Chapter 23. Perspective on Gastric Cancer

Theoretical Background

Dissection of regional lymph nodes had been controversial until recently, but an accumulation of evidence has led us to conclude that D2 dissection should be the standard surgery for potentially curable advanced gastric cancer in most of the world, including Europe. Details of these studies are included in the Chap. 22. It is worth emphasizing the importance of the quality of surgical treatment in clinical trials. The initial three trials—the South African, the Hong Kong, and the Medical Research Council (MRC) trials1–3—had problems with the quality of the surgery. The first two trials were single institutional studies and, although hospital mortality was not high, survival data were poor. In the MRC study, there was no serious quality control for D2 surgery, and both hospital mortality and survival results for the D2 arm were poor. In the Dutch study, their quality effort made the results better than in the MRC study, but hospital mortality was nearly 10% after D2 dissection.4 Regarding overall long-term survival, D2 dissection was better than D1, although not systematically significantly. In 15-year follow-up, they demonstrated significantly better local control and disease-specific overall survival (OS).5 These results are not clear evidence but strongly suggest the benefit of D2 dissection in a Western population. Even in the Dutch study, most of participating surgeons had a quite limited experience with D2 surgery before the study and had quite low hospital volumes throughout the study. We would suggest that the quality of the surgery and postoperative care was not sufficient. All the meta-analyses comparing D1 and D2 resection are therefore unreliable. The Taiwanese study, a single institutional study comparing D1 versus D2, demonstrated significantly better OS after D2 dissection in Asian patients.6 Later, Hundahl et al, both in the INT-0116 and the Dutch studies, reported that insufficient nodal dissection reduces the OS of gastric cancer patients.7,8.

Safe Surgical Margin

According to the recent Japanese gastric cancer treatment guidelines of the Japan Gastric Cancer Association (JGCA),9 a proximal margin of at least 3 cm is recommended for T2 or deeper tumors with an expansive growth pattern, and 5 cm is recommended for those with infiltrating growth pattern. When these rules cannot be applied, a frozen-section examination of the resection margin (in such cases as those invading the esophagus) is recommended. For T1 tumors, a gross resection margin of 2 cm should be obtained. However, tumor borders of T1 tumors are often unclear; stepwise biopsies are often appropriate preoperatively.

Laparoscope-Assisted Gastrectomy for Gastric Cancer

In Japan, many surgeons are performing laparoscope-assisted gastrectomy (LAG) for stage I tumors, although LAG is regarded as an experimental treatment even for stage I lesions in the guidelines. At the moment, there is no evidence demonstrating equivalence of long-term survival with this procedure. There are two large randomized controlled trials (RCTs) comparing LAG versus open gastrectomy for stage I tumors. The Korean RCT (KLASS study) is expected to enroll 1400 patients and the Japan Clinical Oncology Group (JCOG 0912) study will enroll 920 patients. In many respects, LAG has limitations: lack of tactile sensation; difficulty or impossibility of widely spreading the membranes, which is essential for proper D2 dissection; a compromised reconstruction technique; and a much larger variation in surgical skill. Gastric cancer has a high incidence of peritoneal recurrence, and with the frequent appearance of tumor deposits in fatty tissue surrounding the organ and the preservation of vessels during lymphadenectomy, the risk of increasing recurrence using laparoscopic approaches seems higher than is the case for colorectal cancer surgery.

