We report and compare the results obtained with conventional imaging (mammography and US) and MRI in the study of 46 solid nodular breast lesions verified with histologic, cytologic and/or instrumental follow-up examinations for 12-34 months. The variables we compared were relative to the identification, nature and size of the lesions. MRI, which was performed on the basis of previous mammographic and US findings, detected all the lesions but never modified the diagnosis of conventional imaging methods. Questionable MR diagnoses were fewer than mammographic and US ones (2 versus 11), but its role in correcting the questionable diagnoses of conventional imaging methods was controversial. Particularly, of 11 such cases on mammographic and US images, MRI made 8 correct diagnoses but exhibited 2 false positives and 1 false negative for carcinoma. Such MR mistakes are likely to be related to the non-use of contrast medium. As for size, US was more accurate than mammography and MRI; yet, very few misdiagnoses were make on the whole.