Diagnostic performance of the normal range of gastrin calculated using strict criteria based on a combination of serum markers and pathological evaluation for detecting gastritis: a retrospective study.

Background: The ABC method, which combines the pepsinogen method and anti-Helicobacter pylori antibody titers, has been used for risk screening for gastric cancer in Japan. However, it has been reported that there are cases of gastritis and carcinogenesis risk even in group A, which is considered to be a low-risk group based on the ABC method. Currently, in group A, endoscopic examination is needed to strictly discriminate "patients without gastritis" (defined as true A patients) from those "with gastritis." A simple and minimally invasive diagnostic criterion for gastritis using serological markers is desirable. In this study, we aimed to identify the normal serum gastrin concentrations in normal stomach cases based on pathological diagnosis and investigate the usefulness of serum gastrin concentrations in diagnosing gastritis.

Methods: Patients who underwent endoscopy and blood tests at Hiroshima University Hospital were enrolled in the study and categorized into the "pathologically-evaluated group" and "endoscopically-evaluated group," according to the evaluation method of atrophic gastritis. Initially, we measured serum gastrin concentrations in the normal stomach cases in the pathologically-evaluated group and calculated the normal range of serum gastrin concentrations. We used the upper limit of this normal range of serum gastrin concentrations and performed a validation study to determine its usefulness as a diagnostic marker for distinguishing between cases of gastritis and true A in the endoscopically-evaluated group.

Results: The 95th percentile of serum gastrin concentrations in pathologically-evaluated normal stomach cases was 34.12-126.03 pg/mL. Using the upper limit of this normal range of serum gastrin concentrations, the sensitivity, specificity, positive predictive value, and negative predictive value for gastritis were 52.8%, 92.6%, 97.0%, and 31.0%, respectively. Additionally, the receiver operating characteristic (ROC) curve for the endoscopically-evaluated group showed an area under the ROC curve of 0.80.

Conclusion: The gastrin cut-off value of 126 pg/mL has a good positive predictive value (97.0%) for detecting gastritis positing its use as a marker for cases requiring endoscopy. However, the identification of patients with gastritis having normal serum gastrin concentrations due to insufficient sensitivity remains a challenge for the future.