How I Approach Surveillance of Gastric Intestinal Metaplasia

The problem

Globally, gastric cancer remains the 5^th most common cancer and 3^rd most common cause of cancer related-deaths. Intestinal-type gastric adenocarcinoma, the most common type of gastric cancer, develops from the progression of chronic atrophic gastritis to gastric intestinal metaplasia (GIM) to dysplasia, prior to malignant transformation in a small minority of individuals. Multifocal atrophic gastritis and GIM often coexist and are both associated with an increased risk of intestinal-type gastric adenocarcinoma. GIM is generally considered to be the first irreversible mucosal stage in the stepwise progression to cancer, albeit with some mixed data. GIM is defined histologically by intestinalization of normal gastric epithelium in response to chronic injury, most often colonization with Helicobacter pylori (H. pylori) although other causes such as autoimmune gastritis are recognized. GIM can be further classified according to histologic subtype of the intestinal metaplastic tissue as “complete-type”, which has features of small intestinal epithelium, “incomplete-type”, which has features of colonic epithelium, and “mixed-type”, which has features of both.¹ Because of the slow sojourn time between mucosal stages, endoscopic surveillance offers the opportunity for earlier detection and potentially curative treatment of gastric neoplasia (i.e. dysplasia or cancer).

Despite the established association with increased gastric cancer risk, the management of GIM presents a dilemma for many gastroenterologists. In the United States (US), the overall pooled prevalence of GIM is 4.8% (95% CI: 4.8–4.9%).² However, the marked variability in prevalence must be emphasized, as individual studies have demonstrated that GIM prevalence is significantly higher in certain subgroups, including some racial and ethnic minorities and current or former tobacco users, among others.² Recently, the American Gastroenterological Association provided the first US guidelines focused on the role of endoscopic GIM surveillance for improving gastric cancer-related outcomes. Here, we detail our approach to endoscopic surveillance of individuals diagnosed with GIM and highlight the importance of risk stratification. Specific recommendations for GIM surveillance intervals vary slightly among Western gastroenterological societies and full details are provided elsewhere.^3–6

Evaluate for and eradicate H. pylori in all individuals with GIM

Because H. pylori eradication reduces the risk of gastric cancer, all individuals with GIM should be evaluated and treated for active infection followed by confirmation of eradication.^6,7 If gastric biopsies are negative for H. pylori, we often perform an alternative test for H. pylori, such as urease breath test or H. pylori antigen stool test, since H. pylori colonization of the gastric mucosa might be patchy and is often absent in areas of GIM. Medications that reduce the sensitivity of H. pylori testing should ideally be discontinued prior to testing. There is insufficient evidence to advocate for empiric H. pylori eradication therapy among patients with GIM in the absence of documented H. pylori infection.

Patient Selection for GIM Surveillance

Although GIM is associated with a baseline 0.16% annual risk of incident gastric cancer based on a comprehensive systematic review,⁷ there was insufficient high-quality evidence to unequivocally establish whether GIM surveillance versus no surveillance is associated with improvement in important patient outcomes, such as gastric cancer-related mortality. Some individuals with GIM may exceed this baseline risk, however, and risk determination depends on a number of factors (Table 1). Risk stratification is complicated by the paucity of evidence delineating the attributable risks for each of these factors separately among individuals with GIM; nevertheless, individuals generally could be categorized as high- vs low-risk for gastric cancer.

Table 1.

Main risk factors for noncardia gastric cancer

Main risk factors for noncardia gastric cancer

Older age

Male sex

Racial or ethnic minorities

Immigrants from regions of high gastric cancer incidence

Helicobacter pylori infection

Family history of gastric cancer

Certain hereditary cancer syndromes

Smoking

Autoimmune gastritis

Dietary factors

Among individuals with GIM, factors associated with higher risk of incident gastric cancer*

Extensive GIM (i.e. involvement of corpus)

Incomplete histologic subtype of GIM

Family history of gastric cancer

Persistent H pylori infection

OLGA/OLGIM III or IV

^*The risk for gastric cancer in these individuals is above the baseline 0.16% annual risk of incident gastric cancer associated with GIM.⁶ Despite the lack of high-quality evidence, surveillance endoscopy should be considered in these individuals.³

We advocate for a shared decision-making process regarding GIM surveillance and discuss the following with a patient diagnosed with GIM: 1) what is GIM?; 2) what is his/her estimated risk of gastric cancer (high vs low) based on his/her known risk factors?; 3) the uncertain benefits of endoscopic surveillance vs no surveillance for decreasing gastric cancer-related mortality, and 4) the potential risks, cost, and inconvenience associated with endoscopic surveillance.

Endoscopic techniques to enhance detection of GIM

In the US, GIM is often diagnosed incidentally among individuals undergoing upper endoscopy for a variety of indications unrelated to gastric cancer screening. In such scenarios, it is imperative to adjudicate whether or not the index endoscopy was complete or of sufficient quality to adequately visualize the gastric mucosa. If not, it would be reasonable to consider a repeat exam to rule out suspicious lesions and to determine GIM extent.

