Chapter 22. Gastric Adenocarcinoma and Other Gastric Neoplasms (Except Gastrointestinal Stromal Tumors)

Gastric cancer describes a broad mix of malignant neoplasms derived from the different histological components that make up the stomach. These include adenocarcinoma, lymphoma, carcinoid, and sarcoma. Gastric adenocarcinoma accounts for over 90% of all cases of gastric cancers globally.1,2 The incidence of gastric cancer decreased dramatically in the latter half of the 20th century; however, a recent rise in proximal gastric cancer incidence has been noted. Gastric cancer remains the second leading cause of cancer-related deaths worldwide (Fig. 22-1).1–4 As is the case for many cancers, the epidemiological distribution of gastric cancer demonstrates a marked variation in regional incidence—with as much as a 10-fold difference between the highest- and lowest-risk populations.5 An estimated 900,000–950,000 newly diagnosed gastric cancer cases per year occurred worldwide at the beginning of the 21st century, with the great majority of these cases found in developing countries and China.2,3,5–7 Industrialized nations continue to see a marked decline in the incidence of gastric cancer, particularly in the body and antrum. In the United States, the estimated number of new cases diagnosed in 2009 was 21,130 with the number of gastric cancer–associated deaths estimated to be 10,620.7 These numbers highlight the continued decreasing trend in both gastric cancer incidence and mortality (Table 22-1). In fact, death rates attributed to gastric cancer in the Unites States fell by over 40% for males and 32% for females between the years 1990 and 2005.7

The diagnosis of gastric cancer portends a poor prognosis with reported overall 5-year survival rates between 20 and 25% in most industrialized nations.1,6,8 Stage of disease at time of diagnosis is clearly one of the most important correlates of cancer survival. Patients diagnosed with earlier stages of gastric cancer have a distinct advantage in 5-year survival compared to those with more advanced-stage disease (Fig. 22-2). Although the 5-year survival rate for all cases of gastric cancer in the United States between the years 1996 and 2004 was 25%, it was as little as 3% for patients with distant disease and as high as 61% for those who had only localized disease at time of diagnosis.7 The survival advantage of early diagnosis is best exemplified by Japan's overall 5-year gastric cancer survival rate of 52%. This has been attributed to a high percentage of early-stage diagnosis due to mass photofluoroscopic screening of their population.2,6 In the United States where the relatively low incidence of gastric cancer does not support routine population screening, only about one-quarter of all patients are found to have localized disease at the time of diagnosis.6

The risk of developing gastric cancer is associated with a complex interrelationship between environmental factors and their influence on an individual's genetic and epigenetic make up.1,2,6,9,10 Aside from Helicobacter pylori infection, smoking, and possibly a high dietary salt intake, very few proposed environmental risk factors have been validated through scientific analysis.1–3 The effect of regional environmental influences seems apparent given the marked variation in the incidence of gastric cancer between different geographical regions of the world.6 Although regional and racial genetic variation could account for a portion of this effect, mass global migration of the population and immigrant cancer susceptibility studies indicate a large environmental effect.11–15 Epidemiological studies have shown that immigrants who travel from high- to low-prevalence regions still maintain an overall gastric cancer risk about equal to the region they emigrated from.3 The progenies of these immigrants who are born in the low-prevalence regions, on the other hand, exhibit a prevalence similar to local peoples of comparable ethnic origin.11–15 This effect points to likely environmental factors that impact an individual's risk of gastric cancer development at an early age of exposure.

In addition to environmental factors, a clear impact of genetic susceptibility on the risk of developing gastric cancer has been identified.9 This includes not only familial associated genetic cancer syndromes but the effects of similar common genomic composition in individuals sharing ethnic origins. This may account for the notable ethnic variations in gastric cancer observed among members of a population in the same geographic region.1,7,10 Not only does a geographic subpopulation's ethnicity seem to impact their risk of developing gastric cancer, it also affects their average age of presentation and response to therapy.10 What is not clear is what proportion of these observed affects are secondary to a subpopulations' shared genetic background and what may be due to shared local cultural differences within their geographic regions.

Helicobacter Pylori

H. pylori infection has been demonstrated to be linked to the development of distal gastric cancers but not cancer of the gastric cardia.16,17 The clear association between H. pylori infection and the development of gastric cancer led the International Agency for Research on Cancer (IARC) to classify H. pylori infection as a type I human carcinogen in 1994.18 This has been supported by numerous prospective trials that have estimated that H. pylori infection confers an increased relative risk for the development of gastric cancer of 2.1–20 fold.1,2,19,20 Further, a prospective Japanese study of patients tested for H. pylori by serology found that gastric cancer developed in 2.9% of patients who were H. pylori seropositive, but in none of the patients in the H. pylori–seronegative group.21 Based on early data, Correa proposed a model of gastric carcinogenesis whereby H. pylori initiated an inflammatory cascade leading to the sequential development of chronic gastritis, gastric atrophy, intestinal metaplasia, and dysplasia followed by carcinoma.22 This concept is supported by the linkage between chronic inflammation and cancer development in many other organs.

Smoking

Convincing evidence has been derived from the European Prospective Investigation into Cancer and Nutrition (EPIC) trial showing cigarette smoking as a causal factor in the development of gastric cancer. In this study, there was a marked increased risk for developing gastric cancer in patients who had a history of smoking tobacco products with a hazard ratio (HR) of 1.45. The HR increased to 1.73–1.87 for current smokers with an increased risk of gastric cancer proportional to the duration and intensity of the smoking history. Further, cigarette smoking had a more profound effect on the development of gastric cardia cancer (HR 4.10) than of antral cancer (HR 1.94). Overall, the large population-based EPIC trial found that 17.6% of gastric cancer cases may be attributable to smoking tobacco products.23

Obesity

Obesity has also emerged as a possible significant risk factor for the development of several cancer types, including cancer of the gastroesophageal junction and gastric cardia.1,2,24–27 A Swedish population-based study completed by Lagergren et al demonstrated a 2.3-fold increased risk of gastric cancer in the heaviest-weight quartile of the population compared to the lightest-weight quartile and a 4.3-fold increased risk for patients classified as obese.26 These findings were supported by the results of a large case-control study performed in the United Kingdom that demonstrated a strong association between an increased body mass index and the development of gastric cardia cancer (odds ratio [OR] 1.46), but not noncardia gastric cancer.25 Furthermore, this effect was dose-dependent and seemed independent of the presence of gastroesophageal reflux disease. The mechanism of increased cancer risk in the obese population has not been elucidated but has been hypothesized to be linked to altered metabolism and/or increased gastroesophageal reflux disease.24–26

Diet

Numerous studies have examined the possible association between dietary intake and either the development or prevention of gastric cancer.1,2 Specifically, a diet rich in fruits and vegetables has been proposed to be protective against the development of gastric cancer. Although evidence from several retrospective studies support an association between a high dietary intake of fruits and vegetables and a decreased gastric cancer risk, this association proved not to be statistically significant in prospective trial analyses.2

Another dietary association that has been extensively studied is the possible increased risk of gastric cancer in patients who consume a high intake of salts and nitrates.28,29 The evidence for this association is still heavily debated, though most published cohort and case-control studies have found a strong association between a high dietary salt intake and gastric cancer.28,29 These findings, however, have not been uniform. In fact a well-publicized population-based study out of Norway evaluating a cohort of over 73,000 patients found no association between dietary salt intake and cancer.30 Despite these findings, the majority of studies support a possible association between a high dietary salt intake and the development of gastric cancer in higher-risk populations.

