Robotic-laparoscopic rectal cancer excision versus traditional laparoscopy.

BACKGROUND AND OBJECTIVES: Robotic surgery has been advocated for the radical excision of rectal cancer. Most data supporting its use have been reported from European and Asian centers, with a paucity of data from the United States documenting clear advantages of the robotic technique. This study compares the short-term outcome of robotic versus laparoscopic surgery.
METHODS: Consecutive patients who underwent laparoscopic (group 1) or robotic (group 2) rectal cancer excision at a single institution over a 2-year period were retrospectively reviewed. The main outcome measures were operative time, blood loss, conversion rates, number of lymph nodes, margin positivity, length of hospital stay, complications, and readmission rates.
RESULTS: Forty-two patients were analyzed. The median operative time was shorter in group 1 than that in group 2 (240 minutes vs 260 minutes, P=.04). No difference was noted in blood loss, transfusion rates, intraoperative complications, or conversion rates. There was no difference in circumferential or distal margin positivity. The median length of stay was shorter in group 1 (5 days vs 6 days, P=.05). The 90-day complication rate was similar in both groups (33% vs 43%, P=.75), but there was a trend toward more anastomotic leaks in group 1 (14% vs 0%, P=.23). Similarly, a non-statistically significant trend toward a higher readmission rate was noted in group 1 (24% vs 5%, P=.18).
CONCLUSION: Robotic rectal cancer excision yielded a longer operative time and hospital length of stay, although immediate oncologic results were comparable. The need for randomized data is critical to determine whether the added resource utilization in robotic surgery is justifiable.

INTRODUCTION

Robotic surgery (RS) was adopted by urologists for radical prostatectomies in the treatment of prostate cancer during the late 1990s despite the lack of strong scientific data to support its advantages over the laparoscopic technique.[1] As economic pressures continued to drive the use of RS, more data became available and the advantages of RS were eventually shown, including lower blood loss and transfusion rates while achieving similar oncologic outcomes.[2-4]

Despite the introduction of laparoscopic colorectal excision 2 decades ago, its wide adoption has been slow because of a variety of factors, including technical difficulty and unproven oncologic equivalency to open surgery in the setting of rectal cancer.[5] To date, there are no randomized level I data available from North America comparing the outcome of rectal cancer patients treated with open versus laparoscopic surgery. The much-awaited multicenter randomized American College of Surgeons Oncology Group (ACOSOG) Z6051 trial is currently in its final year of accrual.[6] Until the outcome of this study is published, it is unclear how the laparoscopic technique compares with open conventional excision for rectal cancer. Despite this lack of randomized data, there has been a surge of interest in RS for rectal cancer excision. Numerous recent publications have reported the techniques and short-term results of RS.[7-15] The vast majority of comparative data have been reported from Asia and Europe,[16-21] with a paucity of data from North America.[22,23]

The goal of our study was to review our early experience with robotic excision of rectal cancer and compare it with the outcome of laparoscopic surgery.

METHODS

The study was approved by the Institutional Review Board of Kaiser Permanente Southern California. A prospectively collected database was retrospectively reviewed for all consecutive cases performed either laparoscopically (group 1) or robotically (group 2) during the study period (February 2011 to February 2013). All cases were performed by 1 colorectal surgeon (M.A.A.) at Kaiser Permanente Los Angeles Medical Center, a regional tertiary institution in Southern California serving a population of approximately 3.5 million patients. Once RS was introduced in our institution in May 2012, all minimally invasive pelvic cases from that date forward were performed robotically with the da Vinci S Surgical System (Intuitive Surgical, Sunnyvale, California). All patients were prepared with polyethylene glycol–electrolyte solution the day before their operation and received perioperative intravenous antibiotics within 1 hour of incision (2 g of cefazolin and 500 mg of metronidazole). Deep venous thrombosis prophylaxis was given preoperatively and continued postoperatively with subcutaneous heparin.

The inpatient and outpatient electronic health care records were reviewed. Data collected included patient-related preoperative factors such as demographic characteristics, body mass index (in kilograms per square meter), and relevant comorbidities. Intraoperative data included type of operation, operative time, blood loss and need for transfusions, and intraoperative complications. Oncologic measures comprised pathologic stage, number of lymph nodes harvested, and circumferential and radial margin positivity. Postoperative outcomes included hospital length of stay, intensive care unit admission, narcotic pain medication use, overall complications, need for readmission, and need for reoperation.

