Ultrasound-guided core needle biopsy of the breast: does frozen section give an accurate diagnosis?

Reducing the period of uncertainty between the discovery of a breast tumor and histological diagnosis alleviates the psychological impact of breast cancer to an important degree. We aimed to verify whether histological results obtained with frozen sections of core needle biopsies (CNBs) offer an accurate and reliable tool for minimising this period. In 2619 cases we compared histological diagnosis on frozen sections with those on paraffin sections of CNB and finally with the results of open biopsies. Of the cases 49% were proved malignant and 51% benign. In 99.3% of the malignant lesions preceding CNB was correctly classified as B5 (n = 1185, 92.9%) or at least B4 (n = 82, 6.4%) in frozen and in paraffin sections. There were seven false-negative cases in frozen (false-negative rate = 0.5%) and five false-negative cases (false-negative rate = 0.4%) in paraffin sections of CNB. On frozen sections complete sensitivity was 99.5% and the positive predictive value of B5 was 99.9%. There was one false-positive case in frozen sections and one in paraffin sections. False-positive rate = 0.08%, negative predictive value for B2 = 99.4% for frozen and 99.6% for paraffin sections; full specificity was 85.9 for frozen and 85.8 for paraffin sections of CNBs. Immediate investigation of CNB in frozen sections is an accurate diagnostic method and an important step in reducing psychological strain on patients with breast tumors and may be offered by specialised Breast Assessment Units.