Hereditary Forms of Gastric Cancer

One of the first documented cases of hereditary gastric cancer dates back to the 17th century and was described for the family of the French emperor Napoleon Bonaparte.31 We now know that up to 3% of gastric cancers are the result of hereditary syndromes.4 Although many more are likely to be characterized, several well-studied hereditary syndromes include hereditary diffuse gastric cancer (HDGC), Li-Fraumeni syndrome, hereditary nonpolyposis colon cancer, and BRCA2.9

The majority of known hereditary gastric cancers are due to HDGC that carries a high penetrance and incidence of gastric cancer in the studied kindreds.4,9 These patients usually present at an early age with a diffuse multifocal form of gastric cancer. Between 30 and 40% of kindreds with HDGC demonstrate a germline mutation of a single CDH1 allele, the gene encoding for the structural glycoprotein E-cadherin.32 A complete understanding of the sequence of events that lead to gastric cancer through this single allelic mutation is still unfolding, but there is evidence for loss of heterozygosity though somatic cell epigenetic dysregulation of the normal CDH1 allele via promoter site methylation in several well-studied cases. Because these patients have a 60–90% lifetime risk of developing a diffuse-type gastric cancer, they present an unusual therapeutic challenge.9,33 Some have advocated curative early prophylactic gastrectomy in patients who carry the CDH1 mutation, while others advocate early and routine surveillance endoscopy reserving surgery for patients found to have cancer on surveillance biopsy.33,34

Signs and Symptoms of Gastric Cancer

The signs and symptoms of gastric cancer are nonspecific and commonly found in unaffected individuals in the general population. They include dyspepsia, fatigue, and malaise among others. Other, more concerning symptoms that are often referred to as alarm symptoms, include weight loss, dysphagia, persistent vomiting, gastrointestinal bleeding, anemia, and a palpable abdominal mass.35

Dyspepsia is a very common complaint among patients presenting to primary care physicians.36 Dyspepsia is also a frequent complaint in patients with gastric cancer; however, peptic ulcer disease, gastroesophageal reflux disease, and functional dyspepsia are far more common causes of dyspepsia.35 The presence of alarm symptoms presents a more concerning clinical picture in patients with a history of dyspepsia and should alert the evaluating physician that a more extensive workup to rule out malignancy may be indicated. The presence of alarm symptoms is not specific for malignancy; in fact the incidence of alarm symptoms in dyspeptic patients is high, whereas the incidence of gastric cancer is low.35 Despite this, prospective and retrospective studies have shown that 56–90% of patients with gastric cancer had alarm symptoms at the time of endoscopy.37–40 A meta-analysis of seven prospective endoscopic studies involving over 13,000 patients reported the presence of alarm symptoms in 30% of all patients and in 62% of patients found to have gastrointestinal cancer.41 Of these, no single alarm symptom was present in more than 30% of patients with malignancy.

Another study analyzing patients who underwent urgent endoscopy for the presence of alarm symptoms or dyspepsia unresponsive to empiric therapy found that 3.8% had a gastrointestinal malignancy.42 In this study, the only alarm symptoms that were predictive of cancer were dysphagia and weight loss with ORs of 3.1 and 2.6, respectively. The presence of uncomplicated dyspepsia, on the other hand, was found to be a negative predictor for cancer with an OR of 0.1.

Although the presence of alarm symptoms is poorly predictive for the presence of cancer, when they are present in gastric cancer patients, the presence and number of alarm symptoms has been shown to correlate with an advanced stage of disease.43,44 Therefore, the presence of specific alarm symptoms may be of prognostic value in gastric cancer patients. This notion is supported by a recent study completed by Stephens et al where the presence or absence of alarm symptoms correlated with patient survival time. Here, patients were followed from their initial diagnosis of gastric cancer to their date of death. Those patients who presented with alarm symptoms had a survival range of only 7–11 months, whereas patients without alarm symptoms survived between 24 and 39 months.44

Physical examination abnormalities in early gastric cancer are generally not present. In patients with advanced disease, a palpable supraclavicular mass, generally on the left side, can be a sign of distant nodal metastasis (the Virchow node). A bulky antral tumor or extensive nodal metastases will occasionally lead to jaundice from bile duct obstruction in the hepatoduodenal ligament. A palpable abdominal mass may be found, sometimes from a bulky primary tumor, but more commonly from omental caking with metastases. Abdominal distension and ascites is a finding concerning for peritoneal carcinomatosis, as is the finding of a palpable nodule at the umbilicus (the Sister Mary Joseph node). Rectal examination may identify an anterior mass in the pouch of Douglas related to peritoneal carcinomatosis and drop metastasis to the pelvis (the Blumer shelf). In advanced disease, pallor related to anemia and evidence of weight loss may be present.

In symptomatic patients or those with a history of familial gastric cancer undergoing screening evaluation, the diagnosis of cancer is most commonly made by the finding of a mass lesion or concerning ulceration during upper endoscopy. Other patients presenting with more advanced disease may have the diagnosis of malignancy made by CT scan and biopsy of metastatic lesions. Although commonly used in the past, upper gastrointestinal contrast studies with barium or water-soluble contrast agents have largely been replaced by the complementary nature of the combination of endoscopy and CT. Once the diagnosis of gastric cancer has been made, the patient must undergo a staging workup to determine the extent of the disease and potential for curative resection.

Preoperative Staging and Selection of Patients for Surgery

Accurate preoperative staging is essential for appropriate treatment planning. The workup should include upper endoscopy with or without endoscopic ultrasonographic (EUS) evaluation to assess the extent of local and regional disease. The addition of EUS to the preoperative evaluation may improve the accuracy of preoperative staging because it has been shown to be slightly superior to computed tomography (CT) imaging in assessing tumor depth of invasion and locoregional lymph node involvement.45,46 Because EUS may provide only limited additional information to CT scan in most cases, the recent 2010 Practice Guidelines of the National Comprehensive Cancer Network consider it an optional adjunctive study.4

A CT scan of the abdomen with contrast should be performed on all patients along with pelvic CT or ultrasound in females and thoracic imaging in all patients to evaluate the tumor (T) stage, nodal (N) stage, and distant metastatic disease (M) stage, for the purpose of treatment planning. Though CT scan is a recommended and routine part of the preoperative evaluation, it has a relatively low sensitivity for evaluating tumor depth and the presence of metastatic lymph nodes.4 Newer modalities such as multidetector CT, helical CT, and positron emission tomography CT (PET-CT) have been shown to provide better preoperative staging data; however, their routine use has not yet been advocated as an essential or necessary part of the preoperative staging workup.4

In addition to radiographic staging, a complete history and physical examination, an assessment of exercise tolerance, relevant laboratory testing, and indicated physiological evaluations must be performed. These studies help provide evidence of advanced disease as well as the presence and extent of comorbid conditions that may need to be considered prior to treatment. In some patients, the presence of significant comorbid illness or limited performance status may preclude certain treatment options.

Preoperative laboratory testing should include a comprehensive metabolic panel to assess the patient's nutritional status, renal and hepatic function. Patients with a poor nutritional status may benefit from preoperative nutritional supplementation before considering surgical resection. Those with evidence of reduced renal function or poor hepatic synthetic function may not tolerate radical surgical resection or may be demonstrating signs of advanced disease that will need more extensive staging studies. A complete blood (cell) count (CBC) and basic coagulation studies should be performed to assess the status of known or possible bleeding disorders. Anemia is common in gastric cancer and may represent bleeding from the primary tumor or a vitamin B12 deficiency from associated atrophic gastritis

Patients who have signs and symptoms suggestive of cardiopulmonary disease should have at minimum an electrocardiogram (ECG) and chest x-ray (CXR) performed. Additional more extensive testing will be guided by the extent of the patient's symptoms and the findings of the ECG and CXR.

NCCN guidelines recommend that all patients with a diagnosis of gastric cancer, and particularly those with gastroesophageal junction lesions, should undergo comprehensive review of their staging, findings, and treatment planning options by a multidisciplinary cancer treatment team. Unless the patient is to be enrolled in a treatment study protocol approved by an institutional review board, treatment should be based on the current NCCN Practice Guideline recommendations.4

Staging

In the United States and the majority of the Western world, staging is based on the TNM (tumor-node-metastasis) system jointly developed by the American Joint Committee on Cancer (AJCC) and the International Union Against Cancer.47Table 22-2 summarizes the TNM-based staging system for gastric cancer. This system stratifies patients to a stage of disease that correlates strongly with patient survival. This staging system is based on tumor depth relative to the gastric wall histological layers, presence and number of involved regional lymph nodes, and the presence or absence of distant metastatic disease.

NCCN Practice Guidelines for treatment recommendations for initial therapy are based on the preoperative TNM stage of the disease.4 To be considered a candidate for curative resection, a patient must be found medically fit to withstand a major abdominal surgery, have limited locoregional disease amenable to resection with negative margins, and be free of evidence of distant metastatic disease. Those who have local disease but are found medically unfit to undergo a major surgical procedure may be candidates to undergo endoscopic mucosal resection (EMR) as discussed below. Those with advanced disease who are otherwise good surgical candidates may be candidates for palliative procedures if indicated. Patients with unresectable disease and those with extensive locoregional disease who are medically unfit to safely withstand radical surgery should be treated in a nonsurgical treatment arm based on current treatment guidelines.

