Feasibility of robot-assisted surgery for defining circumferential resection margins for rectal cancer: a retrospective study Authors: Toshihiro Nakao, Mitsuo Shimada, Takuya Tokunaga, Masaaki Nishi, Hideya Kashihara, Chie Takasu, Yuma Wada, Toshiaki Yoshimoto Published: 2024 ABSTRACT AbstractBackgroundLocal recurrence is a major problem after surgery for rectal cancer. Precision of the circumferential resection margin (CRM) has been shown to be an independent predictor of local recurrence. The purpose of this study is to evaluate the usefulness of robotic surgery for defining the CRM.MethodsA retrospective cohort study of patients with rectal cancer who underwent radical surgery at Tokushima University Hospital from January 2012 to April 2023 was included in this study to evaluate risk factors for CRM involvement. Short-term outcomes, including CRM rates, as well as long-term outcomes, were compared between patients who had undergone robot-assisted versus laparoscopic surgery following propensity score analysis.ResultsA total of 223 patients were analyzed in this study. Multivariate analysis demonstrated that lymph node metastasis was the most significant predictive factor for CRM involvement (p = 0.030), and that robot-assisted surgery tended to lead to less CRM involvement (p = 0.085). The CRM involvement rate for robotic and laparoscopic surgery before propensity score matching was 4.7% and 11.7%, respectively. Following propensity score matching, the CRM involvement rate for robotic surgery was 4.5% versus 11.4% for laparoscopic surgery. Disease-free survival (DFS) and overall survival (OS) rates for patients who had undergone robot-assisted and laparoscopic surgery were not significantly different before or after matching (DFS before matching: p = 0.876, DFS after matching: p = 0.805, OS before matching: p = 0.511, OS after matching: p = 0.458).ConclusionRobot-assisted surgery may be useful in defining the CRM for rectal cancer.Keywords: Robot-assisted surgery, Circumferential resection margin, Rectal cancer, Propensity score matching Background Local recurrence poses a significant challenge following surgery for rectal cancer; however, the introduction of total mesorectal excision (TME) has led to a marked reduction in local recurrence rates [1–4]. Maurer et al. reported that the local recurrence rate in patients who underwent TME was significantly lower at 5.9%, compared to 20.8% in those who did not receive TME [3]. Furthermore, the addition of adjuvant radiotherapy has been shown to prevent the spread of rectal cancer [5, 6]. The Dutch Colorectal Cancer Group demonstrated that short-term adjuvant radiotherapy also decreases the risk of local recurrence in rectal cancer patients who have undergone TME [7]. These advancements have led to a reduction in local recurrence rates; however, local recurrence remains a serious issue, as it is often difficult to treat, can cause severe disabling symptoms, and may be fatal. The precision of the circumferential resection margin (CRM) has been identified as an independent predictor of local recurrence [8–10]. Wibe et al. found that cases with a negative CRM greater than 1 mm had lower local recurrence rates compared to those with a positive CRM less than 1 mm [11]. Gosens et al. reported that patients with a positive CRM of 1 mm or more had a local recurrence rate of 43%, which was much higher compared to 8% in patients with a negative CRM of less than 1 mm [12]. Currently, laparoscopy is the standard approach for rectal cancer surgery. One advantage of laparoscopic surgery is the magnified field of view, which facilitates more detailed observation than open surgery. However, the superiority of laparoscopy over open surgery in accurately defining CRM remains unproven [13–15]. In recent years, the use of robot-assisted surgery for rectal cancer has rapidly increased. One benefit of robotic surgery is its provision of a three-dimensional, highly magnified view, allowing for greater surgical precision. While some studies have compared robotic, laparoscopic, and open surgery regarding CRM, significant gaps in understanding persist [16, 17]. The purpose of this study is to evaluate the effectiveness of robotic surgery in defining CRM. Methods Patients A retrospective cohort study of