Background: The 8th edition of AJCC gastric cancer pathological staging system (AJCC8) derived from the IGCA database needs an external validated in cohorts with higher proportion of advanced disease.

Patients And Methods: A total of 5386 gastric cancer patients treated at Chang Gung Memorial Hospital (CGMH) and Veteran General Hospital in Taipei (TVGH) were enrolled. Clinicopathological data of the IGCA series and the CGMH/TVGH cohort were compared. Cumulative survival curves of the CGMH/TVGH cohort as stratified by the AJCC7 and the AJCC8 were compared. Lymph node ratio (LNR) was analyzed in patients with N3b disease.

Results: Patients in the CGMH/TVGH cohort were older and had more advanced tumor stage (stage III, 49% versus 26%, p < 0.001) than those in the IGCA cohort. The median survival of stages IIIA, IIIB, and IIIC as defined by the AJCC 8 were 49, 27 and 15 months, respectively, with narrower 95% confidence intervals, in comparison with 62, 30 and 18 months, respectively, as defined by the AJCC7. The AJCC8 exhibited better homogeneity within stages and discriminatory ability between stages, compared to the AJCC7. Six hundred and four patients with N3b disease were stratified by LNR into three subgroups, and their median survival were 31, 17, and 11 months, respectively (p < 0.001). LNR further appeared as a powerful outcome predictor of N3b disease (HR, 3.1).

Conclusion: The AJCC8 performs well in patients with high proportion of advanced gastric cancer. We recommend that LNR is a supplementary prognostic indicator for N3b disease.

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.ejso.2018.05.036DOI Listing

Publication Analysis

Top Keywords

gastric cancer
12
cgmh/tvgh cohort
12
8th edition
8
cancer pathological
8
pathological staging
8
edition american
4
american joint
4
joint committee
4
committee gastric
4
staging classification
4

Similar Publications

Want AI Summaries of new PubMed Abstracts delivered to your In-box?

Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!