chapter 10
Thyroid
Aspiration Technique and Slide Preparation
Terminology for Reporting Results
Suspicious for a Follicular Neoplasm
Suspicious for a Follicular Neoplasm, Hürthle Cell Type
Atypia of Undetermined Significance (or Follicular Lesion of Undetermined Significance)
The main indication for fine-needle aspiration (FNA) of the thyroid is a thyroid nodule. The prevalence of thyroid nodules depends on how carefully one looks for them. Palpable nodules are found in 4% to 7% of adults,1,2 but the prevalence is much higher (20% to 70% of adults) when nonpalpable nodules are included, like those detected by ultrasonography or at autopsy.3–6
Although a thyroid nodule raises the suspicion of cancer, less than 5% are malignant.4,7 Given the high prevalence of nodules, combined with the impracticality of surgically excising all nodules, FNA plays a vital role as a screening test. Few cytology tests have so effectively decreased unecessary surgery while increasing the yield of malignancy.8–10 At the Mayo Clinic, the percentage of patients requiring thyroid surgery dropped from 67% to 43% 1 year after FNA was introduced (around 1980); the percentage of excised nodules that were malignant rose from 14% to 29%; and the cost of a workup of a thyroid nodule decreased by 25%.9 Since then, the reduction in unnecessary surgery has been even greater: The percentage of excised nodules that are malignant is now 45% to 56%.11,12
Every patient with a palpable thyroid nodule is a candidate for FNA. Palpation is not always accurate in assessing the presence and extent of thyroid nodularity, however. After ultrasound examination, 20% of patients with a palpable thyroid nodule prove not to have a nodule greater than 1 cm, and conversely, additional significant nodules are often found that were not appreciated on palpation.13 For this reason, a thyroid ultrasound is often obtained before (or at the time of) the FNA to confirm that the palpated nodule meets biopsy criteria.14
FNA can be avoided in patients who have a hyperfunctioning (“hot”) nodule (about 5% of all nodules) by also obtaining a serum thyrotropin (TSH) level.14 If the TSH level is normal or elevated, an FNA is usually performed. But if the TSH level is depressed, a radionuclide thyroid scan is obtained. If the scan confirms that the nodule is indeed hot, an FNA is not indicated because a hot nodule is very rarely malignant.15,16
An increasing number of thyroid nodules are being detected incidentally on imaging studies of the neck (thyroid “incidentalomas”). The various imaging modalities include ultrasound (for carotid artery disease), sestamibi scans (for hyperparathyroidism), computed tomography (CT), magnetic resonance imaging (MRI), and positron emission tomography (PET). It is not practical or advisable to perform an FNA on all incidentalomas. Evidence suggests that PET-avid thyroid nodules carry a high risk of malignancy and should undergo FNA. Nodules detected by CT, MRI, and sestamibi, however, should first undergo a dedicated ultrasound examination. Nodules less than 1.0 cm in maximum diameter on the ultrasound image are usually not biopsied unless they have sonographically suspicious features (e.g., microcalcifications).14,17
Aspiration Technique and Slide Preparation
The aspiration can be guided either by palpation or ultrasound. The benefits of palpation guidance include its reduced cost and logistical efficiency. Ultrasound guidance permits the operator to be certain that the nodule of interest is aspirated (Fig. 10.1A), reduces the number of unsatisfactory FNA specimens, and improves accuracy.14,18–20 For these reasons, ultrasound guidance is preferred for nonpalpable nodules, nodules that have a significant cystic component (greater than 25%), and nodules that were previously aspirated and yielded an unsatisfactory sample.14
Figure 10.1 A, Transverse sonogram of a thyroid nodule. The tip of a needle (thin arrows) is in the middle of a nodule (large arrow). (Courtesy Dr. Carol Benson, Brigham and Women’s Hospital, Boston, Massachusetts.) B, Ultrasound gel. Note its purple, meshlike or spider web–like appearance. If not wiped off the skin before aspiration, it can cover large areas of the slide surface and interfere with cellular evaluation (Papanicolaou stain).
