Gastric Cancer—An Enigmatic and Heterogeneous Disease

Gastric cancer is a complex and enigmatic disorder noted for marked global variations in etiology, incidence, natural course, and management. Many factors likely lead to and discriminate among biologically and clinically relevant gastric cancer subsets: antecedent tumorigenic conditions (Helicobacter pylori gastritis,¹ CDH1 mutation,² and other chronic gastric disorders), location of the primary tumor (proximal or distal), subtypes of adenocarcinoma (diffuse, intestinal, or mixed),³ ethnicity of the afflicted population (differing levels of susceptibility and aggressiveness of the tumors), and a predictive biomarker (ERBB2). In essence, gastric cancer is a heterogeneous condition that represents several diseases.

Additional complexities are introduced by the prevailing therapeutic approaches that differ across the globe, including types and extent of resection, including endoscopic approaches, laparoscopic approaches, or open gastrectomy with varying extents of lymph nodal dissection, and types of adjunctive strategies, including perioperative chemotherapy, postoperative chemotherapy, and postoperative chemoradiation. In endemic areas, particularly in Japan, Korea, and China, localized gastric cancer (clinical stage >T1a) is often treated by primary surgery that includes more extensive lymph nodal dissection, referred to as D2 dissection when level 2 lymph nodes are removed.⁴ However, in nonendemic areas, such as the United States, the extent of lymph node dissection is often suboptimal (referred to as D0 or <D1 dissection). Are the outcomes in Asian patients different because of differences in tumor biology and patient genetics, differing therapeutic strategies, or both?

Although adjunctive therapeutic strategies vary by region, each strategy has improved the cure rates compared with surgery alone by approximately 10%,^{5 - 7} and therefore, adjunctive therapies should be offered to all high-risk gastric cancer patients. In the West, most high-risk patients often receive postoperative chemoradiation⁶ or preoperative and postoperative chemotherapy,⁵ whereas in Japan, such patients receive adjuvant S-1 chemotherapy following a D2 dissection.⁷ The Japanese adjuvant strategy with S-1 appears appropriate for patients with gastric cancer in China and Korea because these tumors are biologically similar and the extent of surgery is similar. Whether the level 1 evidence established in Asia⁷ can be applied to Western patients is a quandary. Similarly, Asian gastric cancer experts have reservations regarding the level 1 evidence established in the West,^{5 - 6} primarily due to concerns related to suboptimal results and the extent of surgery.

In this issue of JAMA, the GASTRIC (Global Advanced/Adjuvant Stomach Tumor Research International Collaboration) Group⁸ reports findings from a patient-level meta-analysis of postoperative chemotherapy adjuvant trials. The investigators examined primary patient data from 60% of patients from randomized trials completed over a 30-year period to examine the role of adjuvant chemotherapy following gastric cancer resection. Collective analyses for overall survival and disease-free survival suggest a statistically significant benefit for patients who received chemotherapy after surgery compared with those who were only observed after surgery (hazard ratio, 0.82). The authors also analyzed data by region and type of chemotherapy (single agent vs combination), but the results were not consistent. This could be due to subgroups with small numbers of patients or other differences (eg, tumor biology combined with therapy variables). However, the authors suggest that single-agent adjuvant chemotherapy is justified as a control group (presumably for a future randomized trial).

A number of caveats to their analysis and conclusions should be considered. Meta-analyses are considered hypothesis generating and are not performed to test a hypothesis or establish a standard of care.⁹ Moreover, the exploratory results reported by the GASTRIC group⁸ have the limitations inherent in post hoc subgroup analyses. The authors included trials conducted from the 1970s; trials that were underpowered; trials with flawed designs (some without stratifications or with unrealistic expectations); trials with various chemotherapy agents and combinations; trials conducted in various regions, which had different surgical approaches, treatment durations, and completeness of therapy; and trials with other weaknesses, such as those that allowed incomplete resection. Another limitation is the lack of access to patient data from all qualified trials.

Nevertheless, can a large sample size overcome the outlined heterogeneity in biology, therapy, or both? The answer to this question is not known. However, based on the available data, postoperative adjuvant chemotherapy cannot be recommended as another standard to most Western patients with high-risk gastric cancer. Efforts should be invested in designing and executing well-conceived randomized controlled trials that answer questions for specific subsets of patients (adjuvant trials in patients with ERBB2-positive tumors, trials focusing on histologic subsets, or trials of any emerging biomarker-based strategies).

Another report in this issue of JAMA addresses the incidence of noncardia gastric cancer in specific patient cohorts. Anderson et al¹⁰ performed a carefully designed SEER-based observational study to examine age-specific time trends of the incidence of noncardia gastric cancer in the United States over the past several decades. They identified a subgroup of white US residents aged 25 to 39 years for which the incidence of noncardia gastric cancer has increased by a dramatic 67% over the past 3 decades. Although the absolute incidence of gastric cancer in this age range remains low (presently 0.45 cases per 100 000), this finding is noteworthy considering that for every other age range in white US residents, and for all age ranges in other races in the United States, the incidence of noncardia gastric cancer has declined by 1% to 2% per year.

Although Anderson et al¹⁰ were not able to subcategorize white US patients into Hispanic white and non-Hispanic white individuals for the entire study period, a sensitivity analysis performed from a later time point when Hispanic white individuals were distinguished from non-Hispanic white individuals showed similar age-specific trends, suggesting that at least 1 immigration pattern cannot fully explain this phenomenon. Another issue that warrants clarification is whether another immigration pattern (eg, from Eastern Europe) may provide a possible explanation.

These findings have important implications. The distinction between cardia and noncardia gastric cancer is relevant to the role of H pylori carcinogenesis. The pathogenesis of noncardia cancer follows a multistep progression that is likely initiated by chronic inflammation (eg, H pylori,^{1 ,11} chronic, or autoimmune gastritis). The disease progresses through chronic gastritis, intestinal metaplasia, and dysplasia.^{12 - 13} Although H pylori gastritis contributes to the risk of noncardia adenocarcinoma, it may be protective for proximal adenocarcinoma.^{14 - 15} Furthermore, gene-environmental interaction^{16 - 18} may influence susceptibility to the consequences of H pylori gastritis.

The enigma of gastric cancer continues. It remains the second most common gastrointestinal malignancy internationally and is responsible for the most gastrointestinal cancer–related deaths worldwide.¹⁴ Many challenges lie ahead to better identify subsets of gastric cancer and therefore recommend individualized approaches to prevention, early detection, and therapy.