Preoperative evaluation and predictive value of fine-needle aspiration and frozen section of thyroid nodules

We sought to evaluate the predictive value of preoperative fine-needle aspiration (FNA) on surgical decision making by evaluating the final pathologic diagnosis and comparing it to the preoperative diagnosis. Further, we wished to calculate the predictive accuracy of each of several types of preoperative FNA diagnosis.

A retrospective chart review of 151 thyroid resections between July 1990 and April 1996 at the University of Virginia was undertaken. The mean age was 45 years (range, 11 to 85 years). Preoperative laboratory values, presenting symptoms, imaging studies, and predictive values of preoperative FNA and intraoperative frozen section were analyzed.

Symptomatology was poorly predictive of a benign versus malignant postoperative final pathologic diagnosis. Sensitivity, specificity, and accuracy of frozen section versus FNA was 86% versus 86%; 99% versus 93%, and 96% versus 92%, respectively, if the reading "cancer" or "suspicious" were predicted as positive for malignancy and "benign" or "follicular" were predicted as negative for malignancy. If only the reading "cancer" was predicted as positive for malignancy and only "benign" was predicted as negative for malignancy, sensitivity and specificity for FNA were 100% and 96%, respectively, and 100% and 99%, respectively, for frozen section. Forty-nine "follicular" lesions obtained by preoperative FNA resulted in 46 benign diagnoses after surgical resection.

The use of preoperative FNA is a powerful diagnostic tool in the hands of skilled pathologists. There is increasing evidence that intraoperative frozen section adds little to intraoperative decision making in patients diagnosed with thyroid cancer by preoperative FNA. Less definitive interpretations decrease the sensitivity, specificity, and accuracy of the FNA diagnosis.