PURPOSE
We analysed the ultrasonographic (US) features of atypical ductal hyperplasia (ADH) of the breast diagnosed by US-guided core needle biopsy (CNB) with the aim of identifying factors that affect the underestimation of ADH.
MATERIALS AND METHODS
A total of 134 ADH lesions sampled by US-guided CNB were reviewed retrospectively. All lesions were evaluated for pattern, size, lesion characteristics and margins, and the corresponding surgical outcome or imaging follow-up was obtained. Each patient's clinical and radiological features were analysed to identify factors involved in ADH underestimation.
RESULTS
The prevalence of malignancy in each pattern of lesions following surgical excision was 32/81 (40%) for solid masses, 14/31 (45%) for ductal patterns, 5/17 (29%) for complex cystic lesions and 2/5 (40%) for architectural distortions. Based on the results of surgical and US follow-up, none of the category 3 lesions was proven to be a malignancy. Malignancy was found in 17 (21%) of the 80 BI-RADS (Breast Imaging Reporting and Data System) category 4a lesions, 20 (74%) of the 27 category 4b lesions, 12 (92%) of the 13 category 4c lesions, and four (100%) of the four category 5 lesions. Lesions with a higher US assessment category, lacking circumscribed margins, or a mammographic finding of suspected malignancy were all significantly associated with underestimation (p < 0.05 for each).
CONCLUSION
US is useful in evaluating ADH lesions and in clarifying the indication for biopsy of these lesions. Familiarity with the frequency associated with malignancy for each feature will improve the utility of US in the work-up of these breast abnormalities.