patients with rectal cancer who underwent radical surgery at Tokushima University Hospital from January 2012 to April 2023 was conducted. Inclusion criteria were age older than 18 years, rectal cancer without invasion of other organs or distant metastasis, and performance of robotic-assisted or laparoscopic surgery without lateral lymph node dissection. Exclusion criteria were pathology diagnosis without CRM description, emergency surgery, performance of surgery in different areas at the same time, recurrence, ulcerative colitis or Crohn’s disease, simultaneous multiple primary cancers, open conversion, palliative surgery, and pelvic abscess formation. The protocol was approved by the Ethics Committee of Tokushima University (approval number: 3215-1). Preoperative treatment Adjuvant chemotherapy or chemoradiotherapy (CRT) was recommended for patients with massive tumors. The adjuvant chemotherapy regimen was FOLFOXIRI. FOLFOXIRI regimen consisted of an intravenous infusion of irinotecan at a dose of 150 mg/m2 for 90 min, intravenous infusion of leucovorin at a dose of 200 mg/m2 for 120 min, intravenous infusion of oxaliplatin at a dose of 85 mg/m2 for 120 min, and continuous intravenous infusion of 5-fluorouracil at a dose of 2400 mg/m2 for 448 h7,8. Immediately prior to irinotecan, patients received panitumumab at a dose of 60 mg/kg for 60 min, cetuximab at a dose of 250 mg/m2 for 60 min, or bevacizumab at a dose of 5 mg/kg for 90 min. All patients were pretreated intravenously with palonosetron hydrochloride (0.75 mg), dexamethasone sodium phosphate (6.6 mg), and fosaprepitant meglumine (150 mg) 1 h prior to irinotecan administration. Chemotherapy was administered essentially every 2 weeks. The concurrent chemotherapy regimen was either S-1 alone or SOX combined with bevacizumab. S-1 (80 mg/m2/day) was administered orally on days 1–5, 8–12, 15–19, and 22–26, oxaliplatin (50 mg/m2/day) on days 1, 8, 15, and 22, and bevacizumab (5 mg/kg) on days 1 and 15. Curative surgery was performed 6–9 weeks after preoperative CRT. Surgery Both robot-assisted and laparoscopic surgeries were performed according to the principles of total mesorectal excision. The surgical procedure was selected based on the tumor characteristics and the patient’s status: high anterior resection (HAR), low anterior resection (LAR), intersphincteric resection (ISR), abdominoperineal resection (APR), or Hartmann’s operation, with or without the combination of transanal total mesorectal excision (TaTME). The procedure in which the rectal dissection site on the anal side is above the peritoneal reflection is defined as HAR, and when the dissection site is below the peritoneal reflection, it is defined as LAR. When the internal anal sphincter is partially or completely resected while preserving the anus and external anal sphincter, it is defined as ISR. The procedure in which the rectum is resected along with the anus, and the oral side stump is diverted as a single-stoma colostomy, is defined as APR. Additionally, when the rectum is resected, the anal-side stump is closed, and the oral-side stump is diverted as a single-stoma colostomy, it is also defined as Hartmann’s operation. Histopathologic assessment A pathologic examination of the resected specimen was performed. CRM involvement was defined as the presence of tumor cells ≤ 1 mm from the radial resection margin on pathology. Propensity score matching Propensity score matching analysis was performed for the following factors: sex, age, American Society of Anesthesiologists physical status (ASA-PS), body mass index (BMI), preoperative treatment, tumor location, T classification, N classification, maximum tumor size, surgical procedure, occurrence of TaTME, and lymph node dissection. The matching pairs were created at a ratio of 1:1 based on the nearest neighbor matching algorithm with a 0.2-caliper distance. Statistical analysis The data are presented as the mean ± standard deviation or as the number (percentage) as appropriate. Categorical variables were analyzed using Fisher’s exact test and continuous variables were analyzed using the Student’s t-test. A multivariate analysis was performed using the logistic regression model. Kaplan-Meier curves were used to evaluate the prognostic accuracy of different