Since Miles introduced the technique in 1908, abdominoperineal resection (APR) has been the gold standard treatment option for distal rectal cancer. However, improved surgical techniques, better instrumentation, and increased understanding that a tumor-free margin of 1–2 cm is oncologically safe, together with advances in preoperative chemoradiation therapy, have led to a shift in the treatment of rectal cancer to sphincter-preserving surgery [13]. In particular, coloanal anastomosis (CAA) using a double-stapling technique after ultralow anterior resection (uLAR) or intersphincteric resection (ISR) has been adopted as an alternative to abdominoperineal excision for very low rectal tumors since its introduction [4]. Some papers have reported that this operative technique is a safe procedure for sphincter-saving rectal surgery in selected patients that yields acceptable functional and oncologic outcomes.

In the last decade, laparoscopic surgery has been popular for rectal cancer surgery, including CAA, as in other fields of surgery, with its use conferring advantages such as a magnified view, better cosmesis, shorter recovery times and hospital stays, and comparable long-term oncologic outcomes [57]. However, limitations such as two-dimensional visualization, the fulcrum effect, restricted degrees of motion of the laparoscopic instruments, and amplification of physiologic tremors continue to be drawbacks, and laparoscopic rectal cancer surgery still poses a significant challenge for surgeons.

Recently, the use of robotic surgical systems in the field of rectal cancer surgery has been attempted to overcome the technical limitations of laparoscopy [810]. With robotic surgical systems, an excellent three-dimensional stereoscopic view may be obtained with high illumination. Furthermore, adequate traction and countertraction can be easily achieved in a narrow pelvis using the Endo Wrist function. These benefits may be especially helpful in CAA [1115]. However, studies on the application of this novel technique to uLAR and CAA with or without ISR are lacking, and few studies have compared robotic and laparoscopic surgery. Therefore, we designed this study to compare the short- and long-term outcomes between robotic and laparoscopic uLAR and CAA with and without ISR.

Materials and methods

Patients

Between January 2007 and December 2010, a retrospective chart review was conducted for all patients with a diagnosis of low rectal cancer who underwent curative uLAR and CAA with or without ISR using a robotic or laparoscopic approach in the Yonsei University Health System, South Korea. Mid or low rectal cancer patients were enrolled in this study regardless whether they received neoadjuvant chemoradiation or had involvement of the internal anal sphincter. Patients were excluded from the study if they had undergone open surgery, had tumors invading the levator ani or external sphincter, or had T4 cancers invading the prostate or vagina.

All the patients had preoperative assessment using rectal magnetic resonance imaging (MRI), transanal endorectal ultrasound, and chest and abdominal computed tomography (CT) scan. If the patient received neoadjuvant therapy, the assessment was done before and after chemoradiotherapy. For the patients who received neoadjuvant chemoradiation therapy, we used a standard long-course regimen consisting of 5-fluorouracil (FU) for chemotherapy and a radiation protocol of 45 Gy/25 fractions followed by a 5.4-Gy boost, for a total of 50.4 Gy. Postoperative morbidity was stratified by the Accordion Severity Grading System of surgical complications as recommended by Clavien et al. [16], and only significant surgical complications requiring endoscopy, an interventional procedure, or surgical reoperation with the patient under general anesthesia were considered (grade ≥ 3).

Adjuvant chemotherapy was administered based on the pathologic report as guided by the National Comprehensive Cancer Network (NCCN) guidelines. Patient follow-up assessment was performed routinely every 3 months until 3 years and every 6 months until 5 years. Colonoscopy was performed at the 2- and 5-year visits and within 1 year if preoperative colonoscopy could not be completed due to an obstructive lesion. Chest and abdominal CT scans were used to detect local and systemic recurrence at 6 months, 1 year, and every year until 5 years.