Patients with advanced locoregional disease who are medically fit and have a marked response to a neoadjuvant treatment protocol should undergo complete preoperative restaging once treatment has been completed to determine whether their response to therapy renders them a potential candidate for curative surgical resection.

The 5-year survival rate for gastric cancer remains dismally low even for resectable disease. Surgical resection with curative intent remains the mainstay of therapy; however, the addition of neoadjuvant and/or adjuvant therapy has been shown to improve both disease-free survival and overall survival rates in patients with select stages of disease.4,48–50 The literature is replete with case series, retrospective studies, and a few prospective randomized control trials (RCTs) that sought to investigate the efficacy of various adjunctive therapeutic regimens in patients who have undergone a curative surgical resection for gastric cancer. In general, these studies have been plagued with inconsistent results and complicated by the use of diverse therapeutic regimens, many of which are thought to be suboptimal. Fortunately, a handful of well-designed studies and more recent meta-analysis of these data have demonstrated their therapeutic efficacy.

Neoadjuvant therapy consists of a combination of chemotherapeutic agents or chemotherapy plus radiation therapy instituted in the preoperative setting. The theoretical advantages of preoperative treatment include assessing tumor chemosensitivity preoperatively to help tailor postoperative therapy, potential early treatment of micrometastatic disease, better tolerance of therapeutic side effects, and disease down-staging to improve the number of potentially curative resections.48,49 Preoperative radiation therapy alone or in combination with chemotherapy has been demonstrated to provide a significant increase in down-staging, tumor resectability, and overall 5-year survival rates. Unfortunately, these data were obtained from RCTs of patients with predominantly gastroesophageal junction tumors and therefore may not have similar efficacy with cancers of the gastric body or antrum.48

The MAGIC (Medical Research Council Adjuvant Gastric Infusional Chemotherapy) trial51 demonstrated that the addition of combined preoperative and postoperative chemotherapy consisting of epirubicin, cisplatin, and 5-fluorouricil leads to a significant increase in overall survival and reduced disease progression when compared to patients who received surgery alone, establishing the benefit of neoadjuvant chemotherapy without radiation. Currently, the NCCN guidelines recommend the addition of neoadjuvant chemotherapy or combined chemoradiation therapy for any patient without metastatic disease who is either node positive or staged as T2 or greater in the preoperative setting.4

The role of combined modality chemoradiation adjuvant therapy in the postoperative period has been established to significantly increase patient overall survival rates.4 Until recently, the benefit of adjuvant chemotherapy in the absence of radiation has been more controversial. A recent meta-analysis published in the Journal of the American Medical Association by the Global Advanced/Adjuvant Stomach Tumor Research International Collaboration (Gastric) Group provided the first level I evidence to support the benefits of fluorouracil-based adjuvant chemotherapy.50 They reported a statistically significant increase in both overall survival and disease-free survival, when compared to surgery alone. The current NCCN guidelines recommend adjuvant therapy for patients without distant disease based on their pathological disease stage and resection margin status. The addition of adjuvant therapy is optional for those who undergo an R0 resection with stage T2/N0/M0 or lesser disease. For those who underwent an R1 resection or an R0 resection with greater than T2 or node-positive disease, an adjuvant fluoropyrimidine-based chemoradiation regimen (preferred) or alternative chemotherapy alone is recommended. Patients who undergo an R2 resection, regardless of T or N stage, have the option of undergoing therapy as noted previously for an R1 section or alternatively may be treated with best supportive care in the absence of adjuvant therapy. Unless patients are enrolled in an approved study protocol, we recommend adherence to the most current published treatment guidelines outlined by the NCCN.4

The principles of surgical resection are to obtain resection margins that are grossly at least 5 cm from the visible or palpable mass.52 The specimen should be sent for frozen-section pathological analysis to access margin status. Ideally, margins should be microscopically free from cancer, often referred to as an R0 resection.53 If the surgical margins are not initially free of microscopic disease, an additional resection should be performed if anatomically feasible. Obtaining an R0 resection is the guiding principle in determining whether to perform a distal gastrectomy, a total gastrectomy, or an esophagogastrectomy. The long-term survival rates for patients undergoing a total gastrectomy and a distal gastrectomy are similar provided that negative surgical margins are obtained.54 However, a more limited surgical resection may provide a lower rate of postoperative complications and a higher quality of life.55 Although some authors advocate performing a proximal gastrectomy for limited proximal gastric cancers, the authors of this chapter have not found consistent convincing evidence for its benefit over total gastrectomy. In proximal gastrectomy, the antrum can be preserved; however, it is not highly distensible and provides little capacity advantage compared to total gastrectomy. We have therefore decided not to include the technique for proximal gastrectomy in this chapter but recognize that it is a reasonable surgical option practiced at some centers.

It is important to acknowledge that preoperative staging is not always correct and that some patients will be found to be unable to undergo curative resection at the time of surgery. These include patients found to have previously undiagnosed distant metastatic disease, carcinomatosis, or advanced locoregional disease. It is for this reason that many authors advocate laparoscopic staging prior to advancing to a full laparotomy.4 Relatively asymptomatic patients found to have unresectable disease on laparoscopic evaluation can be referred for nonsurgical treatment options without suffering the attendant risks of a full laparotomy or radical surgical resection. Functional, but symptomatic, patients presenting with marked anemia or obstructive symptoms may benefit from a limited surgical approach to include intestinal bypass or partial gastric resection as a palliative measure. These patients do not necessarily need to undergo laparoscopic staging because they will need a palliative procedure if they are found to be unresectable for cure.

In experienced centers, patients with limited disease may be candidates for laparoscopic gastrectomy for cancer. Recent published case series and one randomized control study have shown patients undergoing laparoscopic gastric resections to have similar oncological outcomes to patients undergoing open gastrectomy techniques.56–58 The data from these studies are limited, and the selection criteria defining patients appropriate for laparoscopic resection have not yet been well defined. Sufficient randomized data have not convincingly demonstrated equivalent or superior survival following laparoscopic resection compared to laparotomy and resection. If applied, the laparoscopic approach should follow the same surgical principles as for open gastric resection. The major differences will lie in the unique requirements of the laparoscopic approach. Because it has not been studied sufficiently to regard it as the standard of care, we have elected not to include a detailed description of nuances of the laparoscopic approach in this chapter.

Extent of Lymph Node Dissection

The importance of the presence and extent of lymph node metastasis is reflected in the system used to stage the disease and provide prognosis for the patient's extent of disease. What is less clear, however, is the impact of surgical resection of regional lymph nodes on survival. The locoregional gastric nodes have been characterized based on their anatomical location relative to the stomach and are described according to their stations 1–11.59 As demonstrated in Fig. 22-3, stations 1, 3, and 5 lymph nodes are located along the lesser curvature of the stomach. Stations 2, 4, and 6 nodes lie along the greater curvature of the stomach. Station 7 nodes are found in the tissue along the left gastric artery, station 8 nodes along the common hepatic artery, station 9 nodes along the celiac artery, and stations 10 and 11 nodes along the splenic artery.

The perigastric lymphatics located at nodal stations 1–6 make up a subset of lymph nodes referred to as N1 nodes.5 Lymph nodes located at nodal stations 7–11 are referred to as N2 nodes. All other nodes encountered in the surgical dissection are considered distant nodes and are thought to preclude a curative resection when positive for metastasis. The extent of surgical lymph node dissection has traditionally been categorized as a D1 resection when care is taken to completely dissect and remove all of the N1 nodes with the surgical specimen, a D2 resection when all N1 and N2 nodes are completely dissected and removed, or a D0 resection when stations 1–6 are not completely removed.4 The extent of resection is important when the Japanese gastric cancer staging system is used; however, it is not used as a prognostic factor in the TNM staging system adopted by Western nations. When the TNM staging system is used, the most important principles of lymph node dissection that have emerged are to remove all grossly involved nodes and to obtain at least 15 perigastric lymph nodes for pathological sampling.47

