Synopsis  

The worldwide annual incidence of new cases of gastric cancer—both sexes—is estimated at nearly 900 000 in the IARC–WHO database Globocan 2000; this accounts for almost 10% of new cancers in the world. Stomach cancer is frequent in Japan and some other Asiatic countries, and in some countries of South America and Eastern and Northern Europe. The overall prognosis of stomach cancer is poor because detection rarely occurs at the preclinical stage.

The stomach is lined with a single layer of columnar epithelium organized in invaginations, called pits and glands. At the cardia there is a very short segment with mucous glands. In the glands of the fundic mucosa there are cells secreting acid and pepsin. More distally, in the antral or pyloric mucosa, there are mucous glands. In adaptation to inflammation and reflux at the junction of the esophagus and stomach, the epithelial squamocolumnar junction tends to migrate into the distal esophagus, with development of a metaplastic columnar epithelium called Barrett's esophagus (BE). Columnar neoplasia occurs in the stomach or in the esophagus with columnar metaplasia. At the esophagogastric (EG) junction and cardia, columnar neoplasia arises either from the metaplastic lower esophagus or from the stomach.

In the upper digestive tract neoplasia occurs on a background of chronic inflammation. This results in striking differences in the mechanisms of progression to premalignant and malignant lesions as compared to the colon. Inflammation and mucosal atrophy are the consequence of exogenous and endogenous factors. Exogenous, aggressive factors include an infectious agent, H. pylori, and dietary factors such as salt, nitrites, and deficiency in antioxidants and vitamins. Endogenous factors include acid and biliary secretions. The diffuse epithelial alterations carry an increased risk of progression to neoplasia as compared to normal, but this occurs only in a minority of cases; those lesions are classified as premalignant conditions. Circumscribed and often non-protruding lesions of the epithelium with cell atypia and architectural disorganization are precursors of gastric cancer and are classified as premalignant lesions. The signature of malignancy is the invasion of the lamina propria. The risk of progression towards malignancy is high only when there is high-grade atypia. Moreover, small, malignant, neoplastic lesions (called de novo carcinoma) may develop in the absence of identified precursors.

When there is malignancy with invasion of the submucosa, Eastern and Western experts use similar criteria for the diagnosis of cancer. Divergences concern intramucosal neoplasia; the same lesion is called premalignant in the West (high-grade dysplasia) and malignant (intramucosal cancer) in Japan. Recently a consensus classification was adopted after a workshop in Vienna. Now the two types of lesions are included in the same group of the classification.

The detection of premalignant and early malignant lesions requires high-quality imaging, because most lesions are neither protruding nor excavated. Gastroscopy is acknowledged as the 'gold standard' procedure and its role is further increased with the development of endoscopic mucosal resection (EMR).

 Definitions   

Gastric carcinoma   

The signature of malignancy in lesions of the gastric mucosa is invasion of the lamina propria. The Lauren classification was based on the epidemiology of gastric cancer and is now confirmed by molecular biology: the tumor is classified as intestinal when there are recognizable glands, or diffuse if there is no cohesion between the malignant cells. The precursor adenomas exist only in the more frequent intestinal type; the diffuse type, in the absence of proven precursor, is called de novo carcinoma. The endoscopic appearance of the tumor is superficial when it is assumed that the muscularis propria is intact; other tumors are called advanced when it is assumed that there is deeper penetration.

The name early cancer suggests a superficial tumor with potential for complete cure after complete resection, i.e. a very low risk for lymph node metastases; this applies to intramucosal and submucosal cancer in the stomach. The depth of invasion of superficial cancer in the stomach wall is classified in the T item of the International TNM classification as intramucosal or submucosal (premalignant lesions are not included in the TNM). In the group of intramucosal cancers high-grade dysplasia without invasion of the lamina propria is classified as carcinoma in situ. In tumor registries the tumors are usually distributed in three groups according to their progression:

• Localized tumors, which are limited to the gastric wall and do not involve the serosa.
  
• Regional tumors, which show invasion of adjacent viscera (pancreas, spleen, peritoneum) or regional lymph nodes, or both.
  
• Distant tumors with metastases, either in lymph nodes or in distant viscera.
  

Premalignant gastric lesions   

Most premalignant gastric lesions develop in a mucosa modified by chronic inflammation, which is a premalignant condition [1–4]. They form circumscribed, and often non-protruding, benign neoplastic lesions of the epithelium, with architectural disorganization and cell atypia. The risk of progression towards malignancy is low in low-grade atypia and fairly high in high-grade atypia.

Premalignant lesions are clinically important. In Western countries benign neoplastic lesions of the columnar epithelium are called 'adenoma' when they are protruding (polypoid) and 'dysplasia' when they are flat or depressed (non-polypoid). In Asian countries, both types of lesions are called 'adenoma' in the stomach, with a distinction between polypoid and flat or depressed adenomas [5–8]. Neoplastic lesions of the digestive mucosa are called 'superficial' when the endoscopic appearance suggests that there is no invasion of the muscularis propria; this includes premalignant and malignant intramucosal lesions.

Gastric polyps as premalignant lesions   

Gastric polyps are frequent but do not play an important role as precursors of gastric cancer; indeed most of them are non-neoplastic lesions [9–22]; the two most frequent types of non-neoplastic polyps are the fundic cystic [14–16] and the hyperplastic lesions [17,18]. However, when a gastric polyp occurs in a context of inflammation and atrophy of the gastric mucosa, it can be associated with a synchronous premalignant or malignant lesion. This justifies a very systematic and complete exploration of the gastric mucosa. A registry of gastric polyps has been established in Bayreuth and very large series have been published, as shown in Fig. 1.

Adenomatous polyps   

These account for approximately 10% of gastric polyps; they have a sessile morphology and some may reach a large size. Their architecture is tubular or villous. A malignant focus is detected in less than 10% of cases but is more frequent when the lesion is over 1 cm in diameter. The terminology flat or depressed adenoma [5–8] is used in Japan to describe non-protruding lesions with an adenomatous architecture; such lesions are less frequent than polypoid adenomas, but have a higher risk of malignant transformation.

Cystic fundic polyps   

These are small and shiny, and characterized by dilation of the neck of the gastric glands with shortened gastric pits; they develop in a well-preserved fundic mucosa. They occur as sporadic or through a germline mutation in familial adenomatous colorectal polyposis (FAP). In the FAP syndrome, fundic cystic lesions are organized as a polyposis syndrome [14]; the phenotypic molecular characters of neoplasia are present in this polyposis (cell atypia and second-hit mutation of the APC gene) but the evolution towards malignancy is quite unusual. Sporadic cystic polyps are not neoplastic; they show a somatic mutation of the a-catenin without progression to malignancy.

Hyperplastic or hyperplasiogenic polyps   

These are detected as isolated or multiple lesions or as a polyposis syndrome. They are characterized by a proliferation of foveolar cells with elongated gastric pits. They occur in a mucosa modified by inflammation and atrophy and may show areas with intestinal metaplasia. Hyperplastic polyps often undergo regression after the eradication of H. pylori[18]. They are non-neoplastic lesions, but malignancy may occur in large polyps in 2–4% of cases; in addition, synchronous neoplastic foci may develop at distance. The systematic removal of large hyperplastic polyps is recommended because neoplastic foci are missed by simple forceps biopsies. Serrated adenomas (combining adenomatous and hyperplastic features) are very rare in the stomach.

Fibro-inflammatory polyps   

These are large non-neoplastic lesions with a congestive surface, even oozing blood; they are characterized by hypertrophy of the lamina propria.

Hamartomas and juvenile polyps   

Gastric hamartomas include gastric polyps in association with polyposis of the small intestine in the Peutz–Jeghers syndrome [21]. This is an inherited cancer syndrome, with a germline mutation of the LKB1gene in 19p. The polyps show muscular fibers in continuity with the muscularis mucosae. The progression to malignancy tends to occur in extraintestinal sites.

Juvenile polyps, isolated or organized in a polyposis syndrome, are also hamartomas [22]; their structure shows dilated cystic glands filled with mucus. They occur as sporadic lesions or through a germline mutation of the (SMAD4/DCP4) gene in 18q in the juvenile polyposis syndrome, which may be associated with juvenile polyposis in the colon. The risk of progression to malignancy occurs only when there is polyposis, but most cases occur in the colon.

Cowden syndrome or multiple hamartoma syndrome is caused by a germline mutation of the PTEN/MMAC1 gene in 10q and does not progress to malignancy.

Other polyps   

In the rare Cronkhite–Canada syndrome, an acquired disease with digestive polyposis, the gastric polyps are premalignant, with a high risk of malignant transformation.

Premalignant conditions in the gastric mucosa   

Premalignant conditions of the gastric mucosa are the target of the primary prevention of cancer. In the stomach, neoplasia occurs against a background of chronic inflammation with exposure to various pro-inflammatory factors, of which H. pylori infection is the principal. The diffuse alterations of the epithelium (atrophy, hyperplasy and/or intestinal metaplasia) are associated with an increased risk of cancer. Metaplasia, the transformation of an epithelium to another type of epithelium with distinct morphology, occurs in the esophagus and stomach [23,24]. Intestinal metaplasia type I (or complete), is comprised of absorptive cells, goblet cells, and a few Paneth cells. Intestinal metaplasia types II or III (or incomplete), show columnar 'intermediate' cells and goblet cells that secrete sialo-(type II) or sulfo-(type III) mucins.

Carditis   

In the proximal stomach, carditis, or inflammation of the epithelium at the cardia, is extremely frequent. Abnormalities include elongated pits with hyperplasia and small areas of intestinal metaplasia. The source of intestinal metaplasia at the cardia is the gastric stem cells [24]. Carditis is not a premalignant condition and there is still debate about the causal factors of neoplasia at this site.

Chronic atrophic gastritis   

In the distal (non-cardia) stomach, chronic atrophic gastritis is a premalignant condition. Inflammation of the lamina propria develops in the oxyntic mucosa of the fundus in the autoimmune gastritis type A. However, H. pylori gastritis is the usual condition; it develops in patchy areas of the pyloric and oxyntic mucosa. Chronic gastritis results in atrophy of the glands, and increased proliferation with loss of differentiation in surface and intestinal metaplasia.

 Histopathological classification of gastric neoplasia   

TNM classification   

The depth of invasion of the tumor in the gastric wall is classified according to the T item of the TNM classification; this applies to cancer and not to premalignant lesions. Superficial cancer is classified either as T m or T sm. Intraepithelial cancer is also called 'in situ'. Advanced cancer is classified as T2 (muscularis propria involved), T3 (the tumor penetrates the serosa), T4 (tumor invades adjacent structures). The invasion of the regional lymph nodes and distant metastases are classified in the N and M items. Stage grouping combines the three items as shown in Fig. 2.

Vienna classification   

The tumor is also classified by the pathologist according to its tubular, acinary, papillary architecture, and to the grade (well, moderately, poorly differentiated). The distinction in two types was established by Lauren in 1965, based upon epidemiology, and is confirmed by molecular biology. Intestinal carcinomas form recognizable glands and occur in a background of intestinal metaplasia. Diffuse carcinomas consist of poorly cohesive malignant cells infiltrating the digestive wall.

Divergences in the classification on intramucosal benign or malignant neoplastic lesions occur between Eastern and Western pathologists [25]. The same lesion is classified as premalignant in the West (high-grade dysplasia) and malignant (intramucosal cancer) in Japan. Recently a consensus classification was adopted and most differences disappear in the revised version of the Vienna classification [26–28] as shown in Fig. 3.

The tendency is to abandon the terminology dysplasia and refer to intraepithelial neoplasia (IEN) or intramucosal neoplasia. This applies not only to the stomach but also to the esophagus and the colon.

