A common feature of human tumor tissue is mutant p53 protein accumulation. Here we evaluate a new "sandwich" immunoassay for p53 protein incorporating modifications to a previously reported method, including the use of microtiter plates coated directly with the anti-p53 monoclonal antibody DO-1, a detergent- and mouse serum-containing sample diluent, and a labeled secondary antibody diluent containing goat serum. The use of CM-1 antiserum to probe the immunocaptured p53 and the detection of bound complexes by a labeled secondary antibody allows coupling to a time-resolved fluorescence detection system. The new assay yielded p53 concentrations comparable with those by the previous assay for breast tumor cytosols (n = 198), nondiseased breast tissues (n = 70), and five transformed cell lines, but showed differences in p53 values measured in sera from patients without cancer (n = 78). These serum differences were found to reflect nonspecific interferences affecting the original method, which implies that the new immunoassay has improved specificity for serum p53 quantification.