目的：探讨具有阳性驱动基因的中枢神经系统（CNS）转移非小细胞肺癌（NSCLC）患者的预后因素。方法：回顾性分析了2016年1月至2020年12月在北京同仁医院接受治疗的103名CNS转移NSCLC患者的临床数据，将患者分为阳性驱动基因组（患有驱动基因突变并接受相应的靶向治疗）和阴性驱动基因组。进行Cox单因素和多因素回归分析以确定影响患者预后的因素，并使用接受者操作特征曲线（ROC）比较4个评分系统（递归分区分析（RPA）类别，诊断特异性分级预测评估（DS-GPA）指数，脑转移基础评分（BS-BM）和肺分子分级预测评估（肺－mol GPA））对患者预后的预测能力。结果：103名患者中，男性48名，女性55名，年龄为（64.6±9.7）岁。103名患者的中位生存时间为24.0（95%CI：20.0-28.0）个月，阳性驱动基因组59名患者的中位生存时间为33.0（95%CI：23.4-42.6）个月，阴性驱动基因组44名患者的中位生存时间为17.0（95%CI：14.4-19.6）个月，两者差异有显著性（χ2=24.69，P<0.001）。Cox多因素分析的结果显示，阴性驱动基因（HR=3.788，95%CI：1.951-7.301，P=0.001）、Karnofsky表现状况（KPS）评分<70（HR=2.613，95%CI：1.185-5.761，P=0.017）和中性粒细胞/淋巴细胞比值（NLR）>3.22（HR=2.714，95%CI：1.157-6.365，P=0.022）是影响CNS转移NSCLC患者预后的独立危险因素。KPS评分<70（HR=3.719，95%CI：1.165-11.876，P=0.027）和未接受放疗（HR=2.032，95%CI：1.033-11.364，P=0.041）是影响具有阳性驱动基因的CNS转移NSCLC患者预后的独立危险因素。ROC曲线分析显示，肺-mol GPA在4个评分系统中的曲线下面积（AUC）值最高（AUC=0.843，95%CI：0.731-0.956，P<0.001），肺-mol GPA联合评分系统的AUC值（AUC=0.904，95%CI：0.816-0.991，P<0.001）高于肺-mol GPA。结论：KPS评分低和未进行颅内放疗是CNS转移NSCLC具有阳性驱动基因患者预后的独立危险因素；肺-mol GPA联合评分系统有助于CNS转移NSCLC具有阳性驱动基因患者的预后评估。

Objective: To investigate the prognostic factors of patients with central nervous system (CNS) metastatic non-small cell lung cancer (NSCLC) with positive driver genes. Methods: The clinical data of 103 patients with CNS metastatic NSCLC admitted to Beijing Tongren Hospital from January 2016 to December 2020 were retrospectively analyzed, and the patients were divided into positive driver gene group (patients with driver genes mutation and receiving corresponding targeted therapy) and negative driver gene group. Cox univariate and multivariate regression analyses were performed to identify the factors affecting patients' prognosis, and the receiver operating characteristic curve (ROC) was used to compare the predictive ability of 4 scoring systems [recursive partitioning analysis (RPA) classes, diagnosis-specific graded prognostic assessment (DS-GPA) index, basic score for brain metastasesn (BS-BM) and (lung-molecular graded prognostic assessment (lung-mol GPA)]on patients' prognosis. Results: Among the 103 patients, 48 were males and 55 were females, and aged (64.6±9.7) years old. The median survival time of the 103 patients was 24.0 (95%CI: 20.0-28.0) months, the median survival time of the 59 patients in the positive driver gene group was 33.0 (95%CI: 23.4-42.6) months, the median survival time of the 44 patients in the negative driver gene group was 17.0 (95%CI: 14.4-19.6) months, and the difference was statistically significant (χ2=24.69, P<0.001). The results of Cox multivariate analysis showed that the negative driver genes (HR=3.788, 95%CI: 1.951-7.301, P=0.001), Karnofsky performance status (KPS) score<70 (HR=2.613, 95%CI: 1.185-5.761, P=0.017) and neutrophil-to-lymphocyte ratio (NLR)>3.22 (HR=2.714, 95%CI: 1.157-6.365, P=0.022) were independent risk factors affecting the prognosis of patients with CNS metastatic NSCLC. KPS score<70 (HR=3.719, 95%CI: 1.165-11.876, P=0.027) and no radiotherapy (HR=2.032, 95%CI: 1.033-11.364, P=0.041) were independent risk factors affecting the prognosis of patients with CNS metastatic NSCLC with positive driver genes. ROC curve analysis showed that the area under curve (AUC) value of lung-mol GPA was the highest among the 4 scoring systems (AUC=0.843, 95%CI: 0.731-0.956, P<0.001), and the AUC value of the lung-mol GPA combined scoring system (AUC=0.904, 95%CI: 0.816-0.991, P<0.001) was higher than lung-mol GPA. Conclusions: A low KPS score and no cranial radiation therapy are independent risk factors for the prognosis of patients with CNS metastatic NSCLC with positive driver genes; the lung-mol GPA joint scoring system is more conducive to the prognostic assessment of patients with CNS metastatic NSCLC with positive driver genes.