Robotic Right Hemicolectomy with Complete Mesocolic Excision and D3 Lymphadenectomy: Perioperative Complications and Oncologic Outcomes.

INTRODUCTION
Rising rates of colon cancer in younger individuals have renewed interest in optimizing colon cancer treatment options, including surgery.1 Early studies showed an inverse association between the number of lymph nodes harvested and the likelihood of cancer recurrence, independent of stage migration.2–5 Complete mesocolic excision (CME) (in which dissection proceeds within the embryological plane between the colonic mesentery and the parietal retroperitoneal fascia) is performed to achieve complete removal of the mesentery, resection of an adequate length of bowel to remove pericolic lymph nodes, and central vascular ligation to remove apical (D3) lymph nodes. CME is associated with a larger area of resected mesentery, a greater number of resected lymph nodes, and higher likelihood of disease-free survival (DFS) compared with conventional resection.6–11 Case-control studies,12 retrospective series,13 and meta-analyses14–17 have found evidence of the benefits of CME with D3 lymphadenectomy. Some randomized trials comparing D2 and D3 lymphadenectomies have reported similar short-term outcomes,18–21 but the randomized trial RELARC and the nonrandomized trial RESECTAT reported higher DFS and OS, respectively, for D3 dissection in node-positive patients.20–23
The possibility that dissection along the root of the vascular pedicle (e.g., along the superior mesenteric vein for a right hemicolectomy) may lead to perioperative complications, including vascular injury, has raised concerns. Results from retrospective studies, systematic reviews, and meta-analyses have been mixed, with reports of higher likelihood of intraoperative injury to the superior mesenteric vein and adjacent organs for CME9 as well as reports of no difference between CME and conventional colectomy in rates of conversion to open surgery, perioperative morbidity, or mortality.8–11,15–17 Clinical trials have also shown mixed results, with reports of higher risk of vascular injury for CME12,20 as well as reports of no difference in perioperative blood loss or morbidity.18–20,22
Robotic surgical platforms, with articulating wristed instruments and three-dimensional optics, are particularly well-suited for technically demanding procedures and dissection along vascular pedicles.24–25 In this study, we analyzed perioperative outcomes, complications, and disease-free survival in cancer patients who underwent robotic right hemicolectomy with CME and D3 lymphadenectomy at our comprehensive cancer center over a 10-year period.

METHODS
For this single-center retrospective cohort study, we searched institutional databases (with approval from the Institutional Review Board) for patients who underwent robotic right hemicolectomy with D3 lymphadenectomy for cancer between January 2014 and June 2024 by either of two experienced, high-volume surgeons. Patients with metastatic disease were excluded. The surgeons used the da Vinci Xi robotic platform (Intuitive Surgical, Sunnyvale, CA) and the same CME technique: initial vascular pedicle ligation, medial-to-lateral mobilization, and en bloc D3 lymphadenectomy along the superior mesenteric artery and vein.26–28
Clinicopathologic and perioperative data obtained from prospective databases—demographics, length of stay, operative time, blood loss, complications, readmission, pathology, follow-up, and recurrence—were supplemented by data from electronic medical records. To fully capture complications, including those treated at other institutions, all postoperative correspondence up to 30 days postsurgery was reviewed. Postoperative imaging was performed only when deemed necessary following patient assessment in urgent care or in the clinic or following remote assessment. Local recurrence was identified using endoscopy, radiography, and biopsy. Distant recurrence was identified by radiography without biopsy.
All statistical analyses were performed using RStudio software. Categorical variables were analyzed using raw numbers and percentages. Continuous variables were analyzed using medians and interquartile ranges (IQR). DFS was analyzed using Kaplan-Meier survival curves.

RESULTS
Of 631 patients who underwent robotic right hemicolectomy with CME and D3 lymphadenectomy for nonmetastatic cancer during the 10-year period, 557 had a colon adenocarcinoma, 2 had a neuroendocrine tumor in the colon, 55 had an appendix tumor, and 17 had a tumor in the terminal ileum (Table 1). Median operative time was 165 (IQR 140–188) min, median blood loss was 25 (IQR 25–50) ml, and median postoperative length of stay was 4 (IQR 3–5) days.

Perioperative Complications
No patients required conversion to open surgery. Of 101 total complications (16.0%), 86 were of Clavien-Dindo grade I or II, and 15 were grade III or higher (Table 2). No patients experienced intraoperative vascular injury. Of 7 patients with an anastomotic leak, 4 underwent reoperation. Three patients (0.5%) experienced thrombosis in the superior mesenteric vein following discharge and received anticoagulation treatment. In two of the three patients, thrombosis was diagnosed by CT after re-presentation for abdominal pain; in the third patient, thrombosis was incidentally found by CT performed for an unrelated trauma. Thirty-eight patients (6.0%) were readmitted within 30 days of surgery (to the same institution in which the operation was performed), most commonly for ileus or small bowel obstruction (15 patients [2.4%]). Two patients (0.3%) died following discharge, but within 30 days from surgery.

