To investigate the accuracy and clinical value of core-needle biopsy (CNB) in the diagnosis of breast space-occupying lesions.

We retrospectively analyzed the clinical data of 691 patients with breast space-occupying lesions in the Department of Breast Surgery, Southwest Hospital, Army Medical University from January 2012 to December 2016. All patients underwent preoperative CNB and surgical treatment. The preoperative CNB results were compared with postoperative pathological results. With the postoperative pathological result as the gold standard, the sensitivity and specificity of CNB were evaluated. Kappa test was used to analyze the consistency of CNB with postoperative pathology. χ² test was used to compare the underestimation rates of high-risk lesions and ductal carcinoma in situ by CNB.

The coincidence rate between CNB and postoperative pathological result was 96.0% (663/691), the false negative rate 4.9% (28/571), the sensitivity 95.1%(543/571), the specificity 100%(120/120), the positive predictive value 100%(543/543), the negative predictive value 81.1%(120/148) and the Yoden index 0.951, indicating a high consistency (Kappa=0.871, P<0.001). The underestimation rates of CNB on ductal carcinoma in situ and high-risk lesions were 27.8% (15/54) and 45.0% (18/40), respectively. There was no significant difference between two groups (χ²=2.992, P=0.084). In all patients, 17 patients were diagnosed with lobular tumor by surgery, and among them, three patients had the accurate diagnosis by preoperative CNB. Additionally, 44.4% (16/36) of patients with atypical ductal hyperplasia were diagnosed with malignant tumor after surgery.

CNB has a high accuracy in the diagnosis of breast space-occupying lesions, but there are false negative cases and histological underestimates. If necessary, a surgical resection biopsy is recommended for a definitive diagnosis.