Catching Up with the World: Pepsinogen Screening for Gastric Cancer in the United States

Gastric cancer is the fifth most common and fourth deadliest cancer worldwide (1). While gastric cancer incidence is lower in the United States compared with certain regions of the world, survival outcomes from this cancer remain exceedingly poor compared with other cancers, such as breast and colorectal cancer, for which the United States has established and effective screening programs (1, 2). Data on screening, early detection, and risk stratification for gastric cancer are urgently needed in the multiethnic U.S. population. One promising noninvasive screening test is the use of serum levels of pepsinogen, a strategy studied in Asian and European populations but for which U.S. data is lacking (3–6).

In this important article, In and colleagues report findings from a case–control study nested within the U.S.-based Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial (7). The authors evaluated pepsinogen levels in previously collected serum from individuals who later developed gastric cancer, in comparison with age/sex/race-matched controls. In this study, individuals were classified as test-positive (PG⁺) if they demonstrated both pepsinogen I ≤ 70 μg/L and pepsinogen I to II ratio ≤ 3.0, with all other individuals classified as test-negative (PG⁻). PG⁺ individuals were at nearly 10-fold increased risk for developing gastric cancer compared with PG⁻ individuals, and this effect remained robust after adjusting for Helicobacter pylori (Hp) status, family history of gastric cancer, education, smoking, and obesity. In subgroup analysis, the predictive ability of the test was particularly robust for noncardia gastric cancers, and nonpredictive of cardia gastric cancers.

Noncardia gastric cancer is thought to develop through an Hp-mediated inflammatory cascade, with progression through intermediary precancerous precursors (including atrophic gastritis, intestinal metaplasia, and dysplasia) prior to cancer development. Pepsinogens are proenzymes for pepsin, a digestive enzyme, and can be classified into two groups: pepsinogen I is secreted by glands in the gastric fundus and body, and pepsinogen II is secreted in the gastric antrum and duodenum. Hp-induced inflammation usually begins in the antrum/incisura early in life, and gradually progress upward toward the body and fundus (Fig. 1; ref. 8). As inflammation and atrophy reach the gastric body, serum levels of pepsinogen I (produced by chief cells found in the body) decrease. By contrast, levels of pepsinogen II (produced both in gastric tissue and proximal small bowel) remain more stable (9), such that the pepsinogen I/II ratio decreases.

In a meta-analysis by Huang and colleagues including 31 studies with 1,520 patients with gastric cancer, PG⁺ demonstrated a pooled sensitivity for gastric cancer of 0.69 [95% confidence interval (CI), 0.60–0.76] and pooled specificity of 0.73 (95% CI, 0.62–0.82; ref. 4). This meta-analysis included studies predominantly from Europe and Asia and used varying cut-offs for PG⁺, most of which were consistent with the cut-offs used by In and colleagues (7). A meta-analysis by Zagari and colleagues examined diagnostic accuracy of a panel test including a combination of serum pepsinogens, gastrin-17, and anti-Hp antibodies for the diagnosis of atrophic gastritis (3); predominantly based on European studies, this meta-analysis found a pooled sensitivity of 74.7% and pooled specificity of 95.6% of this panel for atrophic gastritis.

While PG⁺ has been well-studied in Asian and European countries, there is a noticeable sparsity of data from the United States. A recent single-center U.S. study based on a multiethnic population found PG⁺ to demonstrate poor discrimination for precancerous lesions (atrophic gastritis and intestinal metaplasia), even when combined with Hp antibodies (10). This cohort consisted of a heterogeneous multi-ethnic population drawn from the San Francisco Bay Area (San Francisco, CA). The lower performance of PG⁺ in this study may relate to differences in study population, including a lower prevalence of active Hp infection, and higher rates of proton pump inhibitor use. Also notably, the United States has a higher proportion of autoimmune-mediated gastritis, which may be associated with a different mechanism of progression to gastric cancer (11, 12).

There is an urgent need for improved risk stratification tools for gastric cancer in the United States. 5-year survival for gastric cancer in the United States remains poor (∼30%–35%; ref. 2). By comparison, 5-year survival in South Korea, a nation with a structured population-based screening program, stand at 75% (13). Furthermore, gastric cancer screening programs have improved early-stage gastric cancer diagnosis in high-prevalence nations; after implementation of the national screening program, the proportion of early gastric cancer diagnoses in South Korea have substantially increased (14, 15). Noncardia gastric cancer disproportionately affects racial/ethnic minority groups in the United States and represents an ongoing health disparity that must be addressed. Improved gastric cancer outcomes in the United States are unlikely without improved screening strategies that can lead to early detection.

We congratulate the authors on performing an important and much-needed study to elucidate the role of pepsinogens in gastric cancer screening in the United States. We look forward to future studies building on this work that will contribute to the development of noninvasive tools for early detection and risk stratification of this deadly cancer.

No disclosures were reported.