Accuracy and pitfalls of frozen section during thyroid surgery.

The purpose of this study was to assess the accuracy of frozen sections performed during thyroid surgery and to define any pitfalls they may have. The material includes a series of 190 frozen section examinations of thyroid pathology. The overall accuracy of frozen section in this series was 95%. There were no false positives in our series and false negatives related to only follicular adenomas. In two patients, the report was deferred for permanent slide examination. Nine patients had lesions reported as malignant; however, the distinct cell type was not designated on frozen section. Eight specimens initially reported to have follicular adenoma on frozen section turned out to have angioinvasion or capsular invasion indicative of follicular carcinoma. Minor discrepancies between frozen and permanent sections in the types of benign disease were noted occasionally, but were not clinically significant. There was a discrepancy in the diagnosis of Hashimoto's thyroiditis on three occasions. The major discrepancy in patients with malignant pathology was related to the cell type. The diagnosis of anaplastic thyroid cancer, though suspected, was deferred for permanent sections in all cases. The diagnosis of medullary cancer of thyroid was difficult to make on frozen section. The major problem with frozen section was the diagnosis of follicular adenoma versus carcinoma. The benign diagnosis was changed in eight instances (out of 24 frozen sections) from follicular ademona to follicular carcinoma. Because of this experience, we do not provide the frozen section diagnosis to the patient and we wait until the final diagnosis is available.