variables, and log-rank tests were used to assess statistical significance. Overall survival (OS) was defined as the interval between the date of curative surgery and all-cause mortality or the last follow-up. Disease-free survival (DFS) was defined as the time from the date of curative surgery to the recurrence of the tumor or death. All statistical analyses were performed with EZR (Version 1.54) (Saitama Medical Center, Jichi Medical University, Saitama, Japan), a graphical user interface for R (version 4.03) (R Foundation for Statistical Computing, Vienna, Austria) [18]. EZR is a modified version of R commander (version 2.7-1) designed to add statistical functions frequently used in biostatistics. A p-value < 0.05 was considered statistically significant. Results Patient and tumor characteristics A total of 223 patients were analyzed in this study. The mean follow-up time was 32 (1–60) months. The patient and tumor characteristics are shown in Table 1. Table 1Univariated analysis for CRMTotal(n = 223)CRM (+)(n = 20)CRM (-)(n = 203) p Sex0.813 Male13713124 Femal86779Age, years old67.0 ± 10.269.7 ± 8.066.7 ± 10.40.229ASA-PS0.728 1105897 2971087 320218BMI23.1 ± 3.622.7 ± 4.423.2 ± 3.60.593Preoperative treatment0.331 CRT31130 NAC716 None18518167Tumor location0.236 Rs821171 Ra74569 Rb67463T classification0.002 T1, 280179 T3, 414319124N classification> 0.001 N (-)1414137 N (+)821666Maximum tumor size, mm34.6 ± 19.142.3 ± 12.833.4 ± 19.40.009Surgical approach0.093 Robot86482 Laparoscopy13716121Surgical procedure0.197 HAR43538 LAR12910119 ISR21021 APR24420 Hartmann’s operation615TaTME45 (20.2%)2 (10.0%)43 (21.2%)0.380Lymph node dissection0.053 D1101 D259158 D316319144CRM, circumferential resection margin; ASA-PS, American Society of Anesthesiologists physical status; BMI, body mass index; CRT, chemoratiotherapy; NAC, neoadjuvant chemotherapy; HAR, high anterior resection; LAR, low anterior resection; ISR, intersphincteric resection; APR, abdominoperineal resection; TPE, total pelvic exenteration; Ta TME, transanal total mesorectal excision Univariated analysis for CRM CRM, circumferential resection margin; ASA-PS, American Society of Anesthesiologists physical status; BMI, body mass index; CRT, chemoratiotherapy; NAC, neoadjuvant chemotherapy; HAR, high anterior resection; LAR, low anterior resection; ISR, intersphincteric resection; APR, abdominoperineal resection; TPE, total pelvic exenteration; Ta TME, transanal total mesorectal excision Risk of circumferential resection margin involvement From a total of 223 patients, 20 had CRM involvement (9.0%). The patient and tumor characteristics were compared between the patients with and without CRM involvement (Table 1). CRM involvement was not associated with sex, age, American Society of Anesthesiologists physical status (ASA-PS), body mass index (BMI), preoperative treatment or tumor location. Patients with CRM involvement had a significantly higher T classification (p = 0.020), significantly more lymph node metastases (p > 0.001), and a significantly larger maximum tumor size (p = 0.009). Regarding surgery modality, patients with CRM involvement showed a trend toward having undergone laparoscopic surgery (p = 0.093) or D2 lymph node dissection (p = 0.053) compared with patients without CRM involvement. Surgical procedures such as high anterior resection (HAR), low anterior resection (LAR), intersphincteric resection (ISR), abdominoperineal resection (APR) and Hartmann’s operation, or a combination of transanal total mesorectal excision (TaTME) did not correlate with CRM involvement. Multivariate analysis showed that CRM involvement correlated significantly with lymph node metastasis (p = 0.030) and showed a trend toward correlation with laparoscopic surgery (p = 0.085) (Table 2). Table 2Multivariated analysis for CRMOdds ratio95% CI p T classification T3, 4 vs. T1, 22.990.193–46.4000.433N classsification N (+) vs. N (-)4.241.150–15.7000.030Maximum tumor size1.010.984–1.0400.434Surgical approach Robot vs. Laparoscopy0.340.101–1.1600.085Lymph node dissection D3 vs. D1, 21.030.070–15.4000.981CRM, circumferential resection margin; CI, confidence interval Multivariated analysis for CRM CRM, circumferential resection margin; CI, confidence interval Comparison of characteristics of patients