The Efficacy of Systemic High Dose Methotrexate As Central Nervous System Prophylaxis in Patients with High Risk Diffuse Large B-Cell Lymphoma: A Propensity Score-Matched Analysis

Abstract

Background: While the outcome of patients with diffuse large B-cell lymphoma (DLBCL) has been improved with the introduction of rituximab, central nervous system (CNS) relapse is still associated with poor prognosis. Although intrathecal methotrexate has been widely used for prophylaxis, its role has been questioned. Another way of utilizing the agent, systemic high dose methotrexate (HD-MTX) has been advocated from a few retrospective studies. However, they had no comparison group or they were compared with imbalanced, unmatched group even if existed. With prospectively collected cohort, we performed a propensity score-matched analysis to evaluate the efficacy of IV HD-HTX as CNS prophylaxis in high risk DLBCL patients.Methods: The registry data set of DLBCL patients, collected from January 2010 through March 2015 at a single institute, Asan Medical Center, was retrospectively reviewed. From July 2013, all consecutive DLBCL patients who were considered at high risk for CNS recurrence received systemic HD-MTX with standard R-CHOP therapy. High risk CNS relapse was defined by the involvement of ≥2 extranodal sites and elevated lactic dehydrogenase (LDH); or high-risk CNS international prognostic index (CNS-IPI ≥4); or involvement of specific high-risk extranodal sites including bone marrow, breasts, testes, paranasal sinuses. HD-MTX was administered at a dose of 3-3.5 g/m2 on day 15 of alternating cycles of R-CHOP or delivered 2 to 5 weeks after the completion of the primary therapy. We analyzed the progression-free survival (PFS), CNS relapse-free survival (CNS-RFS) and overall survival (OS) of the patients receiving CNS prophylaxis and compared with the patients who received R-CHOP only. Multivariate Cox regression and propensity score analysis were used to evaluate the treatment effect of systemic HD-MTX on survival with multiple covariates including age, sex, LDH, ECOG score, stage, CNS-IPI, extranodal involvement and involved sites.Results: A total of 197 patients with DLBCL and CNS risk factors who were treated with standard R-CHOP therapy were identified between January 2010 and March 2015. Among them, 47 patients received systemic HD-MTX as CNS prophylaxis. In the R-CHOP only group (n=150), with a median follow-up of 58.88 (1.34-84.3) months, 14 CNS relapses occurred within 2 years. The estimated 2-year PFS, CNS-RFS and OS rates were 58.6%, 89.5%, 71.3%, respectively. In the HD-MTX prophylaxis group (n=47), with a median follow-up of 27.24 (3.38-42.02) months, 3 CNS relapses occurred within 2 years. The estimated 2-year PFS, CNS-RFS and OS rates were 62.3%, 93.1% and 74.5%, respectively. Systemic HD-MTX did not significantly increase the PFS (HR 0.619, 95% CI 0.363-1.057, P = .079), CNS-RFS (HR 0.547, 95% CI 0.156-1.924, P=.347), and OS (HR 0.628, 95% CI 0.331-1.19, P = .154) in the multivariate survival analysis. The results seems to consistently show no survival benefit of systemic HD-MTX regardless of the different propensity score analysis methods except for PFS by weighting HR: PFS (matching HR 0.652, 95% CI 0.36-1.18, P=.157; weighting HR 0.689, 95% CI 0.503-0.945, P=.02),CNS-RFS (matching HR 0.635, 95% CI 0.161-2.499, P=.516; weighting HR 0.587, 95% CI 0.287-1.203, P=.146), OS (matching HR 0.625, 95% CI 0.319-1.227, P=.172; weighting HR 0.726, 95% CI 0.505-1.044, P=.083).Conclusion: Systemic HD-MTX prophylaxis for CNS relapse did not show significant survival improvement in high risk DLBCL patients in this study although there was a trend for better outcomes in patients given HD-MTX, as different statistical methods incorporating stringent propensity score-matched analysis were applied. However, these results are still limited by small cohort size its retrospective nature. Multi-center, prospective randomized studies are necessary to appraise the efficacy of systemic HD-MTX as CNS prophylaxis in high risk DLBCL patients. DisclosuresNo relevant conflicts of interest to declare.