Introduction: Secondary CNS lymphoma (SCNSL) is observed in 2-5% of patients with diffuse large B cell lymphoma (DLBCL). Progression of SCNSL after first line CNS-directed systemic therapy is particularly challenging to treat, without a clear standard of care. Brain radiotherapy (RT) is used for progressive SCNSL in either the chemo-refractory or palliative setting. Historically, the RT field included the whole brain due to concerns about diffuse intracranial involvement. However, treatment fields limited to just involved site(s) (partial brain RT) may provide a favorable balance of disease control and reduced treatment toxicity, particularly in an era of improved CNS-directed systemic therapies. We hypothesized that RT would confer a high rate of intracranial response among patients with chemo-refractory SCNSL. Methods: We identified DLBCL patients with SCNSL who received brain RT at our institution for radiographic or clinical progression after at least 1 line of SCNSL-directed therapy. SCNSL-directed therapy was defined as systemic therapy for active SCNSL (as opposed to prophylaxis). Overall survival (OS) was determined from RT start to death from any cause using the Kaplan-Meier method. OS comparing subsequent therapy was landmarked 2 months post RT to minimize immortal time bias. RT response was evaluated with contrast MRI or CT using International Primary CNS Lymphoma Collaborative Group (IPCG) or Response Assessment in Neuro-oncology (RANO) radiographic criteria for parenchymal or leptomeningeal disease (LMD), respectively. LMD was assessed by cytology and/or imaging. Results: Between 1999-2023, 55 patients received brain RT for progressive SCNSL with median follow up of 3.2 months. At the time of RT, median age was 64 (interquartile range (IQR) 49 - 71), median Karnofsky Performance Status was 70 (IQR, 60 - 80), and 85% had neurologic symptoms. Patients had received a median of 2 lines of SCNSL-directed therapy prior to RT (IQR, 1 - 3), 91% received a methotrexate-based regimen with median interval of 21 days (IQR, 10 - 67) between methotrexate and RT. Median RT dose was 30 Gy in 10 fractions delivered to whole brain (n = 44), partial brain (n = 9), or craniospinal axis (n = 2). 38 patients were evaluable for response; the remainder were lost to follow up/hospice (n = 8), lacked suitable imaging (n = 4), died with progressive neurologic symptoms (n = 4) or other cause (n = 1). The overall response rate was 79%, including a 24% complete response rate. Median time to best response was 60 days (IQR, 30 - 145). 32/55 patients (58%) were successfully bridged to subsequent systemic therapy (n = 22), hematopoietic stem cell transplantation (autologous n = 4, allogeneic n = 1), or CAR T (n = 5), after a median of 34 days from RT (IQR, 20 - 57). Median OS for the entire cohort was 3.5 months, and was improved for those who were bridged to subsequent therapy (no subsequent therapy 0.7 months, subsequent systemic therapy 4.0 months, subsequent cell therapy 8.9 months, p = 0.006). Among patients who received partial brain RT, 7/7 evaluable patients had at least a partial response and 7/9 patients were successfully bridged to subsequent therapy. Conclusion: Patients referred to RT for progressive SCNSL represent a high-risk cohort characterized by rapidly progressive, symptomatic, treatment-resistant disease. RT confers a high response rate in this chemo-resistant population, and can facilitate bridging to subsequent therapy. Preliminary data suggest that partial brain radiotherapy may be an effective bridging strategy among select patients, however a larger sample size will be required to determine this.
Read full abstract