The optimal treatment of primary central nervous system lymphoma (PCNSL), a rare form of extranodal non-Hodgkin lymphoma, has yet to be defined. Whole-brain radiation therapy (WBRT) has limited efficacy as a single therapeutic modality and is associated with a high risk of delayed neurotoxicity. Methotrexate-based chemotherapy regimens yield poor drug penetration across the blood-brain barrier (BBB), thus necessitating administration of high doses with the concomitant risk of increased systemic and neurological toxicity. Combined-modality therapy (WBRT plus chemotherapy) can improve response and survival rates, yet it is associated with a high risk of neurotoxicity. The aim of chemotherapy in conjunction with BBB disruption is to maximize drug delivery to the brain and improve the agent's efficacy, while preserving neurocognitive function and minimizing systemic toxicity. Methotrexate-based chemotherapy regimens administered in conjunction with BBB disruption have shown promising results in PCNSL. Animal models of central nervous system lymphoma and drug neurotoxicity offer new possibilities to study the effects of various treatments on PCNSL and normal brain and can also help understand biological and pathophysiological aspects of this disease. Because the intact BBB is even less permeable to antibodies than it is to drugs, preclinical and clinical studies of monoclonal antibody delivery (for example, rituximab and 90Y ibritumomab tiuxetan) to the brain in conjunction with BBB disruption offer a new possibility to make these novel treatments more efficient against PCNSL. Regarding the evaluation of more sensitive and specific diagnostic imaging tools, iron oxide-based contrast agents for magnetic resonance imaging have shown promise for better differentiation of PCNSL from other white matter diseases.