Operative techniques

The procedures were performed by four expert colorectal surgeons using either a robotic or a laparoscopic approach. For the robotic technique, we used total robotic dissection with either a single- or dual-docking technique depending on the need for splenic flexure mobilization and surgeon preference. Port placement depended on the docking type, and in both situations, we used a right-handed dominant third arm [8, 17].

After colonic mobilization, complete rectal mobilization was performed until we reached the pelvic floor. The need for ISR was evaluated at this point, as well as whether a stapling CAA or a hand-sewn anastomosis was required depending on the tumor height from the anal verge, the involvement of the internal anal sphincter, and the ability to insert the stapling device.

After we had transected the distal rectum at the closed level of the anorectal ring using an endolinear stapler, we extracted the specimen through an abdominal incision and then performed stapling anastomosis with a circular staple device. In cases requiring a hand-sewn anastomosis, we divided the distal rectum transanally above the dentate line with preservation of the sphincter, extracted the specimen through the anus, and then performed hand-sewn anastomosis.

For cases requiring ISR, pelvic dissection was advanced to the intersphincteric plane between the puborectalis muscle and the anal canal intraabdominally. Then, a perineal approach was used to dissect along the intersphincteric plane starting at the white line of Hilton, and the internal sphincter muscle was pushed upward until it met the abdominal dissection plane. After specimen extraction through the anal canal, hand-sewn anastomosis of the colon and anal canal was performed.

The laparoscopic technique was performed in a manner similar to the robotic technique except for port placement. We placed one periumbilical port for the camera and two ports for the surgeon, with the one port located at the right lower quadrant, and the other port located at the right lumbar area in the midclavicular line. Two additional ports were placed at locations mirroring the surgeon’s ports on the left side for the assistant.

Statistical analysis

Data were analyzed using the SPSS statistical program (Statistical Product and Service Solution 18 for Windows; SPSS Inc., Chicago, IL, USA). The Chi square test was used for categorical variables. Student`s t test and the Mann–Whitney U test were used for continuous variables and for statistical comparisons of perioperative clinical and oncologic outcomes. Overall survival and disease-free survival were analyzed using the Kaplan–Meier method, and comparisons were performed using the log-rank test. All p values of 0.05 or lower were considered statistically significant.

Results

In this study, 84 consecutive patients underwent uLAR and CAA for low rectal adenocarcinoma. Of the 84 patients, 47 received robot-assisted surgery, whereas the remaining 37 underwent a laparoscopic procedure. The demographic data for patients in the laparoscopic and robotic groups are summarized in Table 1. No significant differences in age, gender, American Society of Anesthesiologists (ASA) score, body mass index (BMI), tumor height from the anal verge, preoperative chemoradiation, or carcinoembryonic antigen (CEA) level were observed between the two groups.

Table 1 Demographic data

Table 2 shows the operative data for the two groups. Rectal division was primarily performed using a stapling device in both groups (89.2 vs 87.2 %), and ISR was performed for only 10.8 % of the laparoscopic group and 14.9 % of the robotic group (p = 1.000). Most of the patients underwent straight-type anastomosis (89.2 vs 85.1 %; p = 0.748) by the hand-sewn method (86.5 vs 76.6 %, p = 0.279). No differences between the two groups were observed in protective stoma formation (97.3 vs 85.1 %; p = 0.073), operative time (360.7 vs 352.7 min; p = 0.737), or estimated blood loss (302.7 vs 190.9 ml; p = 0.087), but conversion to open surgery was significantly more frequent in the laparoscopic group (16.2 vs 2.1 %; p = 0.020).

Table 2 Operative data

The laparoscopic and robotic groups did not differ significantly in terms of overall postoperative morbidity (27.0 vs 19.1 %; p = 0.439) (Table 3). Most complications were in the mild or moderate category by the Accordion Severity Classification (81.1 vs 87.2 %, p = 0.547). Differences in time to soft diet intake between the groups were not shown, but the hospital stay was significantly shorter in the robotic group (11 vs 9 days; p = 0.011). Two patients in the laparoscopic group were readmitted, whereas five patients were readmitted in the robotic group (5.4 vs 10.6 %; p = 0.690). No mortality occurred postoperatively in either group.

