Enhancing diagnostic accuracy: Role of stomach-specific serum biomarkers in real-world risk-based sequential screening for malignant gastric lesions

Abstract

Objective: A risk-based sequential screening strategy, from questionnaire-based assessment to biomarker measurement and then to endoscopic examination, has the potential to enhance gastric cancer (GC) screening efficiency. We aimed to evaluate the ability of five common stomach-specific serum biomarkers to further enrich high-risk individuals for GC in the questionnaire-identified high-risk population.

Methods: This study was conducted based on a risk-based screening program in Ningxia Hui Autonomous Region, China. We first performed questionnaire assessment involving 23,381 individuals (7,042 outpatients and 16,339 individuals from the community), and those assessed as "high-risk" were then invited to participate in serological assays and endoscopic examinations. The serological biomarker model was derived based on logistic regression, with predictors selected via the Akaike information criterion. Model performance was evaluated by the area under the receiver operating characteristic curve (AUC).

Results: A total of 2,011 participants were ultimately included for analysis. The final serological biomarker model had three predictors, comprising pepsinogen I (PGI), pepsinogen I/II ratio (PGR), and anti-Helicobacter pylori immunoglobulin G (anti-H. pylori IgG) antibodies. This model generated an AUC of 0.733 (95% confidence interval: 0.655-0.812) and demonstrated the best discriminative ability compared with previously developed serological biomarker models. As the risk cut-off value of our model rose, the detection rate increased and the number of endoscopies needed to detect one case decreased.

Conclusions: PGI, PGR, and anti-H. pylori IgG could be jointly used to further enrich high-risk individuals for GC among those selected by questionnaire assessment, providing insight for the development of a multi-stage risk-based sequential strategy for GC screening.

Keywords: Diagnostic model; gastric cancer; sequential screening; serological biomarker.

Figure 1

Flowchart of risk-based sequential screening for gastric cancer.

Figure 2

Receiver operating characteristic curves of the serological biomarker model for malignant gastric lesions in development set (A), internal validation using 10-fold and leave-one-out cross validation (B), subgroup validation across high-risk populations with varying proportions (C), and comparison between the combination of stomach-specific serum biomarkers and G-score and G-score alone (D). PGI, pepsinogen I; PGR, pepsinogen I/II ratio; anti-H. pylori IgG, anti-Helicobacter pylori immunoglobulin G; AUC, area under the receiver operating characteristic curve; 95% CI, 95% confidence interval.

Figure 3

Application performance of the serological biomarker model under different screening coverages among the questionnaire-identified high-risk population.