Interim analysis of short-term outcomes between robotic and laparoscopic surgery for colon cancer: results from the ESSIMIC trial.

Graphical abstract

INTRODUCTION
Although the benefits of minimally invasive surgery for the treatment of colorectal cancer have been extensively demonstrated [1–4], widespread adoption of minimally invasive techniques has remained difficult [5, 6]. One of the main barriers can be attributed to the technical challenges associated with acquiring advanced laparoscopic skills.
The introduction and dissemination of robotic surgical tools in the mid-2000s offered a potential solution to these limitations. Robotic instruments provide multiple degrees of freedom, enabling movements that match or even exceed the dexterity of a surgeon’s hands. Enhanced 3-dimensional vision improves visualization of the surgical field, while the 3 robotic arms, all controlled by the surgeon, reduce reliance on an assistant typically required in laparoscopic procedures.
Nonetheless, despite numerous studies evaluating the potential advantages of robotic over laparoscopic surgery, no definitive evidence has yet established the superiority of robotic surgery in the literature [7–9]. Therefore, objective data are needed to justify the broader implementation of robotics in colorectal surgery. The adoption of robotic techniques should rest on clear evidence of superior perioperative outcomes. In this context, a randomized clinical trial was conducted to determine whether the benefits of robotic surgery can be applied to the treatment of colon cancer.

METHODS

Ethics statement
The trial was registered at ClinicalTrials.gov (identifier: NCT02871960) and received ethical approval from the Ethics Committee of the “Federico II” University of Naples (approval No. 108/17/ES01). Written informed consent was obtained from all patients.

Study design
This was an interventional multicenter randomized open-label, parallel-group trial (ESSIMIC study) designed to evaluate surgical stress in robotic versus laparoscopic colorectal surgery (Supplementary Material 1). An interim analysis was performed to compare short-term outcomes in patients who underwent laparoscopic or robotic surgery for colon cancer. Participating surgeons, from 7 Italian centers, were required to demonstrate experience of at least 100 colorectal cancer surgeries before entering the trial, including a minimum of 30 robotic and 30 conventional laparoscopic procedures. Independent central evaluation of surgical technique and operative skills was also performed. This was ensured by video review of at least one anonymous, unedited surgical intervention for each procedure, randomly selected from both laparoscopic and robotic cases, and assessed using the Global Operative Assessment of Laparoscopic Skills (GOALS) [10].

Participants
All consecutive patients presenting between January 2017 and December 2019 with a diagnosis of primary colon or rectal cancer and scheduled for elective colorectal surgery were screened for eligibility. Only patients undergoing surgery for colon cancer were included in this interim analysis, while rectal resections were excluded. Inclusion criteria consisted of a diagnosis of primary colon or rectal cancer planned for elective colonic resection, age over 18 years, and provision of written informed consent. Patients with locally advanced tumors not suitable for curative resection or requiring en bloc multivisceral resection were excluded. Patients scheduled for open surgery due to severe systemic disease contraindicating minimally invasive surgery and those who declined participation were also excluded. Furthermore, atypical resections (e.g., splenic flexure or mid-transverse colon) were excluded to maintain homogeneity across surgical procedure groups (right colon and left colon resections).

Randomization
Enrolled patients were randomized in a 1:1 ratio to undergo either robotic surgery (intervention) or laparoscopic surgery (control) according to the study protocol (Supplementary Material 1). Randomization was stratified to balance potential confounding variables, including sex, age, body mass index (BMI), Charlson Comorbidity Index (CCI), and tumor stage. Type of resection (right colon, left colon, or rectal resection) was also accounted for during the randomization process. Randomization was performed using a computer-based system with sealed envelopes.

Procedures
All operations were required to be fully minimally invasive. In the robotic arm, the robot was used for every surgical step.

Right colectomy
The terminal ileum and ascending colon were dissected through the avascular plane between the mesocolon and Gerota fascia. Dissection continued along the duodenum and pancreatic head, followed by central vascular ligation, extending to the right colic vessels, gastrocolic trunk of Henle, and middle colic vessels. Resection margins of at least 5 cm were achieved. Intracorporeal linear stapled side-to-side anastomosis was performed, and the specimen was extracted through a 5- to 7-cm Pfannenstiel incision.