The lesions classified as high-grade dysplasia in the West or intramucosal carcinoma in the East are included in the same large group 4, [Figs 4–10] which includes high-grade intramucosal neoplasia (lamina propria respected) and intramucosal cancer (lamina propria involved).

 Epidemiology   

Geographical variations of risk   

There is a trend to declining incidence of gastric cancer worldwide. However, the actual annual numbers of new cases are still expected to increase in many countries, particularly in developing countries, in relation to the aging population, because the magnitude of the risk increases sharply with age. Stomach cancer is still the second largest cause of death from cancer in the world. Data on incidence and mortality from gastric cancer are available in cancer registries [29–35].

The risk in men is approximately twice that observed in women. In Japan, in spite of a declining incidence, the risk is particularly high and stomach cancer is still the leading tumor site [35], accounting in 1985–89 for 31% of all-cancer incidence in males and 22% in females in eight population-based registries. The considerable geographical variations in the risk are explained by environmental factors, as shown in studies examining the risk in Japanese migrants to the USA [36–38].

Proximal and distal gastric cancer   

A distinction should be made between cancer at the esophagogastric (EG) junction, which includes the proximal part of the stomach (cardia), and cancer of the distal part of the stomach, called non-cardia stomach cancer [39–44]. Tumors whose center is located at no more than 2 cm distally or proximally to the junction of the tubular esophagus to the stomach should be classified as in the EG region. At the EG junction, tumors arise either in a very short segment of BE or in the gastric epithelium of the cardia. This duality can be observed in surgical specimens and is confirmed by molecular biology [42,43].

Cancer in the proximal stomach (cancer at the cardia) is less frequent than in the distal stomach, and the proportion of stomach cancers at the cardia varies from less than 5% to more than 35% in tumor registries using the classification of subsites in the stomach.

Based on the analysis of the available data, it is estimated that cancer at the cardia represents, on average, 18% of all cases of stomach cancer. The proportion of distal stomach cancer is higher in females than in males, and higher in the developing world and Japan than in North America, Northern and Western Europe, and Australia/New Zealand.

Causal factors   

Cancer at the EG junction   

At the EG junction two distinct categories of tumors arise from a short segment of the Barrett's epithelium or from the gastric cardia. In a study of tumors located within 2 cm distal or proximal of the EG junction, the proportion of males was higher in tumors classified as esophageal in origin than in tumors classified as gastric in origin [42]. There is still debate about the causal factors of neoplasia at the cardia: trauma from the alimentary passage, acid and bile reflux, or H. pylori infection. Smoking plays a lesser role than in adenocarcinomas arising from BE and the proportion of smokers is lower in gastric than in esophageal tumors [42].

Cancer in the distal or non-cardia stomach   

For adenocarcinoma in the distal stomach, mucosal inflammation and atrophy is the premalignant condition. Autoimmune gastritis type A is the causal factor in a small proportion of cases. In most cases; H. pylori infection is the central risk factor for gastritis and for the promotion of early stages of neoplastic transformation [45–62].

The niche of H. pylori is the stomach acidic milieu and the pathogen impact is modulated by the host response and the virulence of the strain. The interference between bacterial antigens and blood-group substances explains a higher risk for stomach cancer in blood group A.

The prevalence of the infection in adults is high as shown in published data from population-based surveys of healthy adults, prospective (cohort) studies, or case-control studies. In the age range 45–64, the prevalence of infection is estimated at 58% in developed countries and 76% in developing countries.

A higher frequency occurs in African countries. In South Africa, the sero-prevalence is over 80%. In Gambia the high prevalence of the infection results in a high frequency of chronic gastric inflammation, although the prevalence of gastric atrophy is apparently low.

In China the prevalence of the infection is lower than in Africa.

Most infected persons never develop gastric cancer in the absence of other promoting environmental factors, such as a high intake of irritants (salt and nitrates) and a low intake of antioxidants (fruit and vegetable) [63,64].

The relative risk for gastric cancer in H. pylori infection is estimated in prospective studies, using serological tests with a long interval between the initial blood sampling and cancer detection. In the most recent meta-analysis from the Helicobacter and Cancer Collaborative Group, the Odds Ratio for the association between the infection and distal gastric cancer is 5.9 when restricted to studies with an interval of at least 10 years.

An estimation of the proportion of stomach cancer attributable to H. pylori infection requires the exclusion of cancer at the cardia, where H. pylori plays no role. An analysis based upon the average proportion of cancer at the cardia (18%) and the worldwide estimation of the number of new cases of stomach cancer in 2000 in developed and in developing countries indicates that, for the year 2000, there were around 725 000 new cases of non-cardia cancer.

Taking into account the estimated prevalence of the infection and the associated relative risk of 5.9, the proportion of cases attributable to H. pylori infection is slightly higher in developing than developed countries, the figures being in the range of 75–80%. In this way H. pylori infection is confirmed as a premalignant condition.

In conclusion, the etiology of stomach cancer is linked to environmental factors, including H. pylori-induced inflammation and atrophy of the gastric mucosa, and a diet rich in salt and nitrite and poor in fruit and vegetables. The variable incidence in different countries of the world is linked to environmental factors.

Time trends in incidence and mortality from gastric cancer   

A generalized decline of the disease   

The generalized decline of incidence, particularly marked in the USA [65], suggests that regression occurs through changing lifestyle and that geographic differences result from variations in the control of environmental factors. Time trends in the incidence of stomach cancer illustrate the role of environmental factors. Studies on migrant populations coming from a country with a high risk for stomach cancer to a low-risk country have been conducted in Japanese migrants in the USA (Hawaii, California, Washington state, New York City).

The incidence rate in Japanese migrants decreases in successive generations [36–38]; in 1973–86, the incidence rate for Japanese born in Asia was lower than in Japan but higher than that of US white natives (×4.3 for men and ×5.8 for women). For the second generation, US-born Japanese had a further decrease in incidence (×2.8 for men and ×2.9 for women). The high rate observed in the first generation of migrants suggests that environmental factors play a determinant role in early infancy. The incomplete decrease of the risk in the second generation suggests the role of persisting ethnic lifestyle characteristics in the USA.

Time trends in Japan   

The decline in incidence and mortality from stomach cancer also occurs in Japan, with the concurrent intervention of a screening policy. The temporal trends for the period 1975–95 in Japan and USA (a country with no screening policy [66]), show the respective influences of both factors.

At the beginning of the study period, the respective age-adjusted incidence rates in Japan and USA in 1975 were 76.0 and 9.52 per 100 000 for men, and 38.34 and 4.26 for women. From 1975 to 1995 the mortality-to-incidence ratio decreased in Japan from 0.8 in 1975 to 0.6 in 1995 for men and from 0.8 to 0.7 for women. In the USA the ratio remained stable. The decline of the ratio in Japan is explained by the increasing percentage of cases detected at the localized stage in relation to screening.

With respect to mortality rates, a significant cohort effect is confirmed in Japan. When the relative risk for stomach cancer in the 1910 cohort is assessed in reference (RR = 1) to the successive cohorts, the Odds Ratio for stomach cancer decreases in both sexes, and reaches the value 0.2 for the cohorts born after 1950.

 Gastric carcinogenesis   

From inflammation to cancer   

At the molecular and cellular levels, the development of premalignant lesions in the gastric epithelium, and the subsequent evolution to malignancy, is the consequence of pro-inflammatory stress [67–73]. Stress conditions may affect many signaling pathways within the epithelial cells and program them to develop into specific differentiation pathways. Modifications of gene expression in response to stress factors may be reversible, and metaplasia, a central event in gastric carcinogenesis, may result primarily from reversible events.

Irreversible changes are mutations that cause clonal expansion into a lesion with high potential for evolution towards neoplasia. Among 'irreversible' events, the most common is mutation of the tumor suppressor gene TP53. Thus, there is a direct correlation between mutagenesis and inflammation. Prevention strategies (such as eradication of H. pylori) aim at reversing the 'reversible' events, to restore normal epithelial function and differentiation patterns.

Gastric carcinogenesis follows the morphological pathway of H. pylori infection described by Correa [67,68] for the intestinal type of gastric cancer—atrophic chronic gastritis—intestinal metaplasia—dysplasia.

However, H. pylori infection may play a different role in the diffuse type of gastric cancer [69]. The carcinogenic influence of H. pylori is suggested by the higher cell proliferation occurring in the H. pylori-infected stomach [70,71]. The balance between proliferation and apoptosis has been examined by immunofluorescent methods in biopsy samples from patients with a normal gastric mucosa (negative for H. pylori infection) and from patients with chronic gastritis (positive for H. pylori infection). In chronic gastritis, the apoptosis index increases less than the proliferation index and there is an overall decrease of the apoptosis/proliferation ratio. H. pylori infection induces an increased expression of COX 2 and is the source of an increased production of nitric oxide (NO) favoring nitrosation and formation of carcinogens. However, the bacterium is not a direct carcinogen class I.

The APC mutation in gastric carcinogenesis   

In colonic neoplasia, the primary event in the Fearon and Vogelstein sequence is the mutation of the APC gene (5q) and the inactivation of the APC protein; then the accumulation of a-catenin triggers the mutation cascade. In gastric carcinogenesis, the APC mutation is not a primary event, and malignant foci are present only in a small proportion of the polypoid adenomas [72]. The early APC mutation present in benign precursors (polypoid or flat adenomas) is not the key of the progression towards malignancy; moreover the APC mutation is often absent in carcinomas.

In a recent study of gastric neoplastic lesions [73], the APC mutation was present in 77% of polypoid adenomas (n = 35), 74% of non-protruding areas of dysplasia (n = 47), and only 3% of adenomas with focal cancer (n = 35) and 4% of 'intestinal'-type carcinomas (n = 54). Protruding or non-protruding premalignant lesions have the same phenotypic molecular characters and a similar rate of proliferation (K-67 test); a high instability character (MSI-H) may explain the higher risk of malignancy of non-protruding precursors.

Mutagenesis in the Lauren classification   

The Lauren classification of gastric tumors into two categories is now confirmed [69], with distinct pathways in intestinal or diffuse tumors.

In the intestinal type of differentiated carcinoma, precursor lesions (protruding or non-protruding adenomas) play a role. The sequence of mutations includes altered oncogenes (K-ras, a-catenin), suppressor or regulatory genes (APC, TP53), and mismatch repair genes (hMLH1).

In the diffuse type of stomach cancer there are no precursors and the tumor is called a 'de novo'carcinoma. The somatic mutation in the regulatory gene (16q) for the transcription of the E-cadherin plays the major role.

Hereditary gastric cancer   

Hereditary cancer plays a small role in the upper digestive tract. In hereditary syndromes, all of the mucosa is classified as a premalignant lesion. This applies to hereditary stomach cancer, described in limited geographic areas in New Zealand and Portugal. The diffuse type of gastric cancer relates to a germline mutation of the regulatory gene (16q) for E-cadherin [74,75]. The same somatic mutation occurs in sporadic diffuse cancer.

In other hereditary syndromes the increase in the risk for gastric cancer is less and the gastric mucosa is only classified as a premalignant condition. This occurs in the HNPCC syndrome or non-polyposis hereditary colorectal cancer. Stomach or endometrial carcinoma also occurs. In familial adenomatosis polyposis (FAP), gastric polyps of a fundic cystic polyposis show a second-hit mutation of the APC gene [16] and may even be classified as neoplastic lesions which do not progress to malignancy. The risk of upper digestive cancer in FAP is in the duodenum. Malignancy can also occur in the hamatomartous polyps of the Peutz–Jeghers syndrome (germline mutation of the LKB1 in 19p) or in juvenile polyposis syndromes (germline mutation on the SMAD4/DCP4 in 18q).

