Clinical Characteristics and Outcomes of Patients with Hodgkin Lymphoma with Central Nervous System Involvement: An International Multicenter Collaboration

Introduction In comparison to aggressive NHL, CNS involvement is exceedingly rare in patients (pts) with Hodgkin lymphoma (HL). Thus, the clinical features and outcomes are not well described. Patients and Methods For this international retrospective analysis, institutional (n =4), national (n =2) and cooperative group (n =1) databases were reviewed for pts with classical HL who developed histologically proven CNS involvement at any stage during their disease course. We included pts diagnosed with CNS involvement after January 1, 1995 and collected clinicopathological characteristics (at both initial diagnosis and at time of CNS involvement), treatment details and outcomes. Progression-free survival (PFS) and overall survival (OS) were calculated using the Kaplan-Meier method from date of CNS involvement to date of disease progression or death from any cause and death from any cause, respectively. Results We screened 32,047 patients with classical HL in our combined databases and identified 18 pts with histologically confirmed CNS lymphoma, with an estimated crude incidence of 0.05%. These pts had a median age of 40 (range 20 - 84) years at the time of diagnosis of CNS involvement. The characteristics at the time of initial diagnosis of HL are summarized in table 1. CNS involvement was present at the time of initial diagnosis in 10 pts (56%) of whom 3 had isolated CNS involvement and 5 had concurrent systemic involvement (details were unavailable in 2 pts). CNS HL was a feature of relapsed/refractory disease in 8 (44%), 7 of whom had concurrent systemic involvement. The most common presenting symptoms were pyramidal weakness (n=7, 39%), headache (n=5, 28%), sensory change (n=4, 22%) and confusion (n=3, 17%). Lesions were parenchymal (n=10, 62%), leptomeningeal (n=2, 12%) or both (n=4, 25%). The median number of discrete CNS lesions detected by imaging was 1 (range 0 - >10), which affected cortical (n=9, 50%) or subcortical structures (n=6, 33%) and the spinal cord (n=2, 11%). The disease appeared to arise from the dura in 9 cases (50%), and the median size of the largest lesion was 2.5 cm (range 1 - 6). Cerebrospinal fluid (CSF) analysis was performed in 10 (56%) pts; of these, cytology was positive in 4 cases with a median CSF white cell count of 10/uL (range 0 - 136). CSF protein was elevated in 7 cases. Immunohistochemistry for Epstein-Barr virus RNA was done on 6 CNS biopsies, of which 3 were positive. Information regarding therapy was available in 15 pts; 9 received radiotherapy (alone (n=4), with steroids (n=1), with systemic (n=2) or intrathecal chemotherapy (n=2)). A further 3 pts were treated with systemic therapy (alternating cycles of R-IVAC/MTX (n=1), brentuximab vedotin (n=1), MTX, cytarabine and thiotepa (n=1), 2 with steroids alone and 1 with surgical resection alone. Of 13 pts formally evaluated for response, 7 (54%) achieved complete and 1 (8%) partial response, for an overall response rate of 62%. 1 patient underwent consolidative autologous stem cell transplantation. After a median follow up of 3.6 (range 0.8 - 13.2) years from diagnosis of CNS involvement, 9 pts experienced disease progression: 1 in the CNS alone, 3 at systemic sites alone, and 5 with both systemic and CNS sites with a median PFS of 9.5 months (Fig 1A). At last follow up, 11 pts have died (6 of progressive disease, 3 from sepsis and 2 from unknown causes) with a median OS of 37 months (Fig 1B). CNS involvement at initial diagnosis vs at relapse (Fig 1C,D) was associated with favorable PFS and OS, whilst isolated CNS involvement was associated with superior PFS but not OS (Fig 1E,F). Conclusion CNS involvement in HL is exceedingly rare and has a distinct clinical presentation with predilection for parenchymal lesions with dural extension (as opposed to true hematogenous spread), providing an important differential diagnosis for meningioma in this setting. CNS HL presenting as an initial manifestation of disease appears to be associated with favorable outcomes. Table 1. Characteristics of patients at initial diagnosis of HL Characteristics at initial diagnosis of HL data avail. n (%) Male 18 11 (61) Histologic subtype 17 Nodular-sclerosing 13 (77) Lymphocyte-rich 3 (18) Lymphocyte-depleted 1 (6) Stage 18 1 2 (11) 2 5 (28) 3 3 (17) 4 8 (44) B sx 18 10 (56) Pruritus 15 3 (20) Nodal size >10 cm 14 3 (21) Hb 15 x 109/L 13 2 (15) lymphocytes Figure 1. Figure 1. Disclosures Vose: Seattle Genetics, Inc.: Honoraria, Research Funding. Villa:Roche: Research Funding. Connors:Roche: Research Funding; Seattle Genetics: Research Funding.