Patient Selection
The study population consisted of patients who underwent robotic surgery at Shizuoka Cancer Center in Japan between December 2011 and December 2016 for primary lower rectal adenocarcinoma with an inferior tumor edge located below the peritoneal reflection. Patients who had tumor of stage 0 or stage IV, who had synchronous or metachronous colorectal cancer, or who did not undergo preoperative magnetic resonance imaging (MRI) or computed tomography (CT) at our institution were excluded. This study was approved by the institutional review board of Shizuoka Cancer Center (institutional code: J2020-32-2020-1-3). Patient characteristics and surgical and pathological findings were recorded in a prospective database. These data included age, sex, body mass index (BMI), American Society of Anesthesiologists (ASA) score, carcinoembryonic antigen level, distance from tumor to anal verge, presence of preoperative chemoradiotherapy (CRT), operative procedure, presence of combined resection of adjacent organs, presence of lateral lymph node (LLN) dissection, and presence of adjuvant chemotherapy. Preoperative tumor staging was performed by digital examination, colonoscopy, CT, MRI, and barium enema. Patients were staged using the tumor node metastasis (TNM) classification [18]. The radial margin (RM) was defined as the distance between the maximum tumor infiltration depth and the cut surface of the mesorectum [19]. The status of residual tumor was defined as follows: R0, no residual tumor; R1, microscopic residual tumor; and R2, macroscopic residual tumor [19].
Treatment
Robotic surgery was indicated for patients with rectal adenocarcinoma of clinical stage 0-IV. Since it was not covered by national public health insurance in Japan until March 2018, it was a more costly treatment option than laparoscopic or open surgery. After provision of informed consent, robotic surgery was performed in all patients who desired it. The principle for radical surgery for lower rectal cancer is TME. In Japan, preoperative CRT has not been the standard treatment [20]. In our institution, indications for preoperative CRT are limited to avoid its toxicity. Even in patients with rectal cancer invading adjacent organs, we performed surgery without preoperative CRT in patients for whom it was predicted that obtaining a clear resected margin (R0) by TME with combined resection of invaded organs would be possible. Preoperative CRT was performed only if surgeons expected it to be difficult to obtain a clear resected margin (R0) without CRT, or if they anticipated that shrinkage of the tumor by CRT would make anal preservation possible or make it feasible to avoid urinary diversion [21]. The external radiotherapy dose was 45 Gy, administered using a four-field approach in 25 fractions to a large pelvic field over the course of 5 weeks, plus a boost of 5.4 Gy in three daily fractions. Concomitant chemotherapy with the 5-fluorouracil derivate capecitabine (825 mg/m2) was administered orally twice per day, 5 days per week. Surgery was performed 6 to 8 weeks after CRT. Indications for LLN dissection were lower rectal cancer of stage T3-4 or of stage T1-2 with suspected metastasis to LLNs, as described by the Japanese Society for Cancer of the Colon and Rectum Guidelines for the Treatment of Colorectal Cancer [20]. Intersphincteric resection was performed when the rectum could not be divided using linear staplers in the abdominal approach. If the tumor invaded the levator ani muscle, if it was classified as the macroscopic infiltrating type, or if fecal continence was impaired preoperatively, abdominoperineal resection was performed. Among patients with pathological stage III disease, 5-fluorouracil-based adjuvant chemotherapy was administered to those younger than 75 years who did not have any severe comorbidity. All treatment strategies, including operative approaches or procedures, were approved by a multidisciplinary team at our institution.
VFA Measurement
VFA was measured using cross-sectional CT scans obtained at the level of the umbilicus and analyzed using dedicated software (SYNAPSE VINCENT, version 4.6; Fujifilm Medical Systems., Inc., and FUJIFILM Corporation, Tokyo, Japan) according to a previously described protocol [8]. Visceral adipose tissue was identified by a Hounsfield unit range of -50 to -200. The region of visceral fat was determined by automatic counter tracing, and VFA was calculated automatically by the software.
MFA measurement
MFA was defined as the area of mesorectum at the tip of ischial spines measured on MRI according to the protocol proposed by Boyle et al [14]. All MRI exams used a high-resolution MRI 3.0-T system (Achieva 3.0T dStream; Royal Philips Healthcare, Amsterdam, the Netherlands), and all analyses were performed using dedicated software (SYNAPSE version 4.1; Fujifilm Medical Systems., Inc., and FUJIFILM Corporation). The circumferences of the mesorectum and rectum were manually traced at the proposed level with the calibrated cursor of the “freehand region of interest tool.” The areas of the delimited mesorectal and rectal regions were automatically calculated by the software based on the traced lines as shown on Fig. 1. MFA was then calculated by subtracting the rectal area from the mesorectal area. When a T3 or T4 tumor located at the level of the tip of the ischial spine transgressed the rectal wall and it was difficult to readily visualize the outer edge of the muscularis propria, the trace was continued along a line following the contour of the outer muscularis propria [14]. The measurements were made by a single observer (IS) who was blinded to the outcomes at the time of measurement.
Surveillance protocol
Surveillance was performed for 5 years postoperatively. The surveillance protocol at our institution was as follows: an interview, physical examination, and blood tests, including carcinoembryonic antigen and cancer antigen 19 − 9, were performed every 3 months for the first 3 years postoperatively and then every 6 months thereafter. Chest, abdominal, and pelvic CTs were performed every 6 months. Colonoscopy was scheduled annually for the first 3 years postoperatively. Recurrence was confirmed by pathological assessment or progressively increasing tumor size in imaging studies. Local recurrence was defined as a new lesion in the pelvis detected by CT or MRI, and distant recurrence was defined as recurrences other than local recurrence.
Statistical Analysis
Categorical variables are described as numbers and percentages, while continuous variables are presented as medians (range). Overall survival (OS) and relapse-free survival (RFS) rates were calculated using the Kaplan-Meier method. OS and RFS were defined as the time from proctectomy to death and the time from proctectomy to recurrence or death from any cause, respectively. The cumulative incidence was used to estimate local and distant recurrence rates with death as a competing risk. A Cox proportional hazards regression analysis was performed to identify clinicopathological factors related to RFS. Although there are several useful cut-off values for assessing obesity and leanness in the context of BMI and VFA [8, 22, 23], the cut-off value for MFA is still controversial. Therefore, all the anthropometric parameters were analyzed as continuous variables in the univariate and multivariate analysis. Variables with a p-value of < 0.10 in a univariate analysis were subsequently entered into a multivariate Cox model to assess variables independently associated with RFS. The role of MFA as a predictor of RFS was further assessed by classifying patients into small- and large-MFA groups based on a cutoff value of 20.7 cm2, which was reported to reflect surgical difficulty in rectal cancer surgery [16]. Clinicopathological factors and OS, RFS, and cumulative local and distant recurrence rates were compared between the two groups. Categorical variables were compared using the chi-square test, while continuous variables were compared using the Mann-Whitney U test. The log-rank test was used to compare OS and RFS, and Gray’s test was used to compare the cumulative local and distant recurrence rates between groups. A p-value of < 0.05 was considered to be statistically significant. The R package software version 3.4.1 (R Development Core Team, Vienna, Austria) was used for statistical analysis.