“Caudal to cranial” versus “medial to lateral” approach in laparoscopic right hemicolectomy with complete mesocolic excision for the treatment of stage II and III colon cancer: perioperative outcomes and 5-year prognosis

The purpose of this study was to compare the “caudal to cranial” (CC) versus “medial to lateral” (ML) approach for laparoscopic right hemicolectomy. Pertinent data from all patients with stage II and III between January 2015 and August 2017 were entered into a retrospective database. A total of 175 patients underwent the ML (N = 109) or CC approach (N = 66). Patient characteristics were equivalent between groups. The CC group showed a shorter surgical time 170.00 (145.00, 210.00) vs. (206.50 (178.75, 226.25) min) than the ML group (p < 0.001). The time to oral intake was shorter in the CC group than in the ML group ((3.00 (1.00, 4.00) vs. 3.00 (2.00, 5.00) days; p = 0.007). For the total harvested lymph nodes, there was no statistical significance between the CC group 16.50 (14.00, 21.25) and the ML group 18.00 (15.00, 22.00) (p = 0.327), and no difference was found in the positive harvested lymph nodes (0 (0, 2.00) vs. 0 (0, 1.50); p = 0.753). Meanwhile, no differences were found in other perioperative or pathological outcomes, including blood loss and complications. For 5-year prognosis, overall survival rates were 75.76% in the CC group and 82.57% in the ML group (HR 0.654, 95% CI 0.336–1.273, p = 0.207); disease-free survival rates were 80.30% in the CC group and 85.32% in the ML group (HR 0.683, 95% CI 0.328–1.422, p = 0.305). Both approaches were safe and feasible and resulted in excellent survival. The CC approach was beneficial in terms of the surgical time and time to oral intake.


Introduction
Until now, the technique of laparoscopic colorectal surgery has been becoming increasingly common for colon cancer according to several large multicenter trials [1][2][3][4]. In 2009, Hohenberger et al. [5] proposed the concept of complete mesocolic excision (CME) for colon cancer, the core of which is separation of the mesocolic from the parietal plane and true central ligation of the supplying arteries and draining veins right at their roots.
For right side colon cancer, in the early period, there are two main approaches that have been reported in the literature regarding laparoscopic right hemicolectomy (LRH): the lateral to medial (LM) [6][7][8] and medial to lateral (ML) approaches [8][9][10][11][12][13][14][15][16]. However, technically speaking, the LM approach can be extremely difficult under laparoscopic conditions because of the limited operative space as well as the maneuver ability of the forceps, so the LM method is usually performed in open surgery. For the common ML approach, the primary oncologic advantages are the early vessel division and the "no-touch" dissection, and it is considered the mainstream surgical approach for LRH. In 2016, our department first reported the "caudal-to-cranial" technique (CC) for LRH with CME [17]. We believe this technique overcomes some of the shortcomings of the ML approach without destroying the oncologic principle.
In addition to the three approaches described above, Laverd et al. and Matsuda et al. explored the cranial-to-caudal approach to solve the difficulty of dissecting and ligating the vessels at the inferior edge of the pancreas using the ML approach because of a deep view or "operating in the tunnel", especially when the right mesocolon is strongly affixed posteriorly or is broad and short [18,19]. Benz et al. [20] first proposed the uncinate process approach using the duodenum as the entry point. Although there are many different surgical approaches for LRH with CME, in our center, the ML and CC approaches are the two most commonly used approaches. Thus far, we have searched much of the literature to identify comparative studies of different approaches, in which most studies focused on safety and short-term outcomes. Therefore, we performed this retrospective analysis to compare the CC and ML approaches in terms of perioperative outcomes and long-term 5-year survival.

