Utility of frozen section analysis on follicular lesions of the thyroid

The experience of one surgeon (R.H.) with intraoperative frozen sections (FS) performed on thyroid nodules over a 10-year period was studied to assess the utility of FS in follicular thyroid lesions. One hundred and ten patients with dominant or solitary nodules demonstrating a follicular growth pattern were evaluated. The FS slides and subsequently the permanent sections of the nodules were reviewed by the pathologists in the study (M.P.B., VAL.) without knowledge of the original diagnoses. Our results indicate: (1) if the FS was definitively benign (58 patients), the final diagnosis was benign [these lesions consisted of adenomatous nodule, nodular goiter, follicular adenoma, and Hürthle cell adenoma); (2) if an FS diagnosis of malignancy was rendered (4 patients), it was confirmed on permanent sections (follicular variant of papillary carcinoma in all 4); and (3) if the FS diagnosis was deferred (48 patients), the final diagnosis was benign in all but 10 (21 %) (of these 10, 6 had minimally invasive follicular carcinoma [2 of the Hürthle cell type], and 4 had follicular variants of papillary carcinoma). Overall, sensitivity, specificity, and accuracy rates for FS diagnoses were 29, 100, and 91%. Because unilateral lobectomy may be acceptable therapy for well-differentiated thyroid cancers, and because the efficiency of FS evaluation in diagnosing malignancy is low (only 4 malignancies of 110 total patients were diagnosed at FS, or 3.6% overall), we conclude that in this era of cost-containment, FS is not useful in the evaluation of follicular thyroid nodules identified preoperatively as follicular lesions by fine-needle aspiration cytology. Several recommendations concerning the 3 categories of FS diagnosis (i.e., definitively benign, definitively malignant-especially the follicular variant of papillary carcinoma-and deferred) are also put forward.