Short-term and long-term outcomes of robotic rectal surgery—from the real word data of 1145 consecutive cases in China

Background Due to a limited patient sample size, substantial data on robotic rectal resection (RRR) is lacking. Here, we reported a large consecutive cases from the real word data to assess the safety and efficacy of RRR. Methods From September 2010 to June 2017, a total of 1145 consecutive RRR procedures were performed in patients with stage I–IV disease. We conducted an analysis based on information from a prospectively designed database to evaluate surgical outcomes, urogenital function, and long-term oncological outcomes. Results Of three types of RRR performed, 227 (24.2%) were abdominoperineal resections, 865 (75.5%) were anterior resections, and 3 (0.3%) were Hartmann. Conversion to an open procedure occurred in 5.9% of patients. The overall positive circumferential margin rate was 1.3%. Surgical complication rate and mortality were 16.2% and 0.8% within 30 days of surgery, respectively. Mean hospital stay after surgery and hospital cost were 6.3 ± 2.9 days and 10442.5 ± 3321.5 US dollars, respectively. Risk factors for surgical complications included male gender, tumor location (mid-low rectum), combined organ resection, and clinical T category (cT3–4). Urinary function and general sexual satisfaction decreased significantly 1 month after surgery for both sexes. Subsequently, both parameters increased progressively, and the values 1 year after surgery were comparable to those measured before surgery. At a median follow-up of 34.6 months, local recurrence and distant metastases occurred in 2.3% and 21.1% of patients, respectively. Conclusions Robotic rectal resection was safe with preserved urogenital function and arrived equivalent oncological outcomes in a nonselected group of patients with rectal cancer.

A robotic approach with superior dexterity and precise movements of the robotic arms provides the surgeon with better exposure and greater ergonomic comfort during dissection of small anatomical structures [1,2]. Robotic rectal resection (RRR) has been proven to be a valid option with a lower open conversion rate than that of conventional laparoscopy for patients with rectal cancer [3]. Furthermore, robotic total mesorectal excision may overcome some difficulties associated with conventional laparoscopic rectal resection (LRR) [4] primarily in patients with mid-low rectal cancer [5]. However, the benefit of robotic rectal resection for urogenital function protection and oncologic outcomes remains controversial due to the limited sample size in reported studies [6,7].
Previous studies show LRR is associated with similar [8] or higher [9] rates of sexual and urinary dysfunction compared to open rectal resection (ORR). The incidence of urinary and sexual complications after RRR is still not well known. Limited data indicated robotic total mesorectal excision (TME) may allow for better preservation of urinary and sexual functions when compared with both ORR and LRR and Other Interventional Techniques Wenju Chang, Ye Wei, Li Ren and Mi Jian contributed equally to this work. 1 3 [10]. Furthermore, regarding oncological outcomes of RRR, based on evidence from a limited number of cases and initial experiences, substantive data are still lacking [11]. Few centers worldwide have the capacity to perform real world large-scale studies of RRR because of low-volume sample sizes. However, colorectal surgeons at Zhongshan Hospital were early adopters of RRR and have performed more than 1700 such procedures in a nonselected group of patients with rectal tumors.
The purpose of this study was to define the safety and function preservation of the robotic rectal surgery base on a real-world database from a single center over an 8-year period in China. The surgical complication and risk factors, sexual and urinary function, and long-term oncological outcomes are described in this study.

