Long-term Oncological Outcomes of Totally Laparoscopic Colectomy With Intracorporeal Anastomosis for Colon Cancer: Propensity Score Matching Analysis

Background: This retrospective study aimed to compare long-term oncological outcomes between laparoscopic-assisted colectomy (LAC) with extracorporeal anastomosis (EA) and totally laparoscopic colectomy (TLC) with intracorporeal anastomosis (IA) for colon cancers, including right- and left-sided colon cancers. Methods: Patients with stage I–III colon cancers who underwent elective laparoscopic colectomy between January 2013 and December 2017 were analyzed retrospectively. Patients converted from laparoscopic to open surgery and R1/R2 resection were excluded. Propensity score matching (PSM) analysis (1:1) was performed to overcome patient selection bias. Results: A total of 388 patients were reviewed. After PSM, 83 patients in the EA group and 83 patients in the IA group were compared. Median follow-up was 56.5 months in the EA group and 55.5 months in the IA group. Estimated 3-year overall survival (OS) did not differ signicantly between the EA group (86.6%; 95% condence interval (CI), 77.4–92.4%) and IA group (84.8%; 95%CI, 75.0–91.1%; P = 0.68). Estimated 3-year disease-free survival (DFS) likewise did not differ signicantly between the EA group (76.4%; 95%CI, 65.9–84.4%) and IA group (81.0%; 95%CI, 70.1–88.2%; P = 0.12). Conclusion: TLC with IA was comparable to LAC with EA in terms of 3-year OS and DFS. TLC with IA thus appears to offer an oncologically feasible procedure.


Introduction
The advantages of totally laparoscopic colectomy (TLC) with intracorporeal anastomosis (IA) for rightsided colon cancer (RC) have been reported in many studies [1][2][3][4][5][6][7]. TLC with IA can avoid the risk of twisting the mesentery and bowel during anastomotic construction and allows the surgeon to select the optimal site of bowel extraction. A systematic review [6] con rmed better short-term outcomes of TLC with IA for RC compared with laparoscopic-assisted colectomy (LAC) with extracorporeal anastomosis (EA) in terms of results such as bowel function, length of hospital stay, and cosmetic results. In addition, Hanna et al. [8] showed no signi cant difference between LAC with EA and TLC with IA for RC in terms of oncological outcomes.
IA for sigmoid colon and rectal cancer using a double-stapling technique (DST) with a circular stapler is a common procedure. If the tumor is located in the descending colon or proximal sigmoid colon where DST anastomosis is di cult, EA such as functional end-to-end or side-to-side colo-colostomy is usually performed. LAC with EA for left-sided colon cancer (LC) needs mobilization of the splenic exure, whereas TLC with IA might omit that procedure. We have reported that TLC with IA is technically feasible for LC where DST anastomosis is di cult and can be performed with good cosmetic outcomes and decreased time to rst atus [9]. Swaid et al. [10] reported similar results, but few studies have compared TLC with IA to LAC with EA for LC. Moreover, no reports have been published on the long-term outcomes of TLC with IA as compared to LAC with EA for colon cancers including both RC and LC.
The present study aimed to compare long-term outcomes between LAC with EA and TLC with IA for colon cancers including LC, using propensity score matching (PSM) analysis to reduce sample selection bias.

Materials And Methods
We started TLC with IA from June 2013 for patients with early-stage colon cancer, and the indications were gradually and carefully expanded to include advanced cancer. Patients who underwent elective laparoscopic surgery for colon cancer in our hospital between January 2013 and December 2017 were retrospectively evaluated. This study was reviewed and approved by the institutional review board at Osaka Medical College (acceptance number: 2853) and was performed in accordance with the Declaration of Helsinki. The choice of operation, as either LAC with EA or TLC with IA, was determined by the operating surgeon. The standard procedures were similar for all patients and performed by the same surgical team. Exclusion criteria were: DST anastomosis; patients with distant metastasis or simultaneous double cancer; conversion from laparoscopic to open surgery; or R1/R2 resection.

Operative procedure
Patients were placed in the lithotomy position with a 0-15° head-up or head-down tilt during surgery.
Intraabdominal pressure was maintained at 10 mmHg, and pneumoperitoneum was established with heated, humidi ed carbon dioxide gas. Five trocars were used in all procedures, with mobilization of the colon and lymphadenectomy performed laparoscopically by the medial-to-lateral approach. No drainage tube was placed after surgery.