Left colectomy
Mobilization of the left colon was performed along the avascular plane. Central vascular ligation was achieved by ligating the inferior mesenteric artery at its root, and the inferior mesenteric vein was ligated at the lower pancreatic border. Splenic flexure mobilization was performed, with division of the posterior attachments of the distal transverse and descending colon. The distal colon was divided intracorporeally with a resection margin of at least 5 cm using a linear stapler. Proximal transection was completed extracorporeally through a 5- to 7-cm Pfannenstiel incision after mesocolon division to the selected site. Anastomosis was performed using the standard Knight–Griffen technique.

Outcomes
Short-term outcomes of robotic versus laparoscopic colonic resection for cancer were evaluated, including complications, oncological radicality, and recovery after surgery. Short-term complications were defined as events occurring within 90 postoperative days. Postdischarge follow-up consisted of outpatient visits at 30 and 90 days after surgery. Both intraoperative and postoperative complications were included. Intraoperative complications also encompassed conversions to open surgery. Postoperative complications were classified according to the Clavien-Dindo classification [11]. Anastomotic leakage was categorized into grades A, B, and C, following Rahbari et al. [12]: grade A required no active therapeutic intervention, grade B required active intervention without re-laparotomy, and grade C required re-laparotomy. Oncological radicality was assessed by the number of lymph nodes retrieved, specimen length, and distance to the distal margin. Recovery after surgery was assessed using a composite of postoperative outcomes. Successful recovery was defined by achievement of all of the following: complete mobilization, stool passage, and tolerance of a solid diet. Prespecified subgroup analyses assessed short-term outcomes separately for right and left colonic resections.

Sample size
Sample size calculations were reported in the original ESSIMIC study. Based on exploratory findings suggesting a ≥50% difference in serum interleukin-6 levels between robotic and laparoscopic surgery, 142 patients were required per group to achieve 80% power at a 5% significance level. To account for dropouts and technical failures, the planned enrollment was approximately 150 patients per group.

Statistical analysis
All analyses were conducted on an intention-to-treat basis. Two-tailed P-values of <0.05 were considered statistically significant. Statistical analyses were performed using IBM SPSS ver. 28 (IBM Corp). Continuous variables were expressed as mean±standard deviation and compared using the Student t-test or the Mann-Whitney U-test. Nominal variables were reported as counts and percentages and compared using the chi-square or Fisher exact test, as appropriate. Logistic regression with a stepwise approach was used to adjust results for sex, age, BMI, CCI, and tumor stage. Subgroup analyses were performed to examine the impact of surgical approach on specific types of resection.

RESULTS
A total of 593 patients with primary colon or rectal cancer scheduled for elective colorectal surgery were assessed for eligibility between January 2017 and December 2019. Eleven patients declined participation, and 62 did not meet the inclusion criteria. Thus, 520 patients were randomized: 260 to laparoscopic surgery and 260 to robotic surgery. Among these, 39 patients were excluded because of atypical resections or abdominoperineal resections, and 8 patients were excluded for locally advanced cancer. For the interim analysis, 150 patients undergoing rectal resections were also excluded (Fig. 1). Consequently, 323 patients were included in the final analysis: 142 underwent robotic-assisted surgery and 181 underwent laparoscopic surgery. In the robotic group, 78 (54.9%) underwent right colectomy and 64 (45.1%) underwent left colectomy. In the laparoscopic group, 103 (56.9%) underwent right colectomy and 78 (43.1%) underwent left colectomy. Baseline demographic and clinical characteristics were well balanced between the groups (Table 1).