Case selection
From January 2014 to August 2017, we performed LRH for 175 consecutive cases. The patient inclusion criteria were as follows: (1) right sided colon cancer was confirmed by preoperative examination, and the advanced tumor was resectable without distant metastasis; (2) preoperative multidisciplinary treatment assessment suggested no significant contraindications; (3) postoperative pathological indication was stage II or III according to the UICC tumor classification [21]; (4) pathological report suggested adenocarcinoma; and (5) laparoscopic surgery. The exclusion criteria were as follows: patients who transfer to open surgery.
Two patients (one for each group) transfer to open surgery due to the large volume of tumor. 66 cases were included for the CC approach. In another group, 109 cases were included for the conventional ML approach, and both groups underwent D3 or beyond D3 lymph node dissection [22,23].
Preoperative, operative, postoperative, and pathological data were collected, and a comparative study between the two groups was carried out. Patients were not managed according an enhanced recovery pathway. In our center, we made a follow-up plan after the operation for all the patients. For the patient of stage II or III, we followed up every 3 months for the first two years. For postoperative 3 to 5 years, the patients were reviewed every 6 months. Five years after surgery, the frequency was changed to once a year. Each review included a physical examination, enhanced CT in the chest and full abdomen, and tumor markers (carcinoembryonic antigen, CEA; carbohydrate antigen 199, CA199; and alpha fetoprotein, AFP) in the digestive tract. However, Colonoscopy was performed annually. If tumor recurrence is found during follow-up, we recommend a multidisciplinary team (MDT) discussion. For the patients who lost follow-up, we mainly asked about the patient's physical condition and other hospital visits history by telephone.
In our center, all patients complete a preoperative evaluation via an abdominal enhanced CT scan. If the patients had stage T2 or higher or suspected lymph node metastasis, they underwent the D3 or beyond D3 technique. Written consent was obtained before the surgery. As a teaching hospital, patients were informed before surgery that their clinical data will be used anonymously for some clinical researches. Institutional review board approval was obtained prior to initiating the study.

Surgical methods
The technical procedure of the CC group was described in a previous study by Li et al. from our department [17]. Firstly, the assist exposed the root of the ileocecal mesentery; then the surgeon explored the retro-peritoneal plane composed of Toldt's and pre-renal fascia. This procedure was continued until reaching the lateral peritoneum and identifying the pancreatic head and Henle's gastrocolic trunk cephalad. The second step was to identified the ileocolic vessel, the lymph nodes' dissection was made along the superior mesenteric vein (SMV) or artery (SMA) caudally to cranially. The origins of the vessels were identified and then ligated at the root. Thirdly, the division continued rightwards along the greater gastric curvature border until the hepatocolic ligament and hepatic flexure was divided completely. The transverse mesocolon was divided downwards caudally until it joined the previous retro-peritoneal dissection.
For the ML group, the technical methods we adopted complied with Feng et al.'s study [10]. The dissection started at ileocolic vessel and proceeds along SMV or SMA, and exposed the inborn surgical plain composed of Told's and prerenal fascia to uncover head of pancreas and to mobilize duodenum. The exposing range began from the origin of transverse colon mesentery to the peritoneal reflection, and ligation at the origin of the central vessels to dissect the entire mesocolon as a whole. The treatment of hepatocolic ligament and hepatic flexure was similar to that described above. SPSS 23.0 was used for all analyses. Fisher's exact test or χ 2 test for parametric value and Student's t test for continuous value were performed as appropriate. Survival data were analyzed with the Kaplan-Meier method. Statistical analysis was done with the log-rank test. We calculated hazard ratio (HR) and 95% confidence interval (CI) using Cox regression. A p value < 0.050 was considered significant.

Results
There were no significant differences in age, sex distribution, body mass index (BMI), preoperative tumor biomarkers (CEA, CA199, and AFP), preoperative hemoglobin, preoperative albumin, history of abdominal surgery, colorectal cancer family history, rate of emergency operation, preoperative obstruction, preoperative transfusion, or comorbidity between these two groups ( Table 1).
Regarding the operative and pathological characteristics ( Table 2), no deaths occurred during the surgery in either group. The surgical time in the ML group was longer than that in the CC group (206.50 (178.75, 226.25) vs. 170.00 (145.00, 210.00) min), and the difference was significant (p < 0.001). However, blood loss was similar in the ML and CC groups, without statistical significance (p = 0.429). For the total harvested lymph nodes, there was no statistical   22.00) (p = 0.327), and no difference was found in the positive harvested lymph nodes (0 (0, 2.00) vs. 0 (0, 1.50); p = 0.753). There were no significant differences in the American Society of Anesthesiologists (ASA) score, operative transfusion, combined organ resection, the rate of beyond D3 lymph node dissection, specimen length, tumor location, tumor gross type, tumor differentiation, tumor size, T stage, N stage, lymphatic vascular invasion, cancer nodules and nerve invasion between these two groups.
In addition, there was a significant difference in the meantime to the start of oral liquid intake between the two groups (3.00 (1.00, 4.00) days in the CC group vs. 3.00 (2.00, 5.00) days in the ML group; p = 0.007). The mean time to first flatus and the time to urine and drainage tube removal were similar between these two groups. The mean length of total and postoperative hospital stays, and postoperative transfusion were similar between these two groups. There were also no significant differences in the rate of ICU stay (Table 3). Postoperative complications occurred in 9 (13.63%) patients in the CC group versus 17 (15.60%) in the ML group (p = 0.724). In the CC group, one patient with lung infection and one patient with postoperative intraperitoneal hemorrhage in need of blood transfusion (grade II, according to the Clavien-Dindo system [24]) was observed, two patients had gastric paralysis (grade II), and five patients had postoperative ileus (grade II). In the ML group, there is one patient each with incision infection (grade I), gastric paralysis (grade II), chylous leakage (grade II), anastomotic leakage (grade II), and urinary retention (grade II); one patient had intra-abdominal infection requiring interventional puncture (grade IIIa); two patients had postoperative ileus (grade II) and cerebral infarction (grade II); three patients had lung infection (grade II); four patients had postoperative bleeding (grade II). Postoperative ileus was a major complication in the CC group, however, there was no statistical difference between the two groups (p = 0.105).
We set the last follow-up time as August 31, 2022. 9 patients (3 from the beginning and 6 in the midway) did not reach the follow-up end point in the CC group; 13 patients (1 from the beginning and 12 in the midway) did not reach the Table 3 Postoperative characteristics For continuous variable: mean (standard deviation) for normal distribution and median (quartile) for nonnormal distribution; CC "caudal to cranial", ML "medial to lateral", ICU intense care unit  (Fig. 2B); 80.00% in the CC group and 78.72% in the ML group (HR 0.868, 95% CI 0.313-2.411, p = 0.786) for stage III (Fig. 3B). In detail, 13 cases (11.93%) were systemic recurrence and 3 cases (2.75%) were local recurrence in the ML group; 11 cases (18.67%) were systemic recurrence and 2 cases (3.03%) were local recurrence in the CC group.

