Gastric Cancer: An Evolving Disease

Abstract

Purpose of Review:

Prevalence of Helicobacter pylori, the primary risk factor for gastric cancer, is declining globally. Paralleling this trend, gastric cancer incidence is also decreasing. Historically, the populations most affected by this neoplasia have been males, Asians and groups with low socioeconomic status. This review provides an update on recently published literature regarding changes in gastric cancer epidemiology.

Recent Findings:

Gastric cancer incidence trends vary by age, sex, race/ethnicity and tumor anatomical location. Overall incidence appears to be leveling off among young birth cohorts in Western populations, where H. pylori has declined considerably. The changes are more prominent for females and for tumors arising beyond the esophageal-gastric junction.

Summary:

The classical incidence pattern of gastric cancer is evolving. While uncertain, several hypotheses may explain the changing burden of disease. The mix of gastric cancer risk factors appears to be shifting, with H. pylori no longer the sole etiological driver. These changes may eliminate the previous predilection of males and lead to increases in overall gastric cancer rates. Analytical studies addressing known and novel factors related to major societal transitions may provide clues to understanding re-emergence of this serious public health problem.

Introduction

Gastric cancer is a heterogeneous disease with four molecular subtypes described as Epstein-Barr virus (EBV)-positive, microsatellite instability, genomically stable, and chromosomal instability [1]. In terms of anatomic subsites, tumors of the gastric cardia, which account for roughly 12% of cases worldwide, differ in epidemiological and clinical features from tumors arising in other parts of the stomach [2]. The etiology of gastric neoplasia is multifactorial, involving a variable combination of chronic Helicobacter pylori infection, occasional involvement of EBV coinfection, other environmental factors and host characteristics [2]. In particular, seropositivity for H. pylori is positively associated with noncardia gastric cancer but inversely associated with cardia cancer [3]. Furthermore, obesity is a risk factor for gastric cardia cancer only [4]. In addition, the driving factors associated with cardia cancer seem to be different in populations with high vs. low gastric cancer incidence [5].

Globally, gastric cancer incidence and mortality have substantially decreased over recent decades. According to GLOBOCAN [6], over 1 million new cases of gastric cancer were estimated to have occurred in 2018, making it the fifth most common malignancy (6% of all cancers). Gastric cancer is the third leading cause of cancer mortality (8% of all cancer deaths) [6], mainly because of advanced stage at diagnosis. It is predicted that numbers of cases will remain level or even increase in the future, reflecting the aging and growth of high-risk populations. The risk of gastric cancer varies in different parts of the world, with high rates in Central and Eastern Asia, Eastern Europe and Latin America [2]. In general, gastric cancer rates have been higher in males than females for reasons that are largely unknown.

The primary causal infection H. pylori tends to be acquired in childhood, persisting in the stomach lifelong in the absence of intensive antibiotic therapy. Determinants of H. pylori infection are associated with low socioeconomic status, particularly overcrowding and poor sanitation [7]. Although once ubiquitous in the human gastric microbiome, H. pylori prevalence has been declining in successive generations and is now rare in Western Europe, North America, Oceania, and some parts of Asia [8]. Thus, the gradual disappearance of this bacterium seems to be a consequence of economic development and widespread use of antibiotics. Incidence of noncardia gastric cancer in turn is declining in many parts of the world. On the other hand, the worldwide obesity epidemic seems to have been responsible for an increase in the incidence of gastric cardia cancer [9]. This variation in relative burdens of noncardia and cardia cancers complicates the interpretation of geographic and secular trends for gastric cancer overall.

Gastric carcinogenesis is a multistep process [10]. As a result of chronic H. pylori infection, most cases of noncardia gastric cancer are preceded by a sequence of histological lesions of the gastric mucosa known as “Correa’s cascade,” beginning with inflammation (i.e., gastritis) and progressing to atrophy, intestinal metaplasia (IM) and dysplasia. H. pylori eradication has been suggested as a potential strategy to interrupt this sequence and thereby prevent progression to malignancy [11]. However, as outlined below, other factors appear to have growing importance in the etiology of gastric cancer. Thus, elimination of H. pylori infection may not have the intended result of diminishing the serious burden of this disease.

Evolving Disease

Trends in gastric cancer

Although trends overall are generally declining worldwide, age-period-cohort analyses of gastric cancer incidence indicate recent cohort-specific changes in some countries. In particular, overall gastric cancer rates appear to be leveling off among younger birth cohorts in the United Kingdom, Czech Republic, Denmark, Norway, and Iceland [12] as well as in Spain [13]. The changes seem to be more pronounced for certain anatomic subsites and populations. In Korea, for example, where the incidence of gastric cancer is the highest in the world, Eom et al., found increasing trends in both cardia and noncardia tumors, with the highest estimated annual percentage change (EAPC) for noncardia tumors localized to the gastric corpus (3.7%) [14]. Also, Song et al. reported increasing incidence among young Korean women [15]. Reports from the Netherlands [16] and Portugal [17] indicate plateauing trends of noncardia gastric cancer incidence in the last decade.

