Follicular nodules are the most commonly encountered problems in the surgical pathology of the thyroid. These lesions can be classified along the full spectrum of thyroid pathology from hyperplastic nodules to benign follicular adenomas and malignant fol-licular carcinomas.
Sporadic nodular goitre is characterised by numerous follicular nodules with heterogeneous architecture and cytology, features that have suggested a hyperplastic rather than neoplastic pathogenesis (7-10). The gland may be distorted by multiple bilateral nodules and can achieve weights of several hundred to a thousand grams, but this disorder is often identified as a dominant nodule in what clinically appears to be an otherwise normal gland. Histologically, the nodules are irregular; some are poorly circumscribed while others are surrounded by scarring and condensation of thyroid stroma, creating the appearance of complete encapsulation. They are composed of follicles of variable size and shape. Some follicles are large, with abundant colloid surrounded by flattened, cuboidal or columnar epithelial cells, often with cellular areas composed of small follicles lined by crowded epithelium with scant colloid in a small lumen, alone or pushing into large colloid-filled follicles as "Sanderson's polsters" (Figure 1). There may be focal necrosis, haemorrhage with haemosiderin deposition and cholesterol clefts, fibrosis, and granulation tissue; these degenerative changes are usually found in the centre of large nodules, creating stellate scars.
The morphologic classification of cellular follicular nodules in nodular glands can be extremely difficult. Hyperplasia may be extremely difficult to distinguish from neoplasia. Classical guidelines that allow distinction of a hyperplastic nodule from a follicular adenoma include the following: (i) multiple lesions suggest hyperplasia whereas a solitary lesion is likely to be neoplastic, (ii) a poorly encapsulated nodule is likely hyperplastic; a well developed capsule suggests a neoplastic growth, (iii) variable architecture reflects a polyclonal proliferation whereas uniform architecture suggests a monoclonal neoplastic growth, (iv) cytologic heterogeneity suggests hyperplasia; monotonous cytology is characteristic of neoplasia, (v) the presence of multiple lesions in hyperplasia means that areas similar to the lesion in question will be present in the adjacent gland; in contrast, neoplasms have a distinct morphology compared with the surrounding parenchyma, (vi) classically hyperplastic nodules are said not to compress the surrounding gland whereas neoplasms result in compression of the adjacent parenchyma. For the most part, large nodules in multinodular glands tend to be incompletely encapsulated and poorly demarcated from the internodular tissue. However, in some glands, large encapsulated lesions with relatively monotonous architecture
and cytology make distinction of hyperplasia from adenoma difficult. Many pathologists have applied nonspecific terms such as "adenomatoid nodules" to describe such lesions.
The pathophysiology of nodule formation remains poorly understood. The aetiology of this disorder has long remained elusive, since the goitres do not appear to be TSH-dependent (9). The work of Studer suggests that the initial proliferation is a polyclonal one involving cells that are intrinsically more rapidly growing than their neighbours (7,10,11). While the stimulus for growth is not certain, high levels of circulating thyroid growth-stimulating immunoglobulins (TGI) and defects in T suppressor cell function have been documented in patients with sporadic nodular goitre (12,13), implicating autoimmunity in the pathogenesis of this disease. Drexhage and colleagues (12) compared immunoglobulin preparations of patients who have goitrous Graves' disease with those of patients who have sporadic nodular goitre and have found that the former are approximately 10-fold more potent in inducing growth than the latter. It has been postulated that the weaker stimuli result in proliferation of only the most sensitive of the heterogeneous follicular epithelial cell population, hence the nodular-ity, and that "toxic" nodular goitre results from preferential replication of cells which are highly responsive to TSH stimulation (14,15). These data implicating an autoimmune pathogenesis explain the presence of chronic inflammation that is usually focally associated with nodular hyperplasia.
