目的：原發性中樞神經系統淋巴瘤是一種少見的疾病。因為病人數少和資料缺乏，要做大型的研究非常困難，所以我們採用統合分析的方式，試著找出原發性中樞神經系統淋巴瘤的理想治療方法和其預後因子。 材料與方法：利用PUBMED和MEDLINE，從2003年9月之前發表的文章中，找出標題含有“中樞神經系統淋巴瘤”或“中樞神經系統非何杰金氏淋巴瘤”的論文。總共找到16篇文章的標題含有上述2項其中之一，但只有其中4篇（85位病人）符合篩選條件而選進研究中。治療方式包括只有放射線治療、化學治療其中之一，或合併放射線和化學治療。對於病人治療後初始反應和預後的分析，本研究分析的變因包括：年齡、性別、治療方式等，而且採用單一變項分析和多變項分析。單一變項分析是採用log rank test，而多變項分析是採用Cox proportional hazards model，存活率是以Kaplan-Meier計算。 結果：6位病人（7.1%）只接受放射線治療，38位病人（44.7%）只接受化學治療，41位病人（48.2%）接受合併放射線和化學治療。在存活率方面，3年和5年整體存活率（overall survival）分別為47.8%和28.9%。三年和五年的疾病相關存活率（disease-specific survival）分別為 49.1%和34.0%。在單一變項分析中，發現年齡（p = 0.0129）、是否有施行化學治療（p = 0.0141）及化學治療是否有 3 次以上（p = 0.0003）對整體存活率是重要的預後因子。另外對於疾病相關存活率，是否有施行化學治療（p = 0.0077）及化學治療是否有3次以上（p = 0.0001）是重要的預後因子。在多變項分析中，化學治療是否有 3 次以上是唯一對於整體存活率和疾病相關存活率有意義的預後因子（p = 0.018, 0.007）。 結論：在本研究中，接受3次以上化學治療的病人其存活率較佳，故建議在原發性中樞神經系統淋巴瘤的治療中，應包含3次以上的化學治療。

Purpose : Primary CNS lymphoma is a rare disease. Owing to small number of patients and lack of data, large-series studies become difficult. We use a meta-analysis to evaluate the treatment of primary CNS lymphoma and prognostic factor. Materials and Methods : Articles published before September 2003 were searched through PUBMED (National Library of Medicine) and MEDLINE and by the terms of 烢rimary CNS lymphoma? 烋on-Hodgkin掇 lymphoma of CNS? Sixteen potentially suitable articles were found. Among them, four articles matched the selective criteria and were collected into this meta-analysis. There were 85 patient entered the study. The treatment included single and combined modalities of radiotherapy and chemotherapy. The variables, such as age, sex, methods of treatment were analyzed about their influence upon initial response and prognosis. Univariate and multivariate analyses were used. Univariate analysis using the log rank test and multivariate analysis using the Cox proportional hazards model were performed. Overall survival and disease-specific survival were calculated with the method of Kaplan and Meier. Results : Forty-one patients (48.2%) were male, and 44 patients (51.8%) were female. The median age was 62 years. Six patients (7.1%) received radiotherapy alone. Thirty-eight patients (44.7%) received chemotherapy alone. Forty-one patients (48.2%) received combination of radiotherapy and chemotherapy. The rate of initial response were 82.4%. During the time of following up, 35 patients (41.2%) were alive, and 50 patients (58.8%) were expired. The 3-year and 5-year overall survival rates were 47.8% and 28.9%, respectively. The 3-year and 5-year disease-specific survival rates were 49.1% and 34.0%, respectively. In the univariate analysis, age was a significant prognostic factor of overall survival rate (p = 0.0129). Chemotherapy was a significant prognostic factor of overall survival rate and disease-specific survival rate (p = 0.0141, 0.0077 respectively). The number of cycle of chemotherapy was another significant prognostic factor of overall survival rate and disease-specific survival rate (p = 0.0003, 0.0001 respectively). In the multivariate analysis, the number of cycle of chemotherapy is the only significant prognostic factor of overall survival rate and disease-specific survival rate (p = 0.018, 0.007 respectively). Conclusion : In this study, patients who received chemotherapy with 3 cycles or more had better survival outcome. Chemotherapy with 3 cycles or more is recommended in treatment of primary CNS lymphoma.