Table 3 Postoperative mortality and morbidity

Pathologic characteristics are presented in Table 4. The tumor node metastasis (TNM) stages were distributed evenly. The mean number of retrieved lymph nodes was 14.1 in the laparoscopic group and 10.6 in the robotic group (p = 0.061). Tumor sizes, grades of differentiation, proximal and distal resection margins, and lymphovascular invasions did not differ significantly between the two group. Three patients had involvement of the circumferential resection margin (≤1 mm) in the laparoscopic group compared with one patient in the robotic group (8.1 vs 2.1 %; p = 0.316).

Table 4 Pathologic characteristics

The median follow-up period was 31.5 months (range 1.4–61.3 months). Two patients had local recurrence in the laparoscopic group compared with three patients in the robotic group (5.4 vs 6.4 %; p = 1.00) (Table 5). Local recurrence took place at the anastomotic site, mesentery, lateral pelvic sidewall, or ovary. Systemic metastasis during the follow-up period was observed at a rate of 13.5 % in the laparoscopic group and 14.9 % in the robotic group (p = 1.00).

Table 5 Oncologic outcomes

The sites of metastasis in the laparoscopic and robotic groups respectively included the liver (2.7 vs 4.25 %), lung (5.4 vs 8.51 %), and paraaortic lymph node (2.7 vs 2.12 %). In addition, peritoneal carcinomatosis was reported for one case (2.7 %) in the laparoscopic group. The 3-year overall survival rate was 90.7 % in the laparoscopic group and 86.5 % in the robotic group (p = 0.404) (Fig. 1). No statistically significant difference in 3-year disease-free survival was observed between the two groups (81.2 vs 80.6 %; p = 0.914).

Fig. 1
figure 1

Kaplan–Meier survival curves showing the difference between the laparoscopic and robotic groups in a overall survival (p = 0.404) and b disease-free survival (p = 0.914)

Discussion

The last two decades have seen remarkable advances in the field of sphincter-preserving surgery for rectal cancer. These advances can be attributed to multimodal treatment, such as preoperative chemoradiotherapy, and the development of new surgical devices and techniques.

Laparoscopic uLAR is an alternative to conventional open surgery for very low rectal cancer. Park et al. [18] reported the results for a series of 130 consecutive patients who underwent laparoscopic ISR for low rectal cancer compared with 80 patients who underwent open surgery. The laparoscopic approach had results comparable with those for open surgery in terms of overall and disease-free survival during the intermediate-term follow-up evaluation, with favorable short-term postoperative outcomes. These authors concluded that laparoscopic ISR is a technically feasible and safe alternative to laparotomy, with favorable short-term postoperative outcomes.

Although patients benefit from laparoscopic surgery, its ergonomic discomfort and counterintuitive instrumentation still are stressful for surgeons. The laparoscopic approach for ultralow rectal cancer presents a considerable technical challenge, especially considering the competing objectives of achieving adequate oncologic clearance and preserving anal sphincter function.

The first concern regarding laparoscopic uLAR and CAA is the technical complexity of the surgical procedures. Robotic surgery systems were introduced to minimize the difficulty of laparoscopy. Robot-assisted surgery is especially useful and prevalent in surgical fields with procedures that must be performed in narrow spaces, such as in cardiac, gynecologic, and urologic surgery.

For the same reason, robotic surgery has been rapidly adopted for uLAR with or without ISR. The technique is especially helpful for rectal dissection, which can extend as far as the anal canal deep to the level of the puborectalis muscle.

In a study of 29 consecutive patients, Leong et al. [14] reported that robotic ISR for very low rectal cancer is feasible and safe. However, studies evaluating robotic uLAR and CAA with or without ISR have been limited to date, especially studies comparing robot assistance with laparoscopy.

