The tumor-associated glycoprotein 72 (TAG-72) antigen is present on a high percentage of tumor types including ovarian carcinomas. Antibody B72.3 is a murine monoclonal recognizing the surface domain of the TAG-72 antigen and has been widely used in human clinical trials. After our initial encouraging studies (M. G. Rosenblum et al., J. Natl. Cancer Inst., 83: 1629-1636, 1991) of tissue disposition, metabolism, and pharmacokinetics in 9 patients with ovarian cancer, we designed an escalating dose, multi-arm Phase I study of 90Y-labeled B72.3 i.p. administration. In the first arm of the study, patients (3 pts/dose level) received an i.p. infusion of either 2 or 10 mg of B72.3 labeled with either 1, 10, 15, or 25 mCi of 90Y. Pharmacokinetic studies demonstrated that concentrations of 90Y-labeled B72.3 persist in peritoneal fluid with half-lives >24 h after i.p. administration. In addition, 90Y-labeled B72.3 was absorbed rapidly into the plasma with peak levels achieved within 48 h, and levels declined slowly thereafter. Cumulative urinary excretion of the 90Y label was 10-20% of the administered dose which suggests significant whole-body retention of the radiolabel. Biopsy specimens of bone and marrow obtained at 72 h after administration demonstrated significant content of the label in bone (0.015% of the dose/g) with relatively little in marrow (0.005% of the dose/g). The maximal tolerated dose was determined to be 10 mCi because of hematological toxicity and platelet suppression. This typically occurred on the 29th day after administration and was thought to be a consequence of the irradiation of the marrow from the bony deposition of the radiolabel. In an effort to suppress the bone uptake of 90Y, patients were treated with a continuous i.v. infusion of EDTA (25 mg/kg/12 h x 6) infused immediately before i.p. administration of the radiolabeled antibody. Patients (3 pts/dose level) were treated with doses of 10, 15, 20, 25, 30, 35, 40, or 45 mCi of 90Y-labeled B72.3 for a total of 38 patients. EDTA administration resulted in significant myeloprotection, which allowed escalation to the maximal tolerated dose of 40 mCi. Dose-limiting toxicity was thrombocytopenia and neutropenia. Studies of plasma and peritoneal fluid pharmacokinetics demonstrate no changes compared with patients without EDTA pretreatment. Cumulative urinary excretion of the radiolabel was not increased in patients pretreated with EDTA compared with the untreated group. However, analysis of biopsy specimens of bone and marrow demonstrated that bone and marrow content of the 90Y label was 15-fold lower (<0.001% injected dose/g) than a companion group without EDTA. Four responses were noted in patients who received 15-30 mCi of 90Y-labeled B72.3 with response durations of 1-12 months. These results demonstrate the myeloprotective ability of EDTA, which allows safe i.p. administration of higher doses of 90Y-labeled B72.3 and, therefore, clearly warrant an expanded Phase II trial in patients with minimal residual disease after standard chemotherapy or for the palliation of refractory ascites.