Asian Pacific Association of Gastroenterology task force recommendations on surveillance for associated gastric premalignant conditions.

Introduction
Gastric cancer remains the fifth most common malignancy and the fourth leading cause of cancer related death worldwide. According to the latest GLOBOCAN 2022 data,1 there were 968 784 new cases and 659 853 deaths globally. However, there is marked geographic variation in the gastric cancer burden, with >70% of new cases and deaths occurring in the Asia Pacific region. The highest incidence rates are reported in East Asia, particularly in China, Japan and Korea.2
Early stage gastric cancer often has no symptoms. Early detection and timely intervention of early stage gastric cancer can improve 5 year survival rates to >90%.3 With the advancement of gastrointestinal endoscopy, including high definition endoscopy and image enhanced endoscopy, endoscopy has emerged as the most direct and widely used method for early detection of gastric cancer, as well as enhanced detection and staging of gastric premalignant conditions.
Chronic atrophic gastritis (CAG), gastric intestinal metaplasia (GIM) and dysplasia (DYS) are well recognised premalignant conditions in Correa’s gastric carcinogenesis cascade, usually triggered by chronic Helicobacter pylori infection.4 5 In a recent meta-analysis, the global prevalence estimates of CAG, GIM and DYS were 25.4%, 16.2% and 2.0%, respectively.6 The prevalence rates were generally higher in countries with high or medium gastric cancer incidence, and with a higher prevalence among H pylori infected individuals. The pooled global estimates of the progression rate per 1000 person years were 2.09 (95% CI 1.46 to 2.99) for CAG, 2.89 (2.03 to 4.11) for GIM and 10.09 (5.23 to 19.49) for DYS, and progression rates seemed to be similar between countries with low and medium or high gastric cancer incidences.7
In recent years, both European and North American societies have issued guidelines and recommendations on the management of gastric premalignant conditions.810 The updated European guideline (MAPS III)11 recommends structured surveillance intervals based on histopathological risk stratification (eg, the operative link on gastritis assessment (OLGA) and the operative link on gastric intestinal metaplasia (OLGIM) system) and offers a detailed follow-up algorithm. In contrast, the American Gastroenterological Association (AGA)4 12 and the American College of Gastroenterology10 provide broader surveillance recommendations, emphasising individual risk assessment and shared decision making without a unified interval schedule. While some individual Asian countries have developed country specific guidelines,1315 a regional evidence based guideline of recommendations is lacking, despite the high incidence of gastric cancer and high prevalence of gastric premalignant conditions. Hence there is a pressing need for a unified regional recommendation on endoscopic surveillance and management of patients with gastric premalignant conditions or neoplastic lesions in the Asia Pacific region.
This task force was commissioned by the Asian Pacific Association of Gastroenterology (APAGE) to develop guidelines and recommendations on the surveillance of gastric premalignant conditions for the Asia Pacific region. The aims of the task force were to standardise the surveillance approach and management in the region as well as highlight areas for future research.

Methods

Data sources and search strategy
The task force was formed in January 2024. Working group members were invited from various member countries of APAGE, based on local gastric cancer incidences and individual expertise on gastric cancer screening and prevention. Members included 18 gastroenterologists, endoscopists, pathologists and epidemiologists from nine Asia Pacific countries and regions. Moreover, two international experts were invited from Europe (MD-R) and North America (SS) to comment and vote on the drafted statements.
A comprehensive literature search was conducted across multiple databases (PubMed, Cochrane Library and Embase) to identify relevant studies for each topic under consideration, particularly those from Asia. The search was supplemented by manual checking of reference lists from identified studies and relevant systematic reviews.
This consensus used the Grading of Recommendations Assessment, Development and Evaluation (GRADE) system as the standard for evaluating the evidence level and recommendation strength (online supplemental table S1).16 The strength of each recommendation was assessed based on several key factors, including the quality of evidence, alignment between the target population and the study population, extent of clinical benefit and cost considerations, among other factors. The recommendations are categorised into two levels: strong and conditional (weak).
A core group (WKL, YL, TLA, Y-CL and KSC) was first formed to develop the clinical questions pertinent to the focus of surveillance and management of gastric premalignant conditions and neoplastic lesions. Four subgroups were then formed to address four different areas: (1) epidemiology of gastric premalignant conditions and neoplastic lesions, (2) histopathological and endoscopic diagnosis, (3) risk stratification for endoscopic surveillance and (4) management of gastric premalignant conditions. Internal discussion among each subgroup was through online meetings and emails.
The modified Delphi process was adopted in developing the consensus statements. The first round of voting was performed anonymously via a web based platform in early November 2024. Voting was made on the strength of recommendation in five point Likert scales, from strongly agree, agree, neutral, disagree to strongly disagree. Consensus was achieved for all statements under a preset threshold of 80% agreement (strongly agree and agree). After the first round of voting, a hybrid meeting was conducted in November 2024 in Bali, Indonesia, during the Asia Pacific Digestive Week to provide feedback and discuss the statements. It was then followed by two more rounds of online voting in December 2024 and January 2025 after incorporation of suggestions and feedback from all members.
This consensus primarily focuses on gastric premalignant conditions that arise through the H pylori associated Correa cascade, namely CAG, GIM and DYS. Conditions arising from hereditary cancer syndromes, gastric neoplasms arising from H pylori infection naïve gastric mucosa,17 18 gastric polyps or autoimmune gastritis are not discussed in this consensus report. While we acknowledge that these conditions are important, they require distinct diagnostic, surveillance and management strategies, which are covered in other dedicated guidelines.1921 Specifically, the management of gastric polyps (eg, hyperplastic polyps) centred on resection rather than surveillance. Autoimmune gastritis has a distinct pathophysiology and risk profile, including risk of neuroendocrine tumours, which necessitates different management and surveillance strategies.

Consensus statements
A summary of all 28 consensus statements is provided in online supplemental table S2.

Part 1. Epidemiology

Statement 1
The term 'gastric premalignant conditions' refers to mucosal environments characterised by chronic inflammation, leading to specific mucosal changes associated with an increased risk of gastric cancer.

Quality of evidence: low

Strongly agree 80%, agree 20%

The Correa gastric carcinogenesis cascade outlines a sequential process beginning with chronic non-atrophic gastritis, progressing through CAG, GIM and DYS, and culminating in gastric adenocarcinoma.22 Gastric premalignant conditions represent an extensive pathological state predisposing the tissue environment to carcinogenesis, incrementally elevating the long term risk of gastric cancer development without necessarily implying an immediate or direct progression to malignancy. Specifically, CAG is characterised by the loss of gastric glandular cells and their replacement with fibrosis or intestinal-type cells. GIM refers to the transformation of gastric mucosal epithelial cells into structures resembling those found in the small intestine (complete type) or colon (incomplete type), marked by the appearance of goblet and absorptive cells.23 The extensive distribution of GIM in the stomach, such as involvement of the fundus and body of the stomach, also carries a higher risk of malignant transformation than localised GIM.24 The presence of incomplete GIM indicates a further increased risk of malignancy.12 25 However, it remains controversial whether adenocarcinoma mainly develops from GIM.25 26 Recent studies have provided evidence that certain subsets of GIM may act as direct precursors of gastric cancer.2729 For DYS, a detailed histopathological diagnosis is provided in statement 16.
In this report, we would also like to highlight the practical distinction between 'gastric premalignant conditions' and 'neoplastic lesions'. Neoplastic lesions usually refer to endoscopically visible and discrete lesions, which are either dysplastic lesions (that can progress to cancer) or early gastric cancer (EGC), both of which require endoscopic or surgical intervention. On the other hand, premalignant conditions typically require monitoring and surveillance only.

Statement 2

Individuals with advanced chronic atrophic gastritis (OLGA III/IV), advanced gastric intestinal metaplasia (OLGIM III/IV or incomplete-type) and gastric dysplasia are at increased risk for the development of gastric cancer.

