应用竞争模型分析胃食管交接肿瘤术后辅助放化疗后复发风险分析

Wei-Xiang Qi

Jiayi Chen
Shengguang Zhao

1.Department of oncology, Rui Jin Hospital affiliated medicine school of Shanghai Jiao Tong University

目的: Gastric cancer (GC) ranks the fifth most diagnosed malignant tumor (1,000,000 new cases) and the third cause of cancer-related mortality (783,000 deaths) worldwide. Surgical resection with regional lymph node dissection is considered as the only curative treatment for patients with GC. Adjuvant chemoradiotherapy (CRT) have been extensively investigated for the treatment of GC patients after curative surgery in order to improve survival. The impact of pre-treatment factors on long-term survival and treatment failure of patients who had undergone a curative resection of gastric or GE junction adenocarcinoma remains undetermined. The present study aims to identify pre-treatment factors relevant to overall survival (OS), distant metastasis (DM) and Local-regional recurrence (LRR) in surgically resected gastric and gastroesophageal (GE) junction adenocarcinoma after adjuvant chemoradiotherapy (CRT). 
      方法: This is a secondary analysis of the NCT00052910 phase III study which randomized patients with gastroesophageal adenocarcinoma into adjuvant chemoradiotherapy with epirubicin, cisplatin, and fluorouracil versus adjuvant chemoradiotherapy with fluorouracil (FU) and leucovorin (LV) after curative resection. Univariate analysis of pre-treatment factors affecting overall survival was conducted by using Cox regression analysis. Factors with p<0.05 in the univariate analysis were then included in the multivariate analysis. The cumulative incidence rates of DM and LRR were calculated using the Fine and Gray’s competing risk analysis. LRR and death events were treated as a competing risk to calculate the cumulative incidence of DM, while DM and death events were treated as a competing risk when calculating the cumulative incidence of LRR. The effect of categorical variables on the cumulative incidence rate was examined by Gray’s test. A two-tailed P-value <0.05 was considered statistically significant.   

结果: A total of 546 patients from the phase III trial were included in the analysis. Median follow up was 6.5 year. Patients≤60 years comprised 53.8% while non-Hispanic race comprised 81.5%. 97.6% of the study patients have a performance score of 0-1. Male gender comprised 67.9% and 30.2% patients were diagnosed with Gastroesophageal junction (GE junction) adenocarcinoma. Median number of lymph node examined was 16 (range: 0-73). The following factors were prognostic predictors for OS in univariate analysis: No. of LN metastasis, histologic grade, gastroesophageal (GE) junction adenocarcinoma, T stage, N stage and M stage. Multivariate Cox regression analysis incorporating these six factors showed that GE junction (p=0.016), No. of LN metastasis (p=0.038), T4 stage (p=0.005) and N3 stage (p=0.049) predicted worse OS, while only pathological T3/4 stage was significantly correlated with distant metastasis (DM) and local regional recurrence (LRR) development in univariate and multivariate competing risk analyses. The 5-year cumulative incidence rates of DM were 0% for T1 stage, 10% for T2 stage, 27%for T3 stage and 14%for T4 stage, respectively. and the 5-year cumulative incidence rates of LRR were 5% for T1 stage, 14% for T2 stage, 27%for T3 stage and 40%for T4 stage；  

     结论:  Pathological T stage is the most important predictive factor for overall survival and treatment failure among resected gastric and GE junction adenocarcinoma patients after adjuvant CRT within a clinical trial. Lymph node metastasis and GE junction adenocarcinoma are also independent predictors for OS in this patient population.