The aspiration technique is essentially the same whether palpation or ultrasound is used for guidance. A very fine (25 or 27gauge) needle is ideal for most thyroid nodules, because the thyroid gland is very vascular.21 Depending on the circumstances and operator preference, the aspiration can be performed with or without suction applied by a syringe. Local anesthesia obtained by subcutaneous lidocaine injection is commonly used but is optional.21 The skin is wiped clean with an alcohol swab. If ultrasound guidance is used, the needle should not pass through gel, which, when aspirated, obscures cytomorphology (Fig. 10.1B).21 The dwell time of each pass should be 2 to 5 seconds, with rapid back-and-forth oscillations of the needle (three per second) within the nodule.21 If suction is used, it should be released before the needle is withdrawn from the nodule. The number of passes needed to ensure adequacy varies. If a cytotechnologist or pathologist is called to evaluate the specimen for adequacy, one or two passes may be sufficient. On-site evaluation is time-consuming,22 however, and many aspirations that are performed in outpatient settings are too far removed from a laboratory for such evaluation. Most practitioners recommend between two and six passes for each nodule aspirated.13,21,22 Even minor complications such as a transient hematoma are uncommon; site infections are almost never seen. Post-FNA infarction of the nodule occurs to some degree in up to 10% of cases.23 Other histologic alterations in the remaining gland include pseudoinvasion, vascular proliferation, and cytologic atypia, but these are usually easily recognized by virtue of their proximity to the linear biopsy tract.24 Serious complications like needle tract seeding are virtually nonexistent.
Slides are prepared by expelling and smearing the cells on a slide. Alternatively, or as an adjunct to smears, the needle is rinsed, and the resulting cell suspension used for cytocentrifuge, thinlayer, and/or cell block preparations. The advantages of thinlayer (”liquid-based”) preparations over smears include reduced blood; ease in preparation of consistently well-fixed slides, particularly when the FNA is not performed by a pathologist; and a concentrated specimen that requires less screening time. When liquid-based preparation is used, one slide is usually sufficient.25 Architectural features (macrofollicles, microfollicles) are retained,26 and adequacy rates and accuracy are similar for smears and thinlayer slides.27–31 Some adjustment to minor morphologic alterations with liquid-based preparations is needed. For example, chronic lymphocytic thyroiditis is more subtle on thinlayer preparations because the lymphoid cells are intermingled with contaminating blood leukocytes.28
Terminology for Reporting Results
The Bethesda System for Reporting Thyroid Cytopathology has been widely adopted in the United States and abroad for reporting the results of thyroid FNAs.32–34 It recommends that every thyroid aspirate interpretation begin with designation of a general diagnostic category (Table 10.1). Each of the categories has an implied cancer risk and is linked to an evidence-based clinical management guideline: Suspicious and malignant nodules are likely to be resected, whereas patients with a benign result are instructed to return for a follow-up examination at an appropriate interval. There are six general categories. Three of them come with a choice of two names; ideally, a laboratory chooses one of the options and uses it exclusively for reporting results that fall into that category.
TABLE 10.1
THE BETHESDA SYSTEM FOR REPORTING THYROID CYTOPATHOLOGY
Diagnostic Category | Risk of Malignancy (%) | Usual Management∗ |
Nondiagnostic (or unsatisfactory) | 1–4 | Repeat fine-needle aspiration with ultrasound guidance |
Benign | <1–3 | Clinical follow-up |
Atypia of undetermined significance (or follicular lesion of undetermined significance) | ∼5–15 | Repeat fine-needle aspiration |
Suspicious for a follicular neoplasm (or follicular neoplasm); specify if Hürthle cell type | 15–30 | Surgical lobectomy |
Suspicious for malignancy | 60–75 | Surgical lobectomy or thyroidectomy |
Malignant | 97–99 | Near-total thyroidectomy |
∗Actual management may depend on other factors besides the fine-needle aspiration result.