Quality of evidence: high

Strongly agree 85%, agree 15%

Overall, the incidence of gastric cancer in patients with CAG and GIM is relatively low. A previous study reported a pooled gastric cancer incidence rate of 1.24 per 1000 person years for CAG (95% CI 0.80 to 1.76) and 3.38 per 1000 person years for GIM (95% CI 2.13 to 4.85).30 In a 12 year follow-up study, Rugge et al showed that high risk CAG (OLGA stages III–IV) was consistently associated with invasive or intra-epithelial gastric neoplasia, whereas mild atrophy (OLGA stages I–II) was associated with a lower risk.31 Additionally, a prospective, multicentre cohort study from Singapore showed that participants with OLGIM stages III–IV were at the highest risk (adjusted hazard ratio (HR) 20.7; 95% CI 5.04 to 85.60) of developing early gastric neoplasia.32 Moreover, a meta-analysis by Akbari et al demonstrated that incomplete GIM (6.60, 95% CI 2.73 to 11.75) was associated with a higher estimated gastric cancer incidence compared with complete GIM (3.24, 95% CI 0.12 to 9.01).30 The details of OLGA and OLGIM staging are discussed in subsequent statements.

Statement 3

The prevalence of chronic atrophic gastritis and gastric intestinal metaplasia in the Asia Pacific region varies across different Asian countries and populations, with chronic atrophic gastritis more common than gastric intestinal metaplasia.

Quality of evidence: moderate

Strongly agree 60%, agree 35%, neutral 5%

The prevalence of CAG and GIM varies widely across Asia, which is attributed to differences in population characteristics, H pylori prevalence, local risk factors and even study methodologies. Most existing data are from small scale, hospital based or regional studies that used heterogeneous diagnostic criteria and age groups. Due to these uncertainties, it is challenging to provide a single Asia Pacific prevalence rate. Reported rates in the literature typically range from 10% to 30% for CAG and from 8% to 25% for GIM,33 34 but these figures should be interpreted with caution due to potential sampling bias and variability. In general, Southern Asia has a higher CAG prevalence and Eastern Asia has a higher GIM prevalence.35 Data from Oceania countries remain sparse.

Statement 4

The prevalence of chronic atrophic gastritis and gastric intestinal metaplasia in Asia is declining overall. However, the rate of decline varies, due to differences in socioeconomic conditions and access to healthcare.

Quality of evidence: moderate

Strongly agree 65%, agree 25%, neutral 10%

The prevalence of CAG and GIM in the Asia Pacific region has declined in recent decades, likely due to improved public health, socioeconomic development and reduced H pylori infection in younger birth cohorts. A pooled analysis showed that the prevalence of GIM fell from 30% (1991–2000) to 21.1% (2011–22), with little change in CAG.35 This decline was more evident in high income countries with better healthcare systems, widespread screening and national screening programmes (eg, Japan and South Korea). However, higher endoscopy rates may also lead to increased detection.36 37 In contrast, data from low resource countries remain limited38, and the true prevalence may be underestimated due to various diagnostic constraints.

Statement 5

There is usually a strong positive correlation between Helicobacter pylori infection and gastric premalignant conditions in Asia; however, there are exceptions in some regions which may be modulated by host, bacterial and environmental factors.

Quality of evidence: high

Strongly agree 65%, agree 30%, neutral 5%

While H pylori infection has been classified as a group 1 carcinogen, it is usually closely correlated with CAG and GIM, with a previous meta-analysis showing an OR of 2.16 (95% CI 2.09 to 2.22) for CAG and 1.64 (95% CI 1.57 to 1.72) for GIM in Asia.35 East Asian countries, such as China, Japan and South Korea, have a higher prevalence of CAG/GIM than other Asian countries.35 38
However, there are also 'Asian enigma'39 and 'Indian enigma',40 where regions with a high H pylori prevalence have low gastric cancer rates, suggesting that other factors, such as difference in H pylori virulent strains, host genetic polymorphisms and dietary habits may mitigate the risk of progression to malignancy.41

Statement 6

Apart from Helicobacter pylori infection, dietary habits, smoking, genetic factors and lower socioeconomic status are all associated with the development of gastric premalignant conditions

Quality of evidence: moderate

Strongly agree 40%, agree 55%, neutral 5%

Several environmental and genetic factors beyond H pylori infection could also influence the development of gastric premalignant conditions and neoplastic lesions. The International Agency for Research on Cancer (IARC) of WHO has classified diet related factors, including high salt intake, pickled and preserved food, and smoked and grilled foods, as group 1 carcinogens for gastric cancer.42 43 Despite the lack of definitive guidelines or consensus regarding the impact of dietary factors on the development of CAG or GIM, a growing body of evidence suggests that poor dietary habits in some Asian countries are significant risk factors for gastric premalignant conditions.43 An endoscopy based study in South Korea showed that high levels of salt intake were associated with CAG and GIM (OR 2.87, 95% CI 1.34 to 6.14).44 Other dietary risk factors, such as deficiencies in fresh fruits and vegetables, are associated with increased oxidative stress and diminished antioxidant defence mechanisms (eg, flavonoids), resulting in chronic inflammation and damage to the gastric mucosa and causing premalignant conditions and neoplastic lesions.44
Cigarette smoke is rich in free radicals and reactive oxygen species, which penetrate the gastric epithelium and induce oxidative stress. Additionally, nicotine present in cigarette smoke induces vasoconstriction, leading to reduced blood flow to the gastric mucosa, impaired nutrient delivery and diminished mucosal healing capacity.45 Genetic predispositions also has a role, with polymorphisms in genes associated with inflammatory and immune responses (such as tumour necrosis factor (TNF) α, interleukin 22 (IL-22)46 47 and IL-1B48) correlating with heightened susceptibility to premalignant conditions. The prevalence of CAG and GIM decreased significantly with higher income or college education levels in a US population based study,49 and individuals with lower educational attainment or socioeconomic status were at a higher risk of developing gastric premalignant conditions and gastric cancer in the Chinese population.50

Part 2. Endoscopic and histopathological diagnosis

Statement 7

APAGE recommends implementing measures to ensure optimal mucosal visualisation and adequate inspection time, along with adequate photographic documentation, to enhance the quality of endoscopic examination of the stomach.

Quality of evidence: Moderate

Strength of recommendation: Strong

Strongly agree 85%, agree 15%

The endoscopic appearance of gastric premalignant conditions, particularly CAG and GIM, can be inconspicuous, requiring a systematic and high quality examination to maximise diagnostic yield 51 52 (figure 1). Endoscopists should ensure excellent mucosal visualisation by aspirating luminal contents and adequately insufflating the stomach to expand mucosal folds. Pre-procedural consumption or irrigation with mucolytic agents like simethicone, 1% N-acetylcysteine and proteolytic enzymes (pronase) may improve mucosal visualisation versus water irrigation alone.5355 However, theoretical concerns regarding the risk of intraprocedural aspiration may limit the routine adoption of this practice.56
Another crucial aspect of a high quality endoscopic examination is to ensure an adequate inspection time. A longer inspection time has been strongly associated with increased detection of gastric lesions.5759 In a retrospective cohort study involving 837 Singaporean patients, inspection time >7 min was associated with increased detection of CAG and GIM, with an OR of 2.5 (95% CI 1.52 to 4.12).57 In another Japanese retrospective cohort of 55 786 patients, inspection time >7 min was associated with increased detection of gastric neoplastic lesions. Compared with that in the fast group, OR for the neoplastic lesion detection rate in the moderate and slow groups were 1.90 (95% CI 1.06 to 3.40) and 1.89 (95% CI 0.98 to 3.64), respectively.59 Both the AGA and European Society of Gastrointestinal Endoscopy (ESGE) have recommended that the whole inspection time for upper endoscopy should be at least 7 min.56 60 Furthermore, we advocate for standardised training programmes to ensure that endoscopists are adequately equipped to visually identify gastric premalignant conditions or neoplastic lesions.

