HOUT-26. FACTORS ASSOCIATED WITH TREATMENT DELAY IN CENTRAL NERVOUS SYSTEM LYMPHOMA

OBJECTIVE To identify clinico-radiographic characteristics associated with delayed treatment initiation in central nervous system lymphoma (CNSL). INTRODUCTION Clinical and radiographic characteristics of CNSL are often varied with a broad differential diagnosis, potentially leading to delays in diagnosis and treatment. METHODS A single-center retrospective review of clinico-radiographic data was performed at Columbia University Irving Medical Center in patients with pathologically confirmed CNSL diagnosed from 1/2010–12/2018. Descriptive statistics and univariate logistic regression were used to identify variables associated with delayed treatment. Using visual binning, delayed treatment time was designated as >33 days from first presentation to medical attention to first chemotherapy for CNSL. Variables of interest included demographic data, presenting symptomatology, radiographic characteristics, location of initial presentation, and diagnostic and therapeutic interventions performed before biopsy. RESULTS Seventy patients (36 men (51%); median age at diagnosis 70 years, IQR 14.75 years; median time from first presentation to treatment 21 days, IQR 41.25 days) were included. Presentation with cognitive deficits suggested a strong, but not statistically significant, association with delayed treatment (OR=1.93, p=0.20), whereas presentation with focal neurologic deficits suggested protection against delayed treatment (OR=0.25, p=0.05). Initial presentation to a hospital suggested a strong, but not statistically significant, trend against delayed treatment (OR=0.41, p=0.08). Multifocal disease on neuroimaging (OR=7.18, p=0.001), pre-biopsy cerebrospinal fluid (CSF) sampling (OR=5.18, p=0.002), and pre-biopsy immunomodulatory treatment (including high-dose intravenous corticosteroids) for suspected neuroinflammatory disease (OR=6.33, p=0.03) had statistically significant associations with delayed treatment. Antimicrobial treatment before biopsy for suspected CNS infection suggested a trend toward delayed treatment, but the association was not statistically significant (OR=5.1, p=0.06). CONCLUSIONS Multifocal disease and pre-biopsy CSF sampling and immunomodulatory therapy were associated with delayed treatment initiation for CNSL in our single-center cohort. Recognizing factors associated with delayed treatment may allow physicians to circumvent these factors and permit more rapid diagnosis through tissue sampling.