Complications
Fifteen anastomotic leakages (4.6%) were recorded: 6 (1.9%) in the robotic group and 9 (2.8%) in the laparoscopic group (P=0.484). Of these, 2 (0.6%) were classified as grade A, 7 (2.2%) as grade B, and 6 (1.9%) as grade C (Table 2). Twelve conversions to open surgery occurred: 4 (2.8%) in the robotic group and 8 (4.4%) in the laparoscopic group (P=0.560). A total of 97 complications (30.0%) were observed in the short-term period, with 37 (26.1%) in the robotic group and 60 (33.2%) in the laparoscopic group (P=0.503). Of these, 8 (8.2%) occurred intraoperatively and 89 (91.8%) postoperatively.
Overall, 109 complications (33.7%) were documented: 41 (28.9%) in the robotic group and 68 (37.6%) in the laparoscopic group (P=0.503). Twenty complications (18.4%) occurred intraoperatively and 89 (81.6%) postoperatively.
Twelve of the 20 intraoperative complications were conversions to open surgery (4 robotic cases and 8 laparoscopic cases, P=0.560). The remaining 8 were other intraoperative complications (4 robotic cases [2.8%] and 4 laparoscopic cases [2.2%], P=0.734): 2 ileocecal resections for small bowel perforations, splenectomy for acute bleeding, descending colon ischemia requiring the Deloyers procedure, bladder perforation treated with suturing, cyst drainage for rupture of a renal cyst, pancreatic capsule bleeding treated with a hemostatic agent, and a splenic serosal tear managed with hemostatic agent.
Postoperative complications occurred in 33 patients (23.2%) in the robotic group and 56 patients (30.9%) in the laparoscopic group (P=0.134). According to the Clavien-Dindo classification, 20 patients (6.2%) experienced grade I complications (nausea, n=8; fever, n=7; pain, n=5). Grade II complications occurred in 44 cases (13.6%), including nonsurgical complications (n=19), anemia (n=11), wound infection (n=5), anastomotic bleeding (n=5), grade A leakage (n=2), and paralytic ileus (n=2). Grade III complications were reported in 22 cases (6.8%), including grade B leakage (n=7), grade C leakage (n=6), small bowel obstruction (n=4), anastomotic bleeding (n=3), anastomotic stenosis (n=1), and pancreatic fistula (n=1). Two cases (0.6%) of acute renal failure were classified as grade IV, and 1 postoperative death (0.3%) was recorded (Table 3). Grade III complications occurred more frequently in the laparoscopic group (9.4% vs. 3.5%, P=0.045) (Table 3). Specifically, in the robotic group, complications included 4 anastomotic leakages and 1 pancreatic fistula, whereas in the laparoscopic group, 9 anastomotic leakages, 4 small bowel obstructions, 3 anastomotic bleedings, and 1 anastomotic stenosis were observed.
Regression analysis confirmed that sex (P=0.908), age (P=0.287), BMI (P=0.395), American Society of Anesthesiologists (ASA) physical status (P=0.675), and tumor stage (P=0.888) did not influence the occurrence of complications. However, robotic surgery was significantly associated with a lower risk of Clavien-Dindo grade III complications (odds ratio [OR], 0.323; P=0.033), while a high CCI was associated with an increased risk (OR, 3.136; P=0.031).

Oncological radicality
The total number of lymph nodes retrieved at pathology was satisfactory in both groups, with no significant difference between the laparoscopic and robotic groups (19.6±7.1 vs. 18.9±8.7, P=0.117) (Table 3). Specimen length was comparable and acceptable in both groups (27.3±12.3 cm vs. 26.2±11.1 cm, P=0.786). Proximal and distal margin involvement was not observed in any patient. The mean distance from the distal margin was also similar (10.0±5.9 cm vs. 9.6±7.3 cm, P=0.275).

Recovery
Functional recovery after surgery was superior in the robotic group, with a mean recovery time of 45.1±23.9 hours compared with 63.5±29.3 hours in the laparoscopic group (P=0.037). Time to mobilization was significantly shorter in the robotic group (17.5±10.2 hours vs. 21.1±11.5 hours, P=0.023). No significant differences were observed for time to stool passage (41.5±24.7 hours vs. 62.3±28.9 hours, P=0.139) or tolerance of a solid diet (39.8±22.2 hours vs. 54.1±24.9 hours, P=0.724).
Regression analysis showed no evidence of bias related to patient or tumor characteristics. Specifically, multivariate analyses found no significant effects of sex (P=0.139), age (P=0.687), BMI (P=0.223), ASA physical status (P=0.971), CCI (P=0.193), or tumor stage (P=0.171).

Subgroup analyses
The advantages of robotic over laparoscopic surgery varied depending on the surgical setting (right or left colectomy). In the subgroup of 181 patients (56.0%) with right-sided colon cancer, robotic surgery was performed in 78 cases (43.1%). Grade III complications were significantly less frequent in the robotic group compared with the laparoscopic group (1.3% vs. 11.7%, P=0.008) (Table 3). No significant differences in functional recovery were observed (45.8±25.3 hours vs. 63.9±23.4 hours, P=0.883). A higher number of lymph nodes retrieved was noted in the robotic group compared with the laparoscopic group (20.3±10.3 vs. 20.2±6.4, P=0.009), differing from the trend observed in the overall study population.
In the subgroup of 142 patients (44.0%) with left-sided colon cancer, robotic surgery was performed in 64 cases (45.1%). Functional recovery, measured by the composite outcome, was significantly shorter in the robotic group compared with the laparoscopic group (44.3±22.2 hours vs. 61.1±31.1 hours, P=0.009) (Table 3). However, no differences in the incidence of grade III complications were found (6.3% vs. 6.4%, P>0.999).
All other short-term outcomes analyzed in the subgroups were consistent with the results obtained in the overall cohort. Baseline demographic characteristics in both the right and left colectomy subgroups were well balanced between robotic and laparoscopic groups (Supplementary Tables 1, 2).