 Symptoms of gastric cancer   

Digestive symptoms of advanced gastric cancer are either a continuous or periodic pain, or digestive hemorrhage, or dysphagia suggesting stricture at the cardia, or vomiting and stasis suggesting stenosis at the pylorus. Systemic symptoms are the loss of weight and chronic anemia. Distant metastases, in the liver, can be the first clinical manifestation. Early cancer is asymptomatic or associated with a pseudo-ulcerous post-prandial pain, or 'functional dyspepsia'.

 Endoscopy in the diagnosis of gastric cancer   

Methods   

Upper gastrointestinal endoscopy is the gold standard procedure for the detection of gastric cancer. Advanced gastric cancer with a tumor protruding or excavated in the stomach is easily detected. However, advanced diffuse and infiltrative tumors involving all the thickness of the gastric wall (linitis) may have an inconspicuous signature at the surface of the mucosa; a false negative conclusion is possible if the absence of expansion of the gastric wall and its rigidity are not detected.

Nowadays, the focus is on the detection of early neoplastic lesions with a superficial endoscopic pattern suggesting that the lesion is limited to the mucosa or the submucosa. This superficial pattern corresponds either to a premalignant precursor or to a carcinoma. In the stomach, most superficial neoplastic lesions are non-protruding and demand extreme care during endoscopy.

The first step is to detect a suspect zone of the mucosa that is slightly discolored, or elevated, or depressed, or with an irregular microvascular network during the standard exploration.

The next step is the further examination of the abnormal zone after chromoscopy [76–79]. In the stomach, the most commonly employed dye is indigo-carmine solution (0.5–1%), a contrast stain which enhances depressions where the dye accumulates.

At the EG junction   

Neoplastic lesions arise from the esophagus in a short segment of BE, or from the gastric mucosa at the cardia. The detection of superficial neoplastic lesions at the cardia is therefore indissociable from the exploration of the distal esophagus. The careful exploration of the EG region with a retroflexion maneuver is a systematic step of upper GI endoscopy.

In the non-cardia stomach   

The first step in the exploration of the mucosa is the detection of chronic gastritis. Then, the entire surface of the mucosa, classified as a premalignant condition, is examined sector by sector, before and after indigo-carmine chromoscopy, in order to detect any suspect area. Most authors adopt this method but some still use diffuse staining of the gastric mucosa with methylene blue in order to detect intestinal metaplasia [77].

Technological advances in equipment   

The most recent endoscopes based on digital technology have reached a high standard in resolution and color reproduction with the surge of multiple technological advances [76–83]. Two major directions are stressed: magnification with an optical zoom, and processing of the digital image captured by the CCD. Digitization is easier for instruments with the RGB sequential system (series 200 of Olympus), than for the series with a color chip (Exera of Olympus).

Magnification   

Magnification in endoscopy is now available, using either an optical zoom (×30 to ×80) or a combined optical and electronic zoom. Magnification is of considerable help for mapping the distinct types of epithelium (oxyntic, cardial, and intestinal metaplasia) in the stomach and at the EG junction [77–83].

In a study conducted in Portugal [77] magnification in contrast, using methylene blue chromoscopy, was used in 136 patients with previously diagnosed gastric lesions (gastritis, dysplasia, metaplasia). Comparison of endoscopic findings with the biopsies was controlled by five blind observers; the reliability of the procedure for the detection of occult neoplastic areas was confirmed.

In neoplastic areas, magnification has two distinct applications: the analysis of the disorganized architecture of the epithelium (pit pattern) in surface, and of the microvascular network disorganized by inflammation or neoplasia [82,83]. Magnification by the digital processing of the image is a promising technology ensuring a relatively reliable optical biopsy.

Digitization of the image   

The structure enhancement processing system uses a digital filter to enhance the frequencies containing important information for the diagnosis and reduce the non-useful frequencies; it is present in the RGB and color chip series of Olympus.

The color enhancement or index of hemoglobin (IHb) requires the sequential RGB system (model Celera 260 of Olympus); without altering the other colors, it will change the color tone in the band of hemoglobin, shifting to the red the areas with more blood and hemoglobin and to the yellow the areas poor in blood and hemoglobin. The IHb system enhances the color contrast of small neoplastic lesions.

Narrow band imaging (NBI) is linked to magnification. The incident light is restricted to narrow bands in the three fundamental colors (red, green, blue) to obtain distinct superposed images in depth (red), intermediate (green), and extremely superficial (blue), which result in increased relief of the magnified surface architecture, and enhanced view of the microvascular network.

Spectroscopic techniques   

Spectroscopic techniques offer a new perspective in research, and aim to open a new sector of molecular endoscopy; however, they are probably not cost-effective and should not be used in routine diagnosis. UV fluorescence endoscopy, and trimodal spectroscopy (absorption, fluorescence, and elastic dispersion) should contribute more to the study of neoplasia in Barrett's esophagus, than in the stomach.

Spectroscopy in the near infra-red (NIRS) explores the deep layers of the mucosa or the submucosa, and is adapted to the analysis of the microvascular network in neoplastic lesions after injection of indocyanine green as a fluorescent dye. The pooling of fluorescence from indocyanine green is a signature of invasion of the submucosa.

Macroscopic appearance of digestive neoplastic lesions   

The morphological appearance of advanced gastric cancer was classified by Borrmann in 1926 into four distinct types.

• Type 1—polypoid carcinomas, usually attached on a wide base.
  
• Type 2—ulcerated carcinomas with sharply demarcated and raised margins.
  
• Type 3—ulcerated, infiltrating carcinomas without definite limits.
  
• Type 4—non-ulcerated, diffusely infiltrating carcinomas, as shown in Fig. 11, and Fig. 12.
  

The Japanese gastric cancer association added a type 0 corresponding to neoplastic lesions with a superficial extension, limited to the mucosa and the submucosa, and a type 5 for unclassifiable advanced tumors [84–86].

This morphological classification was adopted in Japan and further accepted for the esophagus and the colon. The type 0 lesions, presumed to be superficial, were divided in type 0-I or protruding (polypoid), 0-II or-non-protruding, and 0-III or excavated, as shown in Fig. 13 and Figs 14 and 15.

Type 0-I lesions are divided in type 0-Ip and type 0-Is—pedunculated and sessile. In pedunculated polyps the base is narrow; in sessile polyps the base and the top of the lesion have the same diameter. In non-protruding lesions the surface is either slightly elevated (type 0-IIa) or flat (type 0-IIb) or slightly depressed (type 0-IIc); noteworthy elevated lesions are classified non-protruding when the protrusion from the adjacent mucosa is not more than 2.5 mm; above this figure the lesions are classified protruding and sessile.

The Japanese classification of type 0 neoplastic lesions applies to premalignant lesions (low-grade or high-grade intraepithelial neoplasia), carcinomas, and also to non-neoplastic lesions which may show a neoplastic component (juvenile or hyperplastic polyps). In non-protruding lesions priority is given to the morphological subtypes; malignancy with invasion of the submucosa is more frequent in depressed lesions (type 0-IIc).

Endoscopic classification of superficial neoplastic gastric lesions   

At the EG junction   

Premalignant and early malignant lesions at the EG junction and the cardia have the same morphology as in BE and the proportion of depressed (0-IIc) lesions is less than in the non-cardia stomach.

In the non-cardia stomach   

Most early neoplastic lesions (premalignant or malignant) are non-protruding, and most (70–80%) are depressed (0-IIc) as shown in Fig. 16. The polypoid adenoma is a rare precursor compared to flat or slightly depressed areas. Focal cancer in the adenomatous polyp is rare (8 cases out of a series of 118) and occurs when the size of the polyp is more than 1 cm [72]; on the other hand a complete exploration of the gastric mucosa is necessary because synchronous non-protruding neoplastic areas may be present at distance from the polyp. In the very large histopathological series of gastric polyps from the Institute of Pathology in Bayreuth [9,10,13], polypoid adenomatous lesions are not frequent while non-neoplastic polyps (fundic cystic polyps and hyperplastic polyp) represent 80% of all lesions, as shown in Fig. 1.

Flat or depressed adenomas have a similar frequency of p53 mutations as protruding adenomas; on the other hand they carry a higher risk of malignant transformation [5–7], as confirmed by the higher expression of CEA and Ca19–9.

In conclusion, most premalignant or early malignant lesions in the stomach are either completely flat or slightly elevated and often depressed. This is confirmed by the data on the morphology of type 0 gastric cancer in the National Screening Campaign in Japan [86] (Fig. 16). In this series 70% of type 0 cancers have a non-protruding depressed morphology (type 0-IIc isolated or combined with other type, and only 6.6% have a polypoid morphology.

 Non-endoscopic procedures in the diagnosis of gastric cancer   

Radiological imaging and ultrasound   

Since the development of endoscopy, there are fewer indications for radiology as a diagnostic procedure of stomach cancer: double-contrast radiological imaging proves less effective than endoscopy. However, in screening protocols the miniature photofluorography is still accepted as a useful test. In the preoperative exploration of patients with advanced gastric cancer, other imaging procedures such as the CT scan and transcutaneous ultrasound with high-resolution probes play a major role.

Molecular biology   

The assessment of histology samples classified as negative, indefinite, and low-grade intraepithelial neoplasia is poorly predictive of the clinical future outcome. This applies particularly to Barrett's esophagus and a large number of biomarkers of neoplasia have been proposed [87]. The same biomarkers have also been tested in gastric columnar neoplasia. Biomarkers are tested in the tissue samples obtained during endoscopy.

Proliferative indices   

Ki67, or PCNA proliferating cell nuclear antigen have not been shown to be predictive of the future outcome.

P53 protein and TP53 mutations   

Multiple non-prospective studies confirmed a high prevalence of the mutated p53 protein in high-grade intraepithelial neoplasia and in confirmed cancer, but p53 immunochemical overexpression is not helpful in the prediction of the risk of cancer in premalignant conditions such as chronic atrophic gastritis.

The DNA sequencing of p53 mutations is more reliable, but requires a sophisticated technology not compatible with routine tests. In tumors of the EG region, the type of the p53 mutation differs according to the esophageal or gastric origin of the tumor. The TP53 mutations in adenocarcinoma arising from BE shows a higher frequency of the G:C > A:T at CpG than in any other cancer type.

Cytokeratins   

The ratio of cytokeratins (CK) has been proposed as marker for intestinal metaplasia in the esophagus and stomach [37–41]. CK 7 is more abundant and CK 20 is less abundant in intestinal metaplasia arising from the esophagus. The cytokeratin ratio also helps in the classification of cancer of the EG region.

In a study of 85 adenocarcinomas of the distal esophagus and 67 adenocarcinomas of the cardia, the CK 7+ and CK 20-pattern is characteristic of esophageal cancer. Ninety per cent of esophageal tumors are positive for CK 7 against less than 45% for tumors at the cardia; 17% of esophageal tumors express CK 20 against 55% for tumors at the cardia.

Mucins   

Mucins have also been used as markers of intestinal metaplasia in Barrett's esophagus; there is a distinct pattern of mucins with the MUC 5B, MUC 6 types.

In conclusion, biomarkers illustrate the successive steps in carcinogenesis, and may be positive in tissue samples collected in premalignant conditions (Barrett's esophagus or chronic gastritis with H. pylori infection) or in premalignant lesions with low-grade atypia. However, their predictive value for the clinical outcome and detection of those patients at high risk of developing a cancer is poor.

Staging of gastric cancer   

In advanced cancer the major objective is the decision for a curative or palliative surgical resection, through the detection of invasion in adjacent viscera or the presence of distant metastases.