undergoing robot-assisted or laparoscopic surgery The patient and tumor characteristics were compared between those who had undergone robot-assisted and laparoscopic surgery (Table 3). Robot-assisted surgery was performed on 86 patients, while laparoscopic surgery was conducted on 137 patients. In all patients, there were significant differences in age (p = 0.023), preoperative treatment (p = 0.002), T classification (p = 0.010), maximum tumor size (p = 0.021), surgical procedure (p = 0.019) and combination of TaTME (p > 0.001) between the patients undergoing the two surgical modalities. After propensity score matching, these significant differences in patient and tumor characteristics were resolved. Table 3Proensity score matcing between robot and laparoscopic surgeryBefore propensity score matchingAfter propensity score matchingRobot(n = 86)Laparoscopy(n = 137) p Robot(n = 44)Laparoscopy(n = 44) p Sex0.0661.000 Male46912322 Femal40462122Age,years old65.0 ± 10.768.2 ± 9.80.023ASA-PS0.7871.000 139662222 238591920 391132BMI22.8 ± 3.623.3 ± 3.60.29965.2 ± 9.166.0 ± 10.00.663Preoperative treatment0.0021.000 CRT102156 NAC7000 None691163938Tumor location0.1280.687 Rs30441216 Ra31361912 Rb25571316T classification0.0101.000 T1, 240401818 T3, 446972626N classification0.4790.827 N (-)57842826 N (+)29531618Maximum tumor size,mm38.5 ± 20.332.2 ± 18.00.02134.1 ± 18.335.7 ± 21.10.714Surgical procedure0.0190.944 HAR20231114 LAR45842422 ISR14744 APR61854 Hartmann151.000TaTME32 (37.2%)13 (9,5%)> 0.00111 (25.0%)10 (22.7%)Lymph node dissection0.2661.000 D10100 D227321312 D3591043132ASA-PS, American Society of Anesthesiologists physical status; BMI, body mass index; CRT, chemoratiotherapy; NAC, neoadjuvant chemotherapy; HAR, high anterior resection; LAR, low anterior resection; ISR, intersphincteric resection; APR, abdominoperineal resection; TPE, total pelvic exenteration; Ta TME, transanal total mesorectal excision Proensity score matcing between robot and laparoscopic surgery Age, years old Maximum tumor size, mm ASA-PS, American Society of Anesthesiologists physical status; BMI, body mass index; CRT, chemoratiotherapy; NAC, neoadjuvant chemotherapy; HAR, high anterior resection; LAR, low anterior resection; ISR, intersphincteric resection; APR, abdominoperineal resection; TPE, total pelvic exenteration; Ta TME, transanal total mesorectal excision Comparison of short-term outcomes for patients undergoing robot-assisted or laparoscopic surgery The operative and postoperative outcomes were compared for patients who had undergone robot-assisted or laparoscopic surgery (Table 4). Before propensity score matching, the postoperative hospital stay following robot-assisted surgery was significantly shorter than after laparoscopic surgery (robot-assisted surgery: 13.8 ± 7.2 days, laparoscopic surgery: 17.5 ± 10.8 days, p = 0.006). The total number of complications of Clavien–Dindo grade 3 or higher was significantly lower for patients who had undergone robot-assisted surgery (robot-assisted surgery: 2.3%, laparoscopic surgery: 10.9%, p = 0.019). The CRM involvement rate for robotic surgery was 4.7%, which was lower than the 11.7% observed for laparoscopic surgery (p = 0.093). The operation time was significantly longer in robot-assisted surgery than in laparoscopic surgery (robot-assisted: 295.0 ± 82.1 min, laparoscopic surgery: 254.0 ± 72.4 min, p = 0.015). After propensity score matching, postoperative hospital stay (robot-assisted surgery: 13.7 ± 6.9 days, laparoscopic surgery: 16.4 ± 11.9, p = 0.195), postoperative complications (robot-assisted surgery: 2.3%, laparoscopic surgery: 11.4%, p = 0.202), and CRM involvement rates for robotic and laparoscopic surgery remained similar (robot-assisted surgery: 4.5%, laparoscopic surgery: 11.4%, p = 0.434) to the values obtained prior to propensity score matching, but p-values increased, making the trend less significant. Table 4Short-term outcomesBefore propensity score matchingAfter propensity score matchingRobot(n = 86)Laparpscopy(n = 137) p Robot(n = 44)Laparoscopy(n = 44) p Operation time, min290.2 ± 71.1276.7 ± 83.30.217295.0 ± 82.1254.0 ± 72.40.015Blood loss, ml39.5 ± 51.835.8 ± 114.20.78038.8 ± 58.920.0 ± 37.70.080Postoperative hospital stay, days13.8 ± 7.217.5 ± 10.80.00613.7 ± 6.916.4 ± 11.90.195CRM (+)4 (4.7%)16 (11.7%)0.0932 (4.5%)5 (11.4%)0.434Postoperative complications Total2 (2.3%)15 (10.9%)0.0191 (2.3%)5 (11.4%)0.202 Asatomotic leakage1 (1.2%)6 (4.4%)0.2541 (2.3%)2 (4.5%)1.000 Ileus1 (1.2%)0 (0%)0.3860 (0%)0 (0%)1.000 Perineal wound infection0 (0%)3 (2.2%)0.2860 (0%)0 (0%)1.000 Perineal wound dehiscence0 (0%)1 (0.7%)1.0000 (0%)1 (2.3%)1.000 Intraabdominal abscess0 (0%)1 (0.7%)1.0000 (0%)0 (0%)1.000 Anastomotic bleeding0 (0%)1 (0.7%)1.0000 (0%)0 (0%)1.000 Stomal necrosis0 (0%)1 (0.7%)1.0000 (0%)1 (2.3%)1.000 Stomal perforation0 (0%)1 (0.7%)1.0000 (0%)1 (2.3%)1.000 Neurogenic bladder0 (0%)1 (0.7%)1.0000 (0%)0 (0%)1.000CRM, circumferential resection margin Short-term outcomes CRM, circumferential resection margin Comparison of long-term outcomes for patients undergoing robot-assisted or laparoscopic surgery No significant difference in DFS for patients who had undergone robot-assisted or laparoscopic surgery was observed, either before or after propensity matching (before propensity score matching: p = 0.876, after propensity score matching: p = 0.805) (Fig. 1). The recurrence patterns following robot-assisted and laparoscopic surgery were evaluated (Table 5). Before propensity matching, the local and distant recurrence rates for patients who had undergone robot-assisted surgery were 3.5% and 11.6%, respectively, while the local and distant recurrence rates for laparoscopic surgery were 3.6% and 13.1%, respectively, showing no significant difference in recurrence patterns (local recurrence: p = 1.000, distant recurrence: p = 0.837). After propensity matching, the local and distant recurrence rates for patients who had undergone robot-assisted surgery were 4.5% and 9.1%, respectively, while the local and distant recurrence rates for laparoscopic surgery were 4.5% and 13.6%, respectively, also showing no significant difference in recurrence patterns (local recurrence: p = 1.000, distant recurrence: p = 0.739). Additionally, OS did not differ significantly for patients who had undergone robot-assisted versus laparoscopic surgery, either before or after propensity matching (before propensity score matching: p = 0.511, after propensity score matching: p = 0.458) (Fig. 2). Disease-free survival before and after propensity score matching for patients who had undergone robot-assisted surgery versus laparoscopic surgery Table 5Recurrence patternBefore propensity score matchingAfter propensity score matchingRobot(n = 89)Laparpscopy(n = 137) p Robot(n = 44)Laparoscopy(n = 44) p Local recurrence3 (3.5%)5 (3.6%)1.0002 (4.5%)2 (4.5%)1.000Distant recurrence10 (11.6%)18 (13.1%)0.8374 (9.1%)6 (13.6%)0.739 Recurrence pattern Overall survival before and after propensity score matching for patients who had undergone robot-assisted surgery versus laparoscopic surgery Discussion In this study, univariate analysis demonstrated that T stage of T3 or T4, lymph node metastasis, and larger maximum tumor size were significantly associated with CRM involvement following resection. Resection by robot-assisted surgery and extended lymph node dissection tended to decrease the risk of CRM involvement. Furthermore, multivariate logistic regression analysis demonstrated that lymph node metastasis was significantly associated with CRM involvement, and the use of robot-assisted surgery tended to decrease the risk of CRM involvement. Independent predictors of CRM involvement have previously been reported to include T stage, lymph node metastasis, type of surgical procedure, tumor grade, lymphovascular invasion, and perineural invasion [19–21].There have been few reports comparing robotic surgery with laparoscopic surgery for rectal cancer that evaluate the involvement of the CRM post-resection. Feng et al. reported that the CRM was positive in 4.0% of patients in the robotic surgery group and 7.2% in the laparoscopic surgery group, with a significant difference between the two groups (p = 0.024) [17]. We further compared the use of robotic and laparoscopic surgery using propensity score matching. After matching for patient characteristics, tumor factors, and surgical factors, including predictors of CRM involvement, the CRM involvement rate was 4.5% for robot-assisted surgery and 11.4% for laparoscopic surgery, which was comparable to the rate before