Statement 8

APAGE recommends using high definition white light endoscopy to systematically evaluate the whole gastric mucosa and avoid blind spots

Quality of evidence: low

Strength of recommendation: strong

Strongly agree 70%, agree 30%

A systematic review reported that the rate of missing gastric cancer during upper endoscopy was 9.4% (95% CI 5.7 to 13.1%). Missed cancers were mainly located in the gastric body and were presumably due to blind spots that were not examined or lesions that were not recognised despite being in the field of examination.61 A systematic approach to ensure complete examination of the stomach has been proposed.62 The ESGE recommends precise photographic documentation that encompasses at least one illustrative image of the following anatomical landmarks: duodenum, major papilla, antrum, incisura, corpus, retroflexed view of the fundus, diaphragmatic indentation, upper aspect of the gastric folds, squamocolumnar junction, and distal and proximal oesophagus.60 The AGA and World Endoscopy Organisation (WEO) have similar recommendations.56 63
There is indirect evidence from a multicentre study using an artificial intelligence (AI) system to monitor photographic documentation that reduces the blind spots during upper endoscopy examination.64 An AI assisted upper GI endoscopic system has been launched commercially and, if available, may be used to complement the systematic examination. Early data concerning the use of AI assistance for gastric cancer diagnosis are also promising.65 66
Another study showed that training an endoscopist to perform standardised stomach examination improved the detection rate of gastric neoplasm from 0.2% to 2.3%.67 Additionally, the impact of dedicated teaching on the optimal technique of examination and recognition of gastric lesions using video based platforms has been demonstrated in randomised controlled trials (RCTs) to improve the endoscopic diagnosis of EGC.68

Statement 9

APAGE recommends use of image enhanced endoscopy for the diagnosis of gastric premalignant conditions and neoplastic lesions.

Quality of evidence: moderate

Strength of recommendation: Strong

Strongly agree 70%, agree 30%

Unlike statement 8 that refers to systematic examination of the stomach to detect lesions, this statement refers to the evaluation of detected pathologies. High definition white light endoscopy (WLE) could be complemented by image enhanced endoscopy to improve visualisation of mucosal details (figure 2a–2c). A cross sectional study from Greece showed that high definition WLE had a sensitivity of 75% and specificity of 94% for detection of GIM.69 However, high definition WLE alone seems to be inferior to image enhanced endoscopy. A meta-analysis including 10 studies confirmed that narrow band imaging (NBI) increased the detection of GIM by 32% compared with WLE.70 Blue light imaging (BLI, Fujifilm, Tokyo, Japan) is another form of narrow band width imaging that is similar to NBI.71 There are also data suggesting that newer technologies based on white light, such as linked colour imaging (LCI; Fujifilm) and texture and colour enhancement imaging (TXI; Olympus, Japan) may improve visibility of gastric lesions,7274 not limited to GIM. LCI has been shown to increase the detection rate of gastric neoplastic lesions in RCTs.72 73 75 A prospective RCT in Singapore involving 90 patients showed that LCI significantly increased the absolute detection rate of CAG by 1.6% and GIM by 12.1% compared with WLE.76 Another multicentre prospective RCT of 1924 patients from China also showed that LCI had better accuracy (78.8% vs 68.4%) and shorter examination time in diagnosing EGC than WLE.75 Some studies also suggested that the use of magnifying endoscopy can predict the presence of CAG and GIM with an accuracy of up to 90%.77 Nevertheless, it requires a special magnifying endoscope and training, which may not be available in most centres.78 79

Statement 10

APAGE recommends that the extent of chronic atrophic gastritis and gastric intestinal metaplasia detected during endoscopy be objectively documented.

Quality of evidence: moderate

Strength of recommendation: Strong

Strongly agree 70%, agree 30%

The risk of development of gastric cancer depends on both the severity and extent of CAG and GIM. The estimated 5 year incidence for gastric cancer was 0.7% for mild CAG but up to 10% for severe CAG.80 For GIM, the risk was reported as 5.3% for antral GIM but 9.8% for GIM involving the antrum and corpus.81 Hence the extent of CAG and GIM detected during endoscopy should be objectively documented. The OLGA and OLGIM scoring systems were developed as histopathological reporting frameworks for CAG and GIM that could be easily understood and prognosticate the risk of developing gastric cancer. They are based on a systematic biopsy protocol for documentation of disease extent and use a visual analogue scale for severity assessment (more details in statement 15).
Three endoscopic classification systems have been validated against the gold standard of histology, including the Kimura–Takemoto classification of CAG,82 the Kyoto classification system83 and the endoscopic grading of gastric intestinal metaplasia (EGGIM) (table 1 and figure 2A–2D).84 In particular, the Kyoto classification and EGGIM have been validated against OLGA and OLGIM systems.85 EGGIM is based on image enhanced endoscopy, whereas the other two are based on WLE. A systematic review reported that a Kyoto classification score of ≥4 might indicate an increased risk of developing gastric cancer (OR 7.3, 95% CI 3.6 to 14.7).86 EGGIM can be used to assess a patient’s risk by the endoscopic assessment of GIM in the antrum incisura, and in the corpus with the use of high definition endoscopes with NBI (or by extension, BLI) function. For the diagnosis of OLGIM stages III–IV, sensitivity, specificity and positive likelihood ratio were 9%, 95% and 16.5, respectively, by using a cut-off value of >4.84 In a multicentre Japanese observational study, OLGIM stages III–IV (OR 2.8, 95% CI 1.5 to 5.3), high EGGIM score (OR 1.8, 95% CI 1.0 to 3.1) and open-type Kimura–Takemoto (OR 2.5, 95% CI 1.4 to 4.5) had significant associations with gastric cancer risk.85 Apart from NBI, recent data have also supported the use of an LCI based EGGIM system,87 88 which may have some advantages (eg, reducing the number of biopsy pieces) over NBI based systems for risk stratification. However, it must be acknowledged that these endoscopic scoring systems may not work as well in non-expert hands and there are a lack of data regarding validation in real world practice.

Statement 11

APAGE recommends that if biopsies are to be considered for histopathological diagnosis of chronic atrophic gastritis and gastric intestinal metaplasia, biopsies from the antrum/incisura and corpus should be obtained separately following a systematic protocol to allow histopathological staging, while targeted biopsies should be obtained separately for lesions suspicious for advanced pathology.

Quality of evidence: low

Strength of recommendation: strong

Strongly agree 85%, agree 15%

We acknowledge that there are differences in practice within Asia, as well as differences in expertise and resources available, in terms of obtaining biopsies for histology. It must be remembered that although the data on EGGIM are promising, there remains a need for further real world validation. Under most circumstances, there is still a need for histopathological assessment and confirmation of premalignant conditions, particularly in the absence of endoscopic expertise or diagnostic certainty.89 90 In addition, objective histopathological confirmation of focal DYS or EGC is required as part of pretreatment confirmation.90 The Updated Sydney System facilitates a consistent approach to map out the extent of CAG and GIM.91 A minimum of five biopsies should be obtained with samples from antrum/incisura and corpus placed in separately labelled containers. Any suspicious areas should also be described and biopsied separately.91 92 Incisura biopsy is suggested because it has been shown to increase the number of cases diagnosed in more advanced stages of CAG and GIM,93 94 although admittedly this observation is not universal.95 A real world retrospective study demonstrated that routine adherence to a systematic protocol increased the detection rates of premalignant conditions, including CAG and GIM.96 Such data have prompted published international guidelines and reviews to consistently advocate for such a systematic approach.4 8 9 97

Statement 12

APAGE recommends that the presence of chronic atrophic gastritis and gastric intestinal metaplasia should prompt a diligent examination for the concomitant presence of neoplastic lesions.