In advanced cancer, tumor staging relies on non-endoscopic procedures such as transcutaneous US and CT. The PET scan using fluorodeoxyglucose has proved to be more accurate than the CT scan in detecting lymph nodes. Endoscopic ultrasound (EUS) may be of some help in assessing invasion of regional lymph nodes or of adjacent viscera.

In superficial type 0 tumors, staging of the depth of invasion in the submucosa helps in treatment decisions between curative endoscopic mucosectomy or surgery. EUS with radial echoendoscopes or with miniprobes is widely used in Japan. In this indication high-frequency (20–30 MHz) thin probes introduced through the operative channel of the endoscope are superior to conventional echoendoscopes. This proves particularly helpful for esophageal cancer. The efficacy of staging early gastric cancer (EGC) with a 30 MHz miniprobe is hampered when there is an ulcer scar.

 Clinical relevance of early diagnosis of gastric cancer   

EG tumors are rarely detected at the preclinical stage, and the prognosis is poor. Common sense suggests that there should be greater efforts to increase early preclinical detection. The secondary prevention of cancer is based on the early detection (and treatment) of premalignant or malignant superficial lesions. Of course the prognosis of tumors detected in the preclinical stage is improved when compared to that of symptomatic tumors detected at the advanced stage; however, an evaluation in terms of cost–effectiveness and cost–benefit is legitimate.

Survival after treatment of a superficial cancer detected at the preclinical stage is much longer than for tumors detected at the advanced stage. However, this optimistic figure involves lead-time bias.

Survival at the preclinical stage aggregates two periods: the time elapsed during the preclinical phase plus the time elapsed from the first symptoms until the symptomatic detection.

The survival benefit is obvious if the final date of death is delayed, as compared to detection at the symptomatic period. In addition the lead-time bias plays a role; tumors detected at the preclinical stage often progress very slowly and may even not reach a symptomatic stage before death from another cause.

 Treatment decisions for gastric cancer   

The role of tumor staging   

Treatment decisions are based on our knowledge of the tumor progression in successive stages. After staging the choice is between treatment of the tumor with a curative intent, i.e. a complete destruction of the neoplastic tissue, and palliation using surgery, endoscopic procedures, and chemotherapy.

When the tumor is localized there is a distinction between superficial cancer with a depth limited to the submucosa (T1 in the TNM classification) with a low risk of lymphatic metastases, and advanced cancer where the tumor involves the muscularis propria or the serosa (T2 and T3 in the TNM classification) with a high risk of lymphatic metastases. The risk of metastatic lymph nodes is small, but never nil, for intramucosal neoplasia, and is significant when the depth of invasion in the submucosa is over 500 µm.

Tumors with regional lymphatic metastases or tumors involving the adjacent viscera (T4 in the TNM classification) are classified as regional, but may still prove resectable.

As a rule, in non-localized tumors with distant metastases a complete resection is impossible.

Treatment with curative intent   

When the curative option is selected, the choice is between curative surgery with a R0 resection or curative endotherapy.

Surgery with a R0 resection and lymphadenectomy is the standard treatment of gastric cancer.

Endoscopic mucosal resection (EMR) for superficial cancer is an elective option for a significant proportion of the so-called early gastric cancers (EGC), and the elective treatment of premalignant lesions [88,89]. EMR for gastric cancer is a common indication in Eastern Asia. Fewer indications are foreseen in Western countries with the decreasing incidence of the disease and the low proportion of cases detected at the early stage. The guidelines for selecting the EMR option are as follows:

• Early detection with the most recent models of video-endoscopes and the routine use of chromoscopy (indigo-carmine).
  
• Evaluation of the depth of invasion of the tumor according to its endoscopic morphology in the Japanese endoscopic classification of superficial (type 0-II) neoplastic lesions with subtypes 0-I, 0-IIa, 0-IIb, 0-Iic, and III. Specific care is given to the analysis of depressed lesions (0-IIc) which show an increased risk of deep invasion in the submucosa.
  
• EUS staging with a high frequency miniprobe.
  
• In operable patients a careful evaluation of the alternative management (i.e. surgery).
  

Other therapeutic options   

When low-grade intramucosal lesions are detected in aged persons, expectant management is an acceptable option.

In advanced localized or regional cancer, chemotherapy and radiation protocols are options complementary to surgery.

In non-operable patients, endoscopic palliation aims at the restoration of the digestive lumen with expandable stents.

 Endoscopic treatment with curative intent   

According to the guidelines of the Japanese Gastric Cancer Society [84,85,89], there are increasing indications for endoscopic surgery in patients with premalignant lesions or early malignant gastric lesions.

For protruding lesions (polyps) the elective treatment is resection with a diathermic snare. Solitary polyps (adenomatous or hyperplastic) are systematically resected if their size reaches 1 cm. Hamartomatous polyps of the Peutz–Jeghers syndrome are also resected.

In gastric polyposis, a few polyps are resected for a histopathological diagnosis. Non-protruding lesions can be destroyed by thermal or non-thermal procedures. However, this yields no specimen for pathology. The most powerful instrument for thermal destruction of tumors is the Nd:YAG laser. Other thermal procedures (electrocoagulation with a diathermic probe or the Argon plasma coagulator) have been proposed for very superficial neoplastic lesions. Actually their efficacy is limited to the destruction of small residual neoplastic areas after a more powerful intervention.

Photodynamic therapy, a non-thermal procedure, can be performed after administration of an exogenous fluorophore (photofrin) or a metabolic precursor of protoporphrin IX (5-amino-laevulinic acid). The procedure is adapted to the curative treatment of superficial and poorly delimited neoplastic areas [90–92]. Presently, priority is given to endoscopic mucosal resection [93–108].

Technique of endoscopic mucosal resection (EMR)   

The methodology of EMR for non-protruding neoplastic lesions in the stomach is now established, and is illustrated in Figs 17–21. Most procedures require a single-channel rather than a double-channel endoscope. The technique using a transparent over-tube with an accessory channel for the diathermic snare is useful for the resection of esophageal lesions; it is called the EEMR (endoscopic esophageal mucosa resection) tube technique [106] and a single-channel endoscope is passed in the over-tube. There has been a considerable evolution of the devices offered by various companies. Most of them are presented as EMR kits for single use.

The first step after the detection of the lesion is a careful morphological analysis with chromoscopy, using indigo-carmine as a contrast agent. Treatment applies to lesions with a superficial endoscopic pattern, either premalignant or malignant. If the size and morphology is compatible with an indication for EMR, the limits of the resection are marked on the adjacent mucosa either with a diathermy device or with small clips. Lifting the lesion from the submucosa is of utmost importance; it confirms the superficial character of the lesion and that there is a minimal risk of complications.

The standard procedure for lifting is submucosal injection with a special needle of up to 20 ml of saline with the addition of epinephrine (0.025 mg/ml). In order to prolong the lifting other solutions have been employed [105].

The delay before flattening of the mucosal bleb is short with saline (around 3 min) and can be prolonged (up to 20 min) when using a hyaluronic acid solution. However, the cost is increased. The resected specimen is prepared in the endoscopy room, stretched on a cardboard, and fixed in neutral formalin. Parallel transverse histological sections will assess with precision the depth of tumoral invasion in each sector. The best results are obtained with the en bloc resection of a single specimen. After piecemeal resection a similar analysis requires a reconstitution of the position of the multiple fragments.

EMR with a cap [97]: EMR-C (aspiration method)   

A transparent cap is attached to the tip of a single-channel endoscope and a diathermic snare is prelooped in the rim of the cap. The lesion is sucked into the cap and the snare is passed across the artificial polyp; resection is achieved with blended diathermy current. In case of bleeding, hemostatic clips are used. En bloc resection for lesions up to 15 mm in diameter is possible with the standard cap, and up to 20 mm with a large cap model. Larger lesions can be treated by the piecemeal method.

EMR with a ligating cap [102]: EMR-L (aspiration method)   

In this procedure the transparent cap of an endoscopic ligator-kit operates the suction; the strangulation is ensured by a ligating band. The protruded mucosal bleb is resected with a conventional snare. This procedure is well adapted to resection of lesions in the upper third of the lesser curvature.

EMR with tissue incision [103,105,107,108]   

This procedure is performed either with a double- or a single-channel endoscope. After placing marks around the target lesion, and lifting the lesion, a circular incision of the mucosa is performed with a needle-knife. The lesion raised by the circular incision is resected with a standard snare. The specimen is removed with grasping forceps.

With this procedure, en bloc resection of large segments of the gastric mucosa is possible, but there is a higher risk of hemorrhage. Hemo-clipping and suture clipping of the mucosal break is used. The safety of the incision method is improved when using new models of endoscopic knives. The hook knife, curved at its tip, will not enter to an excessive depth.

The insulated tip needle-knife (IT knife) also prevents excessive depth, but the section is very slow and a preliminary small incision of the mucosa is required for the insertion of the small sphere at the end of the needle. The triangular tip knife (TT-knife) is the most recent device for mucosal incision. The introduction of the triangular tip at the interface of mucosa and submucosa is easy, the progression of the tissue incision is fast, and the knife can be used for hemostasis.

EMR grasp-method [100,103]   

Either an over-tube (EEMR) or a double-channel endoscope is required; after marking the limits and lifting, the lesion is seized with grasping forceps which have been passed across the snare introduced through the other channel; resection is then performed. The four points method is a variant of the grasp-method, where the limits of the lesion are marked with four short clips with modified shortened arms. Then a jumbo grasping forceps lift the lesion with the clips for resection with the snare.

Indications for EMR   

In the stomach most low-grade or high-grade intraepithelial (premalignant) neoplastic lesions have the same appearance as early malignant lesions (intramucosal carcinoma or submucosal carcinoma). EMR is proposed for both types of flat or slightly elevated lesions when their size is not over 2 cm.

If there is a depression on the lesion, the indication is legitimate when the size is not over 1 cm. However, wider indications have also been proposed [96] for well-differentiated cancer less than 30 mm without ulcer or ulcer scar, less than 20 mm if there is an ulcer scar, and less than 10 mm if there is a poorly differentiated cancer.

After pathological assessment, the safety of the endoscopic treatment is confirmed for intramucosal neoplastic lesions, and for submucosal carcinoma when the invasion in the submucosa is less than 500 µm (micrometric method) below the limit of the muscularis mucosae. In other cases, EMR should be followed by surgery.

Results and complications of EMR   

En bloc resection in a single fragment gives better results than piecemeal resection; the proportion of cases with complete resection is higher and the rate of recurrence is lower. This is shown by the follow-up of large series [106]. The proportion of complete resection is higher in en bloc (67%) than in piecemeal resection (11%), and the respective rates of recurrence are 4% and 17%. In addition, the results depend on the topography of the lesion, en bloc resection being easier in the lower third of the stomach.

In a series of EMR for early gastric cancer (EGC) treated at the National Cancer Center in Tokyo [98], intramucosal cancer was confirmed in 405 out of 479 cases, and complete resection was performed in 69%. In 104 of the 127 cases with incomplete resection no subsequent treatment was proposed and there was recurrence in only 17 cases during the surveillance period. Major complications of EMR include severe bleeding and perforation; minor bleeding is frequent. The global risk of complications in large Japanese meta-analyzes has been estimated at 0.5%; but higher figures (around 5%) are presented in most individual series with a precise follow -up.