“Nondiagnostic (ND)” (or “unsatisfactory”) applies to specimens that are unsatisfactory due to obscuring blood, overly thick smears, air-drying of alcohol-fixed smears, or an inadequate number of follicular cells. For a thyroid FNA specimen to be satisfactory for evaluation (and benign), at least six groups of benign, well-visaulaized follicular cells are required, each group composed of at least 10 cells.35–37 Tissue fragments with multiple follicles can be split up and counted as separate and distinct groups.38 The minimum size requirement for the groups (at least 10 cells) allows determination (by the evenness of the nuclear spacing) of whether or not they represent fragments of macrofollicles rather than the more worrisome microfollicles. There are several exceptions to this adequacy requirement. A specimen with abundant colloid is adequate (and benign), even if six groups of follicular cells are not identified. A sparsely cellular specimen with abundant colloid is, by implication, a predominantly macrofollicular nodule and therefore almost certainly benign. Also, whenever a specific diagnosis (e.g., Hashimoto thyroiditis [HT]) can be rendered, and whenever there is any atypia, the specimen is considered adequate. ND results occur in 2% to 30% of cases.11,39,40
Specimens that consist only of cyst contents (macrophages) (Fig. 10.2) are problematic. Many laboratories have traditionally considered a macrophages-only sample unsatisfactory and included them in the ND category. (Because the parenchyma of the nodule has not been sampled, it is not possible to exclude a cystic papillary carcinoma.) In such laboratories, macrophages-only often constituted the great majority of ND cases, with rates that range from 15% to 30%.12,40–42 Other laboratories have considered the risk of a false-negative result negligible and reported macrophages-only as benign.11,41 In the Bethesda System, cyst fluid–only (CFO) cases are a subset of the ND category. The significance (and clinical value) of a CFO result depends on sonographic correlation. If the nodule is almost entirely cystic, with no worrisome sonographic features, an endocrinologist might proceed as if it were a benign result. In a study that segregated CFO cases and analyzed them separately, the risk of malignancy was 4%.40 The risk of malignancy for the ND category (excluding CFO cases) is 1% to 4%.11,39,40
Figure 10.2 Macrophages.
These cells have abundant cytoplasm that may be filled with hemosiderin or red blood cell fragments. They are a nonspecific finding, seen in both benign and malignant thyroid nodules (Papanicolaou stain).
A repeat aspiration with ultrasound guidance is recommended for ND cases.43 The repeat FNA is diagnostic in 50% to 88% of cases,12,35,40,41,44,45 but some nodules remain persistently nondiagnostic. Because about 10% of persistently nondiagnostic nodules are malignant,44 excision is often considered.
A benign result is obtained in about 70% of thyroid FNAs. The most common benign specimen is the benign follicular nodule. The false-negative rate for a benign interpretation is very low (less than 1% to 3%), but patients nevertheless should undergo repeat assessment by palpation or ultrasound at 6- to 18-month intervals.43,46 If the nodule shows significant growth or suspicious sonographic changes, a repeat FNA is considered.43
About 10% to 15% of cases are classified as “suspicious for a follicular neoplasm”; “suspicious for a follicular neoplasm, Hürthle cell type”; or “suspicious for malignancy.”11,12,32,42 Although virtually all patients with specimens classified as suspicious are referred for surgery, the predictive value for malignancy is different for these subcategories (see Table 10.1). The cases that fall into the “suspicious for malignancy” category are further subclassified as “suspicious for papillary carcinoma,” “suspicious for medullary carcinoma,” “suspicious for metastatic carcinoma,” “suspicious for lymphoma,” or other.47
Approximately 3% to 7% of thyroid FNA specimens are interpreted as malignant, most as papillary thyroid carcinoma (PTC).11,12,41,42
Accuracy
The accuracy of a benign thyroid FNA result is difficult to establish, because most patients with a benign result do not have surgery. Those who do undergo surgery represent a highly selected population of patients who have worrisome symptoms, larger nodules, or nodules exhibiting substantial growth. Nevertheless, data indicate that for FNA performed by experienced operators, a benign result is highly reliable. In a long-term follow-up study of 439 patients with benign cytology at the Mayo Clinic, only three proved to have a malignancy, for a false-negative rate of 0.7%, and none of the patients died of their disease.36 Current estimates place the false-negative rate between less than 1% and 3%.32 The most common cause of a false-negative is papillary thyroid carcinoma, followed by follicular carcinoma.12,36,41,48 Errors are due in equal part to sampling and interpretion.48
A malignant interpretation is likewise highly reliable. Experienced practitioners have false-positive rates of 1% to 3%,12,42,49 although higher rates have been reported.41 The most common benign tumors to cause a false-positive result are follicular adenoma (including the follicular adenoma with papillary hyperplasia) and hyalinizing trabecular tumor.12,41,49,50
Ancillary Molecular Testing
The three categories “atypia of undetermined significance (AUS),” “suspicious for a follicular neoplasm (including the Hürthle cell type),” and “suspicious for malignancy” are often referred to by clinicians as the “indeterminate” thyroid FNA categories. Although they are associated with reasonably well defined malignancy risks and management guidelines (see Table 10.1), they have inspired the development of molecular tests to triage patients more effectively for conservative versus surgical management and to further reduce unnecessary surgery for patients who have a benign nodule. Results suggest that molecular testing, particularly in AUS cases, can play a role similar to that of reflex human papillomavirus (HPV) testing for a woman with an atypical Pap test.