In the current study, we compared the short- and long-term results for patients who underwent robotic and laparoscopic uLAR and CAA with and without ISR. Robotic CAA had short- and long-term outcomes comparable with those for laparoscopy. In particular, we found that the patients in the robotic group demonstrated significantly lower rates of conversion to open surgery and had a shorter hospital stay than the patients in the laparoscopic group. A low rate of conversion to open surgery has a very important impact clinically because the conversions often are associated with a high complication rate and a poor prognosis.

The rate of conversion to open surgery in the laparoscopic group of our study was high (16.2 %), similar to the rates of 3–12 % reported in other published series [1921]. On the other hand, the rate of conversion in the robotic group was only 2.1 %, which is especially low considering that this study included our initial experience with the technique. This result suggests that short- and long-term prognoses may differ among patients.

The low rate of conversion in robotic surgery may be due to its advantages of providing an appropriate space by stable traction and excellent dexterity of articulation within the narrow confines of the pelvis. In general, the reasons for conversion were adhesions, fibrosis, and a too-narrow pelvis, but the robotic surgical system is ideally suited to overcome these difficult situations during an operation. The shorter hospital stay also may have been related to the low conversion rate, although the two groups did not differ in terms of postoperative complications. Therefore, robotic surgery for low rectal cancer may be potentially more beneficial than a laparoscopic approach.

The current study showed no difference between the groups in operative time. This is in contrast to most studies, which have shown robotic surgery to require a longer operative time than laparoscopic surgery. In their comparative study between robotic and laparoscopic ISR, Park et al. [22] analyzed the short-term clinical outcomes for 40 patients per group and found that the mean operative time was significantly longer in the robotic group than in the laparoscopic group (235.5 vs 185.4 min). Perhaps no difference was observed in our study because making the already time-consuming and technically challenging dissection easier compensated for the additional time required to perform robot-assisted surgery. This compensation is magnified for a more difficult surgery such CAA compared with a less time-consuming surgery such as anterior resection. Another contributing factor was that the surgeons participating in the study were experts in robotic surgery with a large volume of experience.

When a new technology for operating on cancer is evaluated, an analysis of the oncologic outcomes is an important part of the consideration. In our study, we found comparable long-term results between the robotic and laparoscopic groups even during 3 years of follow-up evaluation. Furthermore, the oncologic results during a 3-year median follow-up period proved to be a valid surrogate marker for the standard end point of 5 years so that comparable 5-year outcomes also could be expected in our study. In particular, total mesorectal excision completeness, including the circumferential resection margin, is a very important factor influencing local recurrence and related to surgical quality [2325]. It is one of the most important benefits that can be expected from robotic assistance. Currently, a multicenter, randomized, controlled trial is in progress to assess total mesorectal excision completeness and local recurrence in Korea, including assessment at our institution, and we are expecting further data on oncologic outcomes that support robotic rectal surgery in the near future.

This study had some limitations that deserve mentioning. First, it was a retrospective study, with the possibility of some bias. Second, the follow-up period was relatively short, and the number of patients was low.

Nonetheless, the current study also had several important advantages. It was a comparative study of robotic and laparoscopic CAA, unlike most previous studies that included only a single arm for evaluating robotic surgery. It also included patients who underwent all types of uLAR rather than ISR alone, and such studies are very rare. Moreover, we showed some potential benefits of the robotic surgical system, such as a low rate of conversion to open surgery and a shorter hospital stay. A randomized prospective study including a larger population is needed in the future, especially a study including an analysis of functional outcomes.

In conclusion, our study demonstrated that robotic uLAR and CAA are feasible and safe to perform for patients with very low rectal cancer. The results are not inferior in terms of short- and long-term outcomes compared with laparoscopic surgery, and robotic assistance also may be superior in terms of conversion to open surgery and length of hospital stay. Future studies with a larger population and a longer follow-up period are clearly necessary to establish the advantage of robotic CAA.