Quality of evidence: moderate

Strength of recommendation: strong

Strongly agree 85%, agree 15%

While CAG and GIM are associated with an increased risk of gastric neoplasia,97 the presence of extensive GIM may partially obscure focal neoplastic lesions. It is important to be cognisant that at the index endoscopy, there may already be concomitant gastric neoplasia (including DYS and gastric cancer) and hence a careful endoscopic examination of the gastric mucosa is crucial.

Statement 13

The presence of any suspicious focal neoplastic lesion on white light endoscopy should be described using the updated Paris classification and further characterised with high definition, image enhanced endoscopy, preferably with magnifying/near focus observation and targeted biopsies obtained for histopathological confirmation.

Quality of evidence: moderate

Strength of recommendation: Strong

Strongly agree 45%, agree 45%, neutral 10%

The updated Paris classification serves as an important means of documenting the morphology of detected focal neoplasia in a consistent manner.98 This will facilitate communication between healthcare providers and, importantly, it also provides an estimate of the risk of submucosal invasion of cancer.98 High resolution image enhanced endoscopy has been shown to be useful in characterising focal lesions and differentiating between EGC and non-cancer (figure 2E, 2F).
For EGC, the diagnostic criteria for the vessel plus surface classification system using magnifying endoscopy with NBI (ME-NBI) are irregular microvascular patterns with a demarcation line or irregular microsurface pattern with a demarcation line.99 A meta-analysis that examined 20 studies involving 7770 patients and 7917 lesions reported a pooled sensitivity for ME-NBI for diagnosing EGC of 0.86 (95% CI 0.80 to 0.90) and specificity of 0.92 (95% CI 0.86 to 0.96).100 The use of BLI has demonstrated comparable performance with NBI.101 102 The overall diagnostic performance for EGC using near focus NBI (NF-NBI) was an accuracy of 87.7%, sensitivity 60.7%, specificity 93.1%, positive predictive value 63.8% and negative predictive value 92.2%.103 A recent phase II RCT compared the third generation NBI with the newer TXI, and high definition WLE. Gastric neoplasia detection rates in the NBI, TXI and WLE groups were 7.3%, 5.0% and 5.6%, respectively.104 The Magnifying Endoscopy Simple Diagnostic Algorithm for Gastric Cancer (MESDA-G) was jointly devised by the Japan Gastroenterological Endoscopy Society, the Japanese Gastric Cancer Association and the WEO as a unified international algorithm for magnifying endoscopy diagnosis of EGC.105 After a focal lesion is detected by WLE, the presence of a demarcation line is first assessed. The lesion is then examined in greater detail with narrow band width imaging, combined with magnification.
It is, however, important to note that many of these diagnostic criteria were developed and validated based on the Japanese classification system, where lesions pathologically defined as high grade dysplasia in the west are typically classified as mucosal adenocarcinoma, regardless of invasion. This distinction and its implications for practice in the wider Asia Pacific region are discussed further in statement 16.

Statement 14

APAGE recommends that pathology reports standardise the description of intestinal metaplasia: to be quantified for each site as mild, moderate and severe. The presence of gastric atrophy in oxyntic mucosa should be assessed where feasible.

Quality of evidence: low

Strength of recommendation: Strong

Strongly agree 45%, agree 45%, neutral 10%

The presence of CAG and GIM (including complete or incomplete subtypes) are important and should be reported in all pathology reports of gastric biopsies. Quantifying the severity is also of prognostic relevance as the extent and severity of the findings have an impact on the risk of developing gastric neoplasia.11 By convention, the grading of severity for CAG and GIM is defined as none (0%), mild (1–30%), moderate (31–60%) and severe (>60% involvement of the biopsy specimen).106 Subclassification of GIM into complete and incomplete is encouraged when feasible, given its higher association with gastric cancer risk.24 Pseudopyloric metaplasia, which marks chronic injury and a reparative response in the stomach, typically arises in the oxyntic mucosa of the gastric body,107 may also be present and should be documented when identified. It must be noted that while the interobserver agreement for GIM is high, it is more variable and can be low for CAG.91 108 109

Statement 15

APAGE recommends that if gastric biopsies are obtained by a mapping protocol, histopathological staging by OLGA/OLGIM can be used to stratify the risk for gastric cancer progression.

Quality of evidence: moderate

Strength of recommendation: Strong

Strongly agree 40%, agree 55%, neutral 5%

The OLGA system was developed to establish a reporting framework for gastric atrophy that is easily understood by clinicians and provides prognostic information.92 The OLGA system uses the biopsy sampling protocol and visual analogue scale recommended in the updated Sydney system.91 Atrophy severity for antrum/incisura and corpus is scored as 0 (no atrophy), 1 (mild), 2 (moderate) or 3 (severe), within the corresponding anatomic locations (antrum/incisura and corpus). The overall antrum/incisura score and overall corpus score are then combined to provide the OLGA stage (stages I–IV), with stages III and IV shown to be predictive for the development of gastric neoplastic lesions.92
Similarly, the OLGIM system was developed to evaluate the severity and extent of GIM on gastric mapping biopsies from the antrum/incisura and corpus in a similar fashion. Stages III and IV were found to be predictive for gastric neoplastic lesions.110 The prognostic values of both OLGA31 111 112 and OLGIM32 have since been validated in longitudinal studies and a meta-analysis. A Singapore study (n=2980) showed increased early gastric neoplasia risk in patients with GIM (adjusted HR 5.36; 95% CI 1.51 to 19.0), particularly in OLGIM stages III–IV (HR 20.7; 95% CI 5.04 to 85.6).32 In a Japanese multicentre observational study, OLGIM stages III–IV were also associated with a higher cancer risk (OR 2.8; 95% CI 1.5 to 5.3).85 OLGIM is favoured over OLGA due to better interobserver agreement9 110 and is endorsed by ESGE (MAPS II–III), Academy of Medicine, Singapore (AMS) and British Society of Gastroenterology.8 9 11 13
However, we recognise that OLGA/OLGIM may not be feasible in all Asian countries due to lack of availability of expertise or resources, or a predominant focus on endoscopic diagnosis. Hence this statement does not mandate histology for diagnosis, but rather recommends that if histopathological assessment is performed, it should follow validated systems, such as OLGA/OLGIM.

Statement 16

APAGE recommends that dysplasia be graded on a two tiered system as low grade and high grade dysplasia, as recommended in the latest WHO classification system. However, APAGE also accepts the Japanese Gastric Cancer Association classification system for the classification of high-grade dysplasia in appropriate settings.

Quality of evidence: low

Strength of recommendation: strong

Strongly agree 40%, agree 55%, neutral 5%

According to the WHO classification system, gastric DYS is defined as an unequivocal neoplastic lesion without evidence of stromal invasion. Currently, DYS is graded dichotomously as low or high grade.113 In a large cohort study from the Netherlands with 92 250 patients, de Vries et al reported that the annual incidence of gastric cancer was 0.1% for patients with CAG, 0.25% for GIM, 0.6% for low grade dysplasia (LGD) and 6% for high grade dysplasia (HGD) within 5 years after diagnosis.114 Notably, within 1, 5 and 10 years of follow-up after initial diagnosis, gastric cancer was diagnosed in 0.3%, 0.6% and 0.8% of patients with CAG; 0.7%, 1.2% and 1.8% of patients with GIM; 2.1%, 3.1% and 3.9% of patients with LGD; and 24.9%, 29.5% and 32.7% of patients with HGD.114 It is worth noting that the study did not exclude gastric cancers diagnosed within the first year of identifying CAG, GIM or DYS. Thus the percentages reported in the study included some prevalent gastric cancers in addition to incident cases.
Notably, there are differences in the diagnostic criteria of superficial gastric neoplastic lesions between the Japanese and western pathologists. Lesions diagnosed as HGD by western pathologists are considered adenocarcinoma by the Japanese Gastric Cancer Association classification based on nuclear and structural atypia, regardless of the presence of invasion.115 116 The pathological criteria adopted by Japan (and South Korea) allow them to correlate better with endoscopically identifiable neoplastic lesions. However, most Asia Pacific countries adhere to their own or international guidelines (such as the WHO criteria) for pathological reporting standards which have implications for training, laboratory accreditation and oncological management. For instance, a single centre observational study from China reported their long term experience with a three tier classification system for gastric dysplasia (mild, moderate or severe), which was based on architectural distortion, nuclear atypia and glandular arrangement. This histopathological framework was also shown to correlate with increasing risk of malignant transformation across the severity spectrum.117 While differences exist, both lamina propria and muscularis mucosae invasion are diagnosed as invasive adenocarcinoma and staged pathologically as pT1a cancer. Irrespective of the differences in diagnostic categories, the approach to treatment is the same, in that resection is indicated and endoscopic resection is preferred when feasible.117

