Cox模型在肿瘤生存资料中的应用

APPLICATION OF COX PROPORTIONAL HAZARD MODEL IN SURVIVAL DATA ANALYSIS OF TUMOR

目的探讨Cox比例风险模型在肿瘤生存资料中的应用及其意义。方法对1014例食管癌切除术患者资料进行研究。选择13个可能对食管癌切除术预后产生影响的因素,通过Cox模型对食管癌切除术患者预后进行多因素分析。根据预后指数(PrognosticIndex,PI)的大小将患者分组,建立其术后生存预测模型。统计分析采用计算机处理。结果全组生存率5年45.9%,10年39.3%。影响预后的主要因素是淋巴结转移、肿瘤分期、肿瘤侵及深度、部位和组织类型(P<0.0001)。PI值小的患者预后较好。结论利用Cox比例风险模型可以分析影响肿瘤患者预后的重要因素,可利用预测模型。

ObjectiveTo study the use and significance of Cox proportional hazard model in the survival data analysis of tumor. MethodsOne thousand and fourteen patients undergoing resection for esophageal cancer, who did not die within 30 days after operation, were included in this study. Thirteen possible factors influencing survival for these patients were selected. A multivariate analysis was performed by Cox proportional hazard model. According to the Prognostic Index (PI), all patients were divided into seven risk groups, and their predicting models of survival rates were established in groups. Statistical analysis was carried out with computer. ResultsThe over all cumulative survival rate was 45.9 % at 5 years and 39.3 % at 10 years. The significant prognostic factors influencing survival of these patients were metastases of lymph nodes ,tumor stage,invasion depth, tumor location and histological type (P< 0.0001 ). The patients with smaller value of PI had a better prognosis. ConclusionThe significant prognostic factors can be screened out by Cox proportional hazard model, and the survival probability can be predicted according to the predicting model of survival rates.