Laparoscopic Technique

A 4-trocar technique was used for laparoscopic rectal excision, including an 11-mm supraumbilical port, a 12-mm right lower quadrant port, and two 5-mm trocars in the right and left mid abdomen. The mesorectal dissection was performed with the Ligasure device (Covidien, Mansfield, Massachusetts). The extraction site was either via a Pfannenstiel incision, a left lower quadrant incision, or anal/perineal wounds for intersphincteric proctectomy or abdominoperineal resection. The rectum was transected laparoscopically with an Endo GIA stapler (Covidien) or, for very low tumors, a TA stapler (Covidien) through the Pfannenstiel incision.

Robotic Technique

For the robotic technique, a 5-trocar technique was used, with three 8-mm robotic trocars in the right lower quadrant, left mid abdomen, and left lateral abdomen; a 12-mm camera port in the right supraumbilical area; and a 5-mm assistant trocar in the right upper quadrant. The mesorectal dissection was performed robotically using both monopolar and bipolar energy. If splenic mobilization was required, it was performed laparoscopically. The specimen extraction and rectal division were similar to the laparoscopic cases.

Statistical Analysis

To analyze the significance of data points between the laparoscopic and robotic groups, 2-tailed P values were calculated by use of the Fisher exact test for categorical variables and paired t test for continuous variables. P < .05 was the criterion for statistical significance. All statistical analyses were performed with SPSS software, version 16.0 (SPSS, Chicago, Illinois).

RESULTS

During the study period, 21 consecutive patients underwent laparoscopic rectal cancer excision (group 1) and 21 subsequent consecutive patients underwent robotic excision (group 2). summarizes the patients' characteristics. There were no differences in group 1 versus group 2 with regard to gender (57% male vs 48% male, P = .76), age (median of 62 years vs 60 years, P = .21), body mass index (median of 27 kg/m2 vs 25 kg/m2, P = .23), and prior abdominal operations (19% vs 29%, P = .71). Both groups were similar in terms of relevant comorbidities, with hypertension, hyperlipidemia, and diabetes being the most common.

highlights the operative findings of the 2 groups. The most common operation in both groups was proctectomy with coloanal anastomosis, and most patients in both groups underwent diversion (ileostomy in 52% in group 1 vs 62% in group 2, P = .76). There was a longer operative time in the robotic group compared with the laparoscopic group (median of 260 minutes in group 2 vs 240 minutes in group 1, P = .04). There was no difference in blood loss (median of 100 mL in group 1 vs 150 mL in group 2, P = .76), need for intraoperative transfusions (5% in each group), number of intraoperative complications (1 major bleeding complication in each group), and rate of conversion to an open procedure.

The oncologic data are presented in . The 2 groups had similar oncologic characteristics in terms of tumor grade and postoperative stage, although group 2 consisted of lower tumors overall (median distance from anal verge to tumor of 6 cm in group 2 vs 8 cm in group 1, P = .02). The number of lymph nodes harvested was clinically similar in the 2 groups, although they differed statistically (median of 15 lymph nodes in group 1 vs 17 lymph nodes in group 2, P = .03). There was no difference in circumferential or distal margin positivity (P > .99).

compares the postoperative outcomes of laparoscopic versus robotic rectal excision. Group 2 had a longer hospital stay (6 days in group 2 vs 5 days in group 1, P = .05) and received intravenous narcotic pain medications for a longer period (5 days in group 2 vs 3 days in group 1, P = .003). There were no differences in overall complications at 90 days, but there was a trend toward more anastomotic leaks in group 1 (14% vs 0%, P = .23). There was also a trend toward more readmissions in the laparoscopic group (24% vs 5%, P = .18).

DISCUSSION

There is a paucity of data from North America comparing the outcome of robotic and laparoscopic rectal cancer excision.[22,23] The aim of this study was to review the early rectal cancer robotic experience at a tertiary center of a colorectal surgeon with significant expertise in laparoscopic pelvic and rectal surgery (10 years' experience with advanced laparoscopy). Starting in May 2012, all minimally invasive rectal pelvic surgery cases (benign and malignant) were conducted robotically, thus minimizing selection bias. The open technique was reserved for patients in need of pelvic exenteration or those undergoing reoperative pelvic surgery. Comparison of the 2 consecutive groups showed no major difference in patient-related factors or tumor characteristics except for a higher number of low rectal tumors in the robotic group. The median operative time was longer by 20 minutes when the procedure was performed robotically. Whether this finding is related to the early experience of the surgeon with RS or inherent to the additional time required to dock and undock the robot is unclear. As our experience with RS continues to grow, we plan to reassess this outcome in the future. The immediate oncologic outcome appeared similar in both groups. A higher number of lymph nodes and a longer specimen length were noted in the robotic group, but these findings can be attributed to the fact that 10% of the robotic patients had a proctocolectomy for rectal cancer in the setting of chronic ulcerative colitis. Surprisingly, a longer length of stay and higher narcotic pain medication use were noted in the robotic group. It is unclear whether this finding is related to our early experience with RS, is related to the small number of patients in the study, or is a true finding inherent to RS possibly because of the significant torque exerted by the robotic arms on the abdominal wall. A higher anastomotic leak rate was noted in the laparoscopic group, which led to a higher readmission rate. This finding was unexpected. We plan to examine this finding further as we gather a large number of robotic cases. If this finding persists, a plausible explanation may be the difference in dissection method in terms of both energy source use and its impact on thermal heat transmission to the anorectal stump and/or visualization.

