OBJECTIVETo evaluate the accuracy and utility of ultrasound-guided core needle biopsy (CNB) in the diagnoses of breast lesions.

METHODSThe clinical data of 2152 consecutive breast lesions examined by CNB were reviewed. The histological agreement between core pathology and subsequent excision pathology was studied. The benign diseases without repeat biopsy were followed up.

RESULTSThere were 1461 cancers in final diagnosis among 2152 breast lesions, 1339 cancers were diagnosed by CNB. The false-negative rate of CNB was 3.5% (51/1461), and the underestimation rate was 4.9% (71/1461). In the repeat biopsy, carcinoma was found in 17 (50.0%) of 34 atypical ductal hyperplasia lesions and 25 (46.3%) of 54 papillary lesions. In 1461 cancers, the false-negative rate of ultrasound-guided CNB (2.1%, 22/1068) was significantly lower than that of free-hand-guided CNB (7.4%, 29/393) (P < 0.05). The false-negative rate of two special doctors for CNB (1.2%, 8/681) was significantly lower than that of other doctors (5.5%, 43/780) (P < 0.05). In 738 of benign lesions, 417 cases were excised and 50 malignant lesions were found, 205 cases were followed up by 2 - 29 months (median, 10.2 months), and one malignant lesion was found.

CONCLUSIONSUltrasound-guided core needle biopsy with histopathological assessment is an accurate method in diagnosis of breast lesions. Excisional biopsy is required to the high-risk lesions in CNB.