A modified subclassification to evaluate the survival of patients with N3 gastric cancer: an international database study

2019 
The eighth TNM classification for gastric cancer categorizes N3 as N3a and N3b in the final pathologic stage. The cutoff for N3a/N3b is defined as 15 metastatic lymph nodes, but the rationale for this cutoff remains unclear. This study aimed to determine the optimal N3a/N3b cutoff and evaluate its prognostic significance. An international database was constructed by combining data from patients with N3 gastric cancer and complete five-year follow-up data from the Surveillance, Epidemiology, and End Results program database (n = 1833) and the Fujian Medical University Union Hospital database (n = 920) (total n = 2753). A log-rank test was performed to determine the optimal N3a/N3b cutoff, and its prognostic significance was confirmed in a two-step multivariate analysis and compared to that of the eighth TNM. A cut-point analysis performed at each metastatic lymph node number identified the greatest survival difference between N3a and N3b at 13 metastatic lymph nodes (χ2 = 157.671, P = 3.65 × 10− 36). In patients with 14–15 metastatic lymph nodes, prognoses were significantly worse than those in patients with 7–13 metastatic lymph nodes (P   15 metastatic lymph nodes (P = 0.078). Therefore, patients with 14–15 metastatic lymph nodes were incorporated into a modified N3b classification. In the two-step multivariate analysis, the eighth N3 classification fell out of the model, while the modified N3 classification remained intact (HR 1.51, P < 0.001). Further analyses demonstrated that the modified TNM classification had superior homogeneity, discriminatory ability, and gradient monotonicity compared to the eighth TNM classification. For improved prognostic stratification, we recommend adjusting the cutoff for subclassification of N3 gastric cancer to 13 metastatic lymph nodes.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    26
    References
    2
    Citations
    NaN
    KQI
    []