There were some limitations to this study. Its small and retrospective nature may inherently introduce bias, and whether the results can be projected to the general population is uncertain. The single colorectal surgeon in this study had extensive prior experience in advanced laparoscopy of the pelvis for both benign and malignant disease. It is unclear whether the results can be duplicated by lower-volume surgeons, those with less laparoscopic experience, and/or those transitioning from open surgery to RS. In addition, this study looked at the immediate oncologic outcome and did not examine the impact of RS on long-term oncologic outcome. Furthermore, no functional data or quality-of-life data were gathered as part of this study.

Despite the acknowledged shortcomings, this study provides valuable insight as to whether RS offers any immediate advantages over laparoscopy for rectal cancer excision. The benefits of RS to the surgeon have been advocated as one of the reasons to increase the use of RS.[24] It is ergonomically more comfortable and offers more degrees of freedom versus traditional laparoscopy and thus may limit physician fatigue. Visualization is improved by the 3-dimensional vision of RS and full surgeon control of the camera from the console. Whether such surgeon benefits translate into better outcomes for the patient remains unknown.

A longer operative time is widely considered to be a significant disadvantage of RS.[25] Our study found a longer operative time in the robotic group compared with the laparoscopy group, a finding previously reported by other authors.[19-21,23] A few studies have actually suggested a shorter operative time with RS, although none of their data regarding operative time reached statistical significance.[16,17,22] In meta-analyses comparing the 2 techniques, operative times were significantly shorter in the laparoscopic group, although the heterogeneity was high.[26,27] As expertise in RS continues to grow and proficiency improves, it is conceivable that the operative time may decrease.

RS must prove at least oncologic equivalency relative to open and/or laparoscopic surgery to justify its continued use. The number of harvested lymph nodes is known to have an important impact on survival.[28] No comparative studies published to date have shown any statistically significant differences in the total number of lymph nodes obtained.[16-23] The same is true regarding radial and distal margins. Our study showed no difference in the rate of circumferential resection or distal margin involvement, and the distal margin distance was comparable between the RS and laparoscopic groups. The meta-analysis of 7 comparative studies performed by Memon et al[26] in 2012 showed no risk difference for circumferential resection margin involvement between the two methods (P = .77) and no difference in mean distal resection margin distances (P = .84). More inclusive meta-analyses from Yang et al[27] from China and Trastulli et al[29] from Italy showed similar oncologic equivalency between the 2 groups. Thus, from an oncologic perspective, early nonrandomized data suggest that RS provides the same immediate oncologic outcome as laparoscopy. However, should the gold standard for comparison be laparoscopic surgery? In reality, we still do not have solid randomized data to conclude whether laparoscopic excision yields the same oncologic outcome as open surgery.[30] It is hoped that the ACOSOG Z6051 study comparing open versus laparoscopic excision for stage II and III rectal cancer will shed some light on the oncologic outcome of laparoscopy.[6] It is interesting to note that the laparoscopic arm of the trial includes a subgroup of robotically operated patients. Whether the number of such patients is large enough to make a meaningful comparison is unclear at this stage, and we will have to await the trial results. The ROLARR (Robotic Versus Laparoscopic Assisted Resection of Rectal Cancer) study is currently under way, and it is hoped that it will provide sufficient comparison between laparoscopy and robotics.[31]

Laparoscopic surgery for colorectal diseases has proven to yield better early postoperative clinical outcomes compared with open surgery,[30] but it is unclear whether RS provides any additional benefit. Our study showed a longer mean hospital length of stay in the robotic group. No other study in the literature has shown this finding. One plausible explanation may be the higher number of abdominoperineal resection and proctocolectomy cases in the robotic group. In the first-ever comparison study, published in 2009, Baik et al[16] found a statistically significant difference in hospital length of stay, favoring a shorter course with RS (5.7 days vs 7.6 days, P = .001), but this result has not been reproduced by any subsequent studies.[17-23] Baik et al attribute their reduction in hospital length of stay to an overall reduction in serious complications (5.4% vs 19.3%, P = .025). However, no other comparative studies including our study have shown a statistically significant difference in overall complication rates or rate of anastomotic leaks.[17-23] It is interesting to note that we found a trend toward more anastomotic leaks in the laparoscopic group.