 Surgery for gastric cancer   

Many cases of early gastric cancer (EGC), according to the Japanese experience, are now treated with less aggressive modalities than gastrectomy, including EMR or laparoscopic surgery; this will contribute to better quality of life of patients with ECG. Laparoscopic surgery provides less pain, faster recovery, and a shorter hospital stay than conventional surgery. On the other hand, surgery remains the treatment for advanced and localized gastric cancer (T2, T3). Various techniques have been proposed, such as ordinary open surgery, laparoscopic-assisted gastrectomy, laparoscopic intragastric surgery, and pylorus-preserving gastrectomy [109–112].

Lymphadenectomy   

Gastrectomy with curative intent is completed by lymphadenectomy. A complete resection of the tumor with safe margins is called a R0 resection. Super-extended lymphadenectomy (D3/D4) involving resection of preaortic lymph nodes is seldom employed and the two current modalities are the lymphadenectomy for perigastric lymph nodes only (D1), or the extensive lymphadenectomy concerning also lymph nodes around the vascular axis (D2), which is the standard for advanced gastric cancer in Japan. Western surgeons still prefer D1 gastrectomy, because the morbidity and mortality is less (in their hands). New trials in Western countries tend to recommend D2 gastrectomy, either when the advanced cancer reaches the serosa or when perigastric metastatic lymph nodes occur in a superficial cancer.

Extent of the resection   

A distal resection (four-fifths of the stomach) is recommended in distal gastric cancer, while total gastrectomy is recommended in proximal gastric cancer and is often completed by splenectomy. When there is a local invasion of an adjacent viscus (liver or pancreas), gastrectomy can be completed by a segmental resection of the region with invasion and omentectomy.

Palliative gastrectomy   

If liver or peritoneal metastases are present, a palliative gastrectomy is still relevant if there is severe obstruction or chronic bleeding and anemia. Unresectable and obstructive tumors causing severe dysphagia or pyloric occlusion tend now to be treated by endoscopic ablation (laser) or by stent placement, rather than by surgery.

 Chemoradiation in advanced gastric cancer   

Chemoradiation protocols (palliation)   

Chemotherapy and radiation therapy have a marginal role. Only a slight benefit (and toxic effects) has been obtained in non-operable gastric metastatic cancer with conventional agents, and 5 fluorouracil (5 FU) remains the principal agent used in association with Methyl-CCNU or Mitomycin C or Epirubicin, and Cisplatinum. Better results are expected either from new modalities of administration (infusional 5 FU) or from new agents such as Oxaliplatin and Irinotecan, and, in the future, growth factor inhibitors and antiangiogenesis agents. Similarly, tumors are not affected by radiation protocols, which can prove toxic for liver and intestine. On the other hand brachytherapy has proved helpful in the palliation of dysphagia from tumors at the EG junction.

Adjuvant chemoradiation protocols   

More positive results are obtained with combined chemoradiation protocols [113,114]. However, there is no room for current neo-adjuvant protocols in the treatment of gastric cancer. Adjuvant chemoradiation protocols after surgery may prove beneficial as shown in the multicenter analysis conducted in North America by MacDonald [113]. The protocol INT 0116 combines radiation (45 Gy in 25 fractions) and two agents (5 fluorouracil and leucovorin). However, the guideline recommending this protocol in North America has not reached consensus in other countries because the percentage of D2 resection in the study was small (10%) and the D1 lymphadenectomy was often incomplete. There have been considerable variations in the benefit afforded. Meanwhile, adjuvant chemoradiation protocol can be recommended for patients with a R0 resection and positive lymph nodes.

 Endoscopic palliation with Nd:YAG laser   

The Nd:YAG laser is a powerful tool adapted to the reduction of the tumor volume. The infrared photon beam of a Nd:YAG laser source is transmitted across a flexible fiber passed through the operative channel of the endoscope, and aimed at the tumor either at distance or in the tissue.

There are two distinct indications for Nd:YAG laser in palliation for non-operable patients: hemostasis and lumen ablation. In patients with continuous blood oozing from the tumor and chronic anemia, repeated sessions of thermal coagulation of the surface of the tumor at intervals of 4 weeks may reduce the blood loss and improve the status of the patient. The other indication concerns severe dysphagia from obstructive fungating tumors at the EG junction. The procedure is concurrent with the placement of an expandable stent, or brachytherapy. Nd:YAG laser ablation has also been combined with brachytherapy [115].

In conclusion, lumen recanalization at the EG junction through thermal destruction is possible. Dysphagia can be relieved by sessions at 4 week intervals, ensuring a correct nutrition; however, thermal destruction has poor results when the stricture is infiltrating and non-exophytic; stent placement is usually required.

 Endoscopic palliation with stents   

Types of stents   

Semi-rigid, plastic, non-expandable stents are used less now than metal expandable stents derived from intravascular stents, constructed in stainless steel or with nitinol alloy (nickel, titanium). The stents are built either in a thin grid or knitted mesh with the metal thread, or in a large grid in the Z model, or in a rigid coil. Depending on their structure the stents may shorten when they expand. Models differ in rigidity and radial expansive force. With respect to prevention of reflux at the EG junction, some stents are equipped with a distal antireflux cuff. Stents are usually delivered as a set mounted on an introducer. Metal stents are either uncovered or covered. The drawback of uncovered stents is the obstruction by tumor in-growth across the mesh. The drawback of covered stents is the risk of migration. Non-metallic expandable stents have recently become available (Polyflex: stent from Rusch, Kernen, Germany). Prototypes of biodegradable stents made of a polymer of lactic acid are preferred in the treatment of bronchial strictures.

Similar results in the relief of dysphagia are obtained with various types of expandable stents; however, the rate of complications appears to be higher with the Z structure and with the coil structure. A randomized study comparing covered and uncovered expandable stents placed at the EG junction [116] has shown similar relief of dysphagia and similar survival, but tumor in-growth requiring re-intervention was more frequent with uncovered stents.

Placement of the stent and indications   

The placement of a stent usually includes both endoscopic and fluoroscopic control. However, placement of self-expanding stents without fluoroscopy has been reported, and placement without endoscopy, under fluoroscopy, is also possible.

The primary indication for stenting a stricture in gastric cancer is tumoral obstruction at the EG junction generating dysphagia in non-operable patients [116,117].

More recently, stenting has been proposed as an alternative to palliative surgery for strictures at the gastric outlet [118–122]. In non-operable patients the objective of re-establishment of the passage of food is the same when there is a prepyloric cancer as when there is duodenal compression by a pancreatic cancer. Metal expandable stents are used. Radiological guidance is often required, either alone, or in association to endoscopy. Endoscopic guidance alone is also possible using delivery systems with an adapted length. Stents are pushed through an interventional endoscope with a large channel, or alongside the endoscope. Special models of expandable stents are available. The Enteral Wallstent is 9 or 10 cm in length, with a diameter of 20 or 22 mm and is passed through the operative channel of an duodenoscope.

Results and complications of stenting   

Results at the EG junction   

For tumors at the EG junction the relief of dysphagia and improvement of nutrition is obtained in most cases in 24 h. The dysphagia score of from 4 (aphagia) to 0 (no dysphagia) is used to evaluate the efficacy in palliation as well as the health status and the quality of life of the patient. The median survival after placement of a stent is estimated in the range of 9–12 weeks [106]. Other options in the palliation of a stricture at the EG junction include brachytherapy and lumen recanalization with a thermal procedure (Nd:YAG laser), or photodynamic therapy.

A recent study [106] compared stenting to tumor ablation with the Nd:YAG laser. The median survival was longer in the thermal treatment group. The health quality of life was impaired in both groups at the baseline, and remained stable in the thermal treatment group while deterioration occurred in the stent group, with more pain. In conclusion, stenting is preferred when the tumor is infiltrative, whereas Nd:YAG laser is preferred for obstructive tumors with a smooth fungating surface. In a second phase such tumors will be stented.

Results at the gastric outlet   

Stenting for malignant strictures of the gastric outlet improves the passage of food, but biliary obstruction may develop later. In patients with stenosis of the gastric outlet from pancreatic cancer, endoscopic palliation with stenting has given results similar to those in operative gastro-jejunostomy

Complications   

The toll of complications after placement of a stent is influenced by many factors, including the stent material, diameter, and its radial force of expansion. Self-expanding stents have less complications than semi-rigid stents. Tumor ingrowth across the mesh of uncovered stents or at the periphery of covered stents, food impaction in the stent, and migration may occur. Tumor ingrowth requires recanalization with thermal procedures (argon plasma or microwaves). Migration requires extraction and placement of another stent. Specific complications after placement at the EG junction are gastroesophageal reflux which can result in further respiratory complications, and migration of the stent in the stomach. Specific complications after placement of a stent in the gastric outlet are distal or proximal obstruction by the tumor growth, and migration of the stent in the intestine.

 Guidelines in surveillance   

Surveillance of the risk of gastric cancer after an index gastroscopy is also based on endoscopy; the general tendency is to avoid repetition at short intervals. Surveillance occurs in two distinct situations:

When endoscopy has shown a premalignant condition (e.g. chronic atrophic gastritis with H. pylori infection and no neoplastic lesion), surveillance is not generally recommended in spite of a recent British study [123] showing a high yield of gastric cancer in patients with chronic gastritis who agreed to annual endoscopic surveillance. Specific situations arise with surveillance after a partial gastrectomy for peptic ulcer, but this procedure has become rare. Endoscopic surveillance after detection of pernicious anemia is recommended in the early period, but must not be unduly prolonged.

When index endoscopy has shown a premalignant lesion with IEN in the stomach, patients should enter a surveillance protocol (when they have not been treated). The risk of progression from high-grade IEN to cancer is very high in the short term (around 80%), while the risk is low with low-grade IEN (less than 15% in 2 years) (in [27]). The progression from low-grade to high-grade IEN and to cancer is altogether inconstant and slow. Taking this into account, lesions with high-grade IEN are generally treated immediately; surveillance applies to lesions with low-grade IEN. Endoscopic exploration is repeated at intervals of 6–12 months and should not be limited to the targeted area, because metachronous lesions may develop. The same strategy also applies to persons who have been treated by EMR for a premalignant or early malignant lesion in the stomach.

 Prevention of gastric cancer   

Primary prevention of gastric cancer is by the control of causal factors. There is a general belief that, in most countries, primary prevention of stomach cancer has priority over secondary prevention. It aims to reduce the risk factors, which are hypochlorhydria from chronic atrophic gastritis with H. pylori infection, and dietary factors, such as an excess of salt, nitrites, and nitrates, and a lack of antioxidants. H. pylori infection is just one of a number of factors in gastric carcinogenesis [124–127].

Prevention and H. pylori infection   

A reduction in the prevalence of chronic atrophic gastritis has been proposed in countries with a high incidence of stomach cancer. The large-scale eradication of H. pylori infection by vaccination has been proposed, or by serological screening, with antibiotic therapy for those testing H. pylori positive. The potential drawback of H. pylori eradication is the promotion of gastroesophageal reflux. At the individual level, eradication is recommended in persons who have chronic atrophic gastritis at the index endoscopy and are positive for H. pylori. This can improve the status of the mucosa and is assumed to reduce the risk of cancer. In a recent study of 22 elderly men positive for H. pylori[127] the scores for atrophy and intestinal metaplasia were stable during the first part (7.5 years) of the follow-up without treatment and regressed during the second part of the follow-up (2.5 years) after eradication.

Prevention through dietary intervention   

A planned dietary intervention through addition of antioxidants (beta-carotene, retinol, alpha-tocopherol, vitamin C) has been attempted in trials where the surrogate endpoint is the occurrence of premalignant gastric lesions. However, a major confusing variable is the H. pylori status, which may differ between tested persons and their controls; therefore eradication of H. pylori in both groups is a requested preliminary. Prevention trials have been conducted in South America and Europe. In the cohort study in Finland [124] daily supplementation for 5–8 years in 29 133 male smokers had no impact on the occurrence of neoplasia. The overall disappointing results of chemoprevention trials contrast with the generalized decrease in the incidence through unplanned primary prevention.

