Central Nervous System Relapses in Patients with Diffuse Large B Cell Lymphoma: Multicenter Retrospective Analysis in Korea

Abstract

Introduction: Central nervous system (CNS) involvement in patients with diffuse large B-cell lymphoma (DLBCL) is a serious complication with a poor prognosis. Secondary CNS involvements in DLBCL are divided into 2 types as follows: CNS involvement at the time of systemic relapse or progression, and isolated CNS relapse despite systemic remission. The frequency of secondary CNS involvement is reported as variable from 4% to 27% in patients with DLBCL who did not receive CNS prophylaxis. However, there is no clear consensus as to the subset of DLBCL benefits from CNS-directed preventive therapy. Furthermore, there is no general agreement on the intensity and type of therapy against secondary CNS involvement. Thus, we analyzed the clinical features and treatment outcomes of secondary CNS involvement in patients with DLBCL.Patients and methods: We analyzed 59 patients who diagnosed with DLBCL and had secondary CNS involvement. Secondary CNS involvement was defined as a brain parenchyma lesion or leptomeningeal involvement or combined during the follow-up or during the treatment.Results: 35 patients were initially treated with CHOP or CHOP-like regimens, and 24 patients with rituximab-CHOP (R-CHOP). The median age of the patients was 54 years old (range 17–77). 37 patients (62.7%) were presented as advanced stage (7 patients with stage III, 30 patients with stage IV), and 40 patients showed elevated serum LDH. Thus, 27 patients belonged to the high-intermediate or high risk of international prognostic index (IPI). The pattern of CNS involvement was as follows: 21 had lesions of brain parenchyma, 30 leptomeningeal involvements, and 8 combined lesions. There was no clinical variable predicting a pattern of CNS involvement. We dichotomized the patients based on the onset of CNS involvement: early CNS involvement occurred within one year from diagnosis (Early CNS group, 42 patients) and delayed CNS involvement after one year (Delayed CNS group, 17 patients). The early CNS group had more advanced stage of patients (26 patients with stage IV) than delayed CNS group (4 patients with stage IV). The presence of bone marrow invasion was also significantly associated with early CNS involvement: 18 patients in early CNS involvement and 1 patient in delayed CNS involvement. In addition, high or high-intermediate risk of IPI was related with early CNS involvement (P < 0.05). The response to the primary treatment against DLBCL was not related with the onset of CNS involvement, thus 46 patients showed objective response including 33 complete responses (CR). The use of rituximab was not associated with the onset of CNS involvement. However, the number of patients with partial response (PR) was higher in the early group (9/46, 19.6%) than delayed CNS group (2/17, 11.7%). Thus, these findings suggest patients with high tumor burden represented by advanced stage, high IPI, and bone marrow invasion might have CNS involvement at the time of diagnosis. However, the overall survival was not significantly different between these two groups, and the median overall survival of patients with secondary CNS involvement was 5.13 months (95% confidence interval: 3.39–6.87 months). When we compared survival based on the type of therapy in patients with isolated brain parenchyma involvement, combined treatment approach including systemic chemotherapy (high-dosage methotreaxate (MTX) chemotherapy) and localized therapy (intrathecal MTX chemotherapy and/or whole brain radiotherapy) showed a tendency of better survival than localized therapy alone. However, the addition of systemic chemotherapy was not translated into survival benefit in patients with isolated leptomeningeal CNS involvement. Furthermore, the role of systemic chemotherapy using high-dosage MTX was not proved in patients with systemic disease progression and secondary CNS involvement although it showed the tendency of better survival.Conclusions: Patients with advanced stage, high or high-intermediate risk of IPI, and the presence of bone marrow invasion might have a high risk of secondary CNS involvement even during the active treatment. Thus, CNS prophylaxis should be considered. However, further prospective study should be warranted to determine the intensity and type of treatment against secondary CNS involvement in DLBCL.