Discussion
The first description of a laparoscopic operation for colon cancer was published in 1991 [25], and many studies, including some large prospective multicenter trials, have proven its short-term safety and efficacy and long-term prognosis for treating colorectal malignant disease [1][2][3][4]. Hohenberger et al. [5], who recently identified the concept of CME in colon cancer surgery, proposed that CME was superior to traditional surgery in terms of 5-year cancer-related survival. This surgical technique is based on oncologic resection with careful dissection of the mesocolon along the embryological tissue planes, resulting in a colon and mesocolon specimen lined by intact fascial coverage of the tumor and containing all blood vessels, lymphatic vessels, lymph nodes, and surrounding soft tissue, which may contain disseminated cancer cells. It has also been shown that transecting the supplying vessels at their origin and removing the entire mesocolon lead to crucial surgical planes for curative colonic resection [5,26,27]. Excision of specimens with an intact mesocolon has been found to be associated with better survival rates compared with excision of specimens with a defective mesocolon [27,28].
There are two main approaches that have been reported in the literature regarding LRH: the LM and ML approaches. However, the LM method is usually performed in open surgery. To date, there have been several reports describing the feasibility of the ML approach [8][9][10][11][12][13][14][15][16]. These reports comprise two multimedia articles [9,11], five retrospective non controlled studies [10,[12][13][14][15] and two retrospective control studies [8,16] that involved 104, 16, 3, 27, 35, 32, 172, 79, and 8 cases. In terms of its technical efficiency, short convalescence and acceptable short-term oncologic results, the European Association for Endoscopic Surgery (EAES) consensus states that the ML technique should be the preferred approach for laparoscopic colon dissection [13,29]. In 2016, Zou et al. and Li et al. from Guangdong Provincial Hospital of Traditional Chinese Medicine were the first to report the feasibility of the CC approach for right sided colon cancer [17,30,31].
In terms of the common ML approach, the primary oncologic advantages are the early vessel division and the "no touch" dissection. In 1968, Turnbull et al. [32] retrospectively reported greatly improved survival rates using "no touch" techniques (5-year survival rates for curable colon cancer, 68.85% vs. 52.13%) and suggested that conventional manipulative resection techniques for cancers of the colon should be abandoned. Hayashi et al. [33] reported that the "no touch" technique could prevent cancer cells from being shed into the portal circulation. The vessels and mesenteric division prior to tumor-containing bowel mobilization could minimize cancer cell spread though the draining vessels. In other words, the value of the "no touch" technique is debated, and the one randomized study on this topic indicated no advantage to support the routine use of the technique [34]. Second, in terms of surgical skill, the longer the lateral abdominal wall attachments of the colon are preserved, the better the exposure and the easier the dissection. As the ascending colon is fixed to the right lateral wall, counter traction can be applied to the mesentery, which facilitates mesenteric dissection. Another advantage of the ML approach is the completeness of D3 lymphadenectomy. Early separation of the white line of Toldt resulted in the distortion of colonic anatomic landmarks and an increase in colon redundancy, which evidently hinders upcoming lymphadenectomy procedures. The ML approach along the vascular anatomy (superior mesenteric vessels) can lead to more complete D3 lymphadenectomy. In the LM approach, vessel ligations are usually performed on the right side of the SMV, while for the ML approach, total exposure of SMA and SMV is typical. The ML approach, however, involves several limitations. For overweight patients, the mesocolon may be too thick for the identification of vessel landmarks, and it is difficult for inexperienced surgeons to enter Toldt's space accurately. The ML approach encounters numerous variations in vessels and anatomic complexity around the pancreatic area and is thought to be a more complex and demanding procedure with longer learning curves compared with other approaches [35][36][37]. There is also a concern of increased bacterial contamination because early ligation of the vasculature lengthens the ischemic time of the colonic segment, but this remains theoretical at this point.