Several unanswered questions remain regarding the use of robotic technology in rectal cancer excision. Long-term oncologic outcome data of robotic total mesorectal excision including recurrence rates and cancer-specific survival are lacking. Data on the impact of RS on quality of life and sexual and urinary function are scarce.[32] Although several studies have looked at short-term outcome, there is a definite need for additional studies to look at long-term outcomes. In addition, we did not study cost differences between the 2 techniques. It appears from the literature that the upfront cost of the robotic technique is higher than that of laparoscopic surgery, but the overall cost-effectiveness has yet to be determined because the assessment should be based on long-term oncologic outcomes and functional results.[33,34] Finally, the issue of training and implementation of RS needs to be further studied. Some publications have addressed the learning curve for robot-assisted rectal cancer excision, showing the learning phase to be achieved after approximately 25 cases when operative time and conversion rates were used as proxies for competency.[35-37] However, the studies were reported from centers with experienced high-volume pelvic surgeons already skilled in laparoscopy. Can these results be duplicated by lower-volume surgeons with less expertise in laparoscopy or by open surgeons transitioning to RS?

CONCLUSION

This comparative study exploring a single surgeon's early experience with RS for rectal cancer showed a similar outcome to laparoscopy. However, the robotic group had a longer operative time, longer length of stay, and higher use of narcotic pain medications. The rate of complications was no different between laparoscopy surgery and RS, although a trend toward a higher readmission rate and more anastomotic leaks was observed in the laparoscopic group. Although this study provides additional data on this newly implemented technology in the field of colorectal surgery, no firm recommendations can be offered for greater implementation of RS at this stage because of the retrospective nature of this study and the limited number of patients. The need for more randomized data is critical to determine whether the added resource utilization in RS is justifiable.

Table 1.

Characteristics of Patients in Laparoscopic and Robotic Groups

	Laparoscopic (n = 21)	Robotic (n = 21)	P Value
Gender: male/female	12 (57%)/9 (43%)	10 (48%)/11 (52%)	.76
Median age (mean, range) (y)	62 (61.9, 48–79)	60 (58.3, 41–73)	.21
Median BMI[a] (mean, range) (kg/m2)	27 (28.4, 20–38)	25 (25.8, 20–37)	.23
Prior abdominal surgery	4 (19%)	6 (29%)	.71
Comorbidities
Hypertension	9 (43%)	8 (38%)	>.99
Hyperlipidemia	10 (48%)	9 (43%)	>.99
Diabetes	8 (38%)	5 (24%)	.5
Smoker	3 (14%)	3 (14%)	>.99
CAD[a]	4 (19%)	1 (5%)	.34
COPD[a]	0	2 (10%)	.48
IBD[a]	0	1 (5%)	>.99
Cirrhosis	1 (5%)	0	>.99

BMI = body mass index; CAD = coronary artery disease; COPD = chronic obstructive pulmonary disease; IBD = inflammatory bowel disease.

Table 2.

Operative Data for Laparoscopic and Robotic Groups

	Laparoscopic (n = 21)	Robotic (n = 21)	P Value
Type of resection
Low anterior resection	6 (29%)	1 (5%)	.09
Proctectomy	13 (62%)	11 (52%)	.76
Intersphincteric proctectomy	1 (5%)	2 (10%)	>.99
Abdominoperineal resection	1 (5%)	5 (24%)	.18
Total proctocolectomy	0	2 (10%)	>.99
Fecal diversion	13 (62%)	19 (90%)	.07
Ileostomy	11 (52%)	13 (62%)	.76
Colostomy	2 (10%)	6 (29%)	.24
Invasion of adjacent organs	2 (10%)	0	.48
Median operative time (mean, range) (min)	240 (236.3, 171–360)	260 (274.8, 189–449)	.04
Median robot dock time (mean, range) (min)	—	10 (10.4, 4–23)
Median robot work time (mean, range) (min)	—	61 (68.5, 37–125)
Median blood loss (mean, range) (mL)	100 (271.4, 50–1200)	150 (252.6, 30–2000)	.76
Intraoperative transfusions	1 (5%)	1 (5%)	>.99
Intraoperative complication	1 (5%)	1 (5%)	>.99
Bleeding (requiring transfusion)	1 (5%)	1 (5%)	>.99
Conversion to open surgery	0	1 (5%)	>.99

Table 3.