The extent of lymph node dissection to be performed and its impact on postoperative survival is perhaps the most debated topic in surgery for gastric cancer. The Japanese and, more recently, several European centers have published studies advocating more extensive D2 and D3 lymph node dissections for patients undergoing surgery with curative intent.60–62 The literature on this topic contains a large number of small case series, retrospective studies, and uncontrolled noncomparative studies on the impact of various degrees of lymph node dissection concurrent with both partial and total gastrectomy for gastric cancer. Only two RCTs from Western centers that had adequate study designs and appropriate statistical analysis of outcomes measures have been published comparing D1 and D2 lymph node dissections in gastric cancer patients.63–67 Both studies evaluated similar primary outcomes measures looking at 5-year survival rates, postoperative morbidity, and postoperative mortality. Both studies concluded that a more extensive D2 dissection provided no significant benefit to 5-year survival, while those patients undergoing D2 lymph node dissections had significantly more postoperative complications and a higher in-hospital mortality rate. Other, nonrandomized single-arm trials evaluating outcomes after D2 dissections completed at specialized centers noted much lower postoperative morbidity and mortality rates, similar to rates reported for D1 dissections. These authors have criticized the data from the previously described RCT, stating that they suffered from lack of surgeon experience and operative standardization.62,68

In 2004 the Cochrane collaboration attempted to determine the superiority of D1 versus D2 lymph node dissections for gastric cancer through a meta-analysis and systematic review of the literature.68 Here they critically analyzed the literature, then evaluated properly conducted studies from both randomized control trials and nonrandomized trials with similar outcomes measures. Based on a meta-analysis of RCT, they concluded there was no survival benefit to patients undergoing a D2 dissection, with the possible exception of patients with T3-positive disease. They also concluded that there was a markedly higher operative mortality rate specifically associated with D2 dissection and concurrent spleen and pancreas resection. They also noted that both RCTs were confounded by a lack of surgeon compliance and inexperience. Based on published nonrandomized comparative studies, they concluded that D2 dissections may provide a survival benefit to patients with intermediate-stage gastric cancer. In addition, based on their analysis of published observational studies, they found that patients undergoing D2 dissections may have an overall survival advantage without a markedly increased operative mortality rate when performed at experienced centers. Overall, they noted the available published data comparing D1 and D2 lymph node dissections to be limited and seriously flawed.68

Since the publication of the Cochrane review, numerous nonrandomized comparative studies and single-armed observational studies have continued to argue the case for the benefit of a D2 dissection over a D1 dissection. The majority of these studies have shown that patients undergoing D2 dissections have superior 5-year survival rates with equivalent or better rates of operative-associated morbidity and mortality. Most authors attribute these findings to improved surgical skill and experience along with performance of D2 dissections without performing concurrent spleen and pancreas resections, as was standard for both of the published RCTs.60–62,69–71 In support of this, the Dutch RCT study group recently published a 15-year follow-up of their trial and noted an increased trend in the overall 15-year survival rate of patients undergoing D2 resections (29%) compared to patients undergoing D1 resections (21%), though the difference proved not to be statistically significant (p = .34). They did, however, find a significantly lower gastric cancer–related death rate and lower rate of locoregional recurrence in the D2 population. Based on these data, and contrary to previous recommendations from this group, they now advocate D2 lymphadenectomy for patients undergoing curative resection when performed at experienced high-volume centers using safer spleen-preserving techniques.71 Despite the increasing trend in the literature advocating D2 lymph node dissections, it is clear that further multicenter RCTs are needed to determine the risks and benefits of more extensive lymph node dissection techniques. Currently, the question of whether to perform a D1 versus a D2 dissection for patients undergoing surgery with curative intent remains to be determined by individual surgeons and their associated multidisciplinary cancer treatment teams (Table 22-3).

The system of lymph node staging in gastric cancer also continues to be a topic of debate. In a recent study of over 700 gastric cancer patients, four lymph node staging systems, including the Japanese and TMN systems were compared for their ability to predict patient outcomes. In this study the system found to be easiest to use and most predictive of postoperative outcomes was a method based on the ratio of nodes positive for metastatic disease to the total number of nodes collected, independent of the total number of nodes sampled.72 In light of data from this and other recent studies and the ongoing intense debate regarding the optimal staging of lymph nodes in gastric cancer, we are likely to see modifications to the current systems in the future.

Endoscopic Submucosal Resection

Endoscopic submucosal resection (ESR) is a minimally invasive resection technique usually performed by gastroenterologists, primarily in Japan and a few specialized centers worldwide.4,73 This procedure is reserved for relatively small (<2 cm) mucosal lesions that have a very low risk of lymph node metastasis and no findings consistent with metastatic disease.74 Recently this technique was extended to include lesions at high risk for lymph node metastasis in patients with a poor performance status or in higher-risk lesions when ESR is accompanied by a laparoscopic lymph node dissection. The results of ESR case series on patients meeting these later criteria are published in the literature, but appropriate controlled studies and scientific analysis of this technique for higher-risk lesions are absent.75,76

Distal Gastrectomy

Laparoscopic exploration of the peritoneal cavity should be considered prior to initiation of a formal laparotomy to ensure absence of carcinomatosis or distant disease that would preclude a curative resection. Patients in whom the yield of laparoscopy may be higher include those with a prolonged duration of symptoms, those with weight loss, and those with equivocal CT findings of metastatic disease. NCCN guidelines suggest that the use of laparoscopy may be helpful in complete clinical staging prior to resection, but the level of evidence supporting its use is not high enough to make it the standard of care for all patients.4 Next, an upper midline or bilateral subcostal incision is made of adequate length to allow placement of a fixed surgical retractor and facilitate adequate operative exposure. The operative sequence that follows may vary depending on the extent of lymph node dissection and the reconstructive technique chosen.

Step 1 is to identify the location of the tumor through manual palpation, or for smaller lesions, visualization of an endoscopically placed tattoo. This first step is essential in order to determine the extent of resection necessary to obtain an adequate resection margin. It is critically important in proximal lesions to ensure that a negative esophageal margin can be obtained before transaction of the duodenum. Next, if there is a high concern for metastatic disease, the duodenum and pancreatic head are mobilized in order to expose the para-aortic lymph nodes that may demonstrate signs of distant nodal spread and help establish whether there is potential for a curative resection. The retroperitoneum is then incised along the lateral border of the second portion of the duodenum. Medial visceral rotation of the duodenum and pancreatic head is performed, exposing the inferior vena cava (IVC) and aorta. The exposed para-aortic lymph nodes located in the aortocaval space are dissected and sampled. If pathologic-appearing nodes are encountered, frozen-section analysis should be performed to exclude distal nodal spread, as this would preclude a curative resection. Once the absence of distal nodal spread is confirmed, we commence with the inferior portion of the dissection.

The gastrocolic ligament is detached from the transverse colon along the avascular plane using electrocautery (Fig. 22-4). The anterior layer of the transverse mesocolon is sharply dissected to the level of the inferior border of the pancreas. This step separates the anterior mesocolonic peritoneum from the underlying vessels and posterior layer, thus skeletonizing the mesocolonic vessels (Fig. 22-5). The exposed right gastroepiploic vessels are ligated and transected. Dissection of the anterior layer of the mesocolon is typically continued to its confluence with the anterior capsule of the pancreas. The dissection will expose the left gastroepiploic vessels, which must be ligated and transected. Continued dissection of the anterior pancreatic capsule is continued to the superior margin of the pancreas, allowing exposure of the celiac, splenic, and hepatic arteries and their associated nodal beds. When a D2 node dissection is to be performed, these nodal beds are cleared of lymphatic tissue. Ideally, dissection of these nodes is delayed until the duodenum has been divided in order to facilitate exposure.

The gastroduodenal junction is palpated and evaluated for distal tumor involvement. If the area appears free of malignancy, the duodenum is divided 1–2 cm distal to the pylorus (Fig. 22-6). If tumor is palpable at the pylorus or proximal duodenal bulb, the duodenum is divided 1–2 cm distal to that point to obtain a microscopically negative resection margin. In this situation, however, care must be taken not to injure the retroduodenal portion of the common bile duct, the minor papilla, or the ampulla of Vater. The gastroduodenal artery serves as a useful landmark as it passes behind the duodenal bulb. The retroduodenal common bile duct usually lies within 1 cm to the right of this vessel. We generally complete the division of the duodenum with a GIA stapler. Others prefer to transect the duodenum between bowel clamps and close the duodenal stump with a running 3-0 absorbable monofilament suture such as PDS. There are no data to support the superiority of one method over the other. Some surgeons invaginate the duodenal staple/suture line with interrupted sutures in a standard Lembert fashion (Fig. 22-7).

If a D2 node dissection is performed, division of the duodenum provides improved exposure to the nodal bearing tissue adjacent to the hepatic, celiac, and splenic arteries located along the superior border of the pancreas. This tissue should be dissected and cleared from the region of the gastroduodenal artery to the basin adjacent to the proximal splenic artery. Care must be taken not to injure the pancreatic parenchyma or the celiac plexus.

