Abstract : We report herein the results we obtained and the limitations we experienced during the production and use of a bank of Epstein-Barr virus (EBV)-transformed human cytotoxic T lymphocytes (EBV-CTLs). To assess the feasibility and toxicity of this strategy, we selected and stored, in liquid nitrogen, 4 billion EBV-CTLs from each of the 13 selected donors. Subsequently, in a multicenter phase I/II study, 11 patients with EBV-associated lymphoma resistant to conventional treatments received 1-3 doses of 5 million EBV-CTLs/kg with 1-3 and 0-4 compatibilities for human leukocyte antigen (HLA)-I and HLA-II, respectively. Except for one event of fever after injection, no immediate or delayed toxicity, no graft versus host disease, and no graft rejection attributable to CTL infusion were observed. Three patients presented complete remission and 1 partial remission after treatment. Considering the clinical options currently available, and the constrains associated with CTL preparation and implementation, we conclude that CTL banks should consist of a reasonably small number of cell lines with documented specificities. This objective could be more easily achieved if the few homozygous donors for the most frequent HLA alleles of the targeted population could be made available for such a project.