To retrospectively examine the effectiveness of methotrexate monotherapy for neurosarcoidosis using clinical and radiographic data. 

Treatment strategies for neurosarcoidosis remain poorly defined, and the data evaluating the role of methotrexate monotherapy is limited.

In this single-center, retrospective analysis, adult patients with neurosarcoidosis were included if they received ≥7.5 mg of methotrexate monotherapy weekly for ≥3 months and had MRI studies performed after ≥3 months of therapy. Clinical disease burden was assessed using the modified Rankin Scale and Expanded Disability Status Scale, and radiographic response was evaluated by comparing pre- and post-treatment MRI findings. Response was defined as 1) clinical and radiographic improvement, 2) clinical stability with radiographic improvement, or 3) clinical improvement with radiographic stability. 

Of 274 patients reviewed, 36 (19 females) met the inclusion criteria. At methotrexate initiation, the average age was 49.3 years, the average disease duration was 37 months, and patients had experienced an average of 1.7 attacks. The most frequent phenotypes included cranial leptomeningitis (16), optic neuritis (12), myelitis (12). Methotrexate was used first-line in 27 patients, with an average peak maintenance dose of 16.6mg weekly. After an average of 36.2 months of monotherapy, clinical and radiographic improvement occurred in 9 and 20 patients, respectively, resulting in 20 (56%) patients being classified as responders. Among phenotypes with ≥5 patients, responders most commonly had spinal leptomeningitis (57%), myelitis (58%), and cranial pachymeningitis (50%), whereas non-responders most commonly had brain parenchymal disease (75%), non-optic cranial neuropathies (60%), optic neuropathy (58%), and cranial leptomeningitis (56%). By the time of the last follow-up of methotrexate use, the average dose of concurrent prednisone was reduced from 31.9mg to 7.1mg daily.  

Methotrexate monotherapy was partially effective for the cohort, noting spinal phenotypes were more likely to respond than cranial ones. These results emphasize the importance of considering phenotypes in treatment decision-making.