应用竞争风险模型探讨睾丸癌患者的预后影响因素

To investigate the prognostic factors of testicular cancer patients by competitive risk model

目的:分别应用竞争风险模型及经典Cox比例风险模型对睾丸癌患者的预后风险因素进行分析,并对筛选结果进行评价。方法:通过SEER数据库提取1973年至2015年睾丸癌患者数据信息,关心事件为患者发生睾丸癌特异性死亡,竞争事件为患者死于其他原因。运用累计死亡率(CIF)方法估算患者的总体风险率,采用Cox比例风险模型和竞争风险模型(原因别风险模型、Fine-Gray模型)进行预后风险多因素分析,并对筛选出的预后风险因素进行评价。结果:共有943例符合条件的患者纳入分析,114例患者因睾丸癌特异性死亡。单因素预后风险分析表明,手术、化疗、甲胎蛋白(AFP)、绒毛膜促性腺激素(HCG)和乳酸脱氢酶(LDH)对睾丸癌患者的癌症特异性死亡有统计学意义。多因素Cox回归分析显示,手术、化疗、AFP、HCG和LDH是影响患者生存的独立预后因素。竞争风险模型结果显示,手术、化疗、AFP、HCG是影响患者生存的独立预后因素,且预后因素的分层有显著差异(P<0.05)。结论:手术、化疗、AFP、HCG是影响患者生存的独立预后因素。当生存分析存在竞争风险时,竞争风险模型比Cox比例风险模型得出的结论更为客观合理。

Objective: To analyze the prognostic risk factors of testicular cancer patients with competitive risk model and classical Cox proportional risk model, and to evaluate the screening results. Methods: Data of patients with testicular cancer from 1973 to 2015 were extracted from THE SEER database. The concern event was the occurrence of testicular cancer-specific death, and the competitive event was the death from other causes. The cumulative mortality(CIF) method was used to estimate the overall risk rate of patients. Cox proportional risk model and competitive risk model(cause-specific risk model, Fine-Gray model) were used to analyze the prognostic risk factors, and the prognostic risk factors were evaluated. Results: Results: A total of 943 eligible patients were included in the analysis, of which 114 died from testicular cancer-specific causes. Univariate prognostic risk analysis showed that surgery, chemotherapy, alpha fetoprotein(AFP), chorionic gonadotropin(HCG), and lactate dehydrogenase(LDH) were statistically significant for cancer specific death in patients with testicular cancer. Multivariate Cox regression analysis showed that surgery, chemotherapy, AFP, HCG and LDH were independent prognostic factors affecting the survival of patients. The results of the competitive risk model showed that surgery, chemotherapy, AFP and HCG were independent prognostic factors affecting the survival of patients. 0.05. Conclusion: Surgery, chemotherapy, AFP and HCG are independent prognostic factors affecting the survival of patients. When there is competition risk in survival analysis, the conclusion of competition risk model is more objective and reasonable than Cox proportional risk model.