For a better operation and shorter learning curves, Zou et al. and Li et al. proposed the CC approach in 2016 [17,30,31], and the current major controversy over the CC approach was that it violated the "no touch" principle mentioned before [32,33]. However, this is also one of the purposes for our study. With this approach, it is easy to enter and explore the proper plane by following the fine areolar tissue, which may help reduce potential injuries to retroperitoneal structures such as the ureter and gonadal vessels. The vascular pedicles are divided after they are exposed dorsally, which may be performed safely and easily. In either the LM approach or ML approach, the peritoneum is incised at the lower edge of the ileocolic pedicle, and the lymph nodes at the root of the pedicle (No. 203 lymph nodes) cannot be swept radically in some cases, especially in patients whose ileocolic artery crosses the SMV dorsally. Therefore, if the ileocolic artery and right colonic artery are cut on the right side of the SMV, 3.2-5.8% of positive lymph nodes along the SMV/SMA (root nodes) are left behind [38]. For dissection with the CC approach, Toldt's space is entered first, and the lymph nodes at this point can be dissected easily and completely via a joint view of the dorsal aspect and ventral aspect. For inadvertent and even massive bleeding, this approach can be managed easily and promptly, as the retroperitoneal space has been dissected and the courses of the right colic artery, veins, gastrocolic trunk, and SMV have been uncovered previously in some cases. This can be helpful for decreasing the conversion rate. The CC approach is characterized by prior retroperitoneal treatment, and the tumor-bearing area is less manipulated with this approach than with the LM approach. In our study, there was no intraoperative mortality. The median surgical time was longer in the ML group than in the CC group; the mean blood loss similar in both groups. These results are comparable with the outcomes of the ML group reported by previous studies [10,15,39] and CC group reported by Li et al. [17] (Table 5). All the above studies [10,17,39] used the CME procedure. The surgical time in the CC group in our study was shorter, the possible reason is that it has more advantages in the treatment of blood vessels as the retroperitoneal space has been dissected. In previous studies using conventional colonic surgery, postoperative complication rates have been reported to range from 9.0% to 22.5% [4,40,41]. Several recent studies using CME have shown that postoperative morbidities range from 5.7% to 19.7% [5,10,42]. For right colon cancer using the ML approach, morbidity ranges from 8.6% to 16.3% [10,15], and it was 7.5% for the CC approach in a previous study [17]. In this study, the rates of postoperative morbidity were 13.64% in the CC group and 15.60% in the ML group (Table 5). Postoperative ileus was a major complication in the CC group, and the rate was higher than previously reported studies (1.5% to 3.6%) [39,43,44].
The range of resection and dissection of lymph nodes is a key factor in terms of the radical principles of oncology and can influence the long-term outcome of colon cancer [45][46][47]. The mean number of harvested lymph nodes in our study was 16.50 (14.00, 21.25) in the CC group and 18.00 (15.00, 22.00) in the ML group. The results were roughly in accordance with the outcomes of other studies with the use of the ML and CC approaches [10,16,17,39]. The mean number of positive harvested lymph nodes was 0 (0, 2.00) in the CC group and 0 (0, 1.50) in the ML group in our study. Kanemitsu et al. reported that the mean positive node yields were 1.3 to 1.7 for the intermediate nodes and 1.3 to 1.5 for the main nodes in stage I to III right colon cancer [48]. In particular, for stage III right colon cancer, in the study by Kobayashi et al., the mean number of lymph nodes harvested was 20 ± 12, and the median number of positive nodes was 3 (range 1-23) [49].
For postoperative recovery, in our study, the time to oral intake in the CC group was shorter than that in the ML group (Table 5). However, there were no statistically significant differences between these two groups in terms of the time to first flatus, time to pull out urine and drainage tube. For ML approach, Feng et al. and Han et al. reported the similar time to first flatus and oral intake (Table 5). A meta-analysis conducted by Li et al. revealed that the LM approach required a shorter postoperative flatus recovery time than both the ML approach and the CC approach [37].