Unplanned prevention   

Unplanned prevention occurs through changes in environmental factors which are supposed to play a major role during infancy, as shown by the progressive decrease in the risk in successive generations for Japanese migrants to the USA [36–38].

The promoting environmental factors, linked to the lifestyle, include a diet rich in salt and poor in fruit and vegetables, and H. pylori infection. This can be corrected by a diet characterized by a reduction of salty foods and an increased consumption of fruit and antioxidant vitamins, and a reduction of the contamination in childhood for the H. pylori infection. The decline in the incidence of stomach cancer in Japan is often attributed to the generalization of a Western style of life under the American influence, just after World War II, when the full impact of the declining trend occurred.

Actually the cause of the decline occurred much earlier, when Japan entered the industrial arena in the first decade of the twentieth century, under the influence of Emperor Meiji, just after the Sino-Japanese and Russo-Japanese wars. Changing environmental factors occurred in the infancy of cohorts born in 1910, and after 1910 as shown in an age-cohort study of the Osaka Cancer Registry [66] during the period 1975–95.

 Secondary prevention of gastric cancer   

Secondary prevention of gastric cancer is based on the detection and treatment of premalignant and early malignant lesions in asymptomatic persons (preclinical disease). However, flat precursors which are difficult to detect play the major role in gastric cancerogenesis.

Gastroscopy and opportunistic screening   

Gastroscopy, the gold standard procedure in the detection of non-protruding neoplastic lesions of the stomach, is proposed for persons complaining from upper-digestive symptoms (gastroesophageal reflux or non-ulcer dyspepsia) or for asymptomatic persons wishing to be reassured. However, in Western countries neoplastic lesions are rarely detected (1–2%) in such patients; this is why good-practice guidelines do not recommend prompt endoscopy in persons aged less than 45 years, and often delay the procedure until after a therapeutic trial.

However, studies on the appropriateness of endoscopy have shown that the rate of over-utilization (in non-appropriate indications) is still less than the rate of under-utilization (in appropriate indications). Anyhow, when gastroscopy is performed in the absence of alarming symptoms, the entire mucosal surface of the esophagus and stomach should be examined carefully with a high-resolution instrument, in order to minimize false negative rates [128]; indeed the detection of non-protruding lesions, which are the major precursors of gastric cancer, deserves special attention.

A study of false negative results for gastric cancer has been conducted in Japan in the Fukui area [129,130]. In 3672 persons with a 'negative gastroscopy' in the regional hospital, selected to repeat the procedure after a delay of 1–3 years, a gastric cancer was found in 32 (less than 1%); this result confirms that the risk of missing a cancer after a negative procedure is very small. When all gastroscopies in the regional hospital were compared to the population-based Tumor Registry 1–3 years after performance of the procedure, gastric cancer had been missed in 155 out of 814 cases (19%) found in the registry; here the false negative rate is much higher. An appreciable proportion of gastric cancers are undiagnosed at the first procedure, even in a country where endoscopy reaches a fairly high standard.

Mass screening   

In Japan   

Mass screening detection of early cancer is appropriate when a simple screening test is available and when the prevalence of the disease is high. Japan is the only country where mass screening for stomach cancer is a national policy, and has been for four decades. The screen-detected cases (either early or silent disease) have had an impact on mortality rates [131–134], in spite of a declining cost-effectiveness [135].

Gastroscopy is performed in persons with a positive miniature radiophotofluorography. As an alternative, serologic biological markers aimed at detecting atrophic gastritis have been proposed. A decreased ratio of pepsinogen I to III confirms atrophy of the oxyntic mucosa, when an appropriate cut-off ratio is selected. The pepsinogen test tends now to be generalized as a screening test [136–141]. Positive serology for H. pylori or its cagA phenotype do not fulfill the criteria of a selection test. The mass screening campaign (combined to opportunistic screening) proved beneficial as shown in Figs 22–26. In Japan around 45% of cases are detected at an early and localized stage. The 5 year relative survival for stomach cancer is over 40%.

A comparative study [66] of time trends on the incidence of gastric cancer in Japan (Osaka Cancer Registry) and in a country with no screening policy (USA) has shown that the incidence rate of stomach cancer declined in Japan in the same way as it did for other countries in the period 1975–95. However, the trend in Japan has two specific characters. The stage specific incidence declines only for regional stage cancer, while the incidence of the localized stage cancer increases. Mortality rates decline more sharply than the incidence rates. The policy of early detection of stomach cancer explains the increased incidence of localized cancer, and the dissociation of the mortality rate.

In other countries   

In the Western world, in the absence of a screening policy, most cases of stomach cancer are detected at the advanced stage, with a correspondingly poor prognosis. Indeed a recent analysis of 57 407 cases in the National Cancer Data Base of the USA [32] shows that most cases are still detected at the advanced stage. In the USA and Europe the 5 year survival rate just reaches 20%.

Strategy of detection worldwide   

Taking into account the generalized declining incidence, screening for gastric cancer does not deserve a public health policy outside Japan, where the toll from stomach cancer is the highest. In other countries, detection depends on opportunistic screening with gastroscopy in persons with non-alarming digestive symptoms (non-ulcer dyspepsia) after the age of 45.

High-grade IEN justifies treatment because most of these lesions will progress to advanced cancer [66]. The secondary objective of the procedure is the detection of polyps. Benign adenomatous polyps may be associated with a metachronous malignant non-protruding area. Gastric polyposis (fundic cystic or hyperplastic polyps) justifies an indication for colonoscopy and a questionnaire for familial hereditary syndromes.

 References  

1 Rugge, M, Correa, P & Dixon, MF et al. Gastric dysplasia: the Padova international classification. Am J Surg Pathol 2000; 254: 167–77. CrossRef

2 Tsukuma, H, Oshima, A & Narahara, H et al. Natural history of early gastric cancer: a non concurrent long term follow up study. Gut 2000; 47: 618–21. PubMed CrossRef

3 Guindi, M & Riddell, RH. The pathology of epithelial premalignancy of the gastrointestinal tract. Best Pract Res Clin Gastroenterol 2001; 15: 191–210. PubMed CrossRef

4 Kapadia, CR. Gastric atrophy, metaplasia and dysplasia: a clinical perspective. J Clin Gastroenterol 2003; 36: S29–36. PubMed CrossRef

5 Nakamura, K, Sakaguchi, H & Enjoji, M. Depressed adenoma of the stomach. Cancer 1988; 62: 2197–202. PubMed

6 Xuan, ZX, Ambe, K & Enjoji, M. Depressed adenoma of the stomach, revisited: histologic, histochemical and immunochemical profiles. Cancer 1991; 67: 2382–9. PubMed

7 Katakura, T, Masuda, T & Ikeda, T et al. Immunohistochemical studies in depressed type adenomas of the stomach: comparison with protruded type adenomas and depressed type well differentiated mucosal gastric cancer (Japanese). Nippon Shokakibyo Gakkai Zasshi 1998; 95: 992–1000. PubMed

8 Tabata, H, Fuchigami, T & Kobayashi, H et al. Difference in degree of mucosal atrophy between elevated and depressed types of gastric epithelial tumors. Scand J Gastroenterol 2001; 36: 1134–40. PubMed CrossRef

9 Stolte, M, Sticht, T & Eidt, S et al. Frequency, location, and age and sex distribution of various types of gastric polyp. Endoscopy 1994; 26: 659–65. PubMed

10 Stolte, M. Clinical consequences of the endoscopic diagnosis of gastric polyps. Endoscopy 1995; 27: 32–7. PubMed

11 Stolte, M. Hyperplastic polyps of the stomach: associations with histologic patterns of gastritis and gastric atrophy. Am J Surg Pathol 2001; 25: 1342–4. PubMed CrossRef

12 Muehldorfer, SM, Stolte, M & Martus, P et al. Diagnostic accuracy of forceps biopsy versus polypectomy for gastric polyps: a prospective multicentre study. Gut 2002; 50: 465–70. PubMed CrossRef

13 Schmitz, GM & Stolte, M. Gastric polyps and precancerous lesions. Gastrointest Endosc Clinics North Am 1997; 7: 29–46.

14 Abraham, SC, Nobukawa, B & Giardiello, FM et al. Fundic gland polyps in familial adenomatous polyposis: neoplasms with frequent somatic adenomatous polyposis coli gene alterations. Am J Pathol 2000; 157: 747–54. PubMed

15 Abraham, SC, Park, SJ & Mugartegui, L et al. Sporadic fundic gland polyps with epithelial dysplasia: evidence for preferential targeting for mutations in the adenomatous polyposis coli gene. Am J Pathol 2002; 161: 1735–42. PubMed

16 Torbenson, M, Lee, JH & Cruz-Correa, M et al. Sporadic fundic gland polyposis: a clinical, histological, and molecular analysis. Mod Pathol 2002; 15: 718–23. PubMed CrossRef

17 Abraham, SC, Singh, VK & Yardley, JH et al. Hyperplastic polyps of the stomach: associations with histologic patterns of gastritis and gastric atrophy. Am J Surg Pathol 2001; 25: 500–7. PubMed CrossRef

18 Ohkusa, T, Takashimizu, I & Fujiki, K et al. Disappearance of hyperplastic polyps in the stomach after eradication of Helicobacter pylori: a randomized, clinical trial. Ann Intern Med 1998; 129: 712–15. PubMed

19 Rashid, A, Houlihan, PS & Booker, S et al. Phenotypic and molecular characteristics of hyperplastic polyposis. Gastroenterology 2000; 119: 323–32. PubMed

20 Rubio, CA. Serrated neoplasia of the stomach: a new entity. J Clin Pathol 2001; 54: 849–53. PubMed

21 Boardman, LA, Thibodeau, SN & Schaid, DJ et al. Increased risk for cancer in patients with the Peutz–Jeghers syndrome. Ann Intern Med 1998; 128: 896–9. PubMed

22 Howe, JR, Mitros, FA & Summers, RW. The risk of gastrointestinal carcinoma in familial juvenile polyposis. Ann Surg Oncol 1998; 5: 751–6. PubMed

23 Hirota, WK, Loughney, TM & Lazas, DJ et al. Specialized intestinal metaplasia, dysplasia and cancer of the esophagus and esophagogastric junction: prevalence and clinical data. Gastroenterology 1999; 116: 227–85. PubMed

24 El-Zimaity, HM, Ramchatesingh, J & Saeed, MA et al. Gastric intestinal metaplasia: subtypes and natural history. J Clin Pathol 2001; 54: 679–83. PubMed

25 Schlemper, RJ, Itabashi, M & Kato, Y et al. Differences in diagnostic criteria for gastric carcinoma between Japanese and Western pathologists. Lancet 1997; 349: 1725–9. PubMed CrossRef

26 Schlemper, RJ, Riddell, RH & Kato, Y et al. The Vienna classification of gastrointestinal epithelial neoplasia. Gut 2000; 47: 251–5. PubMed CrossRef

27 Hamilton, S & Aaltonen, L, (2000) Pathology and Genetics: Tumors of the Digestive System. IARC Scientific Publications. IARC, Lyon.

28 Dixon, MF. Gastrointestinal epithelial neoplasia: Vienna revisited. Gut 2002; 51: 130–1. PubMed CrossRef

29 Parkin, DM, Whelan, SL & Ferlay, J et al. (2002). Cancer Incidence in Five Continents, Vol. VIIII. IARC Scientific Publications, No. 155. IARC, Lyon.