matching. This indicates that robot-assisted surgery leads to a lower rate of CRM involvement than laparoscopic surgery, even when the conditions between the two are equivalent. In addition, the rates of CRM involvement following robot-assisted and laparoscopic surgery did not change after adjustment for factors that influence CRM involvement, suggesting that robot-assisted surgery may influence CRM involvement independently of other factors. Propensity score matching eliminated the statistical significance in CRM involvement rates between robot-assisted and laparoscopic surgery, which may be due to a lack of power caused by a reduction in the number of cases included. Regarding short-term outcomes, the postoperative hospital stay length and postoperative complication rate were significantly lower in the robot-assisted surgery group before propensity score matching, but the significance disappeared after matching. This may be due to a decrease in the number of included cases. In this study, after propensity score matching, the operation time for robot-assisted surgery was significantly longer than that for laparoscopic surgery (Table 4). The ROLARR randomized trial reported a mean operative time of 298.5 min for robot-assisted surgery and 261.0 min for laparoscopic surgery, which was similar to our reported results [16]. Regarding long-term outcomes, there were no significant differences in DFS and OS between patients who had undergone robot-assisted versus laparoscopic surgery in the present study. Previously, Qui et al. analyzed 2593 rectal cancer patients and reported no significant difference in OS and DFS between those patients that had undergone robot-assisted versus laparoscopic surgery, confirming our results [22]. There was also no significant difference in local or distant recurrence for patients who had undergone robot-assisted or laparoscopic surgery. Adequate CRM has been shown to be an independent predictor of local recurrence [8–10]. In this study, robotic surgery showed a tendency for a lower CRM involvement rate compared with that observed following laparoscopic surgery, but there was no difference in local recurrence rate. Before propensity score matching, we observed 3 local recurrences out of all 86 robot-assisted surgeries and 5 local recurrences out of all 137 laparoscopic surgeries; following matching, we observed 2 local recurrences out of all 44 cases in robot-assisted and laparoscopic surgery, respectively. The observation of no significant difference in the local recurrence rate for patients who had undergone robotic versus laparoscopic surgery may be due to the small number of local recurrences. This study has several limitations, including that it was a single-institution, retrospective study with a small sample size, the lack of MRI measurement of the closest distance from the tumor or lymph node to CRM, and the inability to identify the location of the closest margin. A future cohort study using multicenter data should incorporate the distance between the tumor or lymph nodes and the CRM, as well as lesion location, into propensity score matching to determine how much the robot improves visualization and the achievement of difficult clear margin resections compared to laparoscopic surgery. Acknowledgements We thank Shelby King, PhD from Edanz (https://jp.edanz.com/ac) for editing a draft of this manuscript. Author contributions TN and MS were involved in the study design and data interpretation. TT, MN, HK, CT, YW, and TY were involved in the data analysis. All authors revised the manuscript, approved the manuscript to be published, and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. Funding The authors did not receive financial support from any organization for the submitted work. Data availability No datasets were generated or analysed during the current study. Declarations Ethics approval and consent to participate The protocol for this research project was approved by the Tokushima University Ethics Committee (approval number: 3215-1) and was performed in accordance with the provisions of the Declaration of Helsinki. Informed consent was obtained from all patients. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement No datasets were generated or analysed during the current study.