In contrast, molecular studies have indicated that the dominant nodules of multin-odular goitres are monoclonal proliferations, and therefore represent benign neoplasms (8,16,17). It may be that these represent true adenomas arising in the background of a hyperplastic process that is mediated by growth stimulating immunoglobulins. Moreover, most hyperfunctioning nodules are also now thought to represent clonal benign neoplasms with activating mutations of the TSH receptor or Gsa (18-22). The evidence of clonal proliferation in sporadic nodular goitre and the identification of ras mutations as early events in morphologically classified hyperplastic nodules in this disorder (23) indicates that the thyroid is a site for the hyperplasia-neoplasia sequence. Nevertheless, clinical experience has shown us that the vast majority of these lesions remain entirely benign.
Solitary follicular nodules have been unequivocally shown to be monoclonal (24,25,26) and in the absence of invasive behaviour or of markers of papillary carcinoma, these lesions are considered to be benign. Follicular adenomas are described as solitary encapsulated follicular lesions that exhibit a uniform architectural and cytologic pattern. However, the inclusion of nodules in sporadic nodular goitre in this category alters these criteria.
On aspiration cytology, the diagnosis of "follicular lesion" covers both follicular adenoma and follicular carcinoma, which are difficult if not impossible to distinguish because the diagnostic criteria do not rest on cytologic characteristics. The aspirate of a follicular lesion is usually cellular with follicular cells in sheets or microfollicular arrangements. The follicular cells are monotonous with elongated, bland nuclei and micronucleoli. Worrisome features include nuclear crowding, altered polarity, pleo-morphism, macronucleoli and coarse chromatin. The main practical role of cytology is to distinguish a colloid nodule or papillary carcinoma from a follicular neoplasm.
Follicular adenomas are well delineated and usually thickly encapsulated neoplasms that can be classified histologically according the size or presence of follicles and degree of cellularity, each adenoma tending to have a consistent microscopic pattern (Figure 2). The subclassification of follicular adenomas into simple, microfollicular, trabecular, oxyphil, atypical, papillary and signet ring cell types has no prognostic significance.
Atypical adenomas are highly cellular tumours with unusual gross and/or histologic appearances that suggest the possibility of malignancy but these tumours lack evidence of invasion. They may have necrosis, infarction, numerous mitoses or unusual cellularity. Many so-called "atypical adenomas" are indeed papillary carcinomas. The distinction of an encapsulated follicular variant papillary carcinoma from follicular adenoma relies on cytologic characteristics. The presence of the cytologic features of papillary carcinoma described below should indicate that diagnosis, despite lack of invasion. Whether some follicular nodules classified histologically as adenomas have the biologic potential to become carcinoma is not clear; aneuploid cell populations
have been described in a significant percentage of these lesions, suggesting that some of these may represent carcinoma in situ.
Follicular adenoma and most follicular carcinomas are indistinct with respect to their clinical presentation, radiographic appearance, cytologic findings and microscopic features. In most cases, the parenchymal component of both tumour types is essentially the same histomorphologically. The distinction between these two conditions has been considered possible only by recognition of invasion or metastasis. As indicated above, some encapsulated follicular adenomas exhibit evidence of aneuploidy and may in fact represent in situ follicular carcinomas.
Nuclear and cellular atypia and mitotic figures may be present in adenomas as well as in carcinomas and therefore cytologic characteristics are not helpful. Most follicular tumours are composed of cells with nuclei that are round to oval with uniformly speckled chromatin; the nuclei are evenly spaced and lack the crowded, overlapping appearance found in papillary carcinoma. As stated previously, these lesions cannot be
diagnosed as benign or malignant by fine needle aspiration; the diagnosis should be restricted to "follicular lesion".
Follicular carcinoma can only be diagnosed by the pathologist on high quality sections of well-fixed tissues that demonstrate capsular and/or vascular invasion (Figure 3). At the time of intraoperative consultation, frozen section will reveal only a very small number of these lesions, since the likelihood of identifying microinvasive foci on a single frozen section are low. The use of multiple frozen sections is not cost effective in the evaluation of these lesions (27).