Surgical Technique

The basic technique of lymph node dissection is common for all gastrointestinal cancers, but, because of the high incidence of tumor deposits in the adipose tissue and significant tendency of developing peritoneal metastasis in gastric cancer, dissection without destroying the thin membranes surrounding the fatty tissue where all nodes and tumor deposits are imbedded is of paramount importance. To perform a proper lymph node dissection of the stomach, an understanding of the unique anatomical structure is essential. The stomach has two mesenteries: the dorsal mesogastrium and the ventral mesogastrium. Moreover, the distal part of the organ is fed by the ventral and dorsal mesoduodenum. During the rotation of the intestinal system, the ventral mesogastrium becomes the lesser omentum and the dorsal mesogastrium becomes the greater omentum. Arteries originating in the ventral mesogastrium include the right gastric artery from the proper hepatic artery and the gastric branch of the left gastric artery, which becomes quite short during the rotation of the intestinal tract, although the origin of the left gastric artery is located in the dorsal mesogastrium. The location of regional lymph node stations and vessels to the stomach10 is shown in Fig 23-1A. The last part of the antrum (4–6 cm) and the first portion the duodenum (duodenal bulb) are fed by the inferior pyloric vessels in the dorsal mesoduodenum and by the supraduodenal vessels in the ventral mesoduodenum. To treat an antral cancer, proper dissection of both the mesogastrium and also the mesoduodenum is essential. The incidence of metastasis to the infrapyloric node station is nearly 50% for distal cancers that are T2 or more, and more than 40% of those having such metastasis will survive more than 5 years. The greater omentum originally hangs between the dorsal pancreas and the stomach and gradually fuses with the mesocolon. Then it elongates toward and over the ventral pancreas and the duodenum (see Fig. 23-1B, Fig. 23-2A). To carry out complete omentectomy, dissection from the anterior pancreatic fascia is essential. The omentobursectomy is the best way to safely access the root of the right gastroepiploic vein (see Fig 23-2A).

1. About 26% of metastatic nodes are 4 mm or less in largest dimension and they look normal.11 However, they are sometimes swollen and adhere to the membranes wrapping the fatty tissue and perinodal deposits, or directly invade the neural sheath surrounding the major branches of the celiac artery. In the latter case, nodal dissection should extend along the adventitial layer of the major arteries (common or proper hepatic, splenic and celiac arteries), and, in the former case, the nerves should be preserved as well.

   To carry out a D2 dissection without splenectomy, meticulous dissection along the splenic vessels and the splenic hilum is needed. For safe dissection of this area, accurate knowledge of the basic anatomy and its variations is essential. The branch-off point of the posterior gastric artery varies widely; it is sometimes at 3–4 cm from the root of the splenic artery and sometimes close to the splenic hilum. There are three or four short gastric arteries, each of which comes ventrally from the final branches of the splenic artery going into the splenic parenchyma. The left gastroepiploic artery is usually the most caudal branch of the splenic artery coming caudad along the splenic hilum. Often, it has a common trunk with the inferior pole branch to the spleen. As demonstrated in the Fig 23-2B, all nodes are included in the layer between the posterior pancreatic fascia of Toldt and the posterior floor of the Bursa omentalis. In the past, when the majority of gastric cancers were large and accompanied by large nodal metastasis surrounding the left gastric, splenic, and celiac arteries, en bloc resection of the entire tumor required the combined resection of the pancreatic tail with the spleen. This procedure that had been carried out for prophylactic dissection of the splenic artery and hilar lymph nodes was abandoned because of the higher mortality and morbidity with limited survival benefit compared with pancreas-preserving total gastrectomy. Now, this extended surgery is utilized only for T4b tumors invading the pancreas.

2. Our preferred method of reconstruction after a distal gastrectomy is a Roux-en-Y (RY), via a retrocolic route. There are a few reasons. First, anastomotic leak is much less common after gastrojejunostomy (including Billroth type II) than gastroduodenostomy (Billroth type I). Second, there is much less risk of reflux esophagitis caused by bile reflux, which is often seen after gastroduodenostomy. Third, severe gastritis of the remnant stomach is rare, which is often seen after Billroth type I. On the other hand, some authors claim a high incidence of so-called RY stasis and the uncut RY has been proposed to avoid it. As we personally seldom have patients who develop RY syndrome, this is not caused by interruption of the peristaltic signal but by hampered movement of the anastomosed intestine due to kinking or bending that is caused by adhesions to the dissected area above the mesocolon. We construct the gastrojejunostomy using a retrocolic route and fix the distal part of the stomach to the orifice of the hole in the mesocolon, so that the entire jejunum is located below the mesocolon and has less chance to kink or bend.