30. Anonymous. Surveillance, Epidemiology, and End Results (SEER). Program Public-Use, CD-ROM (1973–97), National Cancer Institute, DCCPS, Cancer Surveillance, Research Program, Cancer Statistics Branch, released based on the August 1999 submission.

31 Berrino, F, Capocaccia, R & Estève, J et al. (1999). Survival of Cancer Patients in Europe: The EUROCARE 2 Study. IARC Scientific Publications, No. 151. IARC, Lyon.

32 Hundahl, SA, Menck, HR & Mansour, EG et al. The National Cancer Data Base report on gastric carcinoma. Cancer 1997; 80: 2333–41. PubMed

33 Hanai, A, Tsukuma, H & Hiyama, T et al. (1997). Survival of patients with stomach cancer: results from population based cancer registries. In: Gastric Cancer (eds Sugimura T, Sasako M), pp. 1–30. Oxford University Press, Oxford.

34 Osaka Cancer Registry (1998). Survival of cancer patients in Osaka: 1975–89. Shinohara Publications, Tokyo.

35 The Research Group for Population Based Cancer Registration in Japan. Cancer incidence in Japan 1985–89: re-estimation based on data from eight population based cancer registries. Jpn J Clin Oncol 1998; 28: 54–67. PubMed CrossRef

36 Locke, FB & King, H. Cancer mortality risk among Japanese in the United States. J Nat Cancer Inst 1980; 65: 1149–56. PubMed

37 Nomura, A, Stemmermann, GN & Chyou, PH et al. Helicobacter pylori infection and gastric carcinoma among Japanese Americans in Hawaii. N Engl J Med 1991; 325: 1132–6. PubMed

38 Kamieni, A, Williams, MA & Schwartz, SM et al. The incidence of gastric carcinoma in Asian migrants to the United States and their descendants. Cancer Causes Control 1999; 10: 77–83. PubMed CrossRef

39 Spechler, SJ. The role of gastric carditis in metaplasia and neoplasia at the gastroesophageal junction. Gastroenterology 1999; 117: 218–28. PubMed

40 Ormsby, AH, Goldblum, JR & Rice, TW et al. The utility of cytokeratin subsets in distinguishing Barrett's-related oesophageal adenocarcinoma from gastric adenocarcinoma. Histopathology 2001; 38: 307–11. PubMed CrossRef

41 El-Rifai, W, Frierson, HF Jr & Moskaluk, CA et al. Genetic differences between adenocarcinomas arising in Barrett's esophagus and gastric mucosa. Gastroenterology 2001; 121: 592–8. PubMed

42 Tanière, P, Martel-Planche, G & Maurici, D et al. Molecular and clinical differences between adenocarcinomas of the esophagus and of the gastric cardia. Am J Pathol 2001; 158: 33–40. PubMed

43 Tanière, P, Borghi-Scoazec, G & Saurin, JC et al. Cytokeratin expression in adenocarcinomas of the esophagogastric junction: a comparative study of adenocarcinomas of the distal esophagus and of the proximal stomach. Am J Surg Pathol 2002; 26: 1213–21. PubMed CrossRef

44 Couvelard, A, Cauvin, JM & Goldfain, D et al. Cytokeratin immunoreactivity of intestinal metaplasia at normal oesophagogastric junction indicates its etiology. Gut 2001; 49: 761–6. PubMed

45 Banatvala, N, Mayo, K & Megraud, F et al. The cohort effect and Helicobacter pylori. J Infect Dis 1993; 168: 219–21. PubMed

46 Danesh, J. Helicobacter pylori infection and gastric cancer: systematic review of the epidemiological studies. Aliment Pharmacol Ther 1999; 13 (7): 851–6. PubMed CrossRef

47 Eslick, GD, Lim, LL-Y & Byles, JE et al. Association of Helicobacter pylori infection with gastric carcinoma: a meta-analysis. Am J Gastroenterol 1999; 94: 2373–9. PubMed CrossRef

48 Forman, D, Newell, DG & Fullerton, F et al. Association between infection with Helicobacter pylori and risk of gastric cancer: evidence from a prospective investigation. Br Med J 1991; 302: 1302–5.

49 Helicobacter and Cancer Collaborative Group. Gastric cancer and Helicobacter pylori: a combined analysis of 12 case-control studies nested within prospective cohorts. Gut, 2002; 49: 347–53. CrossRef

50 Huang, J-Q, Sridhar, S & Chen, Y et al. Meta-analysis of the relationship between Helicobacter pylori seropositivity and gastric cancer. Gastroenterology 1998; 114: 1169–79. PubMed

51 Sipponen, P & Marshall, BJ. Gastritis and gastric cancer in Western countries. Gastroenterol Clin North Am 2000; 29: 579–92. PubMed

52 Kimura, K. Gastritis and gastric cancer. Asia Gastroenterol Clin North Am 2000; 29: 609–21.

53 Go, MF. Natural history and epidemiology of Helicobacter pylori infection. Aliment Pharmacol Ther 2002; 16: 3–15. PubMed

54 Blaser, MJ, Perez-Perez, GI & Kleanthous, H et al. Infection with Helicobacter pylori strains possessing cagA is associated with an increased risk of developing adenocarcinoma of the stomach. Cancer Res 1995; 55: 2111–15. PubMed

55 Campbell, DI, Warren, BF & Thomas, JE et al. The African enigma: low prevalence of gastric atrophy, high prevalence of chronic inflammation in West African Adults and children. Helicobacter 2001; 6: 263–7. PubMed

56 Holcombe, C. Helicobacter pylori: the African enigma. Gut 1992; 33: 429–31. PubMed

57 Miwa, H, Go, MF & Sato, N. H. pylori and gastric cancer: the Asian enigma. Am J Gastroenterol 2002; 97: 1106–12. PubMed

58 Everhart, JE. Recent developments in the epidemiology of Helicobacter pylori. Gastroenterol Clin North Am 2000; 29: 559–78. PubMed

59 Zhang, ZF, Kurtz, RC, Klimstra, DS, Yu, GP, Sun, M & Harlap, S et al. Helicobacter pylori infection and the risk of stomach cancer and chronic atrophic gastritis. Cancer Detect Prev 1999; 23: 357–67. PubMed CrossRef

60 Blaser, MJ & Berg, DE. Helicobacter pylori genetic diversity and risk of human disease. J Clin Invest 2001; 107: 767–73. PubMed

61 IARC (1994). Schistosomes, liver flukes and Helicobacter pylori. Monographs on the Evaluation of Carcinogenic Risks to Humans, Vol. 61, pp. 177–240. Lyon, IARC.

62 Ebert, MP, Schandl, L & Malfertheiner, P. Helicobacter pylori infection and molecular changes in gastric carcinogenesis. J Gastroenterol 2002; 37 (Suppl. 13): 45–9. PubMed

63 Hill, MJ. Diet and cancer: a review of scientific evidence. Europ J Cancer Prev 1995; 4 (Suppl. 2): 3–42.

64 Hirohata, T & Kono, S. Diet/Nutrition and stomach cancer in Japan. Int J Cancer Suppl 1997; 10: 34–6.

65 Devesa, SS, Blot, WJ & Fraumeni, JF. Changing patterns in the incidence of esophageal and gastric carcinoma in the United States. Cancer 1998; 83: 2049–53. PubMed

66 Lambert, R, Guilloux, A & Oshima, A et al. Incidence and mortality from stomach cancer in Japan, Slovenia and the USA. Int J Cancer 2002; 97: 811–18. PubMed CrossRef

67 Correa, P & Chen, VW. Gastric cancer. Cancer Surv 1994; 19–20: 55–76.

68 Correa, P & Miller, MJ. Carcinogenesis, apoptosis and cell proliferation. Br Med Bull 1998; 54: 151–62. PubMed

69 Solcia, E, Fiocca, R & Luinetti, O et al. Intestinal and diffuse gastric cancers arise in a different background of Helicobacter pylori gastritis through different gene involvement. Am J Surg Pathol 1996; 20 (Suppl. 1): S8–S22. PubMed

70 Scotiniotis, IA, Rokkas, T & Furth, EE et al. Altered gastric epithelial cell kinetics in Helicobacter pylori-associated intestinal metaplasia: implications for gastric carcinogenesis. Int J Cancer 2000; 85: 192–200. PubMed CrossRef

71 Zhang, Z, Yuan, Y & Gao, H et al. H. Apoptosis, proliferation and p53 gene expression of H. pylori associated gastric epithelial lesions. World J Gastroenterol 2001; 7: 779–82. PubMed

72 Park, DI, Rhee, PI & Kim, JE et al. Factors suggesting malignant transformation of gastric adenoma: univariate and multivariate analysis. Endoscopy 2001; 33: 501–6. PubMed CrossRef

73 Lee, JH, Abraham, SC & Kim, HS et al. Inverse relationship between APC gene mutation in gastric adenomas and development of adenocarcinoma. Am J Pathol 2002; 161: 611–18. PubMed

74 Gayther, SA, Gorringe, KL & Ramus, SJ et al. Identification of germ-line E-cadherin mutations in gastric cancer families of European origin. Cancer Res 1998; 58: 4086–9. PubMed

75 Guilford, P, Hopkins, J & Grady, W et al. E-cadherin germline mutations define an inherited cancer syndrome dominated by diffuse gastric cancer. Hum Mutat 1999; 14: 249–55. PubMed CrossRef

76 Canto, MI, Setrakian, S & Willis, JE et al. Methylene blue staining of dysplastic and non dysplastic Barrett's esophagus: an in vivo and ex vivo study. Endoscopy 2001; 33: 391–400. PubMed

77 Dinis-Ribeiro, M, da Costa-Pereira, A & Lopes, C et al. Magnification chromo-endoscopy for the diagnosis of gastric intestinal metaplasia and dysplasia. Gastrointest Endosc 2003; 57: 498–504. PubMed

78 Guelrud, M, Herrera, I & Essenfeld, H et al. Intestinal metaplasia of the gastric cardia: a prospective study with enhanced magnification endoscopy. Am J Gastroenterol 2002; 97: 584–9. PubMed

79 Lambert, R, Rey, JF & Sankaranarayanan, R. Magnification and chromoscopy with the acetic acid test. Endoscopy 2003; 35: 437–45. PubMed CrossRef

80 Rollins, AM & Sivak, MV. Potential new endoscopic techniques for the earlier diagnosis of premalignancy. Best Pract Res Clin Gastroenterol 2001; 15: 227–47. PubMed CrossRef

81 Tajiri, H, Doi, T & Endo, H et al. Routine endoscopy using a magnifying endoscope for gastric cancer diagnosis. Endoscopy 2002; 34: 772–7. PubMed CrossRef

82 Tobita, K. Study on minute surface structures of the depressed-type early gastric cancer with magnifying endoscopy. Digest Endosc 2001; 13: 121–6. CrossRef

83 Yao, K & Oishi, T. Microgastroscopic findings of mucosal microvascular architecture as visualized by magnifying endoscopy. Dig Endosc 2001; 13 (Suppl.): S27–33. CrossRef

84 Japanese Gastric Cancer Association. Japanese classification of gastric carcinoma: 2nd English edn. Gastric Cancer 1998; 1: 10–24. PubMed CrossRef

85 Aiko, T & Sasako, M. The new Japanese classification of gastric carcinomas: points to be revised. Gastric Cancer 1998; 1: 25–30. PubMed CrossRef

86 National Report of the Group Medical examination for digestive cancer in 1999 (in Japanese). J Gastroenterol Mass Survey 2002; 40: 57–76.

