Pretargeted radioimmunotherapy (PRIT) was first investigated in a series of phase I and phase II studies in patients with adenocarcinoma using a pancarcinoma antibody, NR-LU-10. The principles and schema developed were then applied to an initial study in patients with non-Hodgkin's lymphoma (NHL). The PRIT approach used is a multi-step delivery system in which an antibody is used to target streptavidin to a tumor-associated antigen receptor, and subsequently, biotin is used to target the 90Y radioisotope to the tumor localized streptavidin. In the NHL study, a chimeric, IgG1, anti-CD20 antibody (Rituximab) was conjugated to streptavidin (SA) and administered to patients. Thirty-four hours later, a clearing agent, synthetic biotin-N-acetyl-galactosamine, was administered to remove non-localized conjugate from the circulation. Finally, a DOTA-biotin ligand, labeled with 111In for imaging and/or 90Y for therapy was administered. Ten patients with relapsed or refractory NHL were studied, and seven received 30 or 50 mCi/m(2) 90Y DOTA-biotin. Preliminary studies using 186Re labeled conjugate confirmed that it localized to tumor and that the clearing agent removed >95% of the conjugate from the circulation. Radiolabeled biotin localized well to tumor. Unbound radiobiotin was rapidly excreted from the whole body and normal organs. The mean tumor dose calculated was 29+/-23 cGy/mCi 90Y, and the mean tumor to whole body dose ratio was 38:1. Only grade I/II non-hematologic toxicity was observed. Hematologic toxicity was also not severe; i.e. five of the seven patients who received 30 or 50 mCi/m(2) of 90Y-DOTA-biotin experienced only transient grade III (but no grade IV) hematologic toxicity. Although six of 10 patients developed humoral immune responses to the streptavidin, these were delayed and transient and hence may not preclude retreatment. Six of seven patients who received 30 or 50mCi/m(2) 90Y achieved objective tumor regression, including three complete and one partial response. The estimate of tumor to whole body dose ratio (38:1) achieved with PRIT in these NHL patients is higher than that achieved in other studies using conventional RIT. Toxicity was mild and tumor response encouraging. PRIT clearly deserves additional study in patients with NHL.