PURPOSE
The purpose of this study was to identify the factors that may have an impact on upgrading atypical ductal hyperplasia (ADH) lesions to malignancy.

MATERIALS AND METHODS
Between February 1999 and December 2010, the records of 150 ADH lesions that had been biopsied were retrospectively reviewed. The biopsy types included 11-gauge stereotactic vacuum-assisted biopsy (SVAB) (n=102) and ultrasonography (US)-guided 14-gauge automated biopsy (n=48). The patients were divided into two groups: those who had cancer in the ™nal pathology and those who did not. Variables associated with underestimation of ADH lesions were compared between the groups.

RESULTS
The underestimation rates according to the biopsy types were 41.7% (20/48) for the US-guided 14-gauge automated biopsy and 20.6% (21/102) for the 11-gauge SVAB (P = 0.007). The rate of underestimation was signi™cantly higher in lesions greater than 7 mm than it was in smaller lesions, with both US-guided 14-gauge automated biopsy and 11-gauge SVAB (P = 0.024 and P = 0.042, respectively). The rate of underestimation was signi™cantly higher with the 11-gauge SVAB (P = 0.025) in lesions that were suspicious (R4) and highly suggestive of malignancy (R5) than in those that were probably benign (R3).

CONCLUSION
The underestimation rate in ADH lesions was signi™cantly higher with US-guided 14-gauge automated biopsy compared to the 11-gauge SVAB. The underestimation rate was also signi™cantly higher in lesions greater than 7 mm regardless of the biopsy type, and in lesions biopsied usi ng SVAB that were regarded as suspicious (R4) or highly suggestive of malignancy (R5) on imaging.