Among the most promising single markers is BRAF. A BRAF mutation is found in 44% of papillary thyroid carcinomas and virtually never in benign thyroid nodules.51 Testing AUS nodules for selected genetic markers of thyroid cancer—point mutations in BRAF and (H,K,N)RAS and gene rearrangements in RET/PTC and PAX/PPARγ—has a negative predictive value of 94% and positive predictive value of 88%52 and is available commercially as the Inform Thyroid Panel (Asuragen Inc., Austin, TX)
A gene expression classifier (the Afirma test, Veracyte, Inc., San Francisco, CA) has been developed to optimize the detection of benign nodules. The Afirma test was evaluated in a prospective, multicenter study at 49 clinical trial sites and found to have a negative predictive value of 95% for AUS, 94% for suspicious for a follicular neoplasm, and 85% for suspicious for malignancy.53 It has been estimated that the test could allow up to two thirds of patients with indeterminate FNA results to safely avoid surgery.54
Evaluation of the Specimen
Evaluation of the specimen begins with a review of the slide(s) under scanning magnification, which quickly provides a wealth of information. Most benign follicular nodules are sparsely cellular, consisting predominantly of colloid. Colloid can be very thin and translucent (“watery”); thick and opaque, with sharp outlines; or extremely thick and sticky (“bubble gum” colloid). Smears that have a high ratio of colloid to follicular cells generally indicate a benign thyroid nodule. The benign nature can be confirmed by documenting a predominance of intact macrofollicles and macrofollicle fragments (flat sheets comprised of evenly spaced follicular cells).
Rarely, a thyroid FNA sample is contaminated by nonthyroid cells. Accidental penetration of the larynx or trachea, virtually always announced by a cough reflex, occurs in less than 1% of thyroid FNAs,55 and cytologic samples contain ciliated columnar cells.
Benign Conditions
Benign Follicular Nodules
The cytologic term benign follicular nodule was coined to encompass a group of benign histopathologic entities that have identical cytologic features.56 The most common are the nodule in multinodular goiter (MNG) and the macrofollicular type of follicular adenoma.
MNG is a common thyroid disorder characterized by an enlarged thyroid gland (goiter) with multiple areas of nodularity. Worldwide, it is the most common endocrine abnormality, affecting more than 500 million people.57 Its prevalence varies depending on regional iodine intake: Lower dietary iodine correlates with an increased prevalence of MNG. In the United States, despite iodine supplementation, the prevalence of MNG is roughly 4% to 7%.
