We thank Ferreri and colleagues for their comments and interest in our article 'Trials and tribulations in primary CNS lymphoma'.1 We recognize that in the absence of randomized phase III trials there will be disagreements on proposed treatment algorithms, which are based on lower levels of evidence, institutional experience, and opinion. In this regard, the algorithm proposed by Ferreri et al.2 is certainly a reasonable alternative to the one we suggested.

Our algorithm reflects the treatment approach we have chosen based on our interpretation of the data. In primary central nervous system CNS lymphoma (PCNSL), as in other hematologic malignancies, the level of evidence required to pursue autologous stem-cell transplantation in clinical practice varies across institutions, as well as between Europe and the USA. However, we concur with the authors that enrollment of patients to randomized clinical trials such as the one designed by the International Extranodal Lymphoma Study Group (the IELSG #32 trial) is very important, and preferred over off-study treatment approaches.