Improving the Early Diagnosis of Gastric Cancer

Abstract

Gastric cancer (GC) remains a leading cause of cancer morbidity and mortality worldwide. Outcomes from GC remain poor, especially in Western nations where cancer diagnosis is usually at advanced stages where curative resection is not possible. By contrast, nations of East Asia have adopted methods of population-level screening with improvements in stage of diagnosis and survival. In this review, we discuss the epidemiology of GC in Western populations, highlight at-risk populations who may benefit from screening, overview screening modalities, and discuss promising approaches to early GC detection.

Introduction

Every year 1.2 million persons are diagnosed with and 860,000 persons die from gastric cancer (GC) worldwide,¹ making GC the fifth-leading cause of cancer incidence and third leading cause of cancer mortality, respectively.² Outcomes from GC in most of the world remain poor, including in the United States (US). In the US, GC afflicts 27,000 each year³ and carries a dismal prognosis (5-year survival of 27%).⁴ These statistics reflect the fact that the majority of GCs in the US are diagnosed at advanced stages,⁴ where curative resection is unlikely. Strategies to improve the early diagnosis of GC are therefore crucial to improving survival.

GCs are classified as cardia or non-cardia based on the anatomic location of origin within the stomach. Cardia GCs, which share risk factors and natural history with esophageal adenocarcinomas, constitute approximately one-quarter of GCs worldwide.⁵ Non-cardia GCs constitute three-quarters of GCs worldwide, have witnessed improvements in outcomes following adoption of screening programs in nations of East Asia,^{6, 7} and will be the focus of this review.

Helicobacter pylori and Correa’s Cascade

Development of non-cardia GC has been linked in multiple epidemiologic studies with infection with the gram-negative, micro-aerophilic organism Helicobacter pylori (Hp).^{8, 9} Worldwide Hp prevalence rates range from <40% in industrialized nations of Western Europe and North America to >70% in areas of South America, Africa, Eastern Europe, and East Asia.¹⁰ Hp infection is associated with a three-fold increase in lifetime odds of development of non-cardia GC; moreover, Hp is believed responsible for 75-95% of all GC cases worldwide.^{11, 12} Colonization with Hp induces a state of chronic inflammatory insult which leads to a cascade of mucosa perturbations, termed Correa’s Cascade (Figure 1).¹³ In Correa’s cascade, chronic gastritis is followed progressive atrophy of the oxyntic or antral gastric mucosa, and then eventual replacement by intestinal mucosa consisting of Paneth, goblet, and absorptive cells. Intestinal metaplasia (IM) of the stomach is an important precursor lesion in the pathway to GC,^14-17 and regional prevalence of IM correlates closely with incidence of GC worldwide.¹⁸ Even with decreasing prevalence of Hp, with the secular aging of the global population GC cases and deaths are expected to climb well into the 21^st century.^{19, 20}

Figure 1:

Proposed carcinogenic cascade induced by Helicobacter pylori (Hp) and other environmental insult. Patients with atrophic gastritis, intestinal metaplasia, and dysplasia remain at increased risk for gastric cancer even following Hp eradication.

Gastric Cancer Screening and Outcomes in East Asia

The incidence of GC is significantly higher in nations of East Asia compared to the US. While the incidence of GC is roughly 6 per 100,000 in the US, it is approximately 28 per 100,000 in Japan and 34 per 100,000 in South Korea. Yet while incidence of GC is much higher in these countries, survival from GC is also higher compared to the US or Western Europe (Figure 2). Five-year observed survival from GC exceeds 60% in both South Korea and Japan, compared to below 30% for the US and Western Europe.^{21, 22} These differences in survival are due in large part to differences in stage of diagnosis. While nearly 60% of GCs are diagnosed at a surgically or endoscopically curable stage in South Korea and Japan, fewer than a quarter of GCs are diagnosed at such stages in the West.^{4, 23-25}

Figure 2:

Left panel depicts five-year observed survival following gastric cancer diagnosis in East Asia (South Korea, Japan) and Western nations (United States and Europe). Right panel depicts the proportion of all gastric cancers diagnosed at localized stage based on United States National Cancer Institute summary staging.

