An anti-Thomsen-Friedenreich (T) antibody-based enzyme-linked immunosorbent assay (ELISA) with high efficacy in human breast carcinoma detection is described. Immunoreactive T epitopes occur in approximately 90% of all carcinomata; all humans have anti-T antibodies, naturally occurring anti-carcinoma antibodies, induced by their own intestinal flora. Carcinoma patients, but not control subjects, show alterations of serum anti-T hemagglutinin levels. Human anti-T antibodies are predominantly IgM. In the protocol presented here, anti-T IgM antibodies are quantitated by ELISA using Immulon 2 wells coated with human blood group O erythrocyte-derived T antigen as solid phase; in addition, total IgM in each serum is quantitated by ELISA in parallel with the anti-T IgM. Inter-assay coefficient of variation was 2% for both ELISAs. Although anti-T IgM values alone distinguish between carcinoma patients and control subjects, use of the quotient, QMe, which also considers total IgM, increases this distinction. For a given serum, QMe was obtained by the formula: QMe = (100 x (anti-T IgM)2/total IgM). Sera of 242 subjects, 117 breast carcinoma patients, 36 benign breast disease patients and 89 healthy persons were analyzed. QMe identified 88% of the breast carcinoma patients: it all six (100%) in situ, 11/13 (85%) Stage I, 48/58 (83%) Stage II and III and 38/40 (95%) Stage IV patients. Sera from 83% of the 36 benign breast disease patients were negative, i.e. within normal range; five of the six positive sera originated from patients with increased long-term risk of breast carcinoma, while sera from 11 other patients with increased carcinoma risk were negative. Overall, 90% of the 125 non-carcinoma control subjects were negative by both anti-T IgM and QMe. In preliminary studies, the ELISA protocol detected 11/14 (79%) patients with carcinomata other than those of the breast. The identification of all six in situ breast carcinoma patients by QMe points to its usefulness in carcinoma detection, especially early.