Dosimetry and treatment planning for therapeutic infusions of radiolabeled antibodies are usually performed by extrapolation from the biodistribution of trace-labeled antibody. This extrapolation assumes that the biodistribution of high specific activity antibody will be similar to that seen with trace-labeled antibody. However, high doses of radiation result in rapid depletion of lymphoid and hematopoietic cells in lymph nodes, spleen, and marrow with replacement by blood and plasma. If radiolabeled antibody is cleared slowly from blood, this replacement may result in increased radionuclide concentrations in these tissues following infusions of antibody labeled with large amounts of radionuclide. To examine the influence of deposited radiation on the biodistribution of radiolabeled antibody, we treated mice with a constant amount of antibody that was labeled with varying amounts of 131I. Survival was determined in normal specific pathogen-free AKR/Cum mice (Thy1.2+) after infusion of anti-Thy1.1 antibody labeled with 10 to 6500 muCi of 131I, to determine an appropriate range of 131I doses for further study. The dose producing 50% lethality within 30 days following infusion of 131I-labeled antibody was 530 muCi 131I. Biodistribution, bone marrow histology, and dosimetry were subsequently determined after infusion of 500 micrograms of antibody labeled with 10, 250, 500, or 3500 muCi 131I. The amount of 131I did not influence uptake or retention of antibody in blood, liver, lung, or kidney. In contrast, infusion of antibody labeled with 250 to 3500 muCi of 131I led to a dose-related increase in the concentration of 131I in marrow, spleen, lymph node, and thymus. For example, at 96 h after infusion of antibody labeled with 500 or 3500 muCi 131I, concentrations in marrow were 3- to 4-fold higher than after infusion of trace-labeled antibody. The increase in marrow 131I concentrations was associated with depletion of cells and hemorrhage within the marrow space. As a result, estimated mean absorbed doses to marrow, lymph node, spleen, and thymus were 1.2 to 3.1 times higher than would have been predicted from the biodistribution of trace-labeled antibody. These results suggest that the biodistribution of trace-labeled antibody should be an accurate predictor of the behavior of high specific activity antibody in blood and solid organs such as liver and kidney. In contrast, radiation from antibody labeled with large amounts of radionuclide can result in an alteration of the concentration of radiolabeled antibody in rapidly responding tissues such as marrow.(ABSTRACT TRUNCATED AT 400 WORDS)