The study was performed as a retrospective analysis of patients in whom stoma creation was performed at the National Cancer Center Hospital East, Japan, between January and December 2016. During this period, a stoma was created in 202 patients. After exclusion of 42 patients with an end stoma and 7 who did not undergo postoperative CT, the final study population included 153 patients who underwent a temporary diverting loop ileostomy or colostomy after surgery related to malignant diseases. This study was approved by the ethics committee of National Cancer Center Hospital (no. 2017-350).
Data acquisition and follow-up
Data were collected for patient characteristics (sex, age, body mass index (BMI), abdominal wall thickness, rectus abdominis muscle thickness, comorbidities (diabetes mellitus (DM) or steroid use), history of open laparotomy, history of neoadjuvant chemotherapy (NAC) or preoperative CRT, emergency surgery, and American Society of Anesthesiologists (ASA) score), surgical characteristics (operation time, blood loss, blood transfusion, approach type (laparoscopy or open), stoma type (ileostomy or colostomy), specimen extraction from stoma site, stoma formed through the middle of the rectus abdominis muscle, height of stoma, incision size of stoma, base area of stoma, and volume of stoma).
The thickness of the abdominal wall and rectus abdominis muscle at the umbilical level were measured on preoperative CT by surgeons. The base area of the stoma was calculated from the vertical and horizontal size of the skin level. The stoma volume was calculated from the maximum vertical size, the maximum horizontal size, and the maximum height. These measurements of the stoma size were made postoperatively by the operating nurse. All measurements were made by a ruler. The position of the stoma through the rectus abdominis muscle was checked on postoperative CT, which was regularly performed for follow-up of malignant diseases. The rectus abdominis muscle was defined as the distance from the medial edge of the rectus abdominis muscle sheath to the lateral edge and then dividing the width of this distance into three equal parts. A stoma passing through the middle of the rectus abdominis muscle defined as one with the whole stoma passing through the center third of the rectus abdominis muscle (Fig. 1a). A stoma not passing through the middle of the rectus abdominis muscle was defined as one with the part of the stoma passing through the lateral third of the rectus abdominis muscle (Fig. 1b). All data were extracted from medical and operation reports. The condition of the stoma was observed during hospitalization of the patients by surgeons and wound, ostomy, and continence (WOC) nurses. Thereafter, patients were assessed during follow-up outpatient visits. For a temporary stoma, patients who did and did not receive postoperative adjuvant therapy were scheduled for stoma closure after 6 months and after 3 months, respectively. All patients were followed until closure of the stoma or until December 2017, when data were collected for the study.
Diagnosis of parastomal hernia and other stoma-related complications
Parastomal hernia was diagnosed as any protrusion around the stoma that was detected in physical examinations by surgeons or WOC nurses. Additionally, parastomal hernia was checked on postoperative CT in a supine position, which was regularly performed for follow-up of malignant diseases. The parastomal hernia based on CT was diagnosed according to the presence of a hernia sac (Fig. 1c) . The evaluation of CT examination was retrospectively reviewed by colorectal surgeon who was blinded to the patients’ clinical information. Peristomal skin disorders were diagnosed as a change in the skin, such as erythema, erosion, blister/pustule, and ulcer/tissue overgrowth. The degree of peristomal skin disorders was evaluated using the DET score. The DET score was developed by Martins et al. in 2008 as part of the Ostomy Skin Tool (OST) for evaluation of peristomal skin. The severity of discoloration (D), erosion (E), and tissue overgrowth (T) is evaluated between 0 and 2, and each area is evaluated between 0 and 3. The total DET score is the summation of severity scores and areas for D, E, and T respectively. The score is in the range from 0 to 15, where 0 represents normal skin and 15 is the worst combination of severity and extent [13,14]. DET ≥7 was defined as a case with severe peristomal skin disorders . Stoma outlet obstruction was defined was based on bowel dilation at the ostomy in clinical ileus, excluding other causes for bowel obstruction. Stoma prolapse was defined as a condition in which the full thickness of the bowel protruded through the stoma.
Surgical techniques for stoma creation
Each stoma was preoperatively marked in the middle of the rectus abdominis muscle by surgeons and WOC nurses. According to the lifestyle of patients, the patients were marked in the best site in each of standing, sitting and supine positions. The decision to create an ileostomy or colostomy was made by each operating surgeon. Depending on the surgeon's decision, a specimen was removed from a stoma site or a mini-laparotomy site. After extracting the specimen, the stoma was created using the skin bridge technique [15,16]. The skin bridge technique uses skin instead of a plastic rod to secure and prevent retraction of the stoma. The skin bridge is made by incising the skin at the predetermined site of stoma creation, and the subcutaneous fat is divided. The skin bridge is passed through the avascular window, which is opened in the mesentery at the apex of the chosen intestinal loop, and the skin bridge is sutured to the distal edge of the opening. All stoma had a transperitoneal route and were created in the middle of the rectus abdominis muscle based on preoperative stoma site marking. All operations were performed by surgeons with substantial experience with open and laparoscopic colorectal surgery.
Categorical variables are reported as the number (percentage) of patients, and analyzed by Chi-square or Fisher exact test, while quantitative variables are reported as a median and range, and analyzed by Mann-Whitney U test. Multivariate analysis was performed using logistic regression analysis with the backward stepwise method based on the variables in the univariate analysis required to be <0.05 and previous reported risk factors. All statistical analyses were performed using SPSS 22.0 (SPSS Inc., Chicago, IL, USA). Significance was defined as p < 0.05.