Dissection of the lesser omentum is completed along the lesser curvature of the stomach from the inferior edge of the hepatoduodenal ligament to the right crus of the diaphragm. The retroperitoneal incision created along the lateral border of the second portion of the duodenum is extended superiorly to the confluence with the hepatoduodenal ligament at the inferior aspect of the foramen of Winslow. Next, the left lobe of the liver is retracted superiorly and to the right to expose the region of the diaphragmatic hiatus. The hepatogastric ligament is then incised from the diaphragmatic crus anterior to the gastroesophageal junction and along the hepatic border to the level of the porta hepatis at its confluence with the hepatoduodenal ligament (Fig. 22-8). The incision is then carried inferiorly along the left border of the left hepatic artery to just above the junction with the duodenum, then medically to join with the retroperitoneal incision created previously. The hepatoduodenal ligament is then incised superiorly at the level of the cystic duct, and then reflected medially, exposing the structures of the porta hepatis inferiorly. The superior and inferior resection margins of the hepatoduodenal ligament are carried posteriorly behind the portal vein. The right gastric artery and vein are identified, ligated, and transected. The hepatoduodenal ligament that had been freed circumferentially from the porta hepatis can now be taken along with the nodal bearing connective tissue medial to the portal triad. The retroperitoneal dissection is then continued to the right of the aorta superiorly to the median arcuate ligament. The left gastric artery and vein are then exposed and ligated at their origins. If a D2 dissection is to be completed, dissection and clearance of the nodal bearing tissue around the left gastric artery should be performed at this point. The retroperitoneal dissection is then carried to the left and inferiorly to join the retroperitoneal resection margin along the superior border of the pancreas (Fig. 22-9).

The point of proximal gastric resection must be determined based on the location of the lesion. This requires resection of the entire lesion with a minimum 5-cm margin free of cancer. For a distal gastrectomy, the proximal resection plane is created from approximately 2 cm distal to the esophagogastric junction along the lesser curvature to a point along the greater curvature that will allow for a 5-cm resection margin. Division of the remaining greater omentum is performed to the level of the greater curvature resection point either by dividing between clamps and suture ligating the short gastric vessels or with an appropriate surgical energy source. Care should be exercised to avoid injury to the short gastric vessels located in the unresected greater omentum. With the proximal resection line delineated, the stomach is transected either between clamps or with a surgical stapler (Fig. 22-10). The en bloc specimen should be marked to orient the pathologist to the appropriate margins and sent for frozen pathological analysis to ensure an adequate resection margin free of cancer has been obtained. Failure to obtain a cancer-free resection margin necessitates one or more attempted proximal gastric resections until appropriate margins are obtained if anatomically feasible.

Once the resection is complete, the decision as to which reconstructive technique will be used must be made. We generally perform a Billroth II reconstruction. For details on reconstructive options and surgical techniques, please review the section on operative reconstruction options below.

Total Gastrectomy

The surgical approach for a proximal gastric lesion is very similar to that outlined previously for a distal gastric lesion. The only major variation is completion of the proximal dissection at the gastroesophageal junction and diaphragmatic crura with en bloc removal of the gastric pericardial and paraesophageal lymph nodes (Fig. 22-11). The dissection of the omentum along the greater curvature must also be completed, taking care to divide the remaining short gastric vessels close to the spleen. Once the preceding dissection is completed, the proximal transaction margin is identified on the esophagus, just proximal to the gastroesophageal junction. Esophageal division can be completed with an intestinal stapling device or an angled bowel clamp can be placed proximal to the planned transaction margin using a scalpel to divide the esophagus (Fig. 22-12). As with the distal gastrectomy, frozen-section analysis of the proximal margin must be completed in order to ensure a curative resection. If a cancer-free margin cannot be obtained, the surgeon must determine whether the patient is a candidate for a curative esophagogastrectomy. Once the en bloc resection is complete and frozen-section pathological analysis has confirmed adequate operative margins, the intestinal reconstruction must be completed through the use of a Roux-en-Y esophagojejunostomy. The surgical approach to this reconstructive method is described in detail in the following text.

When determining the appropriate reconstructive method to restore intestinal continuity after gastric resection, it is important to choose a technique that will minimize long-term postoperative nutritional deficiencies.77,78 The most common of these complications include marked weight loss and dumping syndrome.77–79 Some authors have asserted that this is best accomplished by restoring gastroduodenal integrity through construction of a jejunal interposition graft after total or subtotal gastrectomy.80–82 Although numerous case reports and case series have been published on various jejunal interposition techniques, there is currently no convincing evidence to support their use or a consensus on a standardized or optimal technique. Given the lack of adequate scientific evidence to support the merits of jejunal interposition grafts, we do not currently recommend their use. What does seem clear from published studies is that the most important concepts of reconstruction are to choose a technique that restores gastrointestinal continuity while reducing the incidence of bile reflux and anastomotic strictures.

Intestinal Reconstruction after Distal Gastrectomy

Billroth II Reconstruction

Given its technical ease, reasonable long-term patency rate, and good functional outcome, we generally recommend the use of a Billroth II reconstruction after distal gastrectomy. This is achieved by identifying the jejunal origin at the ligament of Treitz, by tracing the Billroth II reconstruction distally to identify the shortest amount of jejunum necessary to create a tension-free anastomosis, roughly 15 cm from the ligament of Treitz. A shorter limb is thought to reduce the incidence of afferent limb syndrome. Once this point has been identified, it is marked with a suture to facilitate ease of future identification. Next, it must be decided whether to bring the jejunal limb to the proximal gastric remnant through a retrocolic or antecolic approach. Although there are advocates of both approaches, neither has been shown to have a true functional advantage over the other. We prefer the antecolic approach when the jejunal limb can easily reach in this manner, as it does not carry the attendant risk of retrocolic internal herniation. If limb length is an issue, the retrocolic approach may shorten the distance involved for a tension-free anastomosis. In this setting, we prefer to bring the gastric remnant down through the mesocolic defect so that the anastomosis is completely inframesocolic. This may reduce the incidence of afferent limb obstruction.

The gastrojejunal anastomosis is then created by placing the segment of the jejunal limb previously marked with suture adjacent to and in parallel with the proximal gastric remnant along its posterior-inferior margin. Once the location of the gastrojejunal anastomosis has been determined, a posterior row of Lembert-type sutures is placed to join the jejunum to the gastric wall. This is accomplished using either 3-0 Vicryl or silk-interrupted sutures along the entire posterior aspect of the anastomosis. Electrocautery is then used to create a full-thickness defect in the gastric wall anterior to the row of the posterior Lembert sutures that is long enough to facilitate a 5-cm anastomotic opening. A similar full-thickness defect is made in the adjacent segment of jejunum. An anastomosis is created using 3-0 or 4-0 PDS beginning at the posterior-middle segment. Two 3-0 PDS sutures are placed immediately next to each other and run in opposite directions until they meet in the anterior aspect of the anastomosis. The two PDS sutures are then tied together to complete the anastomosis. Next, an anterior row of Lembert-type sutures are placed using either 3-0 Vicryl or silk suture (Fig. 22-13). If a retrocolic approach was used, the defect in the transverse mesocolon must be closed by sutures between the mesocolon and the stomach to avoid internal herniation.

Roux-En-Y Reconstruction

An acceptable alternative reconstruction method is Roux-en-Y gastrojejunostomy. The Roux technique has the advantage of eliminating bile reflux into the gastric remnant but has the disadvantages of two anastomoses and the possibility of Roux-stasis syndrome. The Roux technique is approached by identifying the origin of the jejunum at the ligament of Treitz. The jejunum is traced distally to approximately 10–15 cm. A defect is created in the jejunal mesentery just below the mesenteric border of the jejunum. The jejunum is then divided either between bowel clamps, or, as we prefer, with a gastrointestinal stapler (Fig. 22-14). The mesentery is divided enough to permit the limb to reach to the gastric remnant while avoiding bowel devascularization. This usually includes division of the first anastomotic arcade of the jejunum. Care should be taken with transillumination of the mesentery to understand the vascular anatomy and preserve blood supply to both limbs of the jejunum. The distal segment of the transected jejunum is brought to lie along the posterior-inferior aspect of the gastric margin as with the Billroth II reconstruction above. A posterior row of Lembert-type sutures is place to attach the jejunum to the gastric wall. This is accomplished using either 3-0 Vicryl or silk-interrupted sutures along the entire posterior aspect of the anastomosis. Electrocautery is then used to create a full-thickness defect in the gastric wall anterior to the row of the posterior Lembert sutures that is long enough to facilitate a 5-cm anastomotic opening. A similar full-thickness defect is made in the adjacent segment of jejunum. An anastomosis is created using 3-0 or 4-0 PDS beginning at the posterior-middle segment. Two 3-0 PDS sutures are placed immediately next to each other and run in opposite directions until they meet in the anterior aspect of the anastomosis. The two PDS sutures are then tied together to complete the anastomosis. Next, an anterior row of Lembert-type sutures are placed using either 3-0 Vicryl or silk suture (Fig. 22-15).

