Objective: To assess the value of the serum carcinoembryonic antigen (CEA), tissue polypeptide antigen (TPA) and breast cancer-associated antigen CA15.3 (CEA-TPA-CA15.3) tumour marker panel in selecting from a group of patients with equivocal bone scintigraphy, chest X-ray or liver echography, those with skeletal, thoracic or liver metastases. Clinical data of 427 breast cancer patients submitted to an intensive follow-up after mastectomy between January 1986 and December 2000 were retrospectively reviewed. Methods: Among the 427 patients operated on for breast cancer, 221 patients with a total of 332 equivocal instrumental examinations (bone scintigraphy, n = 286; chest X-ray, n = 29; liver echography, n = 17) were reviewed. All 221 patients were followed up clinically, biochemically and instrumentally until there was a clear definition of their condition, metastatic or not, for an average time of 35 months. Positive and negative predictive values of the tumour marker panel in patients with equivocal bone scintigraphy, chest X-ray and liver echography were evaluated; concomitant clinical symptoms were also taken into consideration. Results: Among the 221 patients with equivocal bone scintigraphy, chest X-ray and liver echography, tumour markers showed a positive predictive value of 69, 93 and 83% and a negative predictive value of 98, 86 and 91%, respectively, for the indication of the metastatic or benign origin of the equivocal instrumental imaging. Clinical symptoms were not helpful in predicting metastatic disease (sensitivity, specificity and accuracy of 60, 53 and 54%, respectively). Conclusions: These data suggest that a short monitoring with the CEA-TPA-CA15.3 tumour marker panel is an important tool to confirm or exclude metastatic disease in those patients who are suspected to have metastases following common instrumental investigations, and it is particularly important to avoid false positive diagnoses.