Surgical management of adenocarcinoma of the transverse colon: What should be the extent of resection?

Abstract Transverse colon, owing its origin to midgut and hindgut and harbouring a flexure at both ends, continues to pose a surgical challenge. When compared to the rest of the colon, transverse colon adenocarcinoma is relatively uncommon. These cancers usually present late and lie in close proximity to the stomach, omentum, and pancreas. Adequate lymphadenectomy entails dissection around and ligation of the middle colic vessels. Hence, resectional surgery for transverse colon carcinoma is considered difficult. This is more so because of the variation of arterial and venous anatomy. From this perspective, the surgeon is tempted to perform a more radical operation like extended right or left hemicolectomy to secure an adequate lymphadenectomy. Such a cancer has also been dealt with a more limited transverse colectomy with colo‐colic anastomosis. For all these reasons, patients with transverse colon adenocarcinoma were excluded from randomised trials which compared laparoscopic resection with traditional open operation. Surgical literature is yet to establish a definite operation for transverse colon cancer and the exact procedure is often dictated by surgeon's preference. This is primarily because this is an uncommon cancer. The rapid adoption of laparoscopic operation favoured extended colectomy as transverse colectomy can be difficult by minimally invasive technique. However, in the recent past, cohort studies and meta‐analyses have shown equivalent oncological outcome between transverse colectomy and extended colectomy. It is time to resurrect transverse colectomy and consider it equivalent to its radical counterpart for cancers around the mid‐transverse colon.


| INTRODUC TI ON
The transverse colon, the longest segment of the large intestine, is rather unique. With the caecum and right colon, it arises primarily from the midgut and its mucosa is exposed to a similar concentration of biliary salts and bacterial composition as the right colon. But, despite its mucosal surface area being about 2.5 times that of the right and sigmoid colon, the age and mucosal-surface standardized incidence rate of transverse colon cancer is the least among other sites of colon. 1 In terms of gross pathology, only 10% of colonic cancers arise in the transverse colon. 2,3 As its proximal two-third is derived from the midgut while the distal one-third is hindgut in origin, these two segments are supplied by middle colic and left colic artery, respectively. Hence lymphatic spread of transverse colon adenocarcinoma may occur in different directions. This coupled with the fact that the transverse colon is completely intraperitoneal, covered with greater omentum, situated close to liver, stomach, pancreas and spleen, and whose operations entail mobilization of the hepatic and splenic flexures, means that surgery for transverse colon cancer poses considerable challenges. While there is no ambiguity that hepatic flexure adenocarcinoma should be treated with extended right hemicolectomy (Figure 1), 4 splenic flexure cancers have been resected by left hemicolectomy, segmental colectomy, and subtotal colectomy (Figures 2 and 3). [5][6][7] For all these reasons, transverse colon cancers had been excluded from the prospective randomized trials comparing laparoscopic with open colectomy. [8][9][10] This review focuses on the literature surrounding the surgical treatment of transverse colon cancers as there is are no unanimity about the extent of colonic resection and nodal clearance.

| LITER ATURE RE VIE W
In one of the earliest publications in 1939, Mayo  patients. The hospital mortality in the second and third group was 11.1% and 43.7%, respectively. With the adoption of the Halstedian model of cancer progression from the primary to the lymph nodes along the draining vessels, colectomy became more radical, but survival of transverse colon cancer patients continued to be poor. 11 Standard textbooks advocate transverse colectomy with the ligation of middle colic pedicle as the treatment of choice for such cancers. 12 However, it has also been managed by extended right hemicolectomy and subtotal colectomy (Figures 1-3). 13 16 Extended colectomy (EC) was performed in 69 (67%) patients while 34 (33%) underwent transverse colectomy (TC). The hepatic flexure was mobilized in 20%, splenic flexure in 18%, while in 1% of patients both the flexures were mobilized. There was no mention about the level of ligation of vascular pedicle, and mean lymph node harvest was less than five in both the groups (Table 1).
For TNM stages I-II, the 5-year overall survival (OS) was slightly higher for the hemicolectomy group (65% vs 55%) but this was not  15 In the EC arm, all underwent extended right hemicolectomy.
The mid-transverse colon was defined as the middle one-third of the transverse colon. During transverse colectomy only partial colonic resection was performed. The operations were performed with D2 or D3 lymphadenectomy and adhered to the Japanese Society for Cancer of the Colon and Rectum Guidelines. 17 The cT1/T2 tumours, with no nodal disease, are dealt with D2 transverse colectomy during which the right and left branch of middle colic artery are ligated separately and the middle colic trunk is spared. During this the para-and pericolic nodes and intermediate nodes are removed. D3 dissection, which in addition removes the main nodes, is performed for cT3/T4 disease and/or nodal metastases. This entails ligation of ileocolic, right colic, and middle colic artery at their respective origin during extended right hemicolectomy, while for transverse colectomy this would mean ligation of the main trunk of middle colic artery. There was no difference in the incidence of T3/ T4 cancers in either group but average nodal harvest was 26 vs 12 (P = .000). There was no postoperative mortality. The 5-year disease free survival (DFS) (92.4% vs 95.7%, P = .593) and 5-year OS (90.3% vs 79.6%, P = .638) were similar between the two groups. In the extended right hemicolectomy group, there was no nodal metastasis along the right colic or the ileocolic vessels; the single patient with nodal recurrence developed metastases in the para-aortic and supraclavicular nodes. Despite the omission of lymphadenectomy along the right colic vessels, the transverse colectomy group did not develop any local recurrence. It is an interesting finding as the right colic and middle colic artery may arise as a common trunk in 4% of the population 18 (Table 1), there is survival equivalence between extended colectomy and transverse colectomy.

Parameters
Rongen et al 16  To date, five such meta-analyses have been published (Table 2). 24 5-year disease-free and overall survival were also equivalent. 25,27,28 However, the quality of evidence is dilute as the studies were ret-

| SURG I C AL TECHNI QUE S
In the medial-to-lateral technique by Fujita et al, the transverse colon is lifted up and the ventral aspect of the caudal portion of the superior mesenteric vein (SMV) is exposed 32 (Figure 4A,B). Further cranial dissection over this vein brings the surgeon to the origin of the middle colic vein (MCV) which is then divided. The superior mesenteric artery (SMA) is exposed on its left side and the origin of the middle colic artery (MCA) dissected and divided.
The MCV can also be approached by incising the omental bursa and entering the lesser sac 33,34 (Figure 5A,B The "pincer technique" developed by Egi et al is a slight variation of the above technique. 36 The surgeon starts by standing on the right side of the patient, separates the greater omentum from TMC and enters the lesser sac. The root of TMC is freed from the inferior border of the pancreas thereby identifying the MCV. Dissection

ACK N OWLED G EM ENTS
Authors would like to thank the Indian Association of Surgical

Gastroenterology and Japanese Society of Gastroenterological
Surgery for providing an opportunity to present part of this paper at the 74th annual meeting in Tokyo, 17-19 July 2019. The authors also acknowledge the contribution of Mrs I Roy for her help in preparing this manuscript.

CO N FLI C T S O F I NTE R E S T
The authors declare no conflicts of interests for this article.