DISCUSSION
Robotic surgery was associated with fewer severe complications and improved postoperative recovery compared with laparoscopic surgery, as shown by the reduced rate of grade III complications (3.5% vs. 9.4%, P=0.045) and faster recovery (45.1±23.9 hours vs. 63.5±29.3 hours, P=0.037).
Minimally invasive surgery is considered the gold standard treatment for colon cancer. In high-income countries, the trend in colorectal surgery has shifted toward broad adoption of minimally invasive techniques [13]. The technical advantages of robotics may enhance the precision of minimally invasive colonic resections [14]. However, general surgeons have been slower to adopt the robotic platform, largely due to concerns about higher intraoperative costs [15]. The first reports of robotic-assisted colorectal interventions were described by Weber et al. [16]. A meta-analysis demonstrated advantages of robotic surgery in reducing anastomotic leakage rates, overall complication rates, conversion to open surgery, and time to resumption of diet [17]. Nevertheless, these findings cannot be considered conclusive, as the literature remains preliminary and the evidence quality is low [9]. Two recent meta-analyses specifically addressing right and left colectomy confirmed the low quality of available evidence [18, 19]. Only 1 randomized clinical trial on colonic surgery has been reported [20]. A trial on right colon cancer by Park et al. [20] concluded that robotic-assisted laparoscopic right colectomy was feasible but did not demonstrate sufficient benefit to justify its higher cost. Ultimately, the justification for robotic surgery should rest on its ability to provide superior perioperative outcomes. The results of our trial, showing fewer severe complications and faster postoperative recovery with the robotic approach, suggest a meaningful advantage that could support its broader use in colorectal surgery.
In our study, robotic surgery was associated with a significantly lower incidence of Clavien-Dindo grade III complications (3.5% vs. 9.4%, P=0.045) and improved functional recovery, as measured by the composite outcome (45.1±23.9 hours vs. 63.5±29.3 hours, P=0.037). Specifically, robotic surgery reduced the time to achieve complete mobilization (17.5±10.2 hours vs. 21.1±11.5 hours, P=0.023), while no significant differences were noted for time to first stool passage (41.5±24.7 hours vs. 62.3±28.9 hours, P=0.139) or tolerance of solid diet (39.8±22.2 hours vs. 54.1±24.9 hours, P=0.724).
Subgroup analyses further clarified these findings in the context of right versus left colectomy. The advantages of the robotic platform were most evident in right colectomy, where the incidence of grade III complications was significantly lower compared with laparoscopic surgery (1.3% vs. 11.7%, P=0.008). An additional oncological benefit was also identified in right-sided resections, with a higher number of lymph nodes retrieved in the robotic group (20.3±10.3 vs. 20.2±6.4, P=0.009), a difference not observed in the overall population. Conversely, in left colectomy, the benefits of robotics were more limited. Robotic surgery was associated only with faster recovery according to the composite outcome (44.3±22.2 hours vs. 61.1±31.1 hours, P=0.009), without significant differences in complication rates or oncological radicality.
This study has several limitations. No blinding of treatment allocation was possible, although the primary outcomes were objective and unlikely to be influenced by this limitation. Variability in surgical skill is an inherent limitation of multicenter surgical trials. Nonetheless, we attempted to minimize this by including only surgeons experienced in both laparoscopic and robotic procedures. Differences in patient characteristics and surgical techniques across centers may also limit generalizability. However, randomization ensured balanced allocation across treatment arms, and multivariate analysis was performed to confirm the results and exclude potential confounders.
Although additional studies are required to strengthen the evidence, our interim analysis of this interventional multicenter randomized open-label, parallel-group trial (ESSIMIC study) supports the conclusion that robotic surgery provides better short-term outcomes than laparoscopic surgery for colon cancer. These findings may assist in clinical decision-making, particularly for right-sided colon cancer, where the advantages of robotic surgery appear most pronounced.