Statement 17

In the absence of formal OLGA and OLGIM staging, advanced stage chronic atrophic gastritis/gastric intestinal metaplasia may be indicated by the presence of marked and extensive chronic atrophic gastritis/gastric intestinal metaplasia in multiple biopsies, or incomplete subtype gastric intestinal metaplasia.

Quality of evidence: moderate

Strength of recommendation: Strong

Strongly agree 70%, agree 25%, neutral 5%

Although OLGA and OLGIM staging systems are preferred for comprehensive risk assessment,92 110 routine implementation is limited by the need for multiple biopsies and expert pathology. Other easily assessable risk factors include the extent/severity of CAG and GIM, incomplete GIM subtype, smoking (>20 pack years), persistent H pylori infection and a family history of gastric cancer.118 A systematic review of 12 cohort studies (four in Asia) found that incomplete GIM carried a significantly higher cancer risk than complete GIM (relative risk (RR) 5.16; 95% CI 3.28 to 8.12).119 However, this subtype based classification does not reflect the extent/severity of GIM.
As an alternative, the endoscopic grading system EGGIM evaluates GIM across five gastric regions using NBI. Studies from Portugal, Italy, Japan and China showed that high EGGIM scores correlated with increased gastric cancer risk,84 85 87 120 and LCI-EGGIM may offer an advantage for detection.87 However, the routine application of EGGIM by less experienced endoscopists requires further evaluation.

Part 3. Endoscopic surveillance

Statement 18

Endoscopic surveillance is recommended for patients with moderate to advanced stages of chronic atrophic gastritis and gastric intestinal metaplasia, as well as patients with non-advanced chronic atrophic gastritis and gastric intestinal metaplasia who have additional risk factors for gastric cancer.

Quality of evidence: low

Strength of recommendation: weak (Conditional)

Strongly agree 75%, agree 25%

Endoscopic screening and surveillance are widely implemented in Asia due to high H pylori prevalence and elevated gastric cancer risk. Although non-invasive tests (eg, pepsinogen) are preferred for risk identification, they lack accuracy in predicting EGC, particularly after H pylori eradication, as irreversible mucosal damage may persist and affect pepsinogen reliability.121124 Direct endoscopic and histopathologic evaluations remain the most reliable for risk stratification. Surveillance is not routinely recommended for non-advanced CAG/GIM unless risk factors exist, such as poor baseline endoscopy, first degree family history, smoking, persistent H pylori or incomplete GIM (figure 3).12 13 125 Despite the lack of RCTs comparing surveillance versus no surveillance, simulation studies suggest surveillance is cost effective in high risk populations.126 Even in US populations, 5 year surveillance for incidentally detected GIM reduced gastric cancer incidence/mortality and was cost effective.127 Nonetheless, due to assumptions and variability in parameter inputs within cost effectiveness simulation models, the strength of the recommendation may vary by region.

Statement 19

For patients with advanced stage chronic atrophic gastritis and gastric intestinal metaplasia (OLGA/OLGIM III/IV), endoscopic surveillance every 2 years is recommended

Quality of evidence: low

Strength of recommendation: weak (Conditional)

Strongly agree 55%, agree 40%, neutral 5%

The primary goal of endoscopic surveillance is to detect EGC or DYS at a curable stage. In Japan and South Korea, national policies recommend upper endoscopic screening starting at age 40–50 years, regardless of H pylori status or premalignant conditions.128 129 The screening interval is generally based on an estimated 2 year dwelling time, which reflects the window of opportunity to detect EGC before it progresses to an advanced stage.129 This dwelling time, derived from the prevalence-to-incidence ratio, is consistent with observations from routine clinical practice in Asia. Moreover, geographic location does not appear to influence the progression rate from advanced premalignant conditions to gastric cancer, or from EGC to advanced disease.
Patients with OLGA/OLGIM stages III–IV are at increased risk of gastric cancer, with similar progression rates reported globally.101 Although RCTs are lacking, we recommend a 2 year surveillance interval for these high risk individuals, primarily based on cost effectiveness analyses and in alignment with policies implemented in high incidence Asian countries.13 15 129 This interval, which is more intensive than the 3 year interval recommended by other international guidelines for endoscopic surveillance,811 is intended to balance the increased likelihood of early detection with resource efficiency. In South Korea, universal biennial endoscopic screening was found to be cost effective, partly due to lower procedure costs.130 In Japan, biennial surveillance following H pylori eradication was also cost effective, taking into account the residual risk of gastric cancer.131 In the US, a cost effectiveness analysis stratified by race and ethnicity found that endoscopy with mapping biopsies and biennial surveillance for advanced premalignant conditions remained cost effective among Asians.132 However, we acknowledge that if future evidence shows a longer interval provides comparable effectiveness in reducing gastric cancer mortality, the surveillance interval could be lengthened to improve cost-efficiency.

Statement 20

For patients with intermediate stage chronic atrophic gastritis and gastric intestinal metaplasia (OLGA/OLGIM stage II), endoscopic surveillance every 3 years is suggested. Different surveillance intervals may be considered depending on the population risk and available resources.

Quality of evidence: low

Strength of recommendation: weak (Conditional)

Strongly agree 65%, agree 30%, neutral 5%

Evidence regarding the increased risk of gastric cancer in individuals with intermediate stage CAG and GIM (OLGA/OLGIM stage II) is limited.8 9 12 For patients in this category, surveillance every 3 years may be considered, but supporting evidence remains insufficient, and risk assessment should be individualised, particularly when OLGA/OLGIM is not available (figure 3). A cohort study from Singapore which involved mainly Chinese individuals aged ≥50 years demonstrated a stepwise increase in the risk of early gastric neoplasia (ie, HGD and gastric adenocarcinoma), with incidence rates of 12.5 per 1000 person years in the no GIM group, 21.5 in OLGIM I, 108.8 in OLGIM II and 543.8 in OLGIM III–IV. Notably, the intermediate category (OLGIM II) accounted for approximately a quarter of subsequent cases of early gastric neoplasia.32 These results implied that patients with intermediate stage CAG and GIM within a high risk population may benefit from endoscopic surveillance after considering individual factors, such as poor quality baseline endoscopy, first degree family history of gastric cancer, smoking, persistent H pylori infection or the presence of incomplete-type GIM. However, it may not apply to regions with lower cancer incidences. A long term follow-up study in Italy found an incidence of gastric cancer of 1.48 (95% CI 0.48 to 4.58) per 1000 person years for OLGA II patients.133 It is important to note that this recommendation also may not apply to patients with autoimmune gastritis, that are at least OLGA/OLGIM stage II134 and could have distinct malignant potential, including increased risk of neuroendocrine tumours, but is outside the scope of our recommendation.

Statement 21

For patients with mild chronic atrophic gastritis and gastric intestinal metaplasia (OLGA/OLGIM stage I), endoscopic surveillance may not be justified but can be considered based on individual and population risks, as well as resources available.