Data from Refs.^{4, 23-25}

In Japan, a national screening program for GC was first introduced in 1983. This consisted of radiography-based screening of all adults ≥40 years old, with endoscopic examination performed on individuals with abnormal radiographic results.²⁶ Based on the results of several rigorous observational studies, the national screening program was amended in 2016 to allow for either endoscopic or radiographic screening for adults ≥50 years old on a biennial basis.²⁶ While endoscopic screening is rapidly being adopted throughout Japan, radiographic screening is still the predominant screening modality in most prefectures.²⁷

South Korea initiated a biennial screening program consisting of either endoscopic or radiographic screening for adults ≥40 years old in 2002.²⁸ In practice, endoscopic screening has been the predominant modality practiced in South Korea due to patient preference. Since the initiation of the national screening program, the proportion of gastric cancers diagnosed as early gastric cancer(defined as tumor with invasion limited to mucosa or submucosa) has increased from 39% in 2001 to 73% in 2016.²⁹ Moreover, observed five-year survival has increased from 46% to 75%.²⁹

Efficacy and Safety of Screening Modalities

Radiographic Screening

Radiographic screening involves the ingestion of a contrast agent (often barium), and subsequent fluoroscopic imaging of the gastric lumen. Contrast radiography allows for the detection of luminal pathology including ulcers, polyps, and masses. However, compared to modern endoscopy, radiography has both limited sensitivity and specificity.³⁰ Cancer registry data suggests that the sensitivity of radiographic screening ranges from 60-80%, and specificity from 80-90%.⁶ It should also be noted that for early GC (where a luminal prominence or depression may be minimal), the sensitivity of radiography has been reported to be significantly lower (14-36%).^31,32

The efficacy of radiographic screening has been assessed in several observational studies, including both cohort and case-control studies,²⁶ though notably no randomized control trial has been conducted comparing radiographic screening with standard of care. From cohort studies from Japan^{33, 34} comparing radiographic screening with no screening over long-term follow-up (ranging from 11-13 years), receiving radiographic screening was associated with both reductions in GC-specific mortality (with relative risk ranging from 0.52-0.54) as well as all-cause mortality (relative risk 0.71-0.83). However, during the period of these studies radiography was also a standard test for assessment of gastrointestinal symptoms, introducing the possibility of confounding by indication. Moreover, receipt of subsequent screening (such as by endoscopy) in the follow-up period was not ascertained possibly causing overestimation of the effect size. The safety profile of radiographic screening is generally favorable, with mild risk of constipation or ileus and rare cases of aspiration pneumonia.⁶ In a report of over 3 million radiographic screening procedures performed, only a single death was attributed to an adverse event related to screening.²⁶ Radiation exposure from photofluorographic screening is in the range of 0.6 mSv (by comparison a standard chest x-ray exposure is approximately 0.1 mSv).

As a relatively safe and inexpensive modality, radiography may continue to serve a role for GC screening in resource-limited settings. However, as the primary motivation for screening is to improve the detection of early-stage cancers, radiography has limited utility compared to modern, high-resolution gastrointestinal endoscopy. Moreover, when an abnormality is detected through radiography a confirmatory upper endoscopy is required for visualization and tissue acquisition.