Survival is still controversial. Lacy et al. [3] showed advantages of laparoscopic over open surgery in terms of 5-year cancer related survival for stage I to III colon cancer. The results of the randomized controlled trials of COLOR (5-year prognosis) [4], COST (5-year prognosis) [1] and MRC CLASSIC (3-year prognosis) [2] suggested that RFS and OS were similar between these two groups for stage I to III colorectal cancer. For right sided colon cancer, when OS was compared between the laparoscopic and open groups, no difference was found [50]. A similar tendency was observed for stage I to III patients in terms of the rates for RFS and OS; the median follow-up period was 83 months in the laparoscopic group and 105 months in the open group [40].
There have been several studies on the prognosis of LRH for right sided colon cancer (Table 4) [15,39,43,44,51]. In our study, 5 years OS rates were 75.76% in the CC group and 82.57% in the ML group; 5 years DFS rates were 80.30% in the CC group and 85.32% in the ML group. Our results are comparable with the survival  [39] 133 ± 36 min Wang et al. [15] 113.5 ± 34.4 min Li et al. [17] 78.3 (150-215) min Blood loss Our study 50.00 (50.00, 100.00) ml 80.00 (50.00, 100.00) ml Feng et al. [10] 80 (50-300) mL Han et al. [39] 94 ± 34 ml Wang et al. [15] 74.2 ± 28.1 mL Li et al. [17] 81.6 (50-200) ml Morbidity Our study 13.64% 15.60% Feng et al. [10] 8.6% Wang et al. [15] 16.3% Li et al. [17] 7.5% Time to oral intake Our study 3.00 (1.00, 4.00) days 3.00 (2.00, 5.00) days Feng et al. [10] 3 (3-5) days Han et al. [39] 3.2 ± 0.5 days Time to first flatus Our study 3.00 (2.00, 4.00) days 3.00 (2.00, 4.00) days Feng et al. [10] 2 (1-4) days Han et al. [39] 2.1 ± 0.7 days outcomes reported before. The recurrence rate is a crucial indication for evaluating the long-term outcomes of surgery. Some researchers have shown that the 5-year recurrence rate of LRH varies from 14.7% to 29.5% [39,44,51]. Generally, it should be subdivided into 5-year local recurrence and systemic recurrence for patients with stage I to III disease. The local recurrence for stages I to III varies from 2.8% to 6.8% [39,44,51]; the systemic recurrence for stages I to III varies from 11.9% to 22.7% [39,44,51]. As demonstrated in our results, the rates of systemic recurrence were 11.93% in the ML group and 18.67% in the CC group, the rates of local recurrence were 2.75% in the ML group and 3.03% in the CC group for stages II to III during the 5-year follow-up period. Despite the limitations inherent to a retrospective study in a single institute, our findings support the concept that the CC approach can be safely performed through a laparoscopic technique with a satisfactory 5-year prognosis compared to the traditional ML approach. According to previous reports, an experienced minimally invasive colorectal surgeon acquires proficiency in laparoscopic CME for right colon cancer after performing 24-33 cases [52]. Unlike previous studies, the surgeon who carry out the CC approach at our center possess mature surgical skills of laparoscopic ML approach, and in the early exploratory research [17] including 80 patients, the results suggest that the CC approach may be a safe alternative to the ML approach. The main advantages are easy access to the retroperitoneal space by protection of the ureter, safe dissection of lymph nodes along the SMV, and a potentially shortened learning curve. But so far, no studies have been conducted on the learning curve of the CC approach. Furthermore, we did not compare the mesenteric integrity of specimens between these two different surgical approaches.

Conclusion
In this retrospective study, we present the CC versus ML approach in LRH with CME for the treatment of stage II and III right-sided colon cancer. Both approaches for LRH with CME were safe and feasible and resulted in excellent 5-year survival. The CC approach was beneficial in terms of surgical time and the time to oral intake. Several reasons, including the single-center and retrospective study, limit the high quality of this article. A randomized control study with a greater number of patients is necessary for the objective evaluation of the relatively new CC technique.
Funding No funding.

Declarations
Conflict of interest All the authors have no conflict of interest or financial ties to disclose.

Research involving human participants and/or animals and Informed consent
Written consent was obtained before the surgery. As a teaching hospital, patients were informed before surgery that their clinical data will be used anonymously for some clinical researches. Institutional review board approval was obtained prior to initiating the study.