The OLGA-OLGIM staging and the interobserver agreement for gastritis and preneoplastic lesion screening: a cross-sectional study

Diagnosing and grading gastric atrophy and intestinal metaplasia using semi-supervised deep learning on pathological images: development and validation study

OBJECTIVE

Patients with gastric atrophy and intestinal metaplasia (IM) were at risk for gastric cancer, necessitating an accurate risk assessment. We aimed to establish and validate a diagnostic approach for gastric biopsy specimens using deep learning and OLGA/OLGIM for individual gastric cancer risk classification.

METHODS

In this study, we prospectively enrolled 545 patients suspected of atrophic gastritis during endoscopy from 13 tertiary hospitals between December 22, 2017, to September 25, 2020, with a total of 2725 whole-slide images (WSIs). Patients were randomly divided into a training set (n = 349), an internal validation set (n = 87), and an external validation set (n = 109). Sixty patients from the external validation set were randomly selected and divided into two groups for an observer study, one with the assistance of algorithm results and the other without. We proposed a semi-supervised deep learning algorithm to diagnose and grade IM and atrophy, and we compared it with the assessments of 10 pathologists. The model's performance was evaluated based on the area under the curve (AUC), sensitivity, specificity, and weighted kappa value.

RESULTS

The algorithm, named GasMIL, was established and demonstrated encouraging performance in diagnosing IM (AUC 0.884, 95% CI 0.862-0.902) and atrophy (AUC 0.877, 95% CI 0.855-0.897) in the external test set. In the observer study, GasMIL achieved an 80% sensitivity, 85% specificity, a weighted kappa value of 0.61, and an AUC of 0.953, surpassing the performance of all ten pathologists in diagnosing atrophy. Among the 10 pathologists, GasMIL's AUC ranked second in OLGA (0.729, 95% CI 0.625-0.833) and fifth in OLGIM (0.792, 95% CI 0.688-0.896). With the assistance of GasMIL, pathologists demonstrated improved AUC (p = 0.013), sensitivity (p = 0.014), and weighted kappa (p = 0.016) in diagnosing IM, and improved specificity (p = 0.007) in diagnosing atrophy compared to pathologists working alone.

CONCLUSION

GasMIL shows the best overall performance in diagnosing IM and atrophy when compared to pathologists, significantly enhancing their diagnostic capabilities.

Correlation of chronic atrophic gastritis with gastric-specific circulating biomarkers

BACKGROUND AND STUDY AIMS

It has been suggested that the combined detection of multiple serum biomarkers can effectively screen out the high-risk population of chronic atrophic gastritis in the general population. Therefore, it is necessary to establish an effective predictive model of chronic atrophic gastritis.

PATIENTS AND METHODS

Serum biopsies were assessed using five stomach-specific circulating biomarkers pepsinogen I (PGI), PGII, PGI/II ratio, anti- H. pylori antibody, and gastrin-17 (G-17) to identify high-risk individuals and evaluate the risk of developing chronic atrophic gastritis.

RESULTS

In the cross-sectional analysis, PGII, the PG ratio, G17, anti- H. pylori IgG were positively associated with the presence of chronic atrophic gastritis, and combined prediction of the five biomarkers was more accurate than single-factor prediction ((0.692 vs 0.54(PG1), 0.604 (PGⅡ), 0.616(PGI/II ratio), 0.629(G-17)).

CONCLUSION

The combination of PGI, PGII, the PGI/II ratio, G17, and anti-H. pylori antibodies for serological analysis are helpful to screen chronic atrophic gastritis high-risk subjects from the general population and recommend that these people carry out further endoscopy and biopsy.

Comparison of different risk stratifications for gastric cancer and establishing a simplified risk-scoring model based on the Kyoto classification

BACKGROUND AND AIM

The study aims to assess the value of different risk stratifications in diagnosing early gastric cancer (GC) and explore risk factors based on Kyoto gastritis classification.

