recently reported phase III trial (four-arm randomization) in newly diagnosed patients with aggressive non-Hodgkin’s lymphoma (NHL) and assessment of CNS relapse. I would like to make several comments regarding the report and the need for risk stratification regarding CNS-directed therapy in newly diagnosed patients with aggressive NHL. Five key issues appear relevant with respect to CNS prophylaxis in aggressive NHL. First, is there a consensus that CNS disease (mostly lymphomatous meningitis) is clinically relevant? Approximately 55% of all diffuse large B-cell NHL relapse systemically as contrasted with only 2% to 4% relapse in the CNS. 1-6 Among patients with CNS relapse, half have poorly responding or refractory systemic disease. Therefore, the number of patients with NHL-isolated CNS is comparatively small. Second, is there a consensus regarding identification of high-risk (for CNS relapse/disease) patients with aggressive NHL? A number of studies have suggested low International Prognostic Index, two or more extranodal sites of disease and site specific disease (ie, testes, breast, parameningeal [sinus, orbit], and bone marrow) defined as large-cell involvement not small cleaved cell define a substantial group of patients at risk for CNS disease. 1-6 Notwithstanding this recognition of high-risk features, no study of newly diagnosed aggressive NHL has attempted to stratify patients with respect to CNS risk and CNSdirected therapy. Third, and as mentioned by Bernstein et al, 1 high-risk patients are often found at diagnosis (approximately 25%), if evaluated by CSF flow cytometry, to have contamination of the CSF by lymphoma. 7 This suggests that a subset of patients with aggressive NHL and without neurological symptoms has occult lymphomatous meningitis at diagnosis. Importantly, CSF flow cytometry appears superior to CSF cytology with appropriate caveats regarding laboratory methodology for demonstrating lymphomatous involvement of the CSF. Notwithstanding this subset of patients with occult lymphomatous meningitis, the majority of patients with CNS relapse have negative CSF evaluation at diagnosis. Nevertheless, the study by Hegde et al 7 is compelling for mandatory CSF evaluation in patients defined as at high risk for CNS relapse at time of up-front therapy.