Patients
From September 2010 to June 2017, a total of 1211 consecutive patients are slated to undergo robotic rectal resection (RRR) with the Da Vinci S or Da Vinci Si Robot Surgical System. Of those, the robotic procedure was canceled in 66 cases after laparoscopic exploration revealed severe abdominal adhesions and intraperitoneal tumor dissemination. Ultimately, data from 1145 patients who underwent RRR were analyzed for this study. Patients were admitted to the study regardless of sex, age, AJCC/UICC (American Joint Committee on Cancer/Union for International Cancer Control) stage, type of intervention performed, or history of previous abdominal surgery. Data from patient hospital records were prospectively collected in a predesigned Excel file. Other clinical data were collected from computerized and archived patient charts. Postoperative data as well as surgical complications (30-day morbidity and mortality), sexual function, urinary function, and long-term oncological outcomes were assessed during follow-up.
A positive circumferential resection margin was defined as ≤ 1 mm from the specimen surface to the primary tumor or any tumor deposit. [12] Dissection planes are registered according to the description by Quirke. [13] The postoperative complications were defined as adverse events that occurred within 30 days after surgery. Complications were diagnosed and categorized according to patients' symptoms, with the aid of laboratory and radiological evaluation to confirm clinical suspicions. Grading of complications was scored based on the detailed tables of the Surgical Complications Severity Scoring System proposed by Mazeh et al. [14].
The study was approved by the Fudan University Ethics Committee, and all patients were asked to provide informed consent. Urinary and sexual dysfunctions affecting quality of life (QoL) were assessed by means of specific self-administered questionnaires in all patients undergoing robotic TME. For evaluating urinary tract symptoms and the impact on QoL, the International Consultation on Incontinence Male/Female Lower Urinary Tract Symptoms questionnaires (ICIQ-FLUTS and ICIQ-MLUTS) were used [15]. Each module uses a common question format. Most questions use 5-point Likert scales to assess the presence or absence of a symptom and its severity, followed by a scale to assess the associated degree of bother, which is measured by a visual analog scale. For assessing male sexual function, the International Index of Erectile Function (IIEF) questionnaire [16] was adopted, and for female sexual function, the Female Sexual Function Index (FSFI) questionnaire [17] was adopted. These are 15-item self-administered questionnaires that analyze five factors: erectile function (sexual function for female), orgasmic function, sexual desire, intercourse satisfaction, and overall satisfaction. All of these questionnaires can effectively assess urinary and sexual function after prostate cancer [18] and rectal surgery [19,20] as reported previously.

Robotic rectal resection
The single-docking technique with four or five ports was used as described in our previous studies [21]. We performed the high dissection and low ligation for lymph node dissection and preserved the left colic artery in most patients. The splenic flexure of the colon was not routinely mobilized, depending on the tension of the anastomosis. Once the sigmoid colon, mesocolon, entire rectum, and mesorectum were mobilized completely, anterior resection (AR) with the double-staple technique or abdominoperineal resection (APR) was performed accordingly.
We routinely used a standardized approach to prevent anastomotic leakage (AL) for mid-low rectal tumors, which included preserving the left colonic artery to improve the blood supply to the anastomosis. A transanal drainage tube was placed to reduce anastomotic tension in patients with robotic LAR. A diversion stoma was not routinely performed except in those patients at a substantial risk of AL. If the anastomosis was below the peritoneal reflex, the dissected pelvic peritoneum was sutured to avoid leakage of feces, which could cause intraperitoneal peritonitis following AL. At the same time, a double cannula was placed near the anastomosis to monitor the occurrence of AL. When AL occurred, the double cannula located near the anastomoses could be used to wash and drain the feces, promote anastomotic healing, and avoid a salvage ileostomy.

Statistical analysis
Statistical analysis was conducted with IBM SPSS Statistics software version 19 (SPSS Inc., IBM, Chicago, IL, USA). Categorical variables were analyzed using the χ 2 test, and continuous variables were analyzed using the Student t test. The nonparametric Kruskal-Wallis test was employed to compare the qualitative variables for urinary and sexual function analysis. One-way analysis of variance with least significant difference multiple comparisons was used for analysis of quantitative differences between multiple groups. A logistic regression was used for multivariate analysis. Overall survival was calculated using the Kaplan-Meier method. P values less than 0.05 were considered statistically significant.

Perioperative and pathological data
The 1145 patients who underwent robotic rectal resections (RRR) are summarized in Table 1 Tumor locations are detailed in Table 2. There were 367 (32.1%) patients with upper rectal tumors, 423 (36.9%) with middle rectal tumors, and 355 (31.0%) with low rectal tumors. Anterior resection (AR), low anterior resection (LAR), and abdominoperineal resection (APR) procedures were performed in 365 (31.9%), 500 (43.6%), and 277 (24.2%) patients, respectively. The mean operative time and estimated blood loss were 166.8 ± 31.6 min (range 106-720) and 73.8 ± 30.5 mL (range 5-400), respectively. Overall blood transfusion events totaled 16 (1.4%) within 30 days of surgery (Due to improve preoperative serious anemia (hemoglobin < 70 g/L) in 14 patients, and 2 patients suffered from major bleeding after operation). A diverting stoma was completed in 3 of 500 patients with LAR. The number of conversions to open procedures was 68 (5.9% conversion rate), of which, 60 cases were due to combined organ resection, and 8 for major bleeding or difficult tumor dissection. Following laparoscopic exploration, robotic procedures were canceled for 66 patients with severe abdominal adhesions or intraperitoneal tumor dissemination. The adjusted conversion rate was 11.1% (134/1211) if these 66 patients were calculated. The 133 (11.6%) patients accepted for combined organ resections are summarized in Table 1. A total of 181(23.7%) patients with T4 or N2 mid-low rectal cancer accepted preoperative radiotherapy or chemoradiotherapy. After surgery, 406 patients at high risk of relapse accepted adjuvant chemoradiotherapy, another 90 patients with unrespectable distant metastases accepted chemotherapy/target therapy after the primary tumor resection. The mean time of liquid diet and first flatus passage after surgery were 1.6 ± 0.1 days and 2.1 ± 0.5 days, respectively. The mean hospital stay after surgery and total hospital cost were 6.3 ± 2.9 days and 10442.5 ± 3321.5 US dollars, respectively ( Table 2).
The mean number of harvested lymph nodes was 17 ± 10.5 (range 5-54). The positive rates of circumferential margin (CRM) and distant margin (DRM) were 15 (1.3%) and 6 (0.5%), respectively. The surgical quality of mesorectal excision calculated in 777 patients who underwent LAR or APR, according to Quirkes' criteria, with "complete" in 706 (90.2%) patients and "near complete" in 71 (9.1%). No case was the mesorectum defined as "incomplete" by the pathologist. The pathological data are summarized in Table 3.