LAC with EA
A small incision was extended for the trocar incision at the umbilicus or the left lower quadrant port site. After providing protection with Wound Protector, the bowel was externalized. The ileum or colon was resected with 60-mm linear staples (Fig. 1a) and ileo-colostomy/colo-colostomy was performed using a 60-mm linear stapler (Fig. 1b). The enterotomy was then closed using the 60-mm linear stapler (Fig. 1c, d).

TLC with IA
The procedure for TLC with IA for RC or LC has been described in our previous reports [9,11]. In summary, the ileum or colon was divided intracorporeally with 60-mm linear staplers (Fig. 1e), and the enterotomy was closed using the Albert-Lembert method after fashioning side-to-side ileo-colostomy/colo-colostomy with a 60-mm linear stapler ( Fig. 1f-h). The specimen was then extracted through a mini-laparotomy over the trans-umbilical port site, the lower quadrant port site, or via Pfannenstiel laparotomy.

Statistical analysis
Statistical analysis was performed using JMP 14.2 software (SAS Institute Inc., Cary, NC, USA). To minimize the in uence of potential confounders on selection bias, propensity scores of patient characteristics and pathological results were generated using binary logistic regression. One-to-one matching between two groups was accomplished using the nearest-neighbor matching method, performed without replacement and using a caliper width of 0.2 standard deviations of the logit of the estimated propensity score. After propensity score matching, the two matched groups were handled as unpaired independent groups. Data are expressed as median with interquartile range or mean ± standard error. The statistical signi cance of other data was determined using one-way analysis of variance, Fisher's test, the chi-squared test, or Student's t test. Univariate analyses of overall survival (OS) and disease-free survival (DFS) rates were performed using the Kaplan-Meyer method. Differences when comparing survival curves were analyzed using the log-rank test. Cox regression analysis was performed to identify possible prognostic factors. Results are reported as hazard ratios and 95% con dence intervals (CIs). Values of P < 0.05 were considered statistically signi cant.

Patients
Of the 691 eligible patients, 303 patients were excluded, and 388 patients were analyzed in this study (Fig. 2). We identi ed signi cant differences in age, body mass index, American Society of Anesthesiologists score, and tumor location, and PSM was performed. Patient background characteristics and pathological results in both groups were closely balanced by the PSM, resulting in 83 matched pairs (Table 1). Finally, 83 patients in the EA group and 83 patients in the IA group were compared.   (Fig. 3a). Estimated 3-year DFS likewise did not differ signi cantly between the EA group (76.4%; 95%CI, 65.9-84.4%) and the IA group (81.0%; 95%CI, 70.1-88.2%; P = 0.12) (Fig. 3b).

Cox regression analysis after PSM
Neither EA nor IA was associated with poor DFS (Table 3). Tumor location, lymph-node metastasis, and Stage III were likewise not associated with poorer DFS.