In our recent analysis based on North American Association of Central Cancer Registries (NAACCR) data for 1995-2013, noncardia gastric cancer incidence rates among non-Hispanic Whites in the United States have been declining among older adults but rising for younger persons [18]. Moreover, analyses of subsite-specific trends indicate a shifting distribution by anatomic subsite, with a statistically significant increase of gastric corpus cancer. In the population younger than age 50 years, the rise of noncardia gastric cancers was more pronounced in women than in men (EAPC, 2.6% vs. 0.2%). On average, the rising rates were mainly seen in the group of counties with less than 20% prevalence of poverty, implying that factors other than H. pylori infection may be driving the trends. Noncardia gastric cancer incidence trends in Hispanics also showed a statistically significant age interaction, with slightly rising vs. falling rates for individuals aged <50 and 50+ years, respectively. However, these birth-cohort effects were not seen in African Americans or Asian Americans.

With increasingly detailed reporting of anatomic subsite assignment, gastric cancer of unspecified subsite has declined over time, which conceivably could generate artefactual trends for cases at specific subsites. However, even assuming extreme patterns of their redistribution, the numbers of cases that were previously unclassified are insufficient to account for the observed trends at specific subsites. Although trends by histologic subtype may be obscured by lack of standardized pathological reporting, there have been several reports of increases in Lauren diffuse-type and/or signet-ring cell histologies [19-21].

These provocative findings await extension to analyses of descriptive data for other populations, at a minimum separating cardia and noncardia gastric cancer. Heterogeneity by age, sex, and presence of key molecular features would be important to address.

Trends in gastric preneoplastic lesions

IM refers to histological changes of atrophic gastric epithelium that resemble small intestine (known as complete-type) or large intestine (known as incomplete-type), distinguished by the presence or absence of expression of small intestinal digestive enzymes [22]. IM is a recognized premalignant lesion, conferring between 2- to 36-fold increased cancer risk [23]. Prospective studies have documented variable rates of progression to gastric cancer, with annual rates up to 0.4% [24]. IM exhibits genetic instabilities with a complex and dynamic clonal composition [25]. The malignant potential of IM varies by histological phenotype and by anatomical extension of mucosal involvement, with higher risk conferred by incomplete-type IM and by involvement of both the antral and corpus portions of the stomach [26]. While clinical management of IM entails periodic re-examination in Europe and Asia [27, 28], there is a paucity of recommendations regarding endoscopic surveillance in the United States and Latin America.

IM has a patchy distribution and is a relatively common histological finding on upper gastrointestinal endoscopy, present in up to 25% of adults with dyspepsia [29]. Trends of preneoplastic lesions are difficult to analyze due to changing histological definitions and variable adherence to standardized biopsy protocols [30]. Based on limited population-based studies from Hong Kong, Japan, Finland and the Netherlands, the prevalence and incidence of IM may be declining [31-34]. However, no reports have investigated temporal changes in the past decade or potential heterogeneity by age and sex. Prevalence in children and adolescents is largely unknown and the association with H. pylori infection seems to be variable [35, 36]. More consistent documentation of precancerous changes in multiple populations may help in interpreting the observed incidence trends for gastric cancer.

Hypotheses for Changing Trends

Taken together, the above descriptive evidence suggests changes in the epidemiology of gastric cancer, particularly in Western populations. While uncertain, several potential mechanisms may explain these unfavorable trends.

The foremost may be a localized change in microecology of the stomach. The timing of changes in gastric cancer incidence parallels the antibiotic era. The loss of H. pylori infection may open a niche for other microbes with injurious effects on the gastric mucosa [8]. This microbiome dysbiosis, related or unrelated to exposure to antibiotics, could also involve autoimmunity.

In the NAACCR analysis cited above, the largest observed increase was in cancers localized to the corpus of the stomach, the anatomical location where the autoimmune type of gastritis predominates [18]. Thus, it is speculated that autoimmune gastritis may be a plausible reason underlying these trends. Autoimmune gastritis is one manifestation in a constellation of autoimmune phenomena, and frequently co-occurs with other autoimmune conditions including autoimmune thyroiditis, type I diabetes, vitiligo and Addison disease [37]. Its primary clinical manifestation, pernicious anemia, is usually diagnosed through detection of autoantibodies directed against gastric parietal cells and/or intrinsic factor, which may also be important contributors to pathogenesis [38]. The epidemiology of pernicious anemia is poorly characterized but does not seem as restricted to older women of North European decent as once assumed [39].

Epidemiologic data suggest multiple types of allergic and autoimmune diseases have increased in developed countries over recent decades [40] and, in general, are more prevalent in women. The prevalence of autoimmune gastritis is difficult to assess but there are hints that this particular condition may be increasing, especially among younger adults. Based on serosurveys in Northern Sweden over the twenty years between 1990 and 2009, prevalence of serologic atrophy, as defined by low serum pepsinogen I, increased from 22 to 64 per 1000 individuals aged 35 to 44 years, despite declining prevalence of H. pylori seropositivity, implicating gastric autoimmunity as an alternative etiology [41].