87 Reid, BJ, Blount, PL & Rabinovitch, PS. Biomarkers in Barrett's esophagus. Gastrointest Endosc Clin N Am 2003; 13: 369–97. PubMed

88 Leiper, K & Morris, AL. Treatment of oesophago-gastric tumors. Endoscopy 2002; 34: 139–45. PubMed CrossRef

89 Adachi, Y, Shiraishi, N & Kitano, S. Modern treatment of early gastric cancer: review of the Japanese experience. Dig Surg 2002; 19: 333–9. PubMed CrossRef

90. Ell, C & Gossner, L. Photodynamic therapy. Recent Results Cancer Res 2000; 155: 175–81.

91 Ortner, MA, Dorta, G & Blum, AL et al. Endoscopic interventions for preneoplastic and neoplastic lesions: mucosectomy, argon plasma coagulation and photodynamic therapy. Dig Dis 2002; 20: 167–72. PubMed CrossRef

92 Nakamura, H, Yanai, H & Nishikawa, J et al. Experience with photodynamic therapy (endoscopic laser therapy) for the treatment of early gastric cancer. Hepatogastroenterology 2001; 48: 1599–603. PubMed

93 Rembacken, BJ, Gotoda, T & Fujii, T et al. Endoscopic mucosal resection. Endoscopy 2001; 33: 709–18. PubMed CrossRef

94 Shim, CS. Endoscopic mucosal resection: an overview of the value of different techniques. Endoscopy 2001; 33: 271–5. PubMed CrossRef

95 Ahmad, NA, Kochman, ML & Long, WB et al. Efficacy, safety and clinical outcome of endoscopic mucosal resection: a study of 101 cases. Gastrointest Endosc 2002; 55: 390–6. PubMed CrossRef

96 Amano, Y, Ishihara, S & Amano, K et al. An assessment of local curability of endoscopic surgery in early gastric cancer without satisfaction of current therapeutic indications. Endoscopy 1998; 30: 548–52. PubMed

97 Inoue, H, Fukami, N & Yoshida, T et al. Endoscopic resection of esophageal and gastric cancers. J Gastroenterol Hepatol 2002; 17: 382–8. PubMed CrossRef

98 Ono, H, Kondo, H & Gotoda, T et al. Endoscopic mucosal resection for treatment of early gastric cancer. Gut 2001; 48: 225–9. PubMed CrossRef

99 Ohyama, T, Kobayashi, Y & Mori, K et al. Factors affecting complete resection of gastric tumors by the endoscopic mucosal resection procedure. J Gastroenterol Hepatol 2002; 17: 844–8. PubMed CrossRef

100 Tanaka, M & Inatsuchi, S. A four-point fixation method for the resection of early gastric cancer, with particular reference to the analysis of cases of incomplete resection. Surg Endosc 1997; 11: 295–8. PubMed CrossRef

101 Tanabe, S, Koizumi, W & Kokutou, M et al. Usefulness of endoscopic aspiration mucosectomy as compared with strip biopsy for the treatment of gastric mucosal cancer. Gastrointest Endosc 1999; 50: 819–22. PubMed

102 Suzuki, H. Endoscopic mucosal resection using ligating device for early gastric cancer. Gastrointest Endosc Clin N A, 2001; 11: 511–18.

103 Suzuki, H & Ikeda, K. Endoscopic mucosal resection and full thickness resection with complete defect closure for early gastrointestinal malignancies. Endoscopy 2001; 33: 437–9. PubMed CrossRef

104 Tani, M, Takeshita, K & Inoue, H et al. Adequate endoscopic mucosal resection for early gastric cancer obtained from the dissecting microscopic features of the resected specimens. Gastric Cancer 2001; 4: 122–31. PubMed

105 Yamamoto, H, Kawata, H & Sunada, K et al. Success rate of curative endoscopic mucosal resection with circumferential mucosal incision assisted by submucosal injection of sodium hyaluronate. Gastrointest Endosc 2002; 56: 507–12. PubMed CrossRef

106 Lambert, R. Treatment of esophagogastric tumors. Endoscopy 2003; 35: 118–26. PubMed CrossRef

107 Ohkuwa, M, Hosokawa, K & Boku, N et al. New endoscopic treatment for intramucosal gastric tumors using an insulated-tip diathermic knife. Endoscopy 2001; 33: 221–6. PubMed CrossRef

108 Miyamoto, S, Muto, M & Hamamoto, Y et al. A new technique for endoscopic mucosal resection with an insulated tip electrosurgical knife improves the completeness of intramucosal gastric neoplasms. Gastrointest Endosc 2002; 55: 576–81. PubMed CrossRef

109 Bonenkamp, JJ, Hermans, J, Sasako, M & van de Velde, CJ. Extended lymph-node dissection for gastric cancer. Dutch Gastric Cancer Group. N Engl J Med 1999; 340: 908–14. PubMed

110 Lewis, WG, Edwards, P & Barry, JD et al. D2 or not D2? The gastrectomy question. Gastric Cancer 2002; 5: 29–34. PubMed CrossRef

111 Tanimura, S, Higashino, M & Fukunaga, Y et al. Laparoscopic distal gastrectomy with regional lymph node dissection for gastric cancer. Surg Endosc 2003; 17: 758–62. PubMed

112 Zhang, ZX, Gu, XZ & Yin, WB et al. Randomized clinical trial on the combination of preoperative irradiation and surgery in the treatment of adenocarcinoma of gastric cardia (AGC): report on 370 patients. Int J Radiat Oncol Biol Phys 1998; 42: 929–34. PubMed CrossRef

113 Macdonald, JS, Smalley, SR & Benedetti, J et al. Chemoradiotherapy after surgery compared with surgery alone for adenocarcinoma of the stomach or gastroesophageal junction. N Engl J Med 2001; 345: 725–30. PubMed CrossRef

114 Park, SH, Kim, DY & Heo, JS et al. Postoperative chemoradiotherapy for gastric cancer. Ann Oncol 2003; 14: 1373–7. PubMed CrossRef

115 Spencer, GM, Thorpe, SM & Blackman, GM et al. Laser augmented by brachytherapy versus laser alone in the palliation of adenocarcinoma of the esophagus and cardia: a randomised study. Gut 2002; 50: 224–7. PubMed CrossRef

116 Siersema, PD, Schrauwen, SL & van Blankenstein, M et al. Self-expanding metal stents for complicated and recurrent esophagogastric cancer. Gastrointest Endosc 2001; 54: 579–86. PubMed CrossRef

117 Vakil, N, Morris, AI & Marcon, N et al. A prospective, randomized, controlled trial of covered expandable metal stents in the palliation of malignant esophageal obstruction at the gastroesophageal junction. Am J Gastroenterol 2001; 96: 1791–6. PubMed

118 Razzak, R, Laasch, HU & England, R et al. Expandable metal stents for the palliation of malignant gastroduodenal obstruction. Cardiovasc Intervent Radiol 2001; 24: 313–18. PubMed CrossRef

119 Espinel, J, Vivas, S & Munoz, F et al. Palliative treatment of malignant obstruction of gastric outlet using an endoscopically placed enteral Wallstent. Dig Dis Sci 2001; 46: 2322–4. PubMed CrossRef

120 Shand, AG, Grieve, DC & Brush, J et al. Expandable metallic stents for palliation of malignant pyloric and duodenal obstruction. Br J Surg 2002; 89: 349–50. PubMed CrossRef

121 Aviv, RI, Shyalmalan, G & Kahn, FH et al. Use of stents in the palliative treatment of malignant gastric outlet and duodenal obstruction. Clin Radiol 2002; 57: 587–92. PubMed CrossRef

122 Maetani, I, Tada, T & Shimura, J et al. Technical modifications and strategies for stenting gastric outlet strictures using esophageal endoprostheses. Endoscopy 2002; 34: 402–6. PubMed CrossRef

123 Whithing, JL, Sigurdsson, A & Rowlands, DC et al. The long term results of endoscopic surveillance of premalignant gastric lesions. Gut 2002; 5: 378–81. CrossRef

124 Varis, K, Taylor, PR & Sipponen, P et al. Gastric cancer and premalignant lesions in atrophic gastritis: a controlled trial on the effect of supplementation with alpha tocopherol and beta-carotene. The Helsinki Gastritis Study Group. Scand J Gastroenterol 1998; 33: 94–300.

125 Forbes, GM & Threlfall, TJ. Treatment of Helicobacter pylori infection to reduce gastric cancer incidence: uncertain benefits of a community based programme in Australia. J Gastroenterol Hematol 1998; 13: 1091–5.

126 Fendrick, AM, Chernew, ME & Hirth, RA et al. Clinical and economic effects of population based Helicobacter pylori screening to prevent gastric cancer. Arch Intern Med 1999; 159: 142–8. PubMed CrossRef

127 Kokkola, A, Sipponen, P & Rautelin, H. The effect of Helicobacter pylori eradication on the natural course of atrophic gastritis with dysplasia. Aliment Pharmacol Ther 2002; 16: 515–20. PubMed CrossRef

128 Lambert, R. The role of endoscopy in the prevention of esophagogastric cancer. Endoscopy 1999; 31: 180–99. PubMed CrossRef

129 Hosokawa, O, Tsuda, S & Kidani, E et al. Diagnosis of gastric cancer up to three years after negative upper gastrointestinal endoscopy. Endoscopy 1998; 30: 669–74. PubMed

130 Hosokawa, O, Watanabe, K & Hatorri, M et al. Detection of gastric cancer by repeat endoscopy within a short time after negative examination. Endoscopy 2001; 33: 301–5. PubMed CrossRef

131 Kawai, K & Wanatabe, Y. The impact of mass screening on gastric cancer mortality in Japan. Gastrointest Endosc 1998; 47: 320–2. PubMed

132 Oshima, A, Hirata, N & Ubukata, T et al. Evaluation of a mass screening program for stomach cancer with a case-control study design. Int J Cancer 1986; 38: 829–33. PubMed

133 Oshima, A. (1997). Secondary prevention: screening methods in high incidence areas. In: Gastric Cancer (eds Sugimura T, Sasako M), pp. 199–212. Oxford University Press, Oxford.

134 Pisani, P & Parkin, M. (1996). Screening for gastric cancer. In: Advances in Cancer Screening (ed. Miller, AB), pp. 113–19. Kluwer, Boston.

135 Babazono, A & Hillman, AL. Declining cost effectiveness of screening for disease: the case of gastric cancer in Japan. Int J Technol Assess Health Care 1995; 11: 354–64. PubMed

136 Nomura, AMY, Stemmermann, GN & Samloff, IM. Serum pepsinogen I as a predictor of stomach cancer. Ann Intern Med 1980; 93: 537–40. PubMed

137 Stemmermann, GN, Samloff, IM & Normura, AMY et al. Serum pepsinogens I and II and stomach cancer. Clin Chim Acta 1987; 163: 191–8. PubMed CrossRef

138 Miki, K, Ichinose, M & Ishikawa, KB et al. Clinical application of serum pepsinogen I and II levels for mass screening to detect gastric cancer. Jpn J Cancer Res 1993; 84: 1086–90. PubMed

139 Hattori, Y, Tashiro, H & Kawamoto, T et al. Sensitivity and specificity of mass screening for gastric cancer using the measurement of serum pepsinogens. Jpn J Cancer Res 1995; 86: 1210–15. PubMed

140 Yoshihara, M, Sumii, K & Haruma, K et al. The usefulness of gastric mass screening using serum pepsinogen levels compared with photofluorography. Hiroshima J Med Sci 1997; 46: 81–6. PubMed

141 Kitahara, F, Kobahashi, K & Sato, T et al. Accuracy of screening for gastric cancer using serum pepsinogen concentrations. Gut 1999; 44: 693–7. PubMed

Copyright © Blackwell Publishing, 2004