR) were excluded. Intraoperative adverse events, operative time, blood loss, short-term outcomes, and length of hospital stay (LOS) were assessed. Anastomotic leak was defined as a symptomatic leak treated within 30 days of the primary resection, including pelvic infection. The technique was also evaluated based on the full thickness of the resection ring of the colonic and rectal stumps.
Quadruple PSS consisted of double PSS, which used 0 polypropylene for each rectal (Fig. 1a, b) or colonic stump (Fig. 1c), taking full-thickness bites of the tissue through proper bites at the point of 3-5 mm distal to the stumps. Double PSS consisted of single PSS, which used 0 polypropylene for each rectal or colonic stump, taking full-thickness bites of the tissue through proper bites at the point of 3-5 mm distal to the stumps. We defined the morphology of the resection ring after removing 0 polypropylene from the rings as complete (circumferentially intact resection ring with the full thickness of the wall; Fig. 2a: rectal ring; Fig. 2b: colonic ring), nearly complete (resection ring with intact mucosa but partial defect of the muscularis propria; Fig. 2c: rectal ring; Fig. 2d: colonic ring), or incomplete (partial defect of mucosa and muscularis propria of the resection ring; Fig. 2e: rectal ring; Fig. 2f: colonic ring). Ileostomy due to resection rings was defined as ileostomy creation due to nearly complete or incomplete resection rings. Intestinal perfusion of the proximal colon was assessed in the fluorescent imaging mode following intravenous injection of indocyanine green (ICG). If the serosal fluorescent stain of the proximal colon was poor, extended resection of the colon was performed.
This study was reviewed and approved by the ethics committee of the Kagoshima University Hospital.
Surgical procedure
All patients were placed in a modified lithotomy position and subjected to general anesthesia. After setting the self-retaining anal retractor (Lone Star Retractor; CooperSurgical, Trumbull, CT, USA), an access device (GelPOINT path; Applied Medical, Rancho Santa Margarita, CA, USA) was introduced through the anus to the rectum. After temporarily clamping the rectosigmoid using an atraumatic endo bulldog clip (Aesculap AG, Tuttlingen, Germany), the pneumorectum was maintained at 15 mmHg with carbon dioxide using an AirSeal platform (AirSeal System; CONMED, Utica, NY, USA). A double PSS was applied clockwise using 0 polypropylene with a 26-mm rounded needle to tightly occlude the rectum that was 3 cm distal to the tumor.
After irrigation with saline and marking the dissection line by tattooing the rectal mucosa distal to the mucosal folds, mucosal dissection of the rectum was initiated. Full-thickness rectal transection was performed circumferentially. When the tumor’s height was located in the middle of the rectum, rectal transections became challenging. Therefore, we used laparoscopic coagulation shears to dissect the mesorectum. The dissection proceeded toward the presacral plane between the parietal endopelvic fascia and mesorectal fascia in the loose areolar tissue. The dissection proceeded toward the peritoneal reflection between Denonvilliers’ fascia (rectovaginal septum) and the mesorectal fascia on the anterior side. The dissection then proceeded behind the neurovascular bundle (NVB), keeping the layer of Denonvilliers’ fascia intact on both lateral sides and identifying the pelvic nerve using the NVB and endo-pelvic fascia as landmarks.
Before performing PSS, the rectourethral muscle was dissected. Four points using a blue dye were marked at 3, 6, 9, and 12 o’clock. A PSS was performed in a clockwise manner using 0 polypropylene with a 26-mm rounded needle that had 12 motions (Fig. 3a, b) (Supplementary Video). In the quadruple PSS group, another PSS was placed on the distal side of the initial PSS to tightly occlude the rectal stump for future quadruple PSS (Fig. 3c, d) (Supplemental Video).
After performing PSS, the rectosacral fascia on the posterior side and the peritoneal reflection on the anterior and both lateral sides were dissected to connect to the abdominal field in cooperation with the abdominal surgical team. After transabdominal extraction of the specimen, the proximal colon was prepared by inserting a detachable circular stapling anvil (29-mm or 25-mm CDH; Ethicon Endo-Surgery, Cincinnati, OH, USA) and securing a single or double PSS around the center rod. Then, SST with end-to-end coloanal anastomosis (Fig. 4a, b) was implemented using a circular stapler (29-mm or 25-mm CDH) and an anastomotic technique described by Penna et al. [18].
A diverting ileostomy was created when the completeness of the resection ring was graded as nearly complete or incomplete. Other indications for creating the ileostomy were a positive air inflation test result, obese men, poor preparation of feces, stage IV tumors, post neoadjuvant therapy, ultra-low anterior resection (ULAR), lateral lymph node resection, or ASA of 3 (a patient with severe systemic disease). We defined the ULAR as dividing the rectum around the anorectal junction. A suction drain was placed at the deep pelvis through the left lower quadrant of all patients. In addition, another suction drain was placed transanally.
Statistical analysis
Data are presented as median and interquartile range for continuous variables and as the frequency and percentage for categorical variables. When appropriate, the categorical variables were compared with the Chi-square test or Fisher’s exact test. Non-parametric variables are presented as median values and ranges. Wilcoxon’s t-test was used to determine the significance of differences between continuous variables. A P-value <0.05 was considered statistically significant. All statistical analyses were performed using JMP version 16.0 software (SAS Institute, Cary, NC, USA).