Reassessing specimen number and diagnostic yield of ultrasound guided breast core biopsy.

To reassess the minimum number of specimens required for an accurate diagnosis compared to the standard acquisition of five specimens. A total of 190 consecutive breast mass biopsies were performed using a 14-gauge core biopsy needle under ultrasound guidance. Two to six specimens were obtained from each mass and placed in sequential containers. Each specimen was evaluated by a pathologist in the order it was obtained and was labeled as "diagnostic' or 'non-diagnostic'. During the biopsy procedure, the radiologist indicated after which number biopsy he was confident that an adequate diagnostic specimen had been obtained. This was based upon real-time visualization of the needle passing through the lesion and whether the specimen sank or floated in formalin. These observations were compared with the pathologic diagnostic yield according to specimen number. Fifty-eight lesions (30.5%) were malignant, four (2.1%) were atypical, and 128 (67.4%) were benign. Histologic diagnosis was obtained after the first specimen in 157 (82.6%) lesions, the second specimen in 172 (90.5%) lesions, the third specimen in 186 (97.9%) lesions, the fourth specimen in 188 (98.9%) lesions, and the fifth specimen in 190 (100%) lesions. A histologic diagnosis was made in 90% of the malignant lesions after the first biopsy pass, 95% after the second pass, 98% after the third pass, and 100% after the fourth and fifth passes. Nine (4.7%) lesions showed discrepancy between the radiologist's confidence of diagnosis and pathologic diagnostic yield. Of these lesions, seven were benign and two were malignant. A diagnostic yield of 95% was obtained based on operator estimate of the minimum number of required core biopsies. A high diagnostic yield of 98% was achieved after three biopsy passes and 100% after five passes.