Follicular carcinomas are divided into groups that reflect the biology of tumour growth and metastasis. Widely invasive follicular carcinomas, which are usually identifiable as invasive grossly, and certainly are not difficult to recognise as invasive microscopically, carry a poor prognosis with a 25-45% ten year survival (28,29). However, such lesions tend to be insular carcinomas (see below). In contrast, the more common scenario is that of minimal capsular invasion and patients with these tumors have an excellent prognosis. The diagnosis of follicular neoplasms requires very careful and thorough examination of the entire capsule of the follicular neoplasm by the pathologist (30). Minimally invasive follicular carcinoma is identified by invasion through but not widely beyond the capsule. Borderline lesions include those with invasion into the capsule beyond the bulk of the lesion but not through the full thickness of the capsule or situations in which islands of tumor are trapped within a capsule, associated with perpendicular rupture of collagen. The finding of nests, cords, or individual tumour cells within a tumour capsule leads some pathologists to the diagnosis of minimally invasive follicular carcinoma, however, this may represent an artefact in a patient who has undergone fine needle aspiration biopsy, with trapping by fibrosis or displacement of tumour cells into the capsule. The pathologist is therefore advised to carefully search for evidence of fine needle aspiration biopsy in the adjacent tissue. This would include finding focal haemorrhage, deposition of haemosiderin-laden macrophages, the presence of granulation tissue and/or fibrosis, all of which would indicate a needle biopsy site and the possibility of artifactual invasion rather than genuine invasion.
The concept of unencapsulated follicular carcinoma was raised by the identification of tumours that lack a capsule. In one report of four such cases, one patient developed metastases, and this gave rise to citations of a 25% metastatic rate by such lesions (31). However, this has not been substantiated in larger series and this concept has largely been abandoned.
Patients with minimally invasive follicular carcinomas are on average about 10 years younger than those with widely infiltrative carcinomas and since traces of capsule are found in about 24% of widely invasive lesions, it is possible that encapsulated follicular carcinoma is a precursor of the widely invasive lesion (32). Minimally invasive carcinomas have ten year survival rates of 70-100% (33) and therefore some argue that this disease does not warrant the painstaking search for microscopic invasion that distinguishes it from follicular adenoma. Nevertheless, the investigators that have reported these promising data have treated their patients for carcinoma rather than for benign disease (34).
Vasculoinvasive follicular carcinomas are aggressive and require management accordingly. While vascular invasion is more reliable for the diagnosis of malignancy, again the criteria are vague. Vascular invasion cannot be evaluated within the tumour and therefore again the circumference of the lesion is the site that warrants careful examination. Bulging of tumour under endothelium does not qualify as vascular invasion if the endothelium is intact. Nests of tumour cells within an endothelial lumen generally are accepted as representing invasion, however, it is recognised that artefactual implantation of tumour cells into blood vessels can occur during the surgical procedure or sectioning. Therefore, invasive tumour cells infiltrating the wall of an endothelial-lined space and thrombus adherent to intravascular tumour are required to distinguish true invasion from artefact.
Elastin stains are of little value in assessing vascular invasion, since the involved vessels are usually thin-walled veins with little if any elastic tissue. Immunohistochemical markers such as factor-8 related antigen, type IV collagen, CD31 and CD34 can be used to improve the recognition of vascular invasion in follicular carcinoma.