Quality of evidence: low

Strength of recommendation: weak (Conditional)

Strongly agree 60%, agree 40%

In patients with mild CAG or GIM (OLGA/OLGIM stage I), routine surveillance is generally not recommended unless additional risk factors are present. This is primarily because mild CAG or GIM is commonly observed, the risk of developing gastric cancer is low and routine endoscopic surveillance is not cost effective in most countries unless the population’s gastric cancer risk is high6 or the individual has an elevated risk (eg, family history of gastric cancer). Additionally, the progression rate is relatively slower, and risk factors associated with histopathological progression,32 such as persistent H pylori infection and lifestyle risk factors,124 135 can often be treated or modified after the endoscopic examination through interventions like H pylori eradication and lifestyle changes.121 136 137 However, this should not be interpreted as there being no gastric cancer risk for these patients.138 Histopathological grading can also be influenced by factors such as biopsy location, limited number of samples and the patchy distribution of premalignant conditions, as well as variability in histopathological interpretations.90 101 With many uncertainties, clinical decisions should therefore consider both individual and population level risks, as well as the available resources.

Part 4. Management

Statement 22A

APAGE recommends endoscopic resection of visible lesions concerning for, or histopathologically confirmed to be, dysplasia, with the goal being en bloc R0 resection.

Quality of evidence: low

Strength of recommendation: Strong

Strongly agree 80%, agree 20%

Statement 22B

In cases where the biopsy report shows dysplasia but no focal lesions are identified on the repeat high quality endoscopy (so-called non-visible dysplasia), referral to an endoscopist with experience of endoscopic detection and intervention is recommended. Surveillance endoscopy should be performed every 6 months for patients with resected high grade dysplasia and annually for low grade dysplasia, for a minimum duration of 5 years, with relaxed intervals but ongoing surveillance thereafter.

Quality of Evidence: low

Strength of recommendation: weak (Conditional)

Strongly agree 80%, agree 20%

DYS is clinically significant, being associated with synchronous gastric cancer in nearly 10% of cases and an increased risk of progression to gastric cancer.139141 For visible focal gastric lesions with DYS, endoscopic en bloc resection is recommended, as it also allows for proper histopathological assessment. Therefore, en bloc resection is necessary even for LGD to ensure thorough evaluation and management. For lesions ≤15 mm, cap assisted endoscopic mucosal resection is generally effective and straightforward, while for lesions >15 mm, endoscopic submucosal dissection is preferred due to its superior en bloc resection and lower recurrence rates, despite its greater complexity and higher risks related to the procedure, such as bleeding and perforation.142144 Specifically for endoscopic submucosal dissection, the MAPS III guidelines recommend endoscopic submucosal dissection for differentiated lesions staged as dysplastic or intramucosal carcinoma (of any size if not ulcerated and <30 mm if ulcerated),11 which follows the latest Japanese recommendation.145
In cases where HGD or LGD is detected incidentally from random biopsies, but no focal lesions are found on repeat high quality endoscopy, ongoing surveillance is needed.146 A single centre study from China showed that during a median follow-up of 4.2 years, 51.4% of patients with baseline mild-to-moderate DYS showed no detectable DYS.147 Within 5 years of diagnosis, the annual incidence rate of advanced neoplasia and gastric cancer was 0.43% and 0.26%, respectively, which was much lower than previous reports.7 While the results of this single centre study should be interpreted with caution, it also highlights the role of repeated examination, preferably by expert endoscopists and pathologists. Advanced endoscopic techniques, such as image enhanced endoscopy, chromoendoscopy or AI assistance, have been shown to improve detection rates, revealing lesions missed in the initial examinations.64 148 Undetected and unresected DYS is associated with a substantially higher risk of progression to gastric cancer, highlighting the need for careful follow-up.114 Specifically, patients with resected HGD should undergo surveillance endoscopy every 6 months, while those with resected low grade DYS could be monitored annually. This frequent surveillance should continue for at least 5 years.147 In patients with indefinite for dysplasia, where the extent of cellular abnormality is unclear or insufficient to classify as LGD or HGD, a similar strategy of close surveillance and repeated histopathological assessment is recommended.8 9

Statement 23

APAGE recommends test-and-treat for Helicobacter pylori in individuals with gastric premalignant conditions to reduce the risk of histopathological progression and gastric cancer

Quality of evidence: moderate

Strength of recommendation: strong

Strongly agree 85%, agree 15%

A recent meta-analysis of 11 RCTs149 showed that H pylori eradication therapy reduced the risk of gastric cancer and gastric cancer related death by 36% and 22% among H pylori infected individuals without gastric neoplasia, respectively. The pooled results from some observational studies also showed consistent results.150 151 However, for patients with GIM, there was no significant benefit on gastric cancer risk (RR 0.7 6; 95% CI 0.36 to 1.61),152 which is consistent with other systematic reviews and meta-analyses.153 154 In contrast, gastric cancer risk was still reduced by 48% after H pylori eradication therapy among those with gastric neoplasia undergoing endoscopic mucosal resection who were likely having concurrent premalignant changes.149
Histopathological progression of GIM is defined as (1) progression to DYS or non-cardia gastric cancer or (2) increase in histopathological score in terms of Correa’s cascade, OLGA or OLGIM. Correa’s precancerous cascade can be categorised into six stages, from normal to non-atrophic gastritis, multifocal atrophic gastritis, GIM, DYS and gastric cancer.5 22 155 156 Mera et al reported that the histopathological score increased by an average of 0.20 units per year (95% CI 0.12 to 0.28) in individuals with persistent H pylori infection.156 In a meta-analysis conducted within AGA technical review, the RR of histopathological progression after H pylori eradication for all patients in terms of Correa’s histopathological score was 0.91 (95% CI 0.83 to 1.00).152 However, for those with GIM, the RR was 1.17 (95% CI1.01 to 1.36).152
On the other hand, histopathological regression of GIM is defined as (1) regression to normal mucosa, chronic gastritis or CAG or (2) decrease in histopathological score stage.157 According to the study by Mera et al, H pylori eradication was associated with a progressive decline in the score. At the 6 year follow-up, the score for those subjects was 0.13 less than baseline (95% CI 0.11 to 0.15) while at 12 years the score was 0.59 less than baseline (95% CI 0.51 to 0.67).155 The Shandong Interventional Trial158 159 reported the outcome in patients with GIM, showing an RR of 1.56 (95% CI 1.03 to 2.33). Another meta-analysis showed that in terms of regression in CAG and GIM, the pooled OR from four RCTs was 2.61 (95% CI 1.41 to 4.81) and 2.61 (95% CI 1.66 to 4.11), respectively.153 Notably, some meta-analyses153 154 only observed the regression of premalignant conditions in studies with follow-up periods longer than 5 years.
In view of the rising prevalence of antibiotic resistance of H pylori,160 successful H pylori eradication should be confirmed by preferably a non-invasive test (either 13C urea breath test or faecal antigen test). Endoscopy for biopsy can be considered when there is a concomitant indication (eg, follow-up for gastric ulcer).

Statement 24

APAGE recommends test-and-treat for Helicobacter pylori infection for individuals with first degree relatives diagnosed with gastric cancer. Endoscopic screening for detecting both H pylori infection and gastric premalignant conditions and/or cancer can be considered.

