Statistical analysis of fine needle aspiration cytology of the breast. A review of 678 cases plus 4,265 cases from the literature.

Between 1979 and 1984, 678 breast fine needle aspiration (FNA) cytologic specimens were received in our laboratory; tissue follow-up was available for 159 cases (23%). The diagnoses rendered in cases with subsequent tissue study were benign (41%), insufficient (10.5%), atypical and/or suspicious for carcinoma (10.5%) and malignant (38%). Using the tissue diagnosis as the standard, there were four false-negative cytologic results and one false-positive result. Considering only cases with a definitive diagnosis of benign or malignant, the sensitivity of cytologic interpretation for diagnosing malignancy on adequate material was 94% while the specificity was 98%; the overall efficiency of the test was thus 96%. The predictive values were 98% for a positive test and 94% for a negative test. Of the total number of submitted cases, a correct and definitive diagnosis was rendered 76% of the time. Calculation of similar statistics from six other series in which the FNA cytologic diagnoses were compared with the histologic diagnoses yielded data almost identical to our findings. The high degree of uniformity in the data indicates that these statistical parameters can be used as standards for evaluating the efficiency of breast FNA cytology.