To evaluate the clinical and radiologic findings of atypical ductal hyperplasia (ADH) using mammography and ultrasonography, and to correlate the radiologic and histopathologic findings.

Sixty-four pathologically proven lesions in 64 patients who were examined between March 2000 and March 2003 were the subject of this study. Mammography was performed in all 64 cases, and ultrasonography in 30. Two radiologists retrospectively evaluated the radiologic findings, classifying them as one of four types: mass, microcalcification, other finding, and no detected lesion. At mammography, masses were classified according to their shape, margin, and density and microcalcifications according to their shape and distribution. At ultrasonography, masses were evaluated in terms of their shape, margin, internal and posterior echotexture, ductal extension, and parallelism to skin. Geographic correlation between the radiologic and histopathologic findings was classified as direct, near direct, or remote correlation.

Mammography demonstrated 37 cases of microcalcification (57.8%), 14 in which masses were present (21.9%), two in which there were other findings (3.1%), and 11 in which lesions were not detected (17.2%). The "other finding" was ductectasia. Microcalcifications were round in 19 cases, pleomorphic heterogeneous in 16, and branching linear in one. The most common distribution of microcalcification was clustered (29 cases; 78.4%). Masses were oval or round in nine cases and irregular in three, and in seven cases their margin was ill-defined. In 13 cases, the density of the masses was equal to that of breast tissue. Ultrasonography showed that the masses were round or oval in 15 cases and irregular in 14, and that the margin was ill-defined in 16 cases and circumscribed in ten. In 19 cases, the echotexture of the masses was low, and in 20 cases, heterogeneous. Parallel orientation was seen in 25 cases, and ductal extension in 22. Category 4 was the most common final assessed BI-RADS category, found in 75% of cases. Radiologic-histopathologic correlation was direct in 44 cases, near direct in 13, and remote in seven. Clinically, self or clinical examination of the breast revealed no abnormality in 47 cases, a palpable mass in seven, nipple discharge in seven, and breast pain in three.

At mammography, the most common finding of ADH was clustered round or pleomorphic heterogeneous microcalcifications, and at ultrasonography, illdefined, round or oval, or irregular-shaped, hypoechoic masses with parallel orientation and ductal extension. Clinically, most ADH was incidentally discovered at radiologic examination. In this study, 17.2% of ADH cases were not demonstrated by mammography but were detected at ultrasonography, and for the detection of ADH, the use of this latter modality, alongside mammography, is thus feasible.