No Survival Improvement for Central Nervous System Multiple Myeloma By the Use of Novel Anti-Myeloma Agents: The Greek Myeloma Study Group Experience

Abstract Multiple myeloma of the central nervous system (CNSMM) is a rare and severe clinical entity. The efficacy of novel agents in this setting has not been sufficiently explored. Our aim was to describe the incidence, characteristics and outcome of CNSMM in the era of novel agents, to search for prognostic factors of post CNSMM survival and explore the efficacy of novel agent-based combinations (NAC). Data were provided by 8 Centers of the Greek Myeloma Study Group from January 2000 to December 2013. Documentation of CNSMM required: biopsy that proved CNSMM or MRI/CT findings consistent with CNSMM or detection of plasma cells (PCs) and/or monoclonal immunoglobulin (MC) in the cerebrospinal fluid (CSF). Response to CNSMM treatment was defined as improvement/normalization of at least one of the aforementioned variables. Twenty-four (M/F: 10/14; median age: 63.5 years, range 24-97 years) of 3107 newly diagnosed symptomatic MM patients, who were diagnosed in the same centers during the same period of time, developed CNSMM (0.8%); 2 of them were newly diagnosed and 22 had received previous therapies (6/22 had CNS involvement as the sole feature of MM). Fourteen patients had IgG MM, 2 IgA, 3 light chain, 1 IgD and 4 non-secretory MM. The median time to CNSMM diagnosis was 28.5 months (range: 0-98 months). Clinical manifestations included: visual disturbances/diplopia (37%), paresis/paraplegia (29%), lethargy/confusion (25%), headache (16%) and cranial nerve palsy (12%). MRI or CT documentation was available in 20/24 patients and revealed parenchymal lesions (35%), leptomeningeal lesions (10%), direct MM extension (20%), parenchymal/leptomeningeal (10%), parenchymal/direct MM extension (15%) or leptomeningeal/direct MM extension (10%). Diagnostic lumbar puncture was performed in 17 patients and in 10/17 PCs were detected in the CSF (median number of PCs/μL was 172.5/μL, range: 5-2550/μL). LDH and albumin in the CSF was high in all patients, while MC was detected in the CSF of 4 patients. Sixteen patients (67%) had additional extramedullary disease (EMD) or plasma cell leukemia (PCL) prior to CNSMM diagnosis (plasmacytomas: 9 patients, PCL primary/secondary: 3/3 patients, EMD/PCL: 1 patient). Molecular cytogenetics were available in 12/24; 4 patients had high risk cytogenetics. Patients who developed CNSMM after frontline therapy (n=22) had higher LDH at the time of CNS involvement compared to initial MM diagnosis (267 U/L vs 191 U/L; p=0.02). NAC had been given in 17/22 (77%) such patients prior to CNSMM diagnosis (bortezomib-based: 17, IMiD-based: 13, both: 13). The median number of previous MM treatment lines in these patients was 3 (range: 1-5). Treatment of CNSMM included NAC (bortezomib-based 9 and IMiD-based 3 patients), chemotherapy alone (7 patients) and only intrathecal infusions (ITI) with MTX or AraC (3 patients); 7 patients received both systematic therapy and ITI. Additional radiotherapy (RT) was given to 8 patients; one patient underwent ASCT consolidation. Regarding CNSMM response, 10/22 treated patients had improvement or normalization of initial CNS findings and 6 of them relapsed. At the time of evaluation, 2 patients were alive and 22 patients had died (MM progression: 13, sepsis: 6, cerebral hemorrhage: 2, leishmaniasis: 1 and CNSMM progression: 1). The median post CNSMM survival was 3 months (95% CI: 1.9-4.1). The median post CNSMM survival for patients treated with NAC vs. others was 4 (95% CI: 0-8.6) vs. 2 months (95% CI: 0.7-3.3), respectively (p>0.05). Additional RT did not improve survival (p>0.05). In the cox regression analysis, prior treatment with NAC and presence of EMD/PCL prior to CNSMM diagnosis marginally predicted for shorter post CNSMM survival (p=0.05 and 0.068, HzR: 3.03 and 2.57, respectively). The median post CNSMM survival for previously NAC-exposed patients was 2 months (95% CI: 1-2.9) vs. 6 months (95% CI: 0.8-11) for NAC-naive patients (p=0.03), while that of patients with prior EMD/PCL was 2 months (95% CI: 1.3-2.6) vs. 8 months (95% CI: 0-18) of the others (p=0.06). In the era of novel agents the incidence of CNSMM remains low, but yet outcome is extremely poor. In our study, treatment with NAC ± RT and ITI did not offer any survival advantage. Treatment with NAC and presence of EMD/PCL prior to CNSMM diagnosis seems to predict for shorter post CNSMM survival. Patients with EMD/PCL should be monitored more firmly in order to detect early manifestations of CNSMM. Disclosures No relevant conflicts of interest to declare.