Oncologic Data for Laparoscopic and Robotic Groups

	Laparoscopic (n = 21)	Robotic (n = 21)	P Value
Neoadjuvant chemoradiation	13 (62%)	18 (86%)	.21
Median time from treatment to operation (mean, range) (d)	68 (104, 38–517)	68 (71, 40–170)	.87
Location of tumor
Anal verge to 5 cm	4 (19%)	10 (48%)	.10
5.1–10 cm	11 (52%)	10 (48%)	>.99
>10 cm	6 (29%)	1 (5%)	.09
Median (mean, range) (cm)	8 (9.6, 4–20)	6 (6.5, 0–15)	.02[a]
Preoperative tumor grade
Well differentiated	2 (10%)	2 (10%)	>.99
Moderately differentiated	16 (76%)	17 (81%)	>.99
Poorly differentiated	3 (14%)	2 (10%)	>.99
Tumor stage (postoperatively)
No residual disease	3 (14%)	2 (10%)	>.99
Stage I	7 (33%)	5 (24%)	.73
Stage II	4 (19%)	4 (19%)	>.99
Stage III	7 (33%)	9 (43%)	.53
Stage IV	0	1 (5%)	>.99
Median No. of lymph nodes harvested (mean, range)	15 (14.8, 8–21)	17 (19.7, 8–40)	.03[a]
Median No. of positive lymph nodes (mean, range)	6 (5.1, 1–10)	3 (3.7, 1–9)	.93
No. of patients with positive lymph nodes	7 (33%)	10 (48%)	.53
Median size of lesion (mean, range) (cm)	2.5 (2.5, 0–8.1)	1.5 (1.7, 0–6.5)	.34
Median length of specimen (mean, range) (cm)	23 (24.2, 15–35.5)	27.1 (31.8, 17.5–110)	.0001[a]
Lymphovascular invasion	8 (38%)	8 (38%)	>.99
Circumferential margin involvement	1 (5%)	0	>.99
Distal margin involvement	1 (5%)	0	>.99
Median distal margin distance (mean, range) (cm)	5.5 (5.1, 0.5–8)	3.9 (4.6, 1.0–18)	.48

Statistically significant.

Table 4.

Postoperative Outcome of Laparoscopic and Robotic Groups

	Laparoscopic (n = 21)	Robotic (n = 21)	P Value
Median hospital length of stay (mean, range) (d)	5 (6.0, 3–14)	6 (8.7, 4–23)	.05[a]
Intensive care unit admission	0	3 (14%)	.23
Median intensive care stay (mean, range) (d)	None	4 (3.7, 2–5)
Postoperative transfusion	3 (14%)	3 (14%)	>.99
Median transfusion amount (mean, range) (U)	2 (2.3, 2–3)	2 (2.3, 2–3)	.74
Median time receiving PCA[b] (mean, range) (d)	2 (2.8, 1–8)	4 (4.5, 0–10)	.03[a]
Median time receiving IV[b] narcotic (mean, range) (d)	3 (3.2, 2–8)	5 (6.3, 2–16)	.003[a]
Overall complications (90 d)[c]	7 (33%)	9 (43%)	.75
Anastomotic leak and/or pelvic abscess	3 (14%)	0	.23
Dehydration/acute renal insufficiency	3 (14%)	2 (10%)	>.99
Urinary tract infection	1 (5%)	4 (19%)	.34
Wound infection	1 (5%)	1 (5%)	>.99
Small bowel obstruction	1 (5%)	1 (5%)	>.99
Cardiac (myocardial infarction or arrhythmia)	2 (10%)	0	.49
Pneumonia	0	1 (6%)	>.99
Abdominal wall hematoma	0	1 (6%)	>.99
Vaginal bleeding requiring reoperation	1 (3%)	0	>.99
Reoperation at index hospitalization	1 (5%)	0	>.99
Readmission within 90 d (except for stoma reversal)	5 (24%)	1 (5%)	.18
Reoperation within 90 d (except for stoma reversal)	0	0	>.99

Statistically significant.

PCA = patient-controlled analgesia; IV = intravenous.

More than 1 complication per patient is possible.