目的:胃癌常用美國癌症分期聯合委員會與國際抗癌聯盟共同制定的TNM標準分期,本研究嘗試以胃腺癌來比較(AJCC)第六版與第七版病理分期預測預後差異性。方法:採用回溯之世代研究方法,對象為在2002 ~ 2014年彰化基督教醫院診斷及接受手術治療之胃腺癌376 患者。運用Statistical Package of Social Studies software(SPSS for windows, version 22.0, SPSS Inc.)。套裝軟體中單變量與多變項生存分析、ROC (Receiver Operating Characteristic)曲線圖、概度比卡方值、Aikaike Information Criterion (AIC)作為統計分析工具,判斷第六版與第七版病理分期兩模型間哪一模型的適合度較佳。結果:五年特定病因存活率是71.0%,應用Kaplan-Meier 曲線圖呈現第六版病理分期系統能區別五年特定病因存活率能力,IA、IB、II、IIIA、IIIB、IV、病理期別分別為100%、90.8%、66.7%、61.8%、50.2%、40.5%。第六版病理分期系統採用ROC 曲線下面積(AUC)顯示周邊淋巴結(N)與病理期別(Stage)有較高鑑別度。概度比卡方值估計結果得知腫瘤侵犯深度(T)與病理期別(Stage)具有較高同質性。腫瘤侵犯深度(T)和周邊淋巴結(N)與病理期別(Stage)之AIC值,更能適合預測預後能力。結論:我們發現,第六版胃腺癌病理分期較佳預測預後能力。
Purpose: The world's most authoritative staging system for gastric cancer is the TNM staging system jointly developed by the AJCC and UICC. This aims of this study were (1) to perform a retrospective analysis of the differences between the 6th and 7th editions of the AJCC staging system, and (2) to evaluate the accuracy of prediction of prognostic factors, including tumor site and tumor attributes, in patients with gastric cancer. Methods: A total of 376 patients with gastric adenocarcinoma cancer who underwent surgery at Changhua Christian Hospital during the period of January 2002 to December 2014 were enrolled in our analysis. All cases were followed-up to record death until December 31, 2015. Both univariate and multivariate survival analyses were used to assess the impacts of prognostic factors on gastric cancer death. Homogeneity and discriminatory factors were tested using the likelihood ratio Chisquared test and the ROC (Receiver Operating Characteristic) curves in order to compare the 6th edition with the 7th edition, respectively. AIC (Akaike's Information Criterion) was used to evaluate the classification accuracy of the two staging systems. Statistical analyses were performed using SPSS statistical software 22.0 (SPSS Inc., Chicago, IL, USA). Results: The five-year cause-specific survival rate (5YCSSR) for the entire cohort was 71.0%. Our data, obtained by Kaplan- Meier curve analysis, showed a better discriminatory ability of the 6th edition of the staging system in terms of the 5YCSSR, as follows: 100.0% for Stage IA (n = 66), 90.8% for Stage IB (n = 52), 66.7% for Stage II (n = 96), 61.8% for Stage IIIA (n = 76), 50.2% for Stage IIIB (n = 33), and 40.5% for Stage IV (n = 53) (p < 0.001). The 6th staging system had the highest area under the curve of the ROC and higher likelihood ratio Chi-squared scores. The smaller AIC of the 6th edition staging system indicated more accurate prediction of selected prognostic factors in patients with gastric cancer. Conclusions: Based on our results, the 6th edition of the staging system had a better performance in terms of staging classification. Adoption of the 6th edition staging system for gastric cancer as the universal staging system could be suggested.