Central Nervous System Invasion in Multiple Myeloma: A Nationwide Survey of 75 Patients in Japan

Introduction:Central nervous system invasion in multiple myeloma (CNS-MM) is extremely rare (approximately 1% of MM). Prognosis of patients with CNS-MM is generally poor, and the median overall survival (OS) time from its diagnosis (Dx) has been reported to be 2 to 7 months. However, because of its rareness, most available data concerning CNS-MM are based on anecdotal reports on small case series. In this study, we retrospectively collected clinical data of Japanese patients with CNS-MM and analyzed them to reveal clinical features and prognosis of this disease entity.

Methods:We conducted a nationwide multicenter retrospective study involving 107 centers that consist of educational facilities authorized by Japanese Society of Hematology and its related hospitals. CNS-MM was defined based on the previously reported criteria (Br J Haematol 2013;162:483-48). Univariate and multivariate analyses were performed to explore prognostic factors and the suitable treatment of CNS-MM.

Results:From December 1978 to February 2016, 75 patients with CNS-MM were identified. The median age was 58 (range: 32-77 years). CNS invasion was detected at initial Dx of MM in 4% and at relapse in 96%. The median time from Dx of MM to that of the secondary CNS-MM was 1.8 years. Concomitant plasma cell leukemia (18%), skull plasmacytoma (12%), and myeloma cells in the cerebrospinal fluid (63%) were observed at the Dx of CNS-MM. Common symptoms of CNS-MM included consciousness disturbance (37%), cranial nerve palsy (25%), diplopia (18%), headache (16%), spinal nerve disorder (13%), and nausea/vomiting (12%). Intrathecal chemotherapy (IT) was used in 37%, cranial and/or spinal irradiation therapies (RTx) in 45% in addition to various systemic therapies, which included immunomodulatory drugs (IMiDs) (21%), bortezomib (19%), alkylators (21%), dexamethasone alone (6%), autologous stem cell transplant (auto-SCT, 4%), and allogeneic stem cell transplant (allo-SCT, 3%). With a median follow-up of 3.4 months from Dx of CNS-MM and of 31.5 months from that of MM, the median OS time from Dx of CNS-MM was only 3.7 months. The median OS times for the 12 untreated and 61 treated patients were 0.2 and 5.1 months, respectively (P < 0.01), suggesting that the patients who did not receive any treatments (Txs) were in a severe condition and their prognoses were extremely poor. In univariate analyses for the entire group, absence of both atypical lymphocytes and plasma cells in the peripheral blood at Dx of MM and no Tx for MM before Dx of CNS-MM were significantly favorable prognostic factors (P = 0.008 and P < 0.0001, respectively). Regarding Tx for CNS-MM patients who received any Txs (conventional chemotherapies, bortezomib, IMiDs, IT, and RTx), only RTx was correlated with longer OS time (P = 0.047). In multivariate analyses, RTx and IT had a significant impact on longer OS time (>6 months, OR = 3.80, 3.97, P = 0.002, 0.01, respectively). Although the prognosis of CNS-MM in this study cohort was very poor, seven patients could survive for >1 year from Dx of CNS-MM. Among them, RTx, IMiDs, IT, bortezomib, auto-SCT, and allo-SCT were given to 5, 4, 5, 3, 1, and 1 patients, respectively.

Conclusions:Prognosis of CNS-MM is extremely poor in general. Although there must be a selection bias that patients who received various salvage Txs for CNS-MM were in a better condition, the results of our study suggest that multi-modality Txs with RTx, IT, and IMiDs/bortezomib may prolong the survival time in some cases. Prospective studies are needed to confirm our preliminary observations.