Follicular Lesions of the Thyroid: W(h)ither Follicular Carcinoma?

from American Journal of Clinical Pathology

Richard M. DeMay, MD

Fine-needle aspiration (FNA) biopsy has brought about profound changes in the practice of medicine, and there is no better example than the introduction of FNA biopsy into the evaluation of thyroid disease. When FNA biopsy was added to the diagnostic armamentarium, the rate of thyroid surgery was cut in half, which resulted in substantial savings in health care expenditures.[1] Although we can certainly take pride in this success, numerous operations still are performed unnecessarily for benign disease. Many of these operations follow FNA biopsies interpreted as follicular lesions in which follicular carcinoma could not be excluded. But what is the real risk of follicular carcinoma, or of any malignant neoplasm for that matter, in these patients?

For purposes of this discussion, follicular lesions of the thyroid are defined as follicular carcinomas, follicular adenomas, and nodules of goiters. Specifically excluded from the present definition are follicular variant of papillary carcinoma and Hürthle cell tumors. Although follicular variant of papillary carcinoma enters into the differential diagnosis of follicular thyroid lesions, it has its own diagnostic features, primarily related to the nucleus (eg, grooves, pseudoinclusions, "Orphan Annie eyes"). Hürthle cell tumors present a unique set of diagnostic challenges and are not further considered herein.

Diagnosis of follicular lesions, as defined in the pre-ceding paragraph, has long been considered a diagnostic gray area in thyroid cytology. As cytopathologists, we generally admit that we cannot distinguish reliably between follicular adenomas and follicular carcinomas, which we therefore lump together as follicular neoplasms (we should also include cellular nodules of nonneoplastic goiters in this diagnostic gray zone).[1] Owing to our diagnostic limitations, many patients ultimately proven to have benign follicular lesions (goiters, adenomas) are unnecessarily operated on because we cannot exclude the possibility of follicular carcinoma by aspiration biopsy.

Follicular carcinoma traditionally has been considered the second most common malignant neoplasm of the thyroid (after papillary carcinoma). Current texts state that the relative frequency of follicular carcinoma is as high as 20% of cases of thyroid cancer (range, 5%-20%).[2-4] However, a contrary view is that follicular carcinoma of the thyroid is a disease that recently has become quite rare.

Follicular carcinoma is associated with iodine deficiency in the diet. In recent decades, salt has been supplemented with iodine, so that in the United States at least, dietary iodine deficiency is rare. And, there apparently has been a corresponding reduction in the incidence of follicular carcinoma. Unfortunately, however, pathologists continue to overdiagnose this disease. Benign lesions, such as partially encapsulated hyperplastic nodules of goiters, and malignant lesions, particularly follicular variant of papillary carcinoma, frequently are misdiagnosed histologically as follicular carcinoma.[3,5]

It is also possible that follicular adenomas are overdiagnosed histologically, i.e. many so-called adenomas may actually be dominant or pseudoencapsulated nodules of nonneoplastic goiters. At the University of Chicago Medical Center (Chicago, IL), follicular adenomas are diagnosed conservatively. A putative adenoma occurring in a background of any significant damage to the thyroid gland ("nodularity") is likely to be considered a nodule of a goiter rather than a true neoplasm, particularly if the interior of the lesion is histologically variable or shows evidence of degeneration. But whether truly neoplastic or not, adenomas are benign, by definition, and are not precursors of cancer.[2]

A review of the computerized anatomic pathology files of the University of Chicago Medical Center for a 5-year period (January 1, 1994, to December 31, 1998) revealed 680 partial or total thyroidectomies plus 13 thyroid biopsies, for a total of 693 thyroid specimens. There were 197 histologically documented thyroid malignant neoplasms, including 164 papillary carcinomas, 11 anaplastic carcinomas, 6 medullary carcinomas, 5 hematopoietic malignant neoplasms, 4 metastases, and 5 miscellaneous tumors. Seven cases originally were diagnosed as follicular carcinoma; however, after intradepartmental or extradepartmental review of these 7 cases, 5 were reclassified as follicular variant of papillary carcinoma. Of the remaining 2 cases of bona fide follicular carcinoma, 1 was minimally invasive (ie, microscopic vascular or capsular invasion) and 1 was widely invasive with liver metastases. Thus, there were only 2 cases of primary follicular carcinoma among 197 thyroid cancers, for a relative incidence of 1.0%. (Neither patient underwent preoperative FNA biopsy.) In addition, metastatic follicular carcinoma was diagnosed in 4 more patients (total 5), 2 of whom represented outside consultations (age range, 68-80 years). So, follicular carcinoma is a real disease, but its incidence, 1.0% (at least at the University of Chicago Medical Center), is far less than the 5% to 20% that traditionally has been cited in textbooks.

Others have had a similar recent experience with follicular carcinoma (Mark R. Wick, MD, oral communication, October 1998). LiVolsi5 reported that fewer than 2% of thyroid malignant neoplasms seen in her consultation practice were minimally invasive follicular carcinomas and that widely invasive follicular carcinomas were rarer still. Minimally invasive follicular carcinoma has an excellent prognosis, with cure rates of up to 95%, or even 100%.[1,3,4] It is the widely invasive form of follicular carcinoma that has a poor prognosis (approximately 50% mortality). Thus, it seems that when properly diagnosed, follicular carcinoma has become a rare disease.

Now let's consider follicular carcinoma from the point of view of FNA biopsy. Follicular lesions of the thyroid, as defined for this article, typically are divided into 3 cytodiagnostic groups, namely, colloid nodule, cellular nodule, and follicular neoplasm, based largely on the amount of colloid and the cellularity of the specimen.[1] The colloid nodule generally is considered to be a benign cytologic pattern in thyroid FNA biopsy. When this diagnosis is made, the patients usually are followed up clinically without surgery, unless there is some other indication for an operation, such as airway obstruction. On the other hand, most patients with a cytologic diagnosis of follicular neoplasm, as well as many patients with a diagnosis of cellular nodule, undergo surgery because of the presumed risk of follicular carcinoma. The chance of a follicular neoplasm (by cytology) being follicular carcinoma (by histology) traditionally has been given as about 20% to 25% (range, 10%-60%).[1] However, this may substantially overestimate the actual risk of cancer.