Surgical complication and risk factors
The number of overall surgical complications was 187 (16.3%). Grade 1 and grade 2 complications together accounted for 13.8%, while grade 3 and grade 4 complications were 2.0% and 0.4%, respectively. One patient died of hepatic failure after simultaneous hepatectomy for liver metastases within 30 days of the operation. The rehospitalization rate and reoperation rate associated with surgical complications within 90 days of surgery were 26 (2.3%) and 9 (0.8%), respectively.
We evaluated risk factors for surgical complications associated with RRR using a multivariate model, including factors that were statistically significant (P < 0.05) in a univariate analysis ( Table 4). The male gender, tumors located at the mid-low rectum, combined organ resection, and clinical T category (cT3-4) were confirmed as the independent risk factors for surgical complications by multivariate analysis.
Both parameters then increased progressively, and at 1 year after surgery, the values were comparable to those measured before surgery. These data are presented in Table 5.
Concerning urinary function, the grade of incontinence measured 1 year after the intervention was statistically unchanged when compared with the preoperative status for both sexes. These data are summarized in Table 6. In particular, in male patients, we observed no significant deterioration of voiding or incontinence during the entire study period. Filling symptoms and incontinence function in women were both statistically worse 1 month after intervention, 2.3 ± 0.6 versus 3.2 ± 0.8 (P = 0.024) and 1.5 ± 0.6 Table 2 Pathological data AJCC stage indicates the American Joint Committee on Cancer TNM classification, DRM distal resection margin, CRM circumferential resection margin (a positive circumferential resection margin was defined as ≤ 1 mm from the specimen surface to the primary tumor or any tumor deposit) a Total 1145 patients were analyzed, including 1139 patients with malignant tumor and 6 patients with benign tumors b Value expressed by Mean ± SD (range) c According to Quirkes' criteria [13]. Quality of mesorectum in 777 patients undergone LAR or APR were analyzed d Total 111 patients were not accepted radical resection due to unresectable distant metastases (n = 90) and R1 resection of primary tumor (n = 21)    With regards to the impact of urinary symptoms on patient's QoL, no difference was measured in either sex at 1 year compared to preoperative status (Table 7). For example, the scores of voiding QoL were 3.7 ± 1.1 (preoperation) versus 3.3 ± 1.1 (1 year, P = 0.162) for male and 3.7 ± 0.9 (preoperation) versus 3.7 ± 1.0 (1 year, P = 0.955) for female.

