经倾向指数匹配后的GP方案与NP方案治疗非小细胞肺癌的疗效评价

Effect Evaluation between GP and NP Regimens in Treatment of Locally Advanced Non-small Cell Lung Cancer Based on Propensity Score Matching

【提要】目的将非小细胞肺癌患者分为吉西他滨联合顺铂（GP）方案和长春瑞滨联合顺铂（NP）方案两组，用倾向指数匹配法均衡组问的协变量，评价两种治疗方案的效果。方法以分组变量为因变量，协变量为白变量建立logistic回归模型，计算非小细胞肺癌病人的倾向指数，然后按照倾向指数进行组间卡钳匹配，对匹配后的数据进行生存分析。结果匹配前GP组和NP组的中位生存期分别为2．12年和2．18年；其置信区间分别为（1．76～2．51）年和（1．82～2．54）年：GP组和NP组的1年生存率分别为83．98％和85．25％；2年生存率分别为55．68％和58．13％；3年生存率分别为23．47％和24．99％。Logrank检验两组生存曲线间没有统计学差异（P〉0．05）。匹配后中位生存时间分别为1．96年和2．31年；其置信区间分别为（1．56～2．36）年及（1．87～2．66）年；GP组和NP组的1年生存率分别为82．83％和85．23％；2年生存率分别为49．29％和58．04％；3年生存率分别为23．18％和25．26％。Logrank检验结果表明两组生存曲线之间差别无统计学意义（P〉0．05）。结论采用倾向指数匹配法能有效的降低混杂偏倚，GP和NP两种化疗方案均可作为NSCLC的一线治疗方案。

Objective To evaluate the effects between combination of gemcitabine and cisplatin（GP） and combination of vinorelbine and cisplatin（NP） on two groups of non-small cell lung cancer patients after balancing the covariates by propensity score. Methods The logistic regression model was established where the treatment assignment was taken as the dependent varia- ble and the covariates were taken as the independent variables. For each NSCLC patient, the propensity score was calculated for caliper matching and a survival analysis of the matched data were carded out. Results Before matching, the median survival time（MST） of GP and NP were 2. 12 and 2. lg years,and the confidence interval（CI） were 1.76 -2. 51 years and 1.82 -2. 54 years. One-,two-,and three-year survival rates were 83.98% and 85.25% ,55.68% and 58.13% ,23.47% and 24. 99% respec- tively. The Logrank test showed that there was no significant differences in the overall survival rate between the two groups（ P 〉 0. 05 ）. After matching, the MST of GP and NP were 1.96 and 2. 31 years, and the CI were 1.56 - 2. 36 years and 1.87 - 2. 66 years. One-,two-,and three-year survival rates were 82. 83% and 85.23% ,49. 29% and 58. 04% ,23.18% and 25.26% respec- tively. The Logrank test showed that there was no significant differences in the overall survival rate between the two groups（ P 〉 0. 05 ）. Conclusion Propensity score matching can effectively reduce the confounding bias of non-randomized clinical observa- tional data. Both GP and NP regimens could be used as the original chemotherapy for NSCLC.