Colon Adenocarcinoma Pathology
Of 536 patients with colon adenocarcinoma, 152 (28%) had AJCC stage I disease, 197 (37%) had stage II disease, and 187 (35%) had stage III disease; 306 tumors (57%) were pT3, 53 (10%) were pT4, 123 (23%) were N1, and 64 (12%) were N2 (Table 3). Median lymph node harvest was 30.5 (IQR 23–42) nodes, with no difference between the 164 patients with mismatch repair deficiency (MMRd) and the 334 patients with mismatch repair proficiency (median lymph node harvest, 31 nodes for each group). Histologic variants were as follows: adenocarcinoma, 436 patients (81%); medullary carcinoma, 19 patients (4%); mucinous adenocarcinoma, 76 patients (14%); signet ring cell carcinoma, 4 patients (1%); micropapillary carcinoma, 1 patient (0.2%). The tumor was well differentiated in 8 patients (1%), moderately differentiated in 398 patients (74%), and poorly differentiated in 127 patients (24%), with tumor differentiation data missing for the remaining 3 patients (0.6%). Tumor-infiltrating lymphocytes were noted in 174 (32%) specimens. High-risk features included lymphovascular invasion (240 patients [45%]), tumor budding (148 patients [28%]), perineural invasion (104 patients [19%]), and tumor deposits (49 patients [9%]).

Disease-Free Survival in Patients with Colon Adenocarcinoma
Of the 536 patients with colon adenocarcinoma, 207 (39%) underwent adjuvant treatment. Median follow-up was 42.8 months (36.9, 46.9, and 42.3 months for patients with stage I, II, or III disease, respectively). Local (anastomotic) recurrence developed in 2 patients, and distant recurrence developed in 61 patients. The rates of DFS at 3 and 5 years were 86.1% (95% confidence interval [CI], 82.8–89.5%) and 84.7% (CI, 81.1–88.4%), respectively, for the 536 patients: 98.5% (CI, 96.4–100%) each for the 152 patients with stage I disease; 92.0% (CI, 87.8–96.5%) and 89.9% (CI, 84.8–95.3%), respectively, for the 197 patients with stage II disease; and 70.2% (CI, 63.1–78.1%) and 68.6% (CI, 61.1–77.0%), respectively, for the 187 patients with stage III disease. For patients with MMRd tumors and patients with MMR-proficient tumors, 3-year DFS was 94.6% (CI, 90.8–98.6%) and 81.2% (CI, 76.6–86.2%), respectively.