Quality of evidence: high

Strength of recommendation: Strong

Strongly agree 70%, agree 30%

Germline pathogenic variants are closely associated with the risk of gastric cancer161 and familial aggregation of gastric cancer occurs in approximately 10% of cases.162 A meta-analysis showed that among first degree relatives of patients with gastric cancer, the OR of H pylori infection, CAG and GIM were 1.93 (95% CI 1.42 to 2.61), 2.20 (95% CI 1.27 to 3.82) and 1.98 (95% CI 1.36 to 2.88), respectively.163 When compared with those without a family history of gastric cancer, the pooled RR of progression from GIM to gastric cancer among those with a family history of gastric cancer was 4.53 (95% CI 1.33 to 15.46).152
A test-and-treat strategy for H pylori in individuals with first degree relatives diagnosed with gastric cancer is a proactive approach to reduce the risk of developing gastric cancer. Various international and Chinese guidelines9 121 129 164 recommend H pylori screening among individuals at high gastric cancer risk (including those with a family history or first degree relatives having gastric cancer). Familial based testing may also be considered,125 165 particularly in regions with a high prevalence of H pylori infection. A meta-analysis of 12 studies from different countries (including nine from China) showed a nearly threefold higher H pylori eradication success rate when comparing a whole family test-and-treatment strategy with a single patient treatment strategy, and a 70% lower rate of recurrence was noted in the former strategy.166
Results from a cluster RCT found a beneficial effect of H pylori treatment, particularly for those aged 25–45 years, leading to a 27% decrease in the risk of developing gastric cancer.135 A nationwide population based study showed that in patients with and without a family history of gastric cancer, early treatment of H pylori infection at a younger age could maximise gastric cancer prevention, the maximal benefit being observed among those who had H pylori eradication at age <45 years.167
In addition to H pylori screening, individuals with a family history of gastric cancer may also need endoscopic screening to detect gastric premalignant conditions and gastric cancer.121 The recommended starting age of endoscopy screening is 45–50 years9 84 121 or 10 years earlier than the onset age of the first degree relative diagnosed with gastric cancer.10

Statement 25

APAGE suggests against routine empirical use of proton pump inhibitors and potassium competitive acid blockers, particularly long term use, in individuals with gastric premalignant conditions in the absence of compelling indications.

Quality of evidence: low

Strength of recommendation: weak (Conditional)

Strongly agree 85%, agree 15%

A meta-analysis of six RCTs (n=1623) showed that use of proton pump inhibitors (PPIs) was associated with a higher risk of premalignant conditions when used for more than 12 months (RR 2.21; 95% CI 1.47 to 3.33). The RRs for CAG, GIM and DYS with PPI use were 1.50 (95% CI 0.91 to 2.47), 1.93 (95% CI 1.03 to 3.63) and 7.73 (95% CI 0.36 to 165.7), respectively.168 Another meta-analysis showed that the harmful effect of PPIs were limited to non-cardia gastric cancer (RR 2.38; 95% CI 1.90 to 2.98), but not cardia gastric cancer (RR 1.32; 95% CI 0.84 to 2.03).169 A dose–response relationship was reported in a few studies. Abrahami et al showed that a higher dose of omeprazole equivalents was associated with a higher risk of gastric cancer.170 Cheung et al recruited a cohort of H pylori eradicated individuals, showing that at least weekly use of PPIs was associated with a 2.44-fold higher risk of gastric cancer.171 Furthermore, H pylori infection, either current or past, was a more important determinant of gastric cancer risk than PPIs, possibly due to the development of H pylori induced premalignant conditions.171 172
While strong evidence from RCTs is still lacking, one RCT recruiting patients receiving aspirin or rivaroxaban who were assigned to either pantoprazole 40 mg daily or placebo did not reveal an increase in gastric cancer risk.173 However, aspirin is a known chemopreventive agent against many solid organ tumours, including gastric cancer,174 175 and the PPI associated gastric cancer risk may be negated by concomitant aspirin use.176 Moreover, the study investigated multiple outcomes, posing a risk of underpower for detecting gastric cancer occurrence. Since long term use of PPIs is associated with various potential side effects,177 178 the indications for PPIs should be regularly reviewed and de-prescription of PPIs should be attempted when there is no indication.179 180
Potassium competitive acid blockers, with a more potent gastric acid suppressive effect than PPIs, have also been shown to be associated with a nearly twofold higher gastric cancer risk in a population based cohort study of 54 055 patients, with a duration and dose–response relation.181 We therefore advise a similar cautious approach for use of potassium competitive acid blockers, although strong evidence supporting this is lacking.

Statement 26

APAGE recommends against the routine use of non-Helicobacter pylori chemopreventive agents (including aspirin or cyclooxygenase-2 inhibitors) as primary prevention of gastric cancer in individuals with gastric premalignant conditions.

Quality of evidence: moderate

Strength of recommendation: strong

Strongly agree 65%, agree 35%

Meta-analyses showed that aspirin and NSAIDs were associated with a lower gastric cancer risk in observational studies, while post hoc analysis of RCTs showed a trend favouring aspirin use.182 183 However, most studies did not report the status of gastric premalignant conditions and included a heterogeneous population of both H pylori infected and non-infected individuals.
One meta-analysis of five studies found that COX-2 inhibitors reduced gastric cancer risk by 55%.184 Leung et al showed that rofecoxib did not regress GIM and its severity over 2 years in 213 H pylori eradicated individuals with GIM.185 It is noteworthy that rofecoxib was subsequently withdrawn from the market due to cardiovascular adverse events.186 On the other hand, another RCT showed that celecoxib use for 2 years could regress advanced gastric lesions in H pylori infected individuals, but not in H pylori eradicated subjects.187 Given the uncertain benefits of NSAIDs and COX-2 inhibitors in regressing gastric premalignant conditions and cancer prevention, particularly among those with H pylori eradicated, and their side effects,188 189 we do not recommend NSAIDs or COX-2 inhibitors solely for gastric cancer prevention purposes.
We also do not recommend routine use of aspirin for chemoprevention of gastric cancer. However, given its additional cardiovascular and cancer prevention benefits, the use of low dose aspirin should be individualised if other compelling indications exist. A cohort study of H pylori eradicated individuals showed that aspirin users had seven fewer gastrointestinal cancer related deaths but one more bleeding related death per 10 000 person years than non-users.175
Rebamipide, a mucosal protective agent, was shown in various RCTs to reduce the risks of GIM and low grade DYS,190 191 to improve the histopathologic grade of CAG and GIM in the antrum and to alleviate chronic inflammation in the lesser curvature of the corpus after H pylori eradication.192 193 Moluodan, a compound prescription of Chinese herbal medicine, resulted in an 82.8% disappearance rate of DYS in a randomised, double blind, parallel controlled trial.194 However, the recommendations for their routine use require further research.

Statement 27

APAGE does not recommend endoscopic treatment of gastric intestinal metaplasia, such as through endoscopic ablation or resection.

Quality of evidence: low

Strength of recommendation: Strong

Strongly agree 65%, agree 35%

Endoscopic ablation or resection is generally not recommended for eradication of GIM, especially because it is a multifocal condition. As GIM is considered a field effect suggesting a higher risk of cancer development, we emphasise the importance of careful surveillance in patients with GIM while asserting that intervention should be reserved for cases of confirmed DYS.195 Furthermore, endoscopic therapies, such as ablation, have not been shown to be effective in eradicating GIM and may introduce potential risks. A study tested radiofrequency ablation for both GIM and LGD, finding that while it successfully eradicated DYS in all patients with LGD, GIM persisted in most cases.196 Endoscopic resection carries potential complications, such as perforation or bleeding, which can complicate further management.197

Statement 28

APAGE recommends that healthcare providers counsel patients with gastric premalignant conditions to adopt a healthy lifestyle (including smoking cessation and a healthy diet, etc), as this can lower the risk of progression towards gastric cancer.

Quality of evidence: moderate

Strength of recommendation: strong

Strongly agree 40%, agree 55%, neutral 5%

Personal lifestyle choices are closely linked to the progression of gastric cancer. A prospective cohort study including 100 220 individuals found that an unfavourable lifestyle was associated with a twofold higher risk of gastric cancer.198 Even for those with a high genetic risk, a healthy lifestyle could help reduce the risk of gastric cancer by 47%.198 A summary of the management strategies for gastric premalignant conditions is shown in figure 4.