Attention is then turned to creation of the jejunojejunostomy. The proximal jejunal staple line is anastomosed to the distal jejunal segment approximately 45–50 cm distal to the gastrojejunostomy. This distance has previously been shown to be the optimal length of the Roux limb needed to reduce the incidence of bile reflux, while also reducing excessive limb length, which may contribute to stasis and malnutrition. The two segments of jejunum to be anastomosed are aligned parallel to each other in order to create a 5-cm antimesenteric anastomosis. The anastomosis may be created with the use of a gastrointestinal stapler or may be hand-sewn in the same manner a described for the gastrojejunostomy above (Fig. 22-16). Care must be taken to close all mesenteric defects with either 3-0 Vicryl or silk to prevent the development of an internal hernia.

Intestinal Reconstruction after Total Gastrectomy

Total gastrectomy is associated with worse postoperative weight loss and increased dumping symptoms when compared to distal gastrectomy. This is thought to be due to lack of a gastric reservoir. This has resulted in an ongoing debate of whether to create a jejunal pouch either with or without a jejunal interposition technique in order to simulate the gastric reservoir function. The literature in this area has generally been inconclusive due to the lack of appropriate controls, standardized outcomes measures, and poor study design.80–82 A recent meta-analysis and systematic review of the literature has provided level IA evidence to support the use of an inverted J pouch or S pouch in conjunction with a Roux-en-Y reconstruction as a means of improving postgastrectomy-associated dumping, long-term weight loss, loss, and patient quality of life.78 This study did not find evidence to support the benefit of a pouch with a jejunal interposition in order to maintain duodenal passage of enteric contents.

In most patients, a standard Roux-en-Y reconstruction will be the preferred technique to restore intestinal continuity. The procedure will be conducted as described previously for reconstruction after distal gastrectomy with one notable variation. Instead of creating a proximal gastrojejunostomy, the proximal anastomosis will be an end-to-end or end-to-side esophagojejunostomy. This may be performed as a hand-sewn anastomosis as described previously for a gastrojejunostomy or may be performed as a stapled technique using an appropriately sized EEA stapler.

As described previously, the hand-sewn technique entails performing a circumferential reinforcing row of Lembert-type sutures using 3-0 Vicryl or silk. The posterior row is placed after aligning the anastomotic segments, ensuring the jejunal limb is not twisted. The anastomosis is then performed with two full-thickness 3-0 or 4-0 PDS sutures placed immediately next to each other in the posterior segment of the anastomosis. The two sutures are then run circumferentially in opposite directions until they meet in the anterior midline. The sutures are then tied, completing the anastomosis. When feasible, an anterior row of interrupted reinforcing Lembert sutures are placed in the same manner as was completed in the posterior row (Fig. 22-17).

If a stapled technique is used, the anastomosis is created between the transected end of the esophagus and the antimesenteric border of the proximal Roux limb near the staple line. The anastomosis is performed in and end-to-side fashion. The EEA sizers are placed in the esophageal lumen, choosing an anvil size that will allow the largest possible diameter anastomotic lumen without causing undue tension on the esophageal or jejunal wall, preferably a 25- to 28-mm stapler. A purse-string suture is placed circumferentially at the distal end of the esophagus, just superior to the transection border using a 3-0 monofilament suture. The anvil is placed in to the esophageal lumen, and the purse-string suture is tightened and tied snugly around the anvil rod. The stapled end of the Roux limb is opened, and the EEA stapler is placed through the lumen in such a manner as to allow the staple pin to be punctured through the antimesenteric jejunal border several centimeters distally. The EEA anvil is then mated to the stapling device and closed, ensuring that the Roux limb is not twisted and no extraneous tissue is present between the anvil and stapler surface. Once stapling is completed, the device and anvil are removed through the jejunum and the stapler is inspected to ensure that the presence of two completed donuts of tissue are present (Fig. 22-18). The esophageal donut should be marked as “proximal esophageal margin” and sent to pathology for permanent section. The proximal end of the Roux limb must then be closed either with a surgical stapler or hand-sewn technique.

Once the esophagojejunostomy has been performed, attention is turned to creation of the jejunojejunostomy as described for the Roux-en-Y technique for distal gastrectomy previously (Fig. 22-19).

For select patients who are thought to have a good long-term prognosis, use of a jejunal pouch should be considered in an attempt to reduce postoperative weight loss and dumping syndrome. The jejunal S pouch or inverted J pouch are both reasonable choices, although no data exit to prove the benefits of one over the other. The reconstruction is performed just as a standard Roux-en-Y technique with the exception that the pouch is created at the proximal Roux limb prior to creating the esophagojejunostomy. The pouch is created by aligning the proximal jejunum in an inverted J or an S configuration, then by creating a common channel between the overlapping jejunal segments with a GIA stapler (Fig. 22-20). Once the pouch has been formed, a standard Roux-en-Y with esophagojejunostomy is performed as was described previously.

Epidemiology

Gastric lymphoma is the second most common primary malignancy of the stomach, accounting for approximately 5% of gastric cancers.83 Over the past four decades, there has been a nearly 80% increase in the incidence of lymphoma in the United States.84 This marked increase in lymphoma incidence has been especially notable for extranodal lymphomas, where up to 40% of cases present as primary gastrointestinal tract lesions.85–87 Gastric lymphoma accounts for the majority of these cases, representing approximately 50–75 % of primary gastrointestinal lymphomas.83,88,89 Primary gastric lymphoma is typically an extranodal form of a non-Hodgkin's lymphoma (NHL), whereas, Hodgkin's lymphomas, are rarely found to involve the stomach.83,90

Lymphomas represent a diverse and heterogeneous group of neoplasms and, as a result, have been very difficult to classify. Currently, the World Health Organization (WHO) lymphoma classification system is the accepted standard used by most medical professionals worldwide.91 This classification system categorizes lymphomas based on their cell of origin and specific molecular, phenotypic, and genetic characteristics. Its most recent revision also takes into account clinical features such as patient age, site of involvement, and associated etiologic conditions.92 Although the WHO classification system has facilitated treatment approaches and standardization of research protocols, it remains a very complex diagnostic schematic due to the intrinsic heterogeneity of lymphoproliferative disorders. The revised 2008 WHO lymphoma classification system recognizes more than 25 main categories of lymphoma derived from a mature B-cell origin and more than 20 derived from a T cell or NK cell of origin.92,93

Histology

Histologically, up to 98% of primary gastric lymphomas are derived from a B-cell origin.93 Nearly 60% of these are classified as diffuse large B-cell lymphomas (DLBCL) and approximately 38% are marginal zone B-cell lymphomas of the mucosa-associated lymphoid tissue (MALT) (Table 22-4).94 Both DLBCL and MALT-associated B-cell lymphomas are thought to be associated with chronic H. Pylori infection. This linkage is better established for MALT-associated B-cell lymphomas, where as many as 90% of cases are thought to be the result of H. pylori infection and where H. pylori eradication therapy usually leads to a durable remission.95–98 The association between H. pylori infection and DLBCL is more controversial. Approximately 35% of patients with DLBCL are found to be H. pylori positive, with the majority of these patients showing concurrent MALT areas on endoscopic evaluation. Despite these data, up to 63% of patients with DLBCL have a durable treatment response to H. Pylori eradication therapy alone.98

Diffuse large B-cell lymphomas are aggressive high-grade lymphomas that may be derived from MALT-associated B-cell lymphomas.87,88 Diffuse large B-cell lymphomas frequently express high levels of Bcl-6, an oncogene found on chromosome 3.83 There are two recognized subcategories of DLBCL that can be immunohistochemically differentiated: those that resemble germinal center (GC)–type B cells (CD10+, Bcl-6–, and BCL2+/–) and those that are not GC-like (CD10–, Bcl-6+, and BCL2–).83 GC-type B cells are thought to be derived de novo from mature B lymphocytes, while non–GC-type DLBCL is thought to arise from MALT-associated B-cell lymphomas.

