目的：探讨同种异体造血干细胞移植（allo-HSCT）在减轻中枢神经系统白血病（CNSL）不良预后方面的潜力，并评估预防性鞘内注射的意义。方法：对2012年9月至2018年3月在北京大学人民医院接受allo-HSCT的30例有CNSL病史的急性白血病患者（简称CNSL阳性组）进行回顾性队列分析。另外，选取同期接受allo-HSCT且无CNSL病史的90例急性白血病患者（简称CNSL阴性组），根据疾病类型、病情状态进行严格的1∶3匹配、移植型形成对照组。采用Kaplan-Meier分析比较两组患者的预后，并通过Cox比例风险模型识别CNSL移植后复发的高危因素。结果：CNSL阴性组患者的中位年龄显着高于CNSL阳性组患者（32岁vs.24岁，P=0.014）。两组之间的基线数据，包括性别、疾病类型、移植时疾病状态、供受者关系和人类白细胞抗原一致性，均未观察到显着差异。中位随访时间为 568 天（范围：21-1 852 天）。 4年累积复发率（71.4%±20.9% vs. 29.3%±11.5%，P=0.005）和CNSL移植后累积发生率（33.6%±9.2% vs. 1.2%±1.2%，P） <0.001) CNSL 阳性组显着高于 CNSL 阴性组。 CNSL阳性组的4年无白血病生存率显着低于CNSL阴性组（23.1%±17.0%比71.5%±11.6%，P<0.001）。但两组的4年累积移植相关死亡率和总生存率差异无统计学意义（均P>0.05）。多因素分析显示，移植前CNSL病史（HR=25.050，95%CI 3.072-204.300，P=0.003）被认为是移植后CNSL复发的高危因素。相反，半相合移植与移植后CNSL复发风险降低相关（HR=0.260，95%CI 0.073-0.900，P=0.034）。 CNSL阳性组中有7例患者在移植后接受了预防性鞘内治疗，其CNSL复发率显着低于移植后未接受鞘内治疗的23例患者（0/7 vs. 9/23，P= 0.048）。结论：有 CNSL 病史的患者复发风险较高，移植后无白血病生存期较差。预防性鞘内注射的使用显示出降低 CNSL 复发率的希望，尽管需要通过前瞻性研究进一步验证以证实这些观察结果。

Objective: To investigate the potential of allogeneic hematopoietic stem cell transplantation (allo-HSCT) in mitigating the adverse prognosis associated with central nervous system leukemia (CNSL) and to assess the significance of prophylactic intrathecal injection. Methods: A retrospective cohort analysis was conducted involving 30 patients with acute leukemia who had a history of CNSL who underwent allo-HSCT at Peking University People's Hospital between September 2012 and March 2018 (referred to as the CNSL-positive group). In addition, 90 patients with acute leukemia were selected from the same period who underwent allo-HSCT without a history of CNSL (referred to as the CNSL-negative group) and a rigorous 1∶3 matching was performed based on disease type, disease status, and transplantation type to form the control group. The prognosis between the two groups was compared using Kaplan-Meier analysis and the high-risk factors for CNSL relapse post-transplant were identified through Cox proportional-hazards model. Results: The median age of patients in the CNSL-negative group was significantly higher than that of patients in the CNSL-positive group (32 years vs. 24 years, P=0.014). No significant differences were observed in baseline data, including sex, disease type, disease status at transplantation, donor-recipient relationship, and human leukocyte antigen consistency between the two groups. The median follow-up time was 568 days (range: 21-1 852 days). The 4-year cumulative incidence of relapse (71.4%±20.9% vs. 29.3%±11.5%, P=0.005) and the cumulative incidence of CNSL post-transplant (33.6%±9.2% vs. 1.2%±1.2%, P<0.001) were significantly higher in the CNSL-positive group than in the CNSL-negative group. Furthermore, the 4-year leukemia-free survival rate in the CNSL-positive group was significantly lower than that in the CNSL-negative group (23.1%±17.0% vs. 71.5%±11.6%, P<0.001). However, no significant differences were observed in the 4-year cumulative transplant-related mortality and overall survival rates between the two groups (both P>0.05). Multivariate analysis revealed that a history of CNSL before transplantation (HR=25.050, 95%CI 3.072-204.300, P=0.003) was identified as high-risk factors for CNSL relapse post-transplant. Conversely, haploidentical transplantation was associated with a reduced risk of CNSL relapse post-transplant (HR=0.260, 95%CI 0.073-0.900, P=0.034). Within the CNSL-positive group, seven patients received prophylactic intrathecal therapy after transplantation, and their CNSL relapse rate was significantly lower than that of the 23 patients who did not receive intrathecal therapy after transplantation (0/7 vs. 9/23, P=0.048). Conclusions: Patients with a history of CNSL have a higher risk of relapse and experience poorer leukemia-free survival following transplantation. The use of prophylactic intrathecal injection shows promise in mitigating CNSL relapse rates, although further validation through prospective studies is necessary to substantiate these observations.