Endoscopic Screening

Since 2001 in South Korea and 2016 in Japan,²⁹ endoscopic screening has been offered as an alternative modality to radiographic screening. Endoscopic screening offers several advantages to radiographic screening, including the ability to directly visualize the gastric mucosa and tissue sampling of abnormal-appearing tissue or visible lesions. Compared to radiographic screening, endoscopic screening demonstrates both better sensitivity and specificity.^{26, 35} This increased sensitivity is especially important for early GCs that demonstrate only subtle mucosal changes, and which may not have an elevated or depressed component visible on contrast radiography. Techniques to enhance mucosal contrast have been developed to improve detection of subtle lesions, such as narrow-band imaging and chromoendoscopy (Video 1). In narrow-band imaging, conventional white light is filtered into defined wavelengths in order to maximize absorption by hemoglobin, as well as limit penetration of light beyond the mucosal surface. Given this shorter wavelength, the resulting ‘blue’ light penetrates less deeply than conventional white light and may improve contrast of the mucosal surface. Chromoendoscopy also serves to amplify contrast of mucosal lesions through the use of dye-based staining of the gastric mucosa with biologically-compatible agents such as acetic acid or methylene blue.^{36, 37} Application of dilute acetic acid can modify the optical properties of the epithelium by slightly altering the pH or by reversibly altering the structure of cellular proteins to reflect white light. Methylene blue is actively absorbed by small intestinal epithelium but not normal gastric epithelium, enhancing contrast between metaplastic and normal gastric epithelium. Chromoendoscopy may improve the delineation of surface irregularities, which in turn may improve the diagnosis and staging of early GCs.³⁸ Early GC detection may allow for opportunities for endoscopic resection through endoscopic submucosal dissection (ESD, Figure 3). For GCs confined to the mucosa or proximal submucosa (with invasion depth of <500 microns) and without lymph node involvement, ESD offers similar cure rate and fewer rates of adverse events compared to surgical gastrectomy based on retrospective series from East Asia.^{39, 40}

Figure 3:

Top left panel depicts a subtle, flat, erythematous lesion which was biopsied to be gastric adenocarcinoma. Use of narrow-band imaging (top right) enhances visualization and delineation of the lesion. This lesion was staged as an early gastric cancer (tumor invasive to no deeper than mucosa or submucosa), and removed by endoscopic submucosal dissection (bottom left). En bloc resection specimen (bottom right) confirmed tumor confined to mucosal layer, without lymphovascular invasion, and with negative lateral and deep margins consistent with curative resection.

The efficacy of endoscopic screening in decreasing cancer-specific mortality have been evaluated in observational studies form East Asia. A systematic review and meta-analysis of the protective effect of endoscopic screening on cancer-specific mortality identified 10 studies (six cohort studies and four case-control studies) from South Korea, Japan, and China.⁴¹ Receipt of endoscopic screening was associated with an approximate 40% reduction in risk for GC-specific mortality in the pooled estimate, with a robust protective effect found compared both against no screening and radiographic screening controls.⁴¹ When reviewing the existing evidence in support of endoscopic screening, the Japanese Guideline Development Group initially found inadequate observational data to justify population-level endoscopic screening in 2008.^{6, 42} However, based on the results of numerous high-quality observational studies published after 2008, the Japanese guidelines were amended to favor endoscopic screening in 2018 with a evidence score of 2+ (moderate-quality case-control and cohort studies with a low risk for bias, confounding or chance and a moderate probability that the relation is causal).²⁶ Notably the primary endpoint of these studies has been GC-specific mortality (as opposed to overall mortality). Currently no randomized control trial data exists for the benefits of endoscopic screening.

In the US and Europe most upper endoscopies are performed under sedation (either moderate or deep). The risk of cardiopulmonary events related to sedation have been estimated to be between 1 in 170 to 1 in 10,000, with the higher range of estimates incorporating minor events (such as changes in oxygen saturation or heart rate).⁴³ The Japanese Association of Gastroenterological Cancer Screening has found an overall rate of complications of 87 per 100,000 for endoscopic screening and 43 per 100,000 for radiographic screening.²⁶

Serologic Screening

Hp-induced inflammation begins in the antrum and proceeds upward to the corpus with chronic infection. Human pepsinogens are classified into two biologically distinct types, pepsinogen I and pepsinogen II. As inflammation proceeds toward the corpus with chronic Hp infection, levels of pepsinogen I (produced by chief cells in the corpus) decrease whereas levels of pepsinogen II remain more constant.⁴⁴ As such, a decreased level of pepsinogen I and decreased pepsinogen I/II ratio may indicate advanced atrophic gastritis.⁴⁴ Serum pepsinogens in combination with Hp IgG antibody have been evaluated as non-invasive screening tools in East Asian cohorts.^45-47 However, use of these markers demonstrate significant limitations including a high degree of heterogeneity in reported testing characteristics between populations, differing cutoff points, and variability based on proton pump inhibitor use.^{48, 49} These methods are not currently used for population-level screening in either South Korea or Japan. Their use may also be limited in Western populations which differ in prevalence of Hp infection, proton pump inhibitor therapy use, and rates of autoimmune atrophic gastritis.^{49, 50}

