非小细胞肺癌脑转移放射治疗的预后因素分析

Prognostic predictors for non-small cell lung cancer patients with brain metastasis after radiotherapy

背景与目的脑转移最常见于肺癌患者,放射治疗是其有效的常规治疗手段。目前的治疗以姑息止痛和减轻症状为目的,方法差异较大,对肺癌脑转移预后的认识也不尽相同。本研究旨在分析非小细胞肺癌脑转移患者放射治疗的疗效以及影响生存率的预后因素并建立预后模型。方法回顾性分析1995年2月-2006年10月在我科接受全脑放射治疗的82例非小细胞肺癌脑转移患者的临床资料,采用Kaplan-Meier法计算生存率,各亚组生存率差别的比较采用log-rank时序检验,分析影响生存预后的各种因素(单因素和多因素分析),以Cox回归分析获得的有统计学意义的因子组成预后模型。结果随访时间为1-120个月,全组患者诊断脑转移后的中位生存期为10.5个月,6个月、1年和2年的总生存率分别为50.8%、23.7%和5.1%。单因素和多因素分析均显示放射治疗前KPS评分、肺原发肿瘤控制情况、确诊肺癌到脑转移间隔时间、脑转移灶数目是影响生存的独立预后因素,根据上述因素建立预后指数将82例患者分成预后好、中和差三组,其生存曲线的差异有统计学意义(P<0.001)。结论KPS评分、确诊肺癌到脑转移间隔时间、肺原发肿瘤控制情况、脑转移灶数目是影响非小细胞肺癌脑转移放射治疗生存的独立预后因素,预后指数模型能很好地反映预后。

Background and objective Brain metastasis （BM） is often found in the patients with lung cancer. Radiotherapy is regular and effective means of therapy and it aims at palliating symptoms and prolonging survival time. However, now there are different viewpoints on protocols of radiotherapy and prognostic factors. A retrospective analysis is used to evaluate the results of treatment for 82 cases with brain metastasis from non-small cell lung cancer （NSCLC） and explore the prognostic factors to establish a prognostic index （PI） model. Methods From Feb.1995 to Oct. 2006, 82 patients irradiated for BM from NSCLC, with both complete medical charts and follow-up data available, were eligible for this retrospective analysis. A number of potential factors which might affect prognosis after irradiation were evaluated. The significance of prognostic variables in the survival resulted from both univariate analysis by Kaplan-Meier combining with log-rank test and multivariate Cox regression model. The prognostic index （PI） was established based on Cox regression analysis and subgrouping values. Results The follow-up time was 1-120 months. For the entire cohort, the median survival from the start of radiation for BM was 10.5 months, and the actuarial overall survival rate was 50.8%, 23.7% and 5.1% at 0.5, 1 and 2 years respectively. Univariate analysis showed KPS, control of primary tumor, interval from the beginning of diagnostic to BM, extracranial systemic metastasis, counts of lymphocyte and solitary BM were predictors of prognosis. However, in the Cox multivariate analysis, only KPS, control of primary tumor, interval from the beginning of diagnostic to BM and solitary BM were significant prognostic factors. The prognostic index was established based on Cox regression analysis and 82 patients were stratified good, intermediate and poor prognostic sub-groups. The difference of survival rate among 3 subgroups is significant （P〈0.001）. Conclusion Radiotherapy is proved to be effective for NSCLC patients with BM. KPS, control of primary tumor, interval from the beginning of diagnostic to BM and solitary BM are independent prognostic factors. PI model can well predict the prognostic of patients with BM from NSCLC.