METHODS

This study was a single-centered cross-sectional study; all epidemiological data and endoscopic findings were obtained prospectively. To evaluate the proportion of GC in each risk stratification and to compare the diagnostic performance of different methods using the receiver operating characteristic curve, univariable and multivariable analyses were used to explore the correlation between endoscopic findings and GC.

RESULTS

A total of 240 subjects were enrolled, and the diagnostic efficacy of the Kyoto Classification Score was similar to Operative Link on Gastric Intestinal Metaplasia Assessment (OLGIM) stage, and the accuracy was higher than that of the Japanese scoring system and OLGA stage. Moderate atrophy (odds ratio [OR] = 3.52, 95% confidence interval [CI]: 1.52-8.16), severe atrophy (OR = 4.96, 95% CI: 1.75-14.04), map-like redness (OR = 9.89, 95% CI: 1.16-84.15), and xanthelasma (OR = 3.57, 95% CI: 1.15-11.15) were independent risk factors for GC. The simplified Kyoto classification (area under the receiver operating characteristic [AUROC] = 0.76, P = 0.58) based on multivariable analysis demonstrated favorable diagnostic value compared with traditional Kyoto classification score (AUROC = 0.74).

CONCLUSIONS

This study confirms the value of the Kyoto classification score and the OLGIM stage in the risk stratification of GC. Simplified Kyoto classification is also promising in risk assessment of GC but still requires validation in the population.

Autoimmune gastritis: long-term natural history in naïve Helicobacter pylori-negative patients

OBJECTIVE

Autoimmune gastritis (AIG) is an immunomediated disease targeting parietal cells, eventually resulting in oxyntic-restricted atrophy. This long-term follow-up study aimed at elucidating the natural history, histological phenotype(s), and associated cancer risk of patients with AIG consistently tested H. pylori-negative (naïve H. pylori-negative subjects).

DESIGN

Two-hundred eleven naïve H. pylori-negative patients (tested by serology, histology, molecular biology) with AIG (F:M=3.15:1; p<0.001) were prospectively followed up with paired biopsies (T1 vs T2; mean follow-up years:7.5 (SD:4.4); median:7). Histology distinguished non-atrophic versus atrophic AIG. Atrophy was further subtyped/scored as non-metaplastic versus metaplastic (pseudopyloric (PPM) and intestinal (IM)). Enterochromaffin-like-cell (ECL) status was categorised as diffuse versus adenomatoid hyperplasia/dysplasia, and type 1 neuroendocrine tumours (Type1-NETs).

RESULTS

Over the long-term histological follow-up, AIG consistently featured oxyntic-predominant-mononuclear inflammation. At T1, PPM-score was greater than IM (200/211 vs 160/211, respectively); IM scores increased from T1 to T2 (160/211 to 179/211), with no changes in the PPM prevalence (T1=200/211; T2=201/211). At both T1/T2, the prevalence of OLGA-III-stage was <5%; no Operative Link on Gastritis Assessment (OLGA)-IV-stage occurred. ECL-cell-status progressed from diffuse to adenomatoid hyperplasia/dysplasia (T1=167/14 vs T2=151/25). Type1-NETs (T1=10; T2=11) always coexisted with extensive oxyntic-atrophy, and ECL adenomatoid-hyperplasia/dysplasia. No excess risk of gastric or other malignancies was found over a cumulative follow-up time of 10 541 person years, except for (marginally significant) thyroid cancer (SIR=3.09; 95% CI 1.001 to 7.20).

CONCLUSIONS

Oxyntic-restricted inflammation, PPM (more than IM), and ECL-cell hyperplasia/neoplasia are the histological AIG hallmarks. Compared with the general population, corpus-restricted inflammation/atrophy does not increase the GC risk. The excess of GC risk reported in patients with AIG could plausibly result from unrecognised previous/current H. pylori comorbidity.