Tumor location stratifies lymph node metastasis risk and prognosis in gastric cancer: A retrospective analysis of multicenter data and SEER database.

Abstract

Given the significant relationship between the number of metastatic lymph nodes and primary tumor location, which is overlooked in the current pathologic node staging system, this study seeks to optimize pathologic node staging by incorporating this critical factor. This study analyzed patients with primary gastric cancer who underwent radical gastrectomy, comprising 4,926 from a local cohort, 623 from a domestic cohort, and 4,243 from a public database. Patients were stratified by tumor location: upper, middle, or lower gastric subgroups. Restricted cubic spline plots evaluated nonlinear relationships between the number of metastatic lymph nodes and mortality across locations. The Cox proportional hazards model was used to assess the effects of clinical factors on patient prognosis. Univariate analyses assessed metastatic lymph nodes and tumor location effects on overall survival. Survival disparities persist by tumor location within identical pathologic node stages. Patients with upper-third gastric cancer show significantly worse survival than middle- or lower-third groups across pathologic node stages (P < .001 for pathologic node 0-3; P < .05 for pathologic node 4), whereas survival for patients with middle versus lower gastric cancer is similar (P > .05). Mortality risk progressively increases with more metastatic lymph nodes regardless of location. Critically, patients in the upper gastric cancer group exhibit a markedly accelerated mortality risk escalation than middle and lower groups (P for overall < .001, P for nonlinear < .001), a trend consistent across pathologic tumor stages. This study demonstrated that patients in the upper gastric cancer group have significantly worse survival than those in the middle and lower groups with identical metastatic lymph nodes. Leveraging tumor location and metastatic lymph nodes, we established a novel pathologic node staging system that outperformed the conventional pathologic tumor-node-metastasis staging in prognostic discrimination.