Smoking and alcohol
A recent meta-analysis showed that compared with never smokers, the pooled RR of gastric cancer for current and former smokers was 1.53 (95% CI 1.44 to 1.62) and 1.30 (95% CI 1.23 to 1.37), respectively.199 In the Singapore Gastric Epidemiology and Molecular Genetics Programme prospective cohort, OLGIM II–IV patients who smoked ≥20 pack years were associated with a 3.7-fold higher risk of early gastric neoplasia.32
The World Cancer Research Fund International Continuous Update Project highlights strong evidence that consuming three or more alcoholic drinks per day raises the risk of gastric cancer.200 201 A pooled analysis of 20 epidemiological studies showed that heavy drinkers (more than four to six drinks per day) and very heavy drinkers (more than six drinks per day) had a 26% and 48% increase in gastric cancer risk, respectively.202 A Korean study of 202 675 individuals showed that alcohol consumption was associated with both CAG and GIM in a dose–response manner.203

Dietary factors
The third expert report by the World Cancer Research Fund and the American Institute for Cancer Research concluded that there was strong evidence linking salt preserved foods to gastric cancer.204 A meta-analysis involving 2 076 498 participants demonstrated that high salt intake was associated with a twofold higher risk of gastric cancer.205 Another meta-analysis revealed a trend favouring association between salted/salty foods and GIM (OR 1.68; 95% CI 0.98 to 2.90).206
The IARC classified processed meat as a group 1 carcinogen (carcinogenic to humans) and red meat as group 2A (probably carcinogenic to humans).43 A meta-analysis including 43 cohort and case-control studies found that a high intake of processed meat and red meat was associated with a 57% and 41% higher risk of gastric cancer, respectively.207
A meta-analysis of 18 prospective studies indicated that increased consumption of fruits alone and in combination with vegetables was associated with a 13% and 25% lower risk of gastric cancer, respectively.208 On the other hand, the data regarding the effects of milk and dairy products on gastric cancer risk are limited and inconclusive.209 210 A meta-analysis of 60 studies showed a 52% increased risk of gastric cancer with the consumption of pickled foods.211 Another meta-analysis of 18 studies revealed fried food intake was associated with a 1.5-fold higher risk of gastric cancer.212

Physical activity and body mass index
A meta-analysis of 22 observational studies indicated that any form of physical activity, whether occupational, recreational or both, was linked to a 19% lower risk of gastric cancer.213 A retrospective cohort study of 142 832 Korean adults showed that compared with individuals with a body mass index of 18.5–22.9 kg/m2, those with a body mass index of >30 kg/m2 had a 1.48-fold higher risk of GIM, while those with a body mass index of at least 23 kg/m2 had at least 1.1-fold higher risk of CAG.214

Antioxidant consumption
A meta-analysis of 20 RCTs involving the supplementation of beta-carotene, vitamin A, vitamin C, vitamin E and selenium (total n=211 818) revealed that antioxidant supplements overall did not significantly influence the risk of gastric cancer (RR 1.14; 95% CI 0.97 to 1.33).215 While the Shandong Intervention Trial of China found no beneficial effects of either vitamins (C, E and selenium) or garlic (extract and oil) supplements for 7.3 years on gastric premalignant conditions,159 a beneficial effect was noted in terms of reduction of gastric cancer and cancer related mortality after 22.3 years of follow-up.216 Another large trial involving 29 133 male smokers in Finland, the Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study, showed no significant effect of antioxidant supplementation on gastric cancer during and beyond the trial period.217 218 On the other hand, a meta-analysis of 13 studies suggested that a daily dose of 20–30 mg of folic acid for 3–6 months improved the pathological changes associated with gastric premalignant conditions.219 Therefore, evidence is still conflicting to support the routine use of these antioxidant supplementations. Whether a differential effect of antioxidants exists between Asian and non-Asian populations also requires further exploration.

Limitations and future research
We acknowledge that some of the recommendations are based on limited quality evidence and necessitate expert opinions from the region. Specifically, there are no high quality data to support the recommended surveillance intervals, particularly in those with non-advanced premalignant conditions. We recommended more frequent surveillance of 2 years for patients with OLGA/OLGIM stage III or IV, and 3 years for those with stage II disease. This is to a certain extent aligned with the screening interval of upper endoscopy in some East Asian countries which practise nationwide screening endoscopy and is supported by cost effectiveness studies.130132 Comparison with current international guidelines and recommendations on risk stratification and surveillance intervals are shown in table 2. Specifically, there are some differences in the recommended surveillance intervals between the APAGE and the existing recommendations from the US and Europe. Compared with the MAPS III guideline, which establishes structured surveillance intervals based on OLGA/OLGIM staging, our recommendations also adopt a risk stratified framework but incorporate local practical adjustments that reflect existing regional gastric cancer screening practices and diagnostic capabilities. Conversely, the North American guidelines focus more on individualised assessments, particularly for less advanced premalignant conditions.10 12 Moreover, unlike recent guidelines from Europe11 and the USA,10 our consensus did not cover other premalignant conditions, such as gastric polyps and autoimmune gastritis. While we attempted to focus on H pylori associated gastric premalignant changes which are prevalent in this region, this omission represents a limitation of the present consensus, and future revisions and updates are needed to incorporate these different entities to achieve more comprehensive coverage.
Although we aim to harmonise surveillance strategies in Asia, we must admit that significant variations exist in endoscopic practices and healthcare delivery systems among Asian countries. For example, in Japan and South Korea, nationwide gastric cancer screening programmes use endoscopy every 2 years, often without obtaining routine biopsies. Instead, their strategy relies on repeated examinations to identify suspicious lesions and detect neoplastic changes at an early stage. In contrast, some countries in the region lack adequate endoscopic infrastructures for surveillance, a standardised biopsy protocol and even expertise on endoscopic and histopathological diagnosis. These disparities may present a considerable challenge to the implementation of a unified surveillance strategy across Asia and may highlight the necessity of a united call for further funding support and resources allocation to this topic. We also recognise the limitation in the composition of our panel members, as experts from certain high burden Asian countries were not represented (eg, Mongolia). Members also did not include Japanese pathologists, family physicians and primary care providers, which may not have adequately addressed the wider perspectives. Future consensus efforts should strive for more inclusive representation from across the Asia Pacific region with different specialties.
Further data are needed to quantify the optimal surveillance intervals for patients with different risk profiles. In particular, the use of genetic markers together with family history and other modifiable risk factors may further help to precisely risk stratify patients for personal surveillance.161 Some other areas of future research are listed below.
The role of different new image enhanced endoscopy (eg,TXI and LCI) or AI assistance on the detection of gastric neoplastic lesions remains uncertain.

Studies are also needed to identify additional potential chemopreventive agents for individuals with premalignant conditions after H pylori infection, as these individuals are still at risk of developing gastric cancer.

Identification of novel non-invasive gastric mucosal biomarkers or molecular markers for risk stratification and surveillance of patients with gastric premalignant conditions.

A regional registry and clinical trial data to better inform the optimal surveillance intervals in patients with gastric premalignant conditions

Conclusions
This current consensus report provides the first systematic framework for the surveillance and management of gastric premalignant conditions, leveraging the best available evidence, to the Asia Pacific region. Based on current knowledge of gastric cancer risk stratification, diagnostic modalities and surveillance strategies, these recommendations aim to standardise and enhance early detection and intervention for this common and fatal cancer in Asia. Additionally, further refinement of risk stratification models, along with the availability of novel diagnostic and surveillance tools, is essential to improve the precision of recommendations with the goal of personalised risk stratification. While we acknowledge differences within Asian countries in terms of gastric cancer incidence, endoscopy practices and healthcare resources, our current recommendation could help to identify areas for future research and inform policy to address the knowledge gaps and disparities.

Supplementary material
10.1136/gutjnl-2025-335823online supplemental file 110.1136/gutjnl-2025-335823online supplemental file 2