Apart from H. pylori, the carcinogenic role of other constituents of the gastric microbiome (i.e., viruses, bacteria, archaeae, fungi etc.) is an interesting and open-ended question warranting investigation. There is an uncertain and controversial role of the gastric microbiome in the etiology of autoimmune gastritis. H. pylori colonization is variably present and, through molecular mimicry, considered to be a potential trigger for autoimmune activity directed against the gastric mucosa [42]. Parsons et al. found distinct gastric microbiome profiles in patients with autoimmune gastritis compared to those with H. pylori-induced gastritis, with relative increase of Streptococcus spp. [43]. Further studies are needed in different populations of the prevalence of autoimmune gastritis and its association with gastric cancer.

Proton pump inhibitor therapy is common and has increased over time [44]. These drugs have been frequently observed to be associated with gastric cancer, but the etiologic role is uncertain. Some if not all of the association may be explained by reverse causation since premonitory symptoms may lead to prescription [45]. Chronic acid-suppression is well-recognized to exacerbate H. pylori-associated gastritis in the absence of eradication, favoring proximal redistribution of colonization and corpus predominance of gastritis and atrophy [46]. This physiologic alteration conceivably could be related to the increase in incidence of tumors localized to the corpus [45]. Accordingly, Wennerström et al. matched the Danish prescription drug registry with cancer registration data to evaluate the hypothesis that chronic acid suppression therapy and consequent proximal extension of H. pylori infection would increase the proportion of tumors arising in the proximal stomach [47]. The findings did not support this hypothesis, but do not preclude that a proximal shift may emerge in the future with longer duration of exposure.

Episomal EBV is found in the tumor cells of ~9% of gastric adenocarcinomas worldwide [48]. EBV-positive tumors are more common in men and are preferentially localized to the corpus and other non-antral portions of the stomach, which may have relevance for the shifting trends in subsite distribution. There are no reliable data regarding temporal changes in prevalence.

Another potential explanation for the observed trends in gastric cancer would be obesity, which is increasing markedly in the United States and other Western countries [49]. Excess weight is a recognized risk factor for cardia cancer, thought to reflect its association with the inflammatory syndrome and/or gastric reflux [2]. However, a meta-analysis of 10 observational studies does not support an association with noncardia gastric cancer (odds ratio 1.26; 95% confidence interval 0.89 to 1.78) [50]. In particular, there is no specific association with gastric corpus cancer [51].

Age- and sex-specific incidence rates indicate that up to the age of 60 years women typically have had relatively lower gastric cancer risk than men, potentially implicating a protective role for female sex hormones that wanes after menopause [52]. Epidemiological studies of menstrual and reproductive factors are roughly consistent with this hypothesis [53]. However, the sex disparity in gastric cancer incidence has been diminishing in recent birth cohorts [18]. Trends in age at menarche and menopause [54] do not seem to explain the increase in noncardia gastric cancer in young women, but variation in hormonal contraceptive use and other reproductive factors should be further examined.

In the United States, there is no evidence of shifts in cigarette use and salt consumption, well-recognized gastric cancer risk factors, that could explain the profile of changes observed in gastric cancer incidence.

Conclusions

Evolving trends in gastric cancer may be a consequence of modern lifestyles and continuing changes in the environment. Three major types of this disease have been proposed as summarized in the Table. The authors speculate gastric cancer will increasingly affect Western populations, younger individuals, females, and higher socioeconomic groups. These transitions will have a major impact in overall gastric cancer incidence. Human populations worldwide have benefitted from a prolonged period of decline in gastric cancer incidence which may be drawing to a close.

Table.

Changing patterns of gastric cancer epidemiology

		Classic	Modern	Future
Phenotype	Anatomical predilection	Antrum	Cardia	Corpus
	Histology	Intestinal	Both	Diffuse
Population	Race/ethnicity	Asians	Caucasians	Caucasians
	Male predominance	++	+++	−
Risk factors	Old age	+++	++	+
	Low socioeconomic status	+++	+/−	−
	H, pylori	+++	+/−	+/−
	Peptic ulcer disease	++	−	−
	Obesity	−	++	−
	Autoimmunity	−	−	++?
Gastric Microecology	Microbiome diversity	−	+?	++?
	Gastric acidity	+++	++	+

Hatched circles indicate anatomical predilection of each phenotype (Modified from: Atherton and Blaser, 2009 [55]).

The relative contributions of major risk factors for gastric cancer are shifting in the face of a diminishing role for H. pylori, with emerging potential importance of autoimmunity. The identification of specific causative factors will require well-designed studies from multiple disciplines and consideration of factors beyond the current paradigm.

Abbreviations:

EBV

Epstein-Barr virus

EAPC

Estimated annual percentage change

IM

Intestinal metaplasia

NAACCR

North American Association of Central Cancer Registries