We recommend that a careful inspection be performed using high-definition white light endoscopy (HD-WLE). The importance of applying general techniques to improve mucosal visibility, such as the use of defoaming and mucolytic agents and adequate air insufflation, as well as sufficient withdrawal time and photodocumentation, warrants particular emphasis.^8–10 HD-WLE has been shown to have high specificity (>90%) for GIM, although still with suboptimal sensitivity (53-75%) in prior studies.⁴ Conventional chromoendoscopy with application of dyes such as indigo carmine, methylene blue or acetic acid is associated with improved accuracy and sensitivity for GIM, but extends procedural times and can be cumbersome. Image-enhancing technologies such as narrow band imaging (NBI) offer a unique advantage in these regards. We routinely use a combination of HD-WLE and NBI for GIM surveillance exams based on multiple studies demonstrating improved GIM detection with adjunctive NBI use.^11,12 NBI-enhanced HD-WLE can be particularly useful when performing biopsies to determine GIM extent (see below) and allows for more “educated” biopsies, since GIM has a typical appearance on NBI that is characterized by bluish-white areas with an irregular mucosal pattern(Figure 1). Often, the blue-white lines appear on the crests of the epithelial surface in GIM, classically referred to as the “light blue crest” sign, which is more evident on magnification endoscopy. Profound irregular mucosal pattern and architectural distortion is concerning for dysplasia. A detailed discussion of image-enhancing techniques for augmented endoscopic detection of GIM, particularly given the limited or lack of access to these modalities in the US, is beyond the scope of this article; that said, we encourage interested practitioners to refer to the plethora of review articles and other illustrative resources on this topic.

Figure 1.

Endoscopic appearance of gastric intestinal metaplasia (GIM) on high-definition white light endoscopy (HD-WLE) (A1, B1, and C1, with arrows) and narrow band imaging (NBI) (A2, B2 and C2). A1 and A2 are near-focus images. On HD-WLE, the appearance of GIM ranges from mucosal texture irregularity to focal nodularity with various surface patterns, such as a reticular or glandular pattern. NBI allows enhanced evaluation of mucosal surface and vascular patterns of GIM. On NBI, GIM is characterized by bluish-white areas with an irregular mucosal pattern. The blue-white lines outlining the epithelial crests are referred to as the ‘light blue crest’ sign, which is typical of GIM and can be more evident on magnification endoscopy.

Biopsy protocol, histologic reporting, and documentation

Endoscopic GIM surveillance should achieve two main goals: 1) evaluate for and rule out neoplasia, and 2) determine GIM extent and histologic subtype. Extensive GIM, defined as any involvement of the corpus, is associated with at least 2-fold increased risk of incident gastric cancer compared to limited GIM, defined as only antral involvement (+/− incisura).⁷ If local expertise is available, we recommend pathologists perform histologic subtyping of GIM into complete vs incomplete, as the latter is associated with 3.3-fold higher risk of gastric cancer, based on limited but consistent data.¹ The Operative Link on Gastritis Assessment (OLGA) and Gastritis/Intestinal Metaplasia Assessment (OLGIM) staging classifications integrate both the severity (mild, moderate, severe) and extent (antrum/incisura, body) of atrophic gastritis and GIM and range from Stage I (low) to Stage IV (high). While the OLGA/OLGIM staging systems are useful for prognosis, with advanced stages (Stage III/IV) associated with higher risk of neoplastic progression¹³, they are not used routinely in the US. Regarding documentation, it is important to describe the appearance of the mucosa, including any abnormalities such as nodularity or any discrete lesions, as well as the location and number of biopsies, both targeted and non-targeted.

GIM is often patchy and might not be endoscopically discernible on HD-WLE alone. The optimal number of nontargeted biopsies that balances diagnostic yield with procedural time and cost has not been defined. We recommend using the updated Sydney protocol for nontargeted biopsies which requires five biopsies (two from the antrum at the lesser and greater curvature, two from the body at the lesser and greater curvature, and one from the incisura), preferably placed in separate jars (Figure 2).^14,15 If pathology cost dictated by the number of biopsy jars is a concern, we oftentimes place these non-targeted biopsies in two separate jars—that is, one jar labeled “body” and the other labeled “antrum/incisura”. Targeted biopsies using NBI should be obtained from suspicious mucosal lesions or concerning areas and placed in separately labeled jars. This approach has been shown to increase the yield of GIM detection.¹¹

Figure 2.

Gastric mucosal biopsies using the updated Sydney protocol. Biopsies from the five sites shown here are recommended and the specimens should ideally be placed in separate jars. A1, A2: greater and lesser curvatures of the distal antrum; A3: lesser curvature at the incisura angularis; B1, B2: lesser curvature and greater curvature of the gastric body. Targeted biopsies from any suspicious lesions should be obtained and placed in separate jars.

Conclusion

The importance of GIM as a premalignant diagnosis is increasingly recognized in the US, although there is still an unmet need to standardize endoscopic GIM surveillance practices. The magnitude and pace of progress is limited by the dearth and heterogeneity of the literature, particularly when generalizing findings from populations with different gastric cancer incidence and risk factor profiles. As more data are generated, endoscopic surveillance protocols will undergo iterative modifications and become increasingly personalized. Regardless, adjunctive risk reduction interventions, patient selection based on appropriate stratification by gastric cancer risk, and adequate mucosal visualization and assessment will remain foundational elements of GIM management and surveillance.