The capability of the mammography-scintimammography combination to distinguish between benign and malignant isolated clusters of breast microcalcifications is discussed. Scintimammography using Tc 99m-Sestamibi was performed in 97 women with an isolated cluster of microcalcifications on mammograms. Seventy-two women had final histopathologic diagnoses (24 cancer and 48 benign pathology). The other 25 patients had follow-up to 3 years. The results of mammography, scintimammography and mammography-scintimammography combination were divided into five groups, based on the suspicion of malignancy. The sensitivity, specificity, false negative fraction, false positive fraction, predictive positive value, predictive negative value and diagnostic accuracy were calculated varying the diagnostic threshold. The Receiver Operating Characteristic (ROC) statistical technique was employed to compare the diagnostic value of mammography to mammography-scintimammography combination. The area under the ROC curves was calculated by the Wilcoxon statistic without any hypothesis on data distribution. The detected difference between areas under the mammography ROC curve (area=0.854, standard error=0.049) and mammography-scintimammography ROC curve (area=0.897, standard error 0.033) was statistically significant (P>0.05, one tail). The area under a ROC curve represents the probability that a randomly chosen diseased or non-diseased subject could be correctly classified. From this point of view this paper demonstrates that, if properly used, scintimammography can add to mammography in the characterization of an isolated cluster of microcalcifications, even if it is not able to replace FNAB and core biopsy.