GASTRIC CANCER SCREENING IN THE UNITED STATES

At Risk Populations

In the US, GC survival is poor (5-year observed survival of 27%) and the majority of cancers are diagnosed at regional or distant stages.⁴ While overall incidence of GC is modest among the general population (~6 per 100,000), certain high-risk racial (Asians, Alaskan Indians, American Indians, Blacks/African Americans) and ethnic (Hispanics) groups may face significantly higher risk (Figure 4). Very high-risk subgroups such as Japanese and Korean Americans face an incidence six- to eight-fold higher than non-Hispanic Whites.⁵¹ Beyond race and ethnicity, Americans at increased risk for GC include those with a family history or with cancer-predisposing syndromes, recent immigrants form high-incidence regions of the world, those with a history of Hp infection, and those with precancerous changes of the stomach.²⁹ It behooves both clinicians and policy makers to be cognizant of high-risk groups, and to offer appropriate counseling for the role of preventative strategies such as GC screening.

Figure 4:

Crude incidence of gastric cancer in United States (per 100,000; Y-axis) plotted by age group (X-axis). Asians, Blacks, Hispanics, and American Indian/Alaskan Natives face a several-fold increased risk compared to non-Hispanic Whites.

Data from the Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) Research Data (1973-2015), National Cancer Institute, DCCPS, Surveillance Research Program released in April 2018 based on the November 2017 submission.

IM is a critical precursor lesion to GC, and prevalence of IM in populations correlates with GC incidence. In the US, the prevalence of intestinal metaplasia has been estimated to be between 5-10% of the general population.^{52, 53} High-risk subgroups including certain racial and ethnic minorities may have IM prevalence several-fold higher.^{54, 55} Within the US, there appears to be a close association between prevalence of IM with incidence of GC within racial subgroups (Table 1). These data suggest that all racial and ethnic groups are at risk for GC once IM has developed. While gastritis and atrophy may reverse and normalize following Hp eradication, IM often persists.^{56, 57} Moreover, long-term clinical follow-up suggest that patients with IM remain at increased risk for GC even after eradication of Hp.^{58, 59}

Table 1:

Estimated Prevalence of Intestinal Metaplasia and Incidence of Gastric Cancer in Racial/Ethnic Groups in the United States

Racial/Ethnic Group	Prevalence of IM	Incidence of GC	References
Non-Hispanic Whites	7-9%	6-8 per 100,000	4, 21, 52, 54
Non-Hispanic Black	21%	11 per 100,000	4, 21, 55
Hispanics	12-30%	11 per 100,000	4, 21, 54, 55
Chinese	26%	15 per 100,000	51, 54
Koreans	40%	45 per 100,000	51, 54

IM,intestinal metaplasia; GC,gastric cancer.

