Diagnostic accuracy of ultrasonography-guided core needle biopsy for breast lesions.

INTRODUCTION This study aimed to assess the diagnostic accuracy of ultrasonography (US)-guided core needle biopsy (CNB) for breast lesions. METHODS We performed US-guided CNB of 733 lesions in 674 women from January 2003 to December 2005. Surgical excision was performed on 331 lesions. We compared the histopathologic findings of the CNB specimens with those of surgical specimens or with patients' long-term follow-up images. We also calculated the agreement, underestimation, sensitivity and false-negative rates. RESULTS The CNB results showed 334 breast cancers (46%), 28 high-risk lesions (5%) and 367 benign lesions (50%). Four (1%) lesions were categorised as inconclusive. The final diagnosis was breast cancer in 348 lesions. The kappa measure of agreement between the US-guided CNB results and surgical excision findings or follow-up results was 0.861 (p-value < 0.001). The underestimation rate was 40% (10 out of 25) for atypical ductal hyperplasia and 47% (14 out of 30) for ductal carcinoma in situ (DCIS). The CNB false negative rate and sensitivity for malignant lesions was 4% (14 out of 348) and 96% (334 out of 348), respectively. CONCLUSION US-guided CNB is an accurate diagnostic alternative to surgical biopsy in patients with breast lesions detected via US, although the high underestimation rates in DCIS and high-risk lesions are still a concern.