可切除肺癌预后预测模型的构建及生存分析

目的 探讨构建的可切除肺癌预后预测模型在患者生存及预后预测中的价值。方法 选择山西省肿瘤医院2007年1月至2018年9月原发性肺癌患者2 267例,患者均行一次肺癌手术治疗,无第二原发肿瘤。选取性别、年龄、职业、肿瘤部位、病理类型、手术路径、手术方式、肿瘤分期、治疗方案为预后影响因素。采用Cox比例风险模型构建预后指数(PI)方程,计算每例患者的PI值。根据PI值的不同范围,划分低、中、高危预后组,对各组生存情况进行评估。结果 性别(RR=0.684,P=0.001)、年龄(RR=0.591,P<0.01)、职业(RR=1.439,P=0.001)、病理类型(RR=3.694,P<0.01)、手术路径(RR=0.734,P=0.001)、肿瘤分期(RR=0.352,P=0.007)为可切除肺癌患者预后独立影响因素。其中,女性、≤65岁、胸腔镜手术、肿瘤分期Ⅰ期为预后保护因素,其预后不良风险分别降低31.6%、40.9%、26.6%、64.8%;农民、腺鳞癌为预后危险因素,其预后不良风险分别增加43.9%、269.4%。PI方程为:∑βixi=-0.380 X1-0.526 X2+0.364 X31+1.307 X55-0.309 X6-1.045 X81(X1代表性别,X2代表年龄,X31代表职业为农民,X55代表病理类型为腺鳞癌,X6代表手术路径,X81代表肿瘤分期Ⅰ期)。PI<-1为低危组,PI≥-1且≤-0.5为中危组,PI>-0.5为高危组。1、3、5年生存率低危组分别为96.8%、87.0%、77.9%,中危组分别为91.8%、82.2%、61.7%,高危组分别为86.5%、61.7%、50.3%,各组间生存率差异具有统计学意义(P<0.05)。结论 可切除肺癌预后预测模型能够预测可切除肺癌患者的预后风险及相应生存率,帮助临床医师评估预后及制订后续治疗方案。

Construction of prognostic prediction model of resectable lung cancer and survival analysis

Objective To explore the value of the constructed prognostic prediction model of resectable lung cancer in predicting the survival and prognosis of patients.Methods A total of 2267 patients with primary lung cancer in Shanxi Provincial Cancer Hospital from January 2007 to September 2018 were selected.All patients underwent primary lung cancer surgery without a second primary tumor.Gender,age,occupation,tumor site,pathological type,surgical path,surgical method,tumor stage and treatment were selected as the prognostic factors.A Cox proportional hazard model was used to construct a prognostic index(PI)equation to calculate the PI value of each patient.According to the different ranges of PI values,the low-,intermediate-and high-risk prognosis groups were divided,and the survival status of three groups were evaluated.Results Gender(RR=0.684,P=0.001),age(RR=0.591,P<0.01),occupation(RR=1.439,P=0.001),pathological type(RR=3.694,P<0.01),surgical path(RR=0.734,P=0.001),tumor stage(RR=0.352,P=0.007)were independent factors affecting the prognosis of patients with resectable lung cancer.Female,≤65 years old,thoracoscopic surgery,and tumor stageⅠwere prognostic protective factors,and their risks of poor prognosis were reduced by 31.6%,40.9%,26.6%,and 64.8%,respectively.Farmer and adenosquamous carcinoma were prognostic risk factors,and their risks of poor prognosis were increased by 43.9%and 269.4%,respectively.The PI equation was∑βixi=-0.380 X1-0.526 X2+0.364 X31+1.307 X55-0.309 X6-1.045 X81(X1 was the gender,X2 was the age,X31 was the occupation as a farmer,X55 was the pathological type of adenosquamous carcinoma,X6 was the surgical path,X81 was the tumor stageⅠ).PI<-1 was the low-risk group,PI≥-1 and≤-0.5 was the intermediate-risk group,PI>-0.5 was the high-risk group,and the differences of their survival rates were statistically significant(P<0.05).The 1-,3-,and 5-year survival rates for the low-,risk groups were 96.8%,87.0%and 77.9%;the intermediate-risk group were 91.8%,82.2%and 61.7%;the high-risk group were 86.5%,61.7%and 50.3%.respectively.Conclusion The prognostic prediction model of resectable lung cancer can predict the prognosis risk and the corresponding survival rate of patients with resectable lung cancer,and it can help clinicians to evaluate the prognosis and formulate subsequent treatment plans.