Performance of breast fine needle aspiration as an initial diagnostic tool: A large academic hospital experience.

Abstract

The clinical performance of the Yokohama reporting system for breast cytology remains uncertain. In this study, we retrospectively evaluated 318 breast fine needle aspirations (FNABs) from Los Angeles County Hospital over a five-year period, analysing data for breast cytology, histology, and radiology. Among 318 breast FNAB cases, 78.3% (249/318) were benign and 5.3% (17/318) malignant. Of 83 cases with follow-up histology, 14.5% (12/83) were insufficient, 66.3% (55/83) were benign, and 16.9% (17/83) were malignant. Of 55 benign cases, 61.8% (34/55) were fibroadenoma and 9 (9/55, 16.4%) were fibrocystic changes. Two cases were diagnosed as "atypical" but confirmed "benign" on core needle biopsy (CNB). No "suspicious" cases were found. Seventeen malignant cases were confirmed by CNB, including 70.6% (12/17) invasive ductal carcinoma, 11.8% (2/17) invasive lobular carcinoma, and one malignant phyllodes tumour. Receptor studies on cell blocks of three malignant cases showed concordant results with CNB results. In addition, 82.2% (148/180) of lesions with Breast Imaging-Reporting and Data System (BI-RADS) scores of 2 or 3 were benign and 92.3% (12/13) BI-RADS score 5 lesions were malignant on FNAB. Finally, 90% (67/74) of BI-RADS 4a lesions were benign, and 97% (36/37) of fibroadenomas were BI-RADS score 4a. This, by far the largest U.S. breast cytology study, showed 93.3% sensitivity, 100% specificity, 100% positive predictive value, and 98.2% negative predictive value for breast FNAB. Women with breast lesions of BI-RADS score 3 or less have a low risk of malignancy; FNAB would contribute to the reduction of excisional biopsies. FNAB can be considered as an initial diagnostic tool for BI-RADS 4 mass/lesions and satellite lesions, as well as for triaging patients.