MALT-associated B-cell lymphomas are typically multifocal lesions. They arise from gastric mucosal lymphatic tissue, which is thought to occur as a result of chronic H. pylori infection, in most cases. MALT-associated B-cell lymphomas express the CD20 cell surface antigen, generally produce IgG light-chain antibodies, and may express CD43. Three genetic translocations have been identified for MALT-associated B-cell lymphomas that, when combined, may be present in up to 65% of cases. These characteristic translocations include t(11;18)(q21;q21), t(1;14)(p22;q32), and t(14;18)(q32;q21).83,98 Although each of these translocations produces a different direct upstream impact on cellular regulation, they all result in activation of the nuclear factor-kB cell activation pathway.99

Signs and Symptoms

The clinical presentation of patients with primary gastric lymphoma is similar to patients with gastric adenocarcinoma. The signs and symptoms tend to be nonspecific, with dyspepsia, abdominal pain, nausea, vomiting, anorexia, and change in bowel habits being the most common.83,91,98,100 Gastrointestinal bleeding may also occur and is the initial presentation in up to 30% of patients.88 With the exception of weight loss, B symptoms (weight loss, fevers, and night sweats) are rarely present in primary gastric lymphomas.83 A complete history and physical examination must be performed with particular emphasis paid to examination of all accessible nodal beds, including Waldeyer's ring. A detailed abdominal examination should be performed to evaluate for an abdominal masses or organomegaly. The patient's presenting history provides the most important diagnostic clues, as the physical examination will fail to reveal any diagnostic findings up to 60% of the time.83

Diagnostic Workup and Staging

Patients with presenting histories or objective findings on clinical examination concerning for gastric malignancy must be undergo immediate upper endoscopic evaluation. Comprehensive upper endoscopy with biopsy of concerning lesions will make the diagnosis of gastric lymphoma in more than 95% of cases.91 Once the diagnosis of gastric lymphoma has been pathologically confirmed, the patient must undergo staging of their disease to ensure initiation of the appropriate treatment algorithm and provide the patient with prognostic information. Although nearly universal acceptance of the 2008 WHO lymphoma classification system has helped to standardize the lymphoma staging workup, some variations still exist depending on the histological subtype.

All patients should undergo laboratory testing to include H. Pylori serology, a CBC with differential, liver function tests, serum chemistry panel with lactate dehydrogenase (LDH) and B2 microglobulin, and serum electrophoresis to evaluate for M proteins. Additionally, although the bone marrow is rarely involved in primary gastric lymphoma, a bone marrow aspirate and biopsy should be completed. A spiral CT scan of the neck, chest, abdomen, and pelvis is performed to evaluate for addition lesions.91,99 When DLBCL has been conformed by pathological analysis, a PET scan has been demonstrated to increase staging accuracy above that of CT scan alone with a sensitivity of more than 80% and a specificity of more than 90%.101–103 When MALT-associated B-cell lymphoma has been confirmed, additional staging should include a second, more extensive, endoscopy with mapping. Twenty to thirty biopsy specimens should be obtained from both normal- and abnormal-appearing gastric and duodenal mucosa.104,105 This is an important step because MALT-associated B-cell lymphoma is frequently multifocal and MALT-lymphoma containing mucosa may look normal on endoscopy. Repeat endoscopy with extensive biopsy has been shown to diagnose a previously unrecognized DLBCL component in as many as 10% of patients (Table 22-5).91 The addition of colonoscopy and magnetic resonance imaging (MRI) of the salivary and lacrimal glands, in addition to the standard staging practices above, may lead to the finding of multiorgan involvement in as many as 25% of patients.93 As discussed previously for adenocarcinoma, endoscopic ultrasound (EUS) may be of increased value in the locoregional N and T staging.26,106

The data obtained from the staging protocols described previously may then be used to stratify patients into a staging system to facilitate standardized treatment planning and obtain prognostic information. Several staging systems have been developed for primary gastrointestinal lymphomas and include the modified Ann Arbor staging system, the Lugano classification system, and the Paris staging system.91 Although the modified Ann Arbor staging system is the oldest of the three, it is still the most frequently used system in the United States (Table 22-6).107

Treatment

Historically, surgical resection was the primary treatment for gastric lymphoma but is now generally reserved to treat complications of the disease, including perforated viscus, bleeding, and gastrointestinal obstruction. Recent data from RCTs demonstrated comparable durable treatment outcomes for chemotherapy, radiation therapy, or surgery.83,108,109 Furthermore, chemotherapeutic treatment approaches yielded far lower treatment-associated complication rates.99 The evolution of the treatment of primary gastric lymphoma continued into the 21st century as our understanding of lymphoma histological types continued to progress. The current treatment recommendations are radically different for MALT-associated B-cell lymphomas, than for the high-grade DLBCL.83,91,99

Malt-associated B-cell lymphoma treatment arms are generally based on H. pylori status and stage of disease. Early-stage H. pylori–positive disease has a high rate of complete remission with use of antibiotics to eradicate the H. pylori infection.83,96,97 The treatment of early-stage H. pylori–negative MALT-lymphoma and H. pylori–positive antibiotic unresponsive disease remains unclear.99 Although some H. pylori–negative patients will respond to antibiotic therapy, it is generally recommended that these patients be treated with either radiation therapy alone, rituximab, or chemotherapy.83,99,110 Asymptomatic advanced-stage MALT-associated lymphoma is usually observed without treatment, given its indolent nature. Treatment, including chemotherapy and radiation therapy, is usually reserved for the relief of symptomatic disease to include bleeding, organ dysfunction, or gastrointestinal obstruction.83,99

Diffuse large B-cell lymphomas are a much more aggressive lesion than MALT lymphomas. The standard treatment regimen for DLBCL should include rituximab in addition to an anthracycline-based chemotherapy regimen.83,99 Although early studies have suggested that either complete or partial gastrectomy in patients with early-stage DLBCL may have better survival outcomes and a reduced incidence of perforated viscus, obstruction, and bleeding, when compared to chemotherapy alone, more recent data contest these findings.83

Currently, the role for surgical intervention for any stage or histological type of gastric lymphoma is generally reserved for complications of the primary disease or the chemotherapy used to treat it.83,99

Introduction

Carcinoids are an unusual group of neoplasms that arise from cells of neuroendocrine origin.111 These tumors were first characterized as “carcinoids” by Siegfried Orbendorfer in 1907 based on their microscopic similarity to carcinomas, but generally indolent clinical course.112 Carcinoids can occur in nearly any location, with approximately 55% arising from the gastrointestinal tract and 30% from the bronchopulmonary tree.113 Gastric carcinoids represent only 11.7% of all gastrointestinal carcinoids, but their reported incidence has more than tripled over the past 50 years.114,115 Although some of the observed increase may be due to more frequent use of upper endoscopic procedures over the past few decades, the increase in gastric carcinoid incidence is up relative to that of other gastric neoplasms. Overall, gastric carcinoids represent only a small fraction of gastric tumors accounting for about 1.8% of primary gastric neoplasms.116

Unlike carcinoids derived from the embryologic midgut and hindgut, gastric carcinomas do not typically secrete serotonin, characteristic of carcinoids derived from enterochromaffin cells. Gastric carcinoids generally arise from enterochromaffin-like (ECL) cells found in the acid-producing mucosa of the gastric body and fundus. These ECL cells account for approximately a third of all gastric endocrine–type cells and typically produce histamine as a means of regulating gastric acid production.117,118

Classification

Three main types of gastric carcinoids are generally recognized, although a fourth type has recently been reported. Each of the three main types of carcinoids has distinct pathological and physiologically associated conditions, different potentials for metastasis and treatment algorithms.

