Objectives: To explore the prognostic factors of primary central nervous system lymphoma(PCNSL) and to analyze the efficacy of different treatment methods. Methods: Clinical data of 4 812 patients with PCNSL in SEER database from January 1975 to December 2016 were retrospectively analyzed.Among them, 2 831 were male and 1 981 were female, the ratio of male to female was 1.4∶1.0.There were 2 236 cases(46.47%) under 60 years old, 1 718 cases(35.70%) aged 60 to 74 years old, and 858 cases(17.83%) aged 75 years old or above. Two thousand four hundred and seventeen cases(50.23%) had supratentorial tumors, 299 cases (6.21%) had infratentorial tumors, and 554 cases(11.51%) had multiple brain tumors, 1 542 cases (32.04%) were other or unspecified location.Three thousand five hundred and thirteen cases(73.00%) had diffuse large B-cell lymphoma (DLBCL), 234 cases(4.86%) had non DLBCL, 1 065 cases (22.13%) had other or unspecified types of tumor.The treatment included 2 011 cases (41.77%) of biopsy, 61 cases (1.27%) of subtotal resection(STR), 54 cases (1.12%) of gross total resection(GTR), 2 384 cases (49.54%) of biopsy and chemotherapy, 159 cases (3.30%) of STR and chemotherapy, 144 cases (3.00%) of GTR and chemotherapy.Univariate and multivariate Cox regression models were used to analyze the prognostic factors affecting the overall survival of the patients.Fine-Gray test and competitive risk model were used to analyze the prognostic factors affecting cancer-specific survival.Kaplan-Meier method and Log-rank test was used for survival analysis. Results: Univariate and multivariate Cox regression analyses showed that age, race, marital status, tumor site, pathological subtype, surgery, chemotherapy, combined with other malignant tumors, and HIV infection were the independent prognostic factors affecting the overall survival of PCNSL patients.The results of Fine-Gray test and competitive risk model analyses showed that age, race, marital status, tumor location, pathological subtype, surgical method, chemotherapy, combined with other malignant tumors, and HIV infection were independent prognostic factors affecting cancer-specific survival, while gender and radiotherapy had no significant correlation with cancer-specific survival.Compared with biopsy, PCNSL patients may benefit from surgical resection (STR:HR=0.805, 95%CI:0.656‒0.989, P=0.04; GTR:HR=0.521, 95%CI:0.414‒0.656, P<0.01).Kaplan-Meier survival analysis showed that the median survival time of biopsy+chemotherapy group was 28 months (95%CI:24.497‒31.503), 2 months (95%CI:1.756‒2.244) in the biopsy group, 2 months (95%CI:1.410-2.590) in the STR group, 19 months (95%CI:0‒39.311) in the biopsy+chemotherapy group, 67 months (95%CI:46.187-87.813) in the STR+chemotherapy group, 84 months (95%CI:57.448‒110.552) in the GTR+chemotherapy group.The median survival time of patients with different treatment methods was statistically significant (P<0.01). Conclusions: Surgical resection may improve the prognosis of some PCNSL patients.Patients who have access to receive GTR or STR combined with chemotherapy may have prolonged Cancer-specific survival.
目的： 探讨原发性中枢神经系统淋巴瘤（PCNSL）患者的预后因素及不同治疗方法的临床效果。 方法： 回顾性分析1975年1月至2016年12月美国SEER数据库中4 812例PCNSL患者的临床资料。其中男性2 831例，女性1 981例，男女比例为1.4∶1.0；发病年龄<60岁者2 236例（46.47%），60~<75岁者1 718例（35.70%），≥75岁者858例（17.83%）；幕上肿瘤2 417例（50.23%），幕下肿瘤299例（6.21%），脑内多发肿瘤554例（11.51%），其他或未指明部位的脑部肿瘤1 542例（32.04%）；弥漫大B细胞淋巴瘤（DLBCL）3 513例（73.00%），非DLBCL 234例（4.86%），其他或未指明类型的淋巴瘤1 065例（22.13%）；治疗方式为单独活检的2 010例（41.77%），单独部分切除61例（1.27%），单独完全切除54例（1.12%），活检+化疗2 384例（49.54%），部分切除+化疗159例（3.30%），完全切除+化疗144例（2.99%）。采用单因素及多因素Cox回归模型分析影响患者总体生存的预后因素；采用Fine-Gray检验与竞争风险模型研究影响患者肿瘤特异性生存的预后因素；生存分析采用Kaplan-Meier法，并通过Log-rank进行比较。 结果： 单因素及多因素Cox风险回归模型分析结果显示，影响PCNSL患者总体生存的独立预后因素包括年龄、种族、婚姻状态、肿瘤部位、病理学亚型、手术、化疗、合并其他恶性肿瘤、合并HIV感染等。Fine-Gray检验与竞争风险模型分析结果显示，影响患者肿瘤特异生存的独立预后因素包括年龄、种族、婚姻状态、肿瘤部位、病理学亚型、手术方式、化疗、合并其他恶性肿瘤、合并HIV感染，而性别与放疗均与肿瘤特异生存无明显相关性。与活检相比，PCNSL患者可能从手术切除中获益（部分切除：HR =0.805，95%CI：0.656~0.989，P=0.04；完全切除：HR=0.521，95%CI：0.414~0.656，P<0.01）。Kaplan-Meier生存分析结果显示，活检+化疗组中位生存时间为28个月（95%CI：24.497~31.503），单独活检组为2个月（95%CI：1.756~2.244），单独部分切除组为2个月（95%CI：1.410~2.590），单独完全切除组为19个月（95%CI：0~39.311），部分切除+化疗组为67个月（95%CI：46.187~87.813），完全切除+化疗组为84个月（95%CI：57.448~110.552），使用不同治疗方法的患者的中位生存时间的差异有统计学意义（P<0.01）。 结论： 手术切除可能改善部分PCNSL患者的预后。化疗可能使肿瘤完全切除或部分切除患者的肿瘤特异生存时间延长。.
Keywords: Cancer-specific survival; Central nervous system; Lymphoma; Overall survival.