The pathogenesis of MNG is best understood in cases associated with iodine deficiency. Because iodine is required for thyroid hormone synthesis, a deficiency of iodine results in decreased thyroid hormone production and a compensatory elevation in serum TSH levels. Chronically elevated TSH levels stimulate a diffuse follicular cell hyperplasia. Over time, somatic mutations within the follicular cells (possibly as a result of hydrogen peroxide [H2O2] production and free radical generation that occurs with thyroid hormone synthesis) confer a survival advantage over selected clones, which results in nodule formation.58 The mechanism of follicular cell hyperplasia in iodine-sufficient regions is less well understood but is likely to be related to other stimuli (e.g., smoking, radiation, drugs, naturally occurring goitrogens) superimposed on a genetic susceptibility.58
MNG is 5 to 15 times more common in women than in men, and prevalence increases with age. Patients are usually asymptomatic and euthyroid, but 5% to 10% progress to hyperthyroidism (toxic MNG).57 Most cases are discovered incidentally by palpation or imaging studies. MNG varies in severity, from the minimally enlarged, asymptomatic gland with only one or two nodules, to the extremely enlarged nodular gland that extends from the neck into the mediastinum. Patients with large goiters can have compressive symptoms (e.g., shortness of breath, cough, dysphagia), and large goiters can be disfiguring. The growth of nodules is unpredictable and highly variable, but, on average, MNG nodules grow by about 4.5% per year.16
The treatment of MNG is varied and individualized. Choices include thyroid surgery, 131 I (radioactive iodine) therapy, and TSH suppresion with exogenous thyroxine (T4) administration. Surgery is generally recommended for younger patients and those with a large goiter.16
Histologically, the nodules of MNG (commonly referred to as adenomatous or adenomatoid nodules) are usually not encapsulated. The follicles within the nodules vary in size, but most are larger than normal follicles (macrofollicles) and filled with colloid. Some nodules are composed of such enormous macrofollicles that they are virtually all colloid (colloid nodules). Less commonly, a nodule may be very cellular, composed of smaller follicles that contain little colloid. The rapid growth of some nodules leads to hemorrhage, scarring, cystic degeneration, and dystrophic calcification. Uninodular MNG is a controversial entity, but some histopathologists acknowledge its existence.
Unlike MNG, a follicular adenoma is usually a solitary nodule that measures 1 to 3 cm in diameter, but it can be much larger.59 The histologic distinction between an adenomatous nodule in MNG and a follicular adenoma is somewhat arbitrary. In general, follicular adenoma is a solitary nodule with a well-defined fibrous capsule, and the follicular cells within the nodule are morphologically distinct from those in the surrounding gland.59 Capsular and vascular invasion are absent. Follicular adenomas have a wide variety of predominant histologic patterns: macrofollicular (large follicles distended by colloid); microfollicular (follicles smaller than normal); and trabecular (follicular cells are arranged in crowded ribbons).59 Uncommon variants include follicular adenoma with papillary hyperplasia (mostly in children and adolescents), signet ring cell follicular adenoma, mucinous follicular adenoma, lipoadenoma, clear cell follicular adenoma, and follicular adenoma with bizarre nuclei. Only the predominantly macrofollicular follicular adenomas are interpreted as benign follicular nodules by FNA. The others are likely to be interpreted as suspicious because of their unusual architectural patterns.
MNG and macrofollicular follicular adenoma are essentially indistinguishable by FNA, but this is of no clinical consequence. Patients with a benign result are instructed to return for a follow-up examination at 6- to 18-month intervals for at least 3 to 5 years to ensure that the nodule is behaving in a benign fashion.43,46
By definition, benign follicular nodule applies to a sample that is adequate for evaluation and consists predominantly of colloid and benign-appearing follicular cells in varying proportions. The benign follicular nodule has a predominantly macrofollicular architecture. During aspiration and slide preparation, most macrofollicles break into fragments; intact macrofollicles are relatively uncommon (Fig. 10.3A). Instead, one commonly sees flat sheets that represent fragments of macrofollicles (Fig. 10.3B) A macrofollicle fragment is defined by its flatness and the even spacing of follicular cell nuclei. (If there is conspicuous cell crowding and overlapping, the sheet is not a macrofollicle fragment.) Although the term macrofollicle fragment implies large size, and many are indeed large, some macrofollicle fragments are small (10 or even fewer cells). A macrofollicular pattern is defined by architecture and not size. Evenly spaced follicular cells (not crowded or overlapped) constitute a macrofollicle fragment, no matter the size of the cell sheet.
Figure 10.3 Benign follicular nodule.
A, The FNA sample contains small tissue fragments containing intact macrofollicles (Papanicolaou stain). B, Macrofollicles often break into fragments and appear as sheets of various sizes. Although colloid is scant in this case, a predominantly macrofollicular architecture can be inferred from these fragments (Papanicolaou stain).
The macrofollicle fragment is made up of medium-sized cells, each with a round nucleus, coarsely granular chromatin, an inconspicuous nucleolus, and scant or a moderate amount of cytoplasm (Fig. 10.4A and B). Cytologic atypia is generally absent, but some nuclear size variation can be seen, and in some cases the nucleoli are moderate in size. Some follicular cells contain hemosiderin pigment, which is golden brown with the Papanicolaou stain and blue with Romanowsky-type