Type I gastric carcinoids are the most benign and by far the most common gastric carcinoid, accounting for up to 85% of cases.111,119 These lesions develop in the setting of chronic atrophic gastritis with concomitant achlorhydria, occurring in more than 1% of these patients.120 Chronic atrophic gastritis creates a hypoacidic state with development of hypergastrinemia as a result of unabated G-cell stimulation. Chronic hypergastrinemia stimulates ECL-cell upregulation and may lead to ECL-cell hyperplasia that can progress to dysplasia and the development of carcinoid.121 Type I gastric carcinoids tend to be multicentric and are limited to the body and the fundus of the stomach. They are almost always small, benign, polypoid lesions limited to the gastric mucosa and submucosa.111,122 They are generally asymptomatic lesions and have a low metastatic potential, with greater than 90% of lesions found to be confined to the gastric wall at the time of diagnosis.123,124

Type II gastric carcinoids are the rarest of gastric carcinoids accounting for only 5–10% of cases.111 They also occur as a result of ECL-cell stimulation in response to hypergastrinemia; however, in these cases the hypergastrinemic state is the result of unchecked gastrin section from a gastrinoma rather than from chronic atrophic gastritis. The majority of type II gastric carcinoids are found in patients with Zollinger-Ellison syndrome (ZES) secondary to multiple endocrine neoplasia type I (MEN1) and a small minority in patients with sporadic ZES. There is undoubtedly more to the linkage between type II gastric carcinoids and ZES hypergastrinemia because the incidence of type II carcinoids in patients with ZES/MEN1 is up to 37%, whereas it is typically found in fewer than 2% of patients with sporadic ZES.125 Like type I gastric carcinoids, type II gastric carcinoids are frequently multicentric, small tumors (<2 cm) with a relatively indolent clinical course. They occur in the gastric body and fundus and are confined to the gastric mucosa and submucosa in the majority of cases. Despite their relatively low metastatic potential, about 30% of patients have local lymph node involvement and as many as 12% will have metastatic disease at the time of diagnosis.123,126

Unlike types I and II gastric carcinoids, type III gastric carcinoids occur in the absence of hypergastrinemia or other notable pathological conditions and are therefore often described as “sporadic.” Although type III gastric carcinoids are usually found in the gastric body and fundus, they may occur in any portion of the stomach. They are the second most frequent type of gastric carcinoid, accounting for 15–25% of cases.111,119 Type III carcinoids are typically larger (>2cm) lesions that demonstrate a high mitotic rate and nuclear atypia under light microscopy.119 They are aggressive lesions that are found to be metastatic or involve locoregional lymph nodes in over 75% of patients at the time of diagnosis.125 Type III gastric carcinoids carry a poorer long-term prognosis than types I and II gastric carcinoids, with a 5-year survival rate of less than 50%.127 In addition, unlike types I and II carcinoids, type III carcinoids may be functional tumors, productive of histamine. When histamine is actively secreted by these lesions, patients may develop atypical carcinoid syndrome characterized by pruritus, cutaneous flushing, and bronchospasm.126

Presentation and Diagnostic Workup

The diagnosis of gastric carcinoid is typically made in symptomatic patients undergoing diagnostic gastroscopy.128,129 The most common presenting signs and symptoms in patients with gastric carcinoid are abdominal pain, gastrointestinal bleeding, and anemia. Other, less common symptoms include weight loss, reflux, obstruction, pruritus, wheezing, and skin flushing (Table 22-7).111,128,129 Upper endoscopy with biopsy is generally sufficient to make the pathological diagnosis of gastric carcinoid. Lesion(s) must be biopsied and removed, if possible. In addition, multiple biopsies should be obtained from the gastric antrum, body, and fundus along both the greater and lesser curvatures in order to improve diagnostic accuracy and assess for atrophic gastritis.111,130 EUS-guided fine-needle aspiration may be useful for submucosal lesions not amenable to standard endoscopic biopsy techniques.

A full history and physical examination should be performed with careful attention to a history of gastritis, abdominal pain, anemia, flushing, or wheezing. Additionally, a detailed personal and family history must be obtained looking for signs, symptoms, or a history of MEN1 syndrome. As with gastric lymphoma, the physical examination will rarely provide diagnostic clues suggestive of gastric carcinoid. Laboratory analysis can be helpful in establishing the diagnosis and type of carcinoid. All patients should have a CBC, chromogranin A, serum gastrin, serum calcium, and parathyroid hormone (PTH) level. Patients with anemia present on CBC should also have a serum B12 and anti-intrinsic factor level evaluated. Gastric pH testing will be helpful in establishing the type of gastric carcinoid present, as it is expected to be high in type I, low in type II, and normal in type III gastric carcinoids.111 Data combined from the CBC, B12, and anti-intrinsic factor studies in anemic patients help to determine whether pernicious anemia is present, indicative of type I carcinoids. A serum gastrin level will help differentiate between patients with type I/II carcinoids that typically have hypergastrinemia from those with type III carcinoid who will typically have normal gastrin levels. An elevated serum calcium and PTH level may be diagnostic of hyperparathyroidism and should alert the clinician to the potential diagnosis of MEN1 syndrome (Table 22-8). Elevation of chromogranin A is a relatively sensitive and specific marker for neuroendocrine cell tumors. An elevated chromogranin A level may be helpful in narrowing the prebiopsy differential diagnosis of carcinoid, but, more importantly, if elevated, it serves as a marker to follow the therapeutic response or disease progression and recurrence.131

For small (<1 cm) types I and II gastric carcinoids, staging workup beyond endoscopy and biopsy is generally not necessary, given the low potential for spread beyond the submucosa. Tumors larger than 1 cm, however, should undergo EUS to evaluate tumor depth of invasion and characterize possible local lymph node involvement.132 All patients with type III carcinoids and those with large types I and II tumors found to invade the muscularis propria or involve local lymph nodes on EUS should undergo more extensive workup to rule out advanced locoregional or metastatic disease.132 These patients should have a CT scan of the chest, abdomen, and pelvis at minimum with consideration for a somatostatin receptor scintigraphic study. The former is especially useful in detecting small diffuse disease that may be missed by CT.133,134

Treatment

Treatment of gastric carcinoid is based on the type and the extent of disease. Treatment algorithms are similar for both types I and II gastric carcinoids with a few notable exceptions. Type III lesions, on the other hand, are treated far more aggressively because of their relatively high metastatic potential.

Both types I and II carcinoids have a generally indolent course and develop as a result of hypergastrinemia. The treatment for these lesions remains somewhat controversial, but, until appropriate clinical studies have been completed to support less invasive approaches, a more conservative treatment regimen should be pursued. If the lesion(s) are smaller than 1 cm, confined to the mucosa or submucosa and there are fewer than six total lesions in the stomach, endoscopic resection and close endoscopic follow-up is an appropriate treatment approach.111,135 Somewhat more controversial are the recent recommendations of the European Neuroendocrine Tumor Society (ENETS) that is now advocating endoscopic surveillance only for type I gastric carcinoids less than 1 cm.123 These guidelines do not apply to type II carcinoids.

If more than six gastric lesions are present, partial gastrectomy is recommended. Because type I lesions develop as a result of G-cell–induced hypergastrinemia, a surgical antrectomy is also recommended.111 Antrectomy is not indicated for type II carcinoids because the hypergastrinemia is the result gastrinoma. In these cases, surgical resection of the gastrinoma should be completed when feasible. In all cases, if these lesions recur, clear margins are not obtained, or surveillance demonstrates progression of the disease, surgical reexcision or gastrectomy should be performed. For patients undergoing surveillance endoscopy for small type I gastric carcinoids per ENETS guidelines and showing progression of disease, surgical resection of the lesion(s) with antrectomy versus total gastrectomy is indicated, depending on the extent of involvement.

When distant metastatic disease is present, traditionally treatment has involved medical therapy including, chemotherapeutic agents, radionuclides, and/or somatostatin analogues with minor benefit but no notable impact on long-term survival. Surgical therapy for these patients, on the other hand, has generally been reserved to treat limited metastatic lesions or complications of the disease, including gastrointestinal obstruction and bleeding. Recently, a small case series of patients undergoing aggressive surgical resection for metastatic gastric carcinoid had a mean 5-year survival rate of 82% with a reported increase in quality of life.136 These data are intriguing; however, more substantial studies must be completed before this can be recommended as standard of care. Limited metastatic recurrences after gastrectomy may also be treated through surgical resection for favorable lesions or with radiofrequency ablation or chemoembolization if patients are not appropriate surgical candidates.137

Type III gastric carcinoids are far more aggressive lesions and should be approached similar to gastric adenocarcinoma.111 Once extensive metastatic disease has been ruled out, all patients should undergo partial or total radical gastrectomy, depending on the extent of the disease at the time of diagnosis. An extended lymph node dissection has been advocated by some, but data showing any benefit of a D1 versus D2 node dissection for gastric carcinoid are absent.117 Chemotherapy either as an adjunct to surgery or as the sole therapeutic modality in patients with a poor performance status or with widely metastatic disease is recommended, although it has been shown to be of limited benefit.117