It is estimated that the annual rate of progression onto GC from IM to be approximately 0.25%.⁶⁰ However, this aggregation of risk does not capture the variability of presentation in IM histologic severity or topographic distribution. In order to estimate histologic severity, scoring systems such as the operative link for gastritis assessment (OLGA)⁶¹ for atrophic gastritis, and the operative link for gastric intestinal metaplasia (OLGIM) for IM have been developed (Figure 5).⁶² OLGA and OLGIM rely on both an assessment of the topographic extent of disease as well as the percentage of glandular involvement from each biopsy location utilizing a visual-analogue scale.⁶³ The resultant stage score, ranging from 0 (no IM or no atrophy) to 4 (severe, extensive IM or atrophy) have been validated in several observational studies as risk-stratification tools for progression onto subsequent GC.^64-67 Use of the OLGA and OLGIM systems is dependent on consistent sampling of multiple locations of the stomach (including antrum, incisura, and corpus/body) in a systematic manner termed the ‘Sydney Protocol’.⁶³ Another promising method of histologic risk stratification is through distinguishing complete IM from incomplete IM. Complete IM is characterized by well-defined goblet cells and a well-developed brush border, whereas in incomplete IM mucin droplets of varying sizes and shapes can be found and there is an absence of a brush border.⁶⁸ In specialized centers where mucin staining is available, complete IM will be found to display predominantly small intestinal phenotypic markers such as MUC2 and sucrase, whereas incomplete IM only selectively or incompletely expresses small intestinal markers but may express gastric phenotypic markers MUC5AC and large intestinal phenotypic markers such as Das-1.⁶⁹

Figure 5:

Scoring of gastric precancerous lesions using the Operative Link systems. In these scoring systems, biopsies from the gastric antrum and body are individually scored for degree of atrophic gastritis and intestinal metaplasia utilizing a visual-analogue scale (none, mild, moderate, marked). A summary stage for both atrophy and intestinal metaplasia is then assigned.

Current Recommendations

Recommendations for GC screening or precancerous lesion surveillance by US-based professional societies are depicted in Table 2. Currently, the American Society of Gastrointestinal Endoscopy (ASGE) has recommended endoscopic screening for GC in first-generation immigrants from high-risk regions (i.e. Japan, China, Russia, and South America) over aged 40, in particular if there is a family history of GC in a first-degree relative.⁷⁰ Regarding surveillance of patients with precancerous lesions, the ASGE recommends surveillance of patients with atrophic gastritis or IM when there is increased risk of GC due to ethnic/racial background, positive family history, or extensive anatomic distribution of disease.^{14, 70} By contrast, the American Gastroenterological Association (AGA) recommends against the routine use of endoscopic surveillance in patients with IM, but clarifies that this is a conditional recommendation based on very low quality of evidence.⁷¹ The AGA guidelines further state that “Patients with IM at higher risk for GC who put a high value on potential but uncertain reduction in GC mortality, and who put a low value on potential risks of surveillance endoscopies, may reasonably elect for surveillance.” The AGA guidelines identify patients with IM at higher risk for gastric cancer as those with incomplete IM, those with extensive IM, and those with a family history of gastric cancer. The AGA guidelines also identify patients at overall increased risk for GC including racial/ethnic minorities and immigrants from high-incidence regions. Notably neither the ASGE nor the AGA have made recommendation on the optimal interval for surveillance of IM if this strategy is pursued.

Table 2:

Guidelines Issued by United States Professional Societies

Society	Year	Recommendation
Gastric Cancer Screening
American Society of Gastrointestinal Endoscopy70	2015	Endoscopic screening for gastric cancer in first-generation immigrants from high-risk regions (e.g. Japan, China, Russia, and South America) may be considered for those aged 40 years, particularly if there is a family history of gastric cancer in a first-degree relative
Surveillance of Intestinal Metaplasia (IM)
American Society of Gastrointestinal Endoscopy14, 70	2015	Endoscopic surveillance in patients with gastric atrophic gastritis or IM coupled with an increased risk of gastric cancer because of racial/ethnic background, extensive anatomic distribution, or family history
American Gastroenterological Association71	2019	Recommends against routine use of endoscopic surveillance in patients with IM. Conditional recommendation, very low quality of evidence Patients with IM at higher risk for gastric cancer who put a high value on potential but uncertain reduction in gastric cancer mortality, and who put a low value on potential risks of surveillance endoscopies, may reasonably elect for surveillance. Patients with IM specifically at higher risk of gastric cancer include those with: Incomplete vs complete IM Extensive vs limited IM Family history of gastric cancer Patients at overall increased risk for gastric cancer include: Racial/ethnic minorities Immigrants from high incidence regions

Existing recommendations from United States-based professional societies regarding screening of gastric cancer or surveillance of precancerous lesions such as intestinal metaplasia (IM).

