Analyses were performed on 40 patients with TAG-72 expressing metastatic cancer who were entered into three phase II clinical trials. The dose selected was the maximum tolerated dose in phase I studies. Patients all had unresectable metastatic colon or prostate cancer and had recovered from prior therapies. Patients in trials #1 and #2 received 75 mCi/m2 131I-CC49 antibody whereas those in trial #3 received a total of 75 mCi/m2 with equal amounts of 131I-CC49 and 131I-COL-1. The three trials have resulted in a reproducible degree of reversible marrow suppression; 72.5% of patients experienced moderate or severe toxicity. Comparisons were made between demographic, clinical and pharmacokinetical variables and the grade of WBC toxicity, platelet toxicity and the sum of the two as total toxicity. Whole body radiation dose had a statistically significant relationship with platelet toxicity (r = 0.38, p = 0.015) and total toxicity (r = 0.34, p = 0.035). The bone marrow radiation dose is significantly related to all toxicity indicators with correlation coefficients with WBC and platelet toxicities of 0.47 (p = 0.002) and 0.34 (p = 0.033), respectively. Plasma half-life had the strongest correlation with WBC toxicity and combined toxicities. Multivariate models were developed to help describe the simultaneous effect of these variables on toxicity. The results show that the MTD dose was safely given to patients who varied in age, disease burden and degree of marrow compromise. This supports the contention that a fixed dose of radiolabeled antibody per body mass or m2 can be given to a diverse group of non-lymphoma patients with a predictable toxicity range.