Discussion
Our results showed that laparoscopic colectomy with TLC with IA was similar to LAC with EA in terms of 3-year OS and DFS. IA involved exposing the ileum and colon under pneumoperitoneum in the abdominal cavity, raising concerns of increased peritoneal dissemination with IA. In the 1990s, when laparoscopic surgery for colon cancer was started, concern about port-site recurrences was widespread, but was allayed by large prospective series [12][13][14][15]. In the present study including TLC with IA for LC, only 2 patients in the EA group and 3 patients in the IA group experienced disseminated recurrence after PSM.
No patients in either group experienced recurrence at the anastomosis site after PSM. Moreover, neither EA nor IA was associated with poor DFS in Cox regression analysis. No signi cant differences in the number of lymph nodes harvested were identi ed between groups. These results suggested that TLC with IA is an oncologically reasonable procedure.
In this study, the incidence of super cial SSI was signi cantly higher in the IA group (13%) than in the EA group (3.6%; P = 0.047). Exposing the intestinal tract in the abdominal cavity during IA procedures may result in exposure to bacteria, but the frequency of organ-space/deep SSI did not differ signi cantly between groups in our results (4.8% in the EA group vs. 6.0% in the IA group, P = 0.73). We showed that IA procedures did not increase the risk of dissemination, but also did not increase organ-space/deep SSI. Super cial SSI was frequently encountered at the port site through which 60-mm linear staplers were passed for side-to-side ileo-colostomy/colo-colostomy. The port through which the linear stapler was passed for anastomosis may become contaminated with stool, increasing the risk of super cial SSI. Length of postoperative hospital stay is known to be prolonged by the occurrence of SSI [16], and the IA group showed a longer hospital stay than the EA group in our study. Several reports showing that TLC with IA was favorable in terms of SSI have represented the rate of SSI as 1-4.4% [17][18][19], lower than our results. To prevent SSI, we have started applying additional steps, such as cleaning inside the contaminated port and administering chemical preparation with oral antibiotics the day before surgery.
Since introducing such steps, we have encountered no SSIs in 13 consecutive TLCs with IA for RC in 2020 [11].
Good short-term outcomes of TLC with IA for RC have been shown in many retrospective reports [1][2][3][4][5][6][7]. In this study including LC, TLC with IA required a longer operation time and needed a shorter skin incision than LAC with EA. Because of the procedure dissecting the mesentery and anastomosis intracorporeally, TLC with IA took longer, but the smaller skin incision was advantageous in terms of cosmetology and postoperative pain. As another advantage, TLC with IA enables extraction site exibility [17]. This advantage is particularly useful in natural ori ce specimen extraction surgery [20] and holds promise for scarless surgery if natural ori ce transluminal endoscopic surgery can be achieved.
The main potential advantages of IA lie in surgeries for obese patients and patients who have undergone polysurgery. Obese patients may have thick layers of subcutaneous fat and a shortened colonic mesentery, which could present an obstacle to EA and result in a need for longer skin incisions compared to non-obese patients. Although the length of skin incision required for EA and IA are considered similar in non-obese patients, the signi cantly shorter skin incision for IA in our cohort may indicate a difference in obese patients. Patients who have undergone polysurgery might also encounter di culties with EA due to intra-abdominal adhesions, in which cases IA would prove useful. In our institution, IA was performed for patients with advanced cancers such as T4a or subtotal circumferential tumor, but if a skin incision of 8 cm or more was required for specimen removal due to bulky tumor or other-organ invasion, we ruled out the IA procedure, as the advantages of IA were considered to be neutralized. IA was also not indicated for patients with preoperative ileus symptoms because of the high risk of abdominal contamination, but we did perform IA if intestinal decompression proved effective.
Adequate proximal and distal margins are important to guarantee su cient oncological radicality. The IA group obtained a signi cantly longer distal margin in this study. TLC with IA, which determined the dissection line in the abdominal cavity, can secure a wider disease-free margin compared to EA. Scatizzi et al. [4] reported similar results, with resection of a longer specimen potentially reducing the risk of residual colon ischemia.
The present study showed that none of tumor location, lymph-node metastasis, or stage III were associated with poor DFS. Two large population-based studies have reported that RCs show worse prognosis than left-sided colon cancers [21,22], but tumor location was not a prognostic factor in our study. Similarly, lymph-node metastasis and stage III are considered to be associated with poor DFS, but again, not in our study. The bene t of adjuvant chemotherapy (AC) for stage III colon cancer has been clearly established [23,24], and AC is recommended for patients with lymph-node metastasis; in other words, patients classi ed as stage III according to National Comprehensive Cancer Network guidelines [25]. This study had some limitations, including the retrospective, single-center design. However, strengths of the study include the fact that this is the largest retrospective comparison of long-term, oncological outcomes between LAC with EA and TLC with IA, the reduction of selection bias by PSM, and the inclusion of not only RC, but also LC.
In conclusion, the present study showed that TLC with IA was an oncologically feasible procedure with long-term outcomes comparable to those of LAC with EA. A prospective, randomized trial comparing LAC with EA to TLC with IC could validate these ndings.

Declarations
Ethics approval and consent to participate The study protocol was approved by the institutional review board of Osaka Medical College (acceptance number: 2853) and was performed in accordance with the Declaration of Helsinki. Due to the retrospective design of the study, the local ethic committee con rmed that informed consent was not necessary from participants.

Consent for publication
Not applicable.

Availability of data and materials
The datasets generated and analyzed during the current study available from the corresponding author on reasonable request.   Kaplan-Meier curves comparing overall survival (a) and disease-free survival (b) in totally laparoscopic colectomy between intracorporeal anastomosis (IA) and extracorporeal anastomosis (EA).