Scoring model for discriminating gastric cancer risk in patients with negative serum pepsinogen and anti‐Helicobacter pylori antibody results

Abstract

The ABC test measures serum pepsinogen and anti-Helicobacter pylori IgG antibody levels to predict precancerous conditions in the stomach and gastric cancer. However, a limitation of this test is that the gastric cancer risk is not negligible in patients with a negative result. Based on their ABC results, 1157 patients were classified into Groups A (n=392), B (n=479), C (n=247), and D (n=39). In Group A, 24.2% of patients had atrophic gastritis and/or intestinal metaplasia and had thus been incorrectly assigned to Group A. Patients in Group A were then assigned to derivation (n=236) and validation (n=156) cohorts by 3:2 random sampling. Logistic regression analyses were performed to identify the factors discriminating between a correct (true) and incorrect (false) Group A classification. A 4-point discriminative model was constructed based on a high-negative H. pylori IgG antibody titer and the patient's age (50-64 and ≥65years). The areas under the receiver operating characteristic curve for the derivation and validation cohorts were 0.868 and 0.894, respectively. In the validation cohort, the addition of a discriminative model score ≥2 to the ABC method showed a similar accuracy for predicting gastric cancer risk compared with the ABC method alone (93.8% vs. 92.4%). The 4-point discriminative model may help identify patients with a normal serological test who are nonetheless at risk of developing gastric cancer.