OBJECTIVE: To evaluate the accuracy values of 276 fine needle aspriations (FNA) of breast lesions with a subsequent excisional biopsy diagnosis and to make a comparison between 25 studies of the literature using the same criteria to calculate those values. STUDY DESIGN: Cytologic findings were compared with the histologic diagnosis of each mass. The correlation Of results was analyzed by a decision-analysis approach, and the following values concerning diagnostic accuracy were calculated in the present study and in 25 other reports: sensitivity, specificity, positive predictive value, negative predictive value, false positive fraction and false negative fraction. To calculate those values, we eliminated unsatisfactory results and assumed that suspicious and positive cytologic findings represented carcinoma of the breast. RESULTS: Comparing our results with the means in the literature (numbers in parenthesis), FNA detected cancer with a sensitivity of 92.1% (87.7%), specificity of 98.6% (94.7%), positive predictive value of 99.4% (92.8%), negative predictive value of 82.1% (90.7%), false positive fraction of 0.6% (7.1%) and false negative fraction of 17.9% (13.4%); in 6.2% of cases the material was unsatisfactory (13.4%). CONCLUSION All the rates varied enormously between the studies and during the past 13 years. It seems that false positive and false negative fractions tended to diminish and stabilize in more recent years, and specificity and sensitivity underwent a slight increase. The differences between the rates of those studies suggest that FNA of the breast has some unavoidable limitations.