It is obvious that the diagnosis of malignancy in well-differentiated encapsulated follicular tumours rests on subjective criteria. The search for objective markers of malignancy has yielded only one candidate thus far; HBME-1, a marker ofmesothelial cells, is immunohistochemically detected in 40% of thyroid follicular malignancies of papillary or follicular differentiation (35-37) and has been used successfully in cytology studies as well as histopathologic evaluation (Figure 4) of thyroid nodules (36,37). Recent studies have advocated the use of galectin-3 as another marker of malignancy
(38-40). While this marker also stains normal, hyperplastic and inflamed thyroid tissue, positivity in malignancies is more diffuse and strong. These data should limit the application of this technique for cytology but this has not been widely recognized
Another molecular marker with application to follicular carcinoma is a gene rearrangement that involves the thyroid transcription factor Pax 8 and the peroxisome proliferator-activated receptor y (PPARy) gene (42). Normal thyroid follicular cells express Pax 8 at high levels; this transcription factor is essential for thyroid development, involved in regulating expression of the endogenous genes encoding thyroglobulin, thy-roperoxidase, and the sodium/iodide symporter. PPARy, a transcription factor that is implicated in the inhibition of cell growth and promotion of cell differentiation, is also expressed by normal thyroid follicular epithelium. However, this in-frame rearrangement results in a fusion protein that likely interferes with the normal function of both differentiating factors, thereby explaining its potential role in thyroid tumorigenesis. The rearrangement is most reliably detected using fluorescence in situ hybridization (FISH) technology to identify the translocation ofthe two genes that are normally localized on chromosomes 2q13 (Pax 8) and 3p25 (PPARy). The presence ofoverexpressed
protein can also be identified using immunostains for PPARy where strong nuclear staining identifies tumours harboring a translocation (Figure 5). Although follicular carcinomas of thyroid are rare (43), and the numbers of cases studied has been small, it appears to be a useful tool for the diagnosis of malignancy in thyroid follicular lesions, particularly to predict vascular spread and aggressive behaviour (44).
DNA aneuploidy is a well-recognised feature of human malignant tumors and it was initially hoped that ploidy analyses could help to distinguish adenomas from carcinomas of the thyroid. However, it has now been recognised that about 27% of follicular adenomas are aneuploid and about 40% of follicular carcinomas are diploid (45). Therefore such measurements are of limited diagnostic value for the individual patient. In contrast, however, ploidy may be a useful adjunct in determining prognosis.
The significance of this diagnosis must be interpreted in light of clinical data that assess the behaviour of this disorder. The dominant determinant of cause-specific mortality in patients with follicular carcinoma is the presence of distant metastases (46-48). Most studies have indicated that morbidity and mortality for patients with non-metastatic encapsulated follicular carcinoma is very low and correlates better with patient age than with any other parameter. Some have suggested that capsular invasion alone does not alter the incidence of distant metastases or cancer-related death (33).
Since the incidence of follicular carcinoma is low (43), most investigators still advocate total thyroidectomy and radioactive iodine therapy (34,49,50). The rationale for total thyroidectomy is not bilateral carcinoma; multifocal disease in follicular carcinoma is exceedingly rare and the identification of occult papillary carcinoma in the contralateral lobe is not an indication for further surgery (51). The only logical rationale for completion thyroidectomy is to allow selective uptake of radioactive iodine by metastatic tumour deposits rather than by residual thyroid gland. Uptake of radioactive iodine by distant metastases is a favourable prognostic factor and is improved by pre-therapeutic total thyroidectomy, resulting in improved survival (52-54). In contrast, external beam radiotherapy is not thought to be of use in patients with differentiated thyroid carcinoma, apart from those with locally advanced tumours such as widely invasive follicular carcinomas that involve extrathyroidal soft tissues of the neck and cannot be completely resected (54).
The last few decades have seen a decrease in the incidence of follicular thyroid carcinoma, probably due to dietary iodine supplementation (43). However, misdiagnosis of this tumour continues. Benign lesions, such as partly encapsulated hyperplastic nodules or nodules exhibiting pseudoinvasion after fine needle aspiration (55), are often overdiagnosed as malignant; papillary carcinomas with follicular architecture are often misinterpreted as follicular carcinoma. The clinical features, pathophysiology and biological behaviour of follicular cancer differ significantly from those of the entities with which it is often confused. Only careful histopathologic classification will allow correct evaluation of treatment options and prognosis.
Was this article helpful?