FUTURE SCREENING MODALITIES

Promising biomarkers currently under development or validation may revolutionize the field of early GC detection and prevention. MicroRNAs (miRNAs) are small, non-coding molecules involved in biological processes including cell-cycle progression and apoptosis. Given their stability, presence in blood, and role in numerous pathways, miRNAs have been evaluated as potential biomarkers for GC. Five miRNAs have shown particular promise as screening tests for GC: miR-21^72-74 (inhibitor of tumor suppressor genes), miR-106a^{75, 76} (cell proliferation signal), miR-106b^{77, 78} (inhibitor of apoptosis), miR-223^73,79 (cell proliferation and invasion), and miR-421 ^80-82(apoptosis resistance). Notably these miRNAs have mostly been evaluated in cohorts from East Asia where the prevalence of Hp and incidence of GC is much higher. Additional prospective validation studies are required prior to translation of miRNAs to clinical practice.

Risk stratification of IM may also be improved through use of molecular markers. In a prospective cohort study of high-risk Singaporean Chinese patients, IM biopsy samples with shortened telomeres and chromosomal aberrations were found to be associated with subsequent progression to either dysplasia or frank carcinoma. By contrast, IM with normal-like epigenetic patterns were associated with stability or regression.⁶⁵ While requiring validation, such a molecular signature for IM progression may serve as a valuable tool to allow for focused and highly-personalized strategies of surveillance while also avoiding unnecessary endoscopies in low-risk subjects.

SUMMARY

GC remains a devastating disease for the 27,000 Americans diagnosed each year. Compared to East Asia, survival from GC in the US and Europe are lower, reflecting a later stages of diagnosis. In high-incidence nations of East Asia, national screening programs have been adopted. An emerging body of observational data suggests that endoscopic screening may prevent GC-specific mortality in targeted populations. There exist high-risk populations within the US who may benefit from targeted screening, including racial/ethnic groups (American Indians, Alaska Indians, Asians, Blacks, Hispanics), first-generation immigrants from high-incidence regions, and those with a family history of GC. Individuals diagnosed with IM, particularly extensive IM or histologically-severe IM, may benefit from endoscopic surveillance. Emerging molecular technologies may help to identify high-risk individuals who should be screened, as well as stratify IM for risk of cancer progression.

Supplementary Material

Video 1

Video 1: In this upper endoscopy, a subtle erythematous lesion is seen at the incisura which biopsies demonstrated to be gastric adenocarcinoma. White light endoscopic examination of the gastric antrum demonstrate subtle mucosa nodularity suggestive of intestinal metaplasia, which on narrow-band imaging (0:13) can be more clearly defined. Following application of dilute acetic acid (0:30) the areas of intestinal metaplasia become clearly defined. Finally application of methylene blue (0:45) which is selectively absorbed by intestinal epithelium, and not normal gastric epithelium, clearly delineates the areas of intestinal metaplasia.

Key Points:

• Patients are diagnosed with gastric cancer at more advanced stages and have overall lower survival in the United States compared to East Asia

• Observational data from Japan and South Korea, nations with national gastric cancer screening programs, show that endoscopic screening may improve gastric cancer mortality

• In the United States high-risk racial/ethnic groups (Alaskan Natives, American Indians, Asians, Blacks, Hispanics), first-generation immigrants form high-incidence regions, and individuals with a family history may benefit from screening

• Individuals with intestinal metaplasia, particularly extensive or histologically-severe disease, may benefit from endoscopic surveillance

• A video of use of chromoendoscopy to enhance detection of gastric intestinal metaplasia accompanies this article.

Clinical Care Points.

• Racial/ethnic minorities and first-generation immigrants are at increased risk for gastric cancer and may benefit from endoscopic cancer screening

• Patients diagnosed with incomplete, extensive, or severe intestinal metaplasia should be offered endoscopic surveillance

• Use of narrow-band imaging and chromoendoscopy during endoscopy can improve the detection of gastric cancer