DISCUSSION
In this cohort of 631 cancer patients who underwent robotic right hemicolectomy with CME and D3 lymphadenectomy, perioperative complications were relatively rare (with no vascular injuries), recovery was quick in most patients, and DFS rates were generally higher than in studies of D2 lymphadenectomy, indicating that with the use of robotics by experienced surgeons, CME with D3 lymphadenectomy has the potential to be safe and effective.
Concerns with CME and D3 lymphadenectomy have stemmed from reports of longer operative times,7–10,12 higher complication rates, 9,22,29–32 and increased risk of vascular injury.12,20 However, median operative time for robot-assisted surgery in our study was under 3 h, with relatively low variation. The longer operative times and greater variability observed in other studies likely reflect differences in surgeon experience and the effect of the learning curve, particularly when the procedure is performed only on patients with concern for central nodal disease either on preoperative imaging or intra-operatively. The 16% overall rate of complications and the 2.4% rate of grade ≥III complications in our study compare favorably to the 18–26% and 1.2–12% rates, respectively, in other studies.16,17,20,22 The 30-day mortality rate of 0.3% in our study is on the lower end of the 0–5% range reported for right hemicolectomy.33
Differences in surgeon experience may help explain the vascular complications reported in some CME series12,20 but not in others.10–13,20–22 However, comparing studies is difficult due to inconsistent reporting of the extent and specifics of vascular injuries. For example, the RELARC study20 reported a higher rate of vascular injuries for D3 vs. D2 lymphadenectomy, but without higher intraoperative blood loss or other complications. No patient in our study had an operative injury to the superior mesenteric vein, and all 3 cases (0.5%) of partial thrombosis in the superior mesenteric vein were treated with anticoagulation. This rate of thrombosis is lower than the 3–5% rates in other studies.34–35
The 12% rate of recurrence among the 536 patients with colon adenocarcinoma in our study (2 patients with local recurrence and 61 patients with distant metastases) is consistent with the data reported for CME by Hohenberger et al.6 and Bertelsen et al.10 as well as systematic reviews and meta-analyses.31,36–37 In a recent study of patients with stage I, II, or III colorectal cancer in Denmark, the rate of recurrences (distant or local) diagnosed within the first 3 years after curative-intent surgery performed in 2014–2019 (n = 10,470) was 85.4%.38 With median follow-up of 42.8 months in our study, we likely captured the vast majority of recurrences.
Although a prospective multicenter study by Benz et al.22 did not find that overall survival was longer for CME than for standard right hemicolectomy overall, an exploratory analysis in that study indicated a potential benefit in overall survival in patients who underwent CME for stage III disease. Similarly, the prospective randomized trial RELARC found no significant difference in 3-year DFS or overall survival between D3 and D2 resections, but a subgroup analysis revealed that 3-year DFS was higher for D3 resection in younger patients and in patients with stage III disease.23
In the absence of a D2 group in our study, the D2 group in the RELARC trial represents a reasonable comparator group. Considering the fact that RELARC excluded patients with central lymphadenopathy identified on imaging or at surgery, the 3-year DFS rates for patients who underwent D3 resection for stage I, II, or III colon adenocarcinoma in our study (98.5%, 92.0%, and 70.2%, respectively) compare favorably with the 3-year DFS rates in RELARC for both D2 resection (95.7%, 90.2%, and 66.0%, respectively) and D3 resection (95.8%, 91.4%, and 75.8%, respectively). In our study, patients with MMRd tumors had a 13% higher rate of 3-year DFS compared to patients with MMR-proficient tumors, similarly to the 10% and 12% higher rates of 3-year DFS in patients with MMRd tumors in the CME and D2 groups in RELARC, respectively. The higher proportion of MMRd tumors in our study compared with RELARC (31% vs. 19%) may have contributed to the higher DFS rates.
With a hazard ratio of 0.70 and wide confidence intervals due to fewer than expected events, the RELARC trial may have been underpowered to detect a significant difference in DFS between D2 and D3 resections, and more accurate comparison of DFS rates is needed. In the meantime, selection of clinically node-positive patients who may benefit from D3 resection remains challenging, as staging of colon cancer by computed tomography has modest accuracy in detecting nodal disease, with a sensitivity of 72% and a specificity of 63%.17,39 Although D3 lymph nodes are involved in ~10% of patients with metastases, recurrence is often detected late and misdiagnosed as retroperitoneal, and it can be difficult to treat.40
Our data suggest that robotic colectomy with CME may be associated with higher likelihood of DFS compared to standard open or laparoscopic resection. In the COST trial41 5-year DFS was 68.4% for open colectomy and 69.2% for laparoscopic colectomy, and in the COLOR trial42 3-year DFS was 76.2% for open colectomy and 74.2% for laparoscopic colectomy, compared with 3-year DFS of 86.1% and 5-year DFS of 84.7% in our study. The proportion of MMRd tumors was not reported for either the COST trial or the COLOR trial, and it is possible that the 31% proportion of MMRd tumors in our study (with the associated higher 3-year DFS) may have contributed to the higher DFS. The rates of local recurrence after open colectomy and laparoscopic colectomy were 2.6% and 2.3%, respectively, in the COST trial (median follow-up, 84 months) and 4.8% and 4.9%, respectively in the COLOR trial (median follow-up, 53 months), compared with 0.4% in our study (median follow-up, 43 months).
Recently, the randomized REAL trial43 demonstrated the superiority of robotic surgery over laparoscopic surgery in patients with cancer in the middle or low rectum. With median follow-up of 43 months, the rate of locoregional recurrence was lower (1.6% vs. 4.0%) and 3-year DFS was higher (87.2% vs. 83.4%) in patients who underwent robotic vs. laparoscopic surgery. The superior visual acuity and improved instrument dexterity provided by the robotic platform are particularly advantageous for hemicolectomy involving CME and D3 lymphadenectomy and may explain, at least in part, the potentially superior oncologic outcomes and safety profile in our study.
A common limitation of studies of CME is the variability of surgical technique. Terms such as D3 lymphadenectomy, CME, and central vascular ligation are not used consistently in the literature, and a systematic review found that the only universal surgical step for a radical right colectomy was central artery ligation.44 These variations can complicate interpretation of study results. To mitigate this potential limitation, we included in our analyses only the patients of two high-volume surgeons with comparable experience and surgical techniques. Combined with the absence of a comparator group, this may limit the generalizability of our findings. In addition, our analyses of readmission and perioperative complications may have been limited by the degree to which the electronic medical records fully captured relevant data. Functional and quality-of-life outcomes were beyond the scope of the study.
In conclusion, the results of our study indicate that robotic right hemicolectomy with CME and D3 lymphadenectomy can be performed safely, with a low rate of perioperative complications and with potentially superior oncologic outcomes compared to both open and laparoscopic D2 resection.