Development of a modified ABC method among Helicobacter pylori infected but serum pepsinogen test‐negative individuals

Background Although the ABC method for gastric cancer (GC) screening has been widely adopted in Japan, it may not be suitable for other countries due to population heterogeneity and different tumor histology. We aim to develop a modified ABC method to improve GC screening performance, especially among Helicobacter pylori (Hp) infected but serum pepsinogen (sPG) test‐negative individuals. Methods A total of 4745 participants were recruited from Tianjin, China, and were classified into four groups by combined assay for Hp infection and sPG concentrations: Group A (Hp [−], PG [−]), Group B (Hp [+], PG [−]), Group C (Hp [+], PG [+]), and Group D (Hp [−], PG [+]). We used receiver‐operating characteristic (ROC) curves analysis and minimum p value method to determine the optimal cutoff point for PG II in Group B. We performed logistic regressions to examine the risk of GC across different subgroups. In addition to the derivation set, the performance of the modified ABC method was also evaluated in an external set involving 16,292 participants from Liaoning, China. Results In the modified ABC method, we further classified Group B as low‐risk (Group B1) and high‐risk subgroups (Group B2) using optimal sPG II cutoff point (20.0 ng/mL) by ROC curves analysis and minimum p value method. Compared with Group B1, Group B2 had a significantly higher risk of GC (adjusted OR = 2.54, 95% CI = 1.94–3.33). The modified ABC method showed good discrimination for GC (AUC = 0.61, 95% CI = 0.59–0.63) and improved risk reclassification (NRI = 0.11, p