Discussion
We reported here the results of a large consecutive nonselected number of robotic rectal resections from real world to evaluate the safety and efficacy of robotic procedure for rectal cancer in China. In this study, we contributed to followed data analysis: (1) we analyzed the patients' short-term outcomes and identified high-risk patients for robotic rectal surgery. (2) We summarized the sexual and urinary function data, as well as the long-term oncologic outcomes in detail. All this data demonstrated that the robotic procedure was safe and efficacious for patients with rectal cancer, when performed in an experienced medical unit.
This study summarized the complication and identified the risk factor of surgical complications during RRR. The incidence of grade 3 and grade 4 complications together was 2.4% in this RRR study, which was lower than rates reported in previous laparoscopic rectal resection (LRR) studies of 7.4% [22] and 9.5% [23]. A systematic review demonstrated no differences in postoperative morbidity and mortality between LRR and ORR [2]. RRR may help to reduce severe surgical complications; therefore, we recommended patients at a high risk for surgical complications undergo robotic surgery for rectal cancer. More importantly, by applying a standardized robotic approach as described, we effectively reduced the incidence of anastomotic leakage (totally 4.2%) without including a routine diverting stoma. The rate of AL was 0.8% for AR and 6.6% for LAR, which was obviously lower than the reported 10-13% in recent studies [6,8,20]. It is worth noting that previous studies, even multicenter RCT studies, had low operative numbers performed by a single doctor; for example, 30 units completed 739 cases of LRR in the COLOR II study [8] and 40 surgeons completed 237 cases of RRR in the ROLARR study [6]. Limited case volume and inadequate surgical experience may compromise the quality of surgery [24,25], which may be a potential reason possibly responsible for the high rate of AL in previous studies.
Independent risk factors for surgical complications during RRR were male gender, tumors located at the mid-low rectum, combined organ resection, and clinical T category (cT3-4). Risk factors previously reported in LRR or ORR [22,26] studies, BMI ≥ 28, age ≥ 75, history of comorbidity, preoperative radiotherapy (RT) or chemoradiotherapy (CRT), and tumor size ≥ 5 cm, were not significant in our robotic procedure. These differences may be due to the diversity of patient characteristics: (1) The neoadjuvant RT or CRT was associated with an increased rate of anastomotic leakage (AL) as reported previously [26]. In this study, only patients with T4 or N2 disease received preoperative RT or CRT, and other patients with a high risk of relapse received chemoradiotherapy after radical surgery. The number of patients with mid-low rectal cancer who received neoadjuvant RT or CRT was 23.7% (181/778) in this study, unlike a greater proportion of CRT (46-58%) in Western countries [6,8]. (2) Regarding a history of comorbidities, this study only collected data from cardiovascular disease and diabetes; therefore, some data associated with other organ systems were missed. (3) The percentage of patients with a BMI ≥ 28 was about 18% in this study. BMI was significantly lower in Asian people than the patients in western who have a high proportion of obese patients. [27,28].
This study supplied larger cases analysis by questionnaires to evaluate the role of robotic TME on preservation of urogenital function. We recorded no difference in terms of incontinence, filling, or voiding symptoms at 1 year after surgery compared to the preoperative status for both sexes. This result was similar to that of a previous robotic study [10] and better than that of a LRR study, which was associated with a rate of sexual and urinary dysfunction higher [8] or comparable [29] to that following open surgery. Limitations of LRR can be explained by the technical complexities of laparoscopy surgery, including the unstable view of the operative field and the poor ergonomics of the surgical tools that render complex operations more difficult with a higher degree of surgeon fatigue. However, robotic surgery can overcome the limitations of laparoscopic surgery for rectal cancer, offering the surgeon a stable camera platform with a 3-dimensional operative field, precise and dexterous control of the wristed instruments to improve endo-wrist function, and reduced operator fatigue. As reported previously [30], the technical characteristics of a robotic system permits extended lymph node dissection and an accurate dissection of the smaller anatomical structures, thereby protecting the pelvic autonomic nerves.
In terms of the oncologic aspects, we found the 3-years DFS and OS were 81.0% and 87.2%, respectively, which was comparable to recent reports of LRR and ORR results [23,31]. This outcome can be interpreted as confirmation that the therapeutic effectiveness of RRR is equivalent to laparoscopic and open surgery. Interestingly, the local recurrence rate and positive circumferential margin rate were better than that of LRR [7,23] and ORR [31] in previous studies. Lim et al. [7] reported a lower local recurrence rate in RRR (2.7%) than that of LRR (6.3%) for patients with mid-low rectal cancer following neoadjuvant chemoradiation therapy. Baek et al. [32] classified 182 patients who underwent robotic surgery for rectal cancer into easy, moderate, and difficult groups by MRI-based pelvimetry and there was no difference between the groups in terms of operative and pathologic outcomes. Thus, an advantage of RRR may be its ability to overcome challenges associated with difficult pelvic anatomy and allow for a high quality of tumor resection.
The current study has several limitations in that it is single-arm large cases analysis with data from only a single center and is not a head-to-head designed study.

Conclusion
Based on the results, we determined robotic rectal resection to be a safe and adequate technique for the treatment of rectal cancer, with a low incidence of serious complications and anastomotic leakage. Robotic TME allows for preservation of urinary and sexual functions in patients with mid-low rectal cancer.