Zhou Zhangjian, Xie Xin, Hao Nan, Diao Dongmei, Song Yongchun, Xia Peng, Dang Chengxue, Zhang Hao
a Division of Surgical Oncology , The First Affiliated Hospital, Xi'an Jiaotong University , Xi'an , Shaanxi , PR China.
Curr Med Res Opin. 2018 Jun;34(6):963-970. doi: 10.1080/03007995.2018.1429390. Epub 2018 Feb 1.
In addition to the traditional TNM N staging system, lymph node ratio (LNR) and log odds of metastatic lymph nodes (LODDS) staging methods were developed in cancers. This study aimed to examine their relative prognostic performance in patients with adenocarcinoma of the esophagogastric junction (AEG).
Patients who underwent surgical resection for AEG were identified from the Surveillance, Epidemiology, and End Results (SEER) Program and the First Affiliated Hospital of Xi'an Jiaotong University as the training and validation sets, respectively. The Akaike's Information Criterion (AIC), Harrell's C statistic, and ROC curves were utilized for comparison.
A total of 735 patients were involved in the training set. LODDS and LNR staging systems had better prognostic performance than the TNM N staging systems (when considered as a categorical variable: C index = 0.728 and 0.712 vs 0.671; AIC: 6247.537 and 6265.996 vs 6320.045; AUC: 0.762 and 0.719 vs 0.692. For the continuous model: C index = 0.675 and 0.686 vs 0.658; AIC = 6243.740 and 6261.027 vs 6355.077; AUC = 0.778 and 0.733 vs 0.693). In the validation set of 183 patients, the TNM N staging scheme outperformed the LODDS and LNR staging systems (C index = 0.788 vs 0.779 and 0.767; AIC = 1014.702 vs 1026.899 and 1025.288; AUC = 0.806 vs 0.787 and 0.791) when considered a categorical variable. However, when considered a continuous variable, the LODDS and LNR staging systems were better than the TNM N staging system (C index = 0.724 and 0.733 vs 0.747; AIC = 1018.075 and 1025.803 vs 1026.085; AUC = 0.811 and 0.810 vs 0.806).
The LNR and LODDS staging schemes could be considered new options for prognostic prediction of AEG with respect to lymph node status, especially when considered as continuous variables.
除了传统的TNM N分期系统外，癌症中还开发了淋巴结比率（LNR）和转移淋巴结对数优势（LODDS）分期方法。本研究旨在探讨它们在食管胃交界腺癌（AEG）患者中的相对预后性能。
分别从监测、流行病学和最终结果（SEER）计划以及西安交通大学第一附属医院中识别出接受AEG手术切除的患者作为训练集和验证集。利用赤池信息准则（AIC）、哈雷尔C统计量和ROC曲线进行比较。
训练集共纳入735例患者。LODDS和LNR分期系统的预后性能优于TNM N分期系统（作为分类变量时：C指数=0.728和0.712 vs 0.671；AIC：6247.537和6265.996 vs 6320.045；AUC：0.762和0.719 vs 0.692。对于连续模型：C指数=0.675和0.686 vs 0.658；AIC=6243.740和6261.027 vs 6355.077；AUC：0.778和0.733 vs 0.693）。在183例患者的验证集中，当视为分类变量时，TNM N分期方案优于LODDS和LNR分期系统（C指数=0.788 vs 0.779和0.767；AIC=1014.702 vs 1026.899和1025.288；AUC=0.806 vs 0.787和0.791）。然而，当视为连续变量时，LODDS和LNR分期系统优于TNM N分期系统（C指数=0.724和0.733 vs 0.747；AIC=1018.075和1025.803 vs 1026.085；AUC=0.811和0.810 vs 0.806）。
LNR和LODDS分期方案可被视为预测AEG淋巴结状态预后的新选择，尤其是当视为连续变量时。
