Fine needle aspiration cytology of breast masses: an evaluation of its accuracy and reasons for diagnostic failure.

A study of 1458 patients who had undergone breast aspiration cytology was conducted to determine the diagnostic accuracy of the technique. The effect of tumour histology and size on the unsatisfactory aspirate and false negative rate was examined. Seven hundred and thirty one patients (50%) had histological diagnoses. The sensitivity of aspiration cytology for malignancy was 64% for the first aspiration, but was 91% in patients who had had 3 aspirates. The specificity was 56%, this low figure was almost entirely due to inadequate or unsatisfactory cytological preparations. The positive and negative predictive values of aspiration cytology were 99.4% and 85% respectively demonstrating high diagnostic accuracy given a satisfactory aspirate. Invasive lobular carcinoma yielded a significantly higher unsatisfactory rate than invasive ductal carcinoma (P less than 0.001) and fibroadenoma yielded a significantly lower unsatisfactory rate than fibroadenosis (P less than 0.001). Mass size influenced the unsatisfactory rate for invasive ductal carcinoma (P less than 0.05) and fibroadenoma, but not for invasive lobular carcinoma or fibroadenosis. Only 2 of the 32 false negatives were due to misinterpretation, the remainder resulted from the aspiration needle missing the mass. We conclude that aspiration cytology is an accurate preoperative diagnostic procedure for the evaluation of breast masses. Unsatisfactory or negative aspirates should be regarded as 'non-results' if there is clinical or radiological suspicion of malignancy.