Prognostic value of metastatic lymph node ratio in adenocarcinoma of the gastroesophageal junction.
- Author:
Hong-dian ZHANG
1
;
Zhao MA
;
Peng TANG
;
Xiao-feng DUAN
;
Peng REN
;
Jie YUE
;
Zhen-tao YU
Author Information
- Publication Type:Journal Article
- MeSH: Adenocarcinoma; diagnosis; pathology; Adult; Aged; Aged, 80 and over; Esophagogastric Junction; Female; Humans; Kaplan-Meier Estimate; Lymph Nodes; pathology; Lymphatic Metastasis; Male; Middle Aged; Neoplasm Staging; Prognosis; Retrospective Studies; Young Adult
- From: Chinese Journal of Gastrointestinal Surgery 2013;16(9):822-826
- CountryChina
- Language:Chinese
-
Abstract:
OBJECTIVETo compare the prognostic value of AJCC/UICC pN stage with metastatic lymph node ratio (MLR) and the prognostic difference between the tumor-node-metastasis (TNM) stage and tumor-ratio-metastasis (TRM) stage in patients with adenocarcinoma of the gastroesophageal junction.
METHODSClinical data of 414 patients with adenocarcinoma of the gastroesophageal junction undergoing curative resection at the Tianjin Medical University Cancer Institute and Hospital from January 2000 to June 2007 were retrospectively reviewed. Spearman correlation analysis was performed to examine the correlations between pN, MLR and retrieved nodes. Univariate Kaplan-Meier survival analysis and multivariate Cox proportional hazard model analysis were performed to analyze the effects of pN, MLR, TNM and TRM stage on the prognosis of these patients. The area under the ROC curve (AUC) was plotted to compare the value of these stages and to predict the 5-year survival rate.
RESULTSThe median number of retrieved nodes was 17 (4-71) per patient, and the median number of positive nodes was 4 (0-67) per patient. The number of metastatic lymph node was positively correlated with that of retrieved nodes (P<0.01), but MLR was not correlated with the number of retrieved nodes (P>0.05). Univariate and multivariate survival analysis showed that either pN or MLR could be used as an independent risk factor for survival (P<0.01) and the hazard ratio of MLR stage was larger than that of pN stage (1.573 vs 1.382). While pN and MLR were entered into the Cox hazard ratio model as covariates at the same time, MLR remained as the independent prognostic factor (P<0.01), but pN lost significance (P>0.05). The AUC of MLR and pN staging was 0.726 and 0.714, and of TRM and TNM staging was 0.747 and 0.736, respectively, however the differences were not statistically significant (all P>0.05).
CONCLUSIONSMLR is an independent prognostic factor for patients with adenocarcinoma of the gastroesophageal junction. The value of MLR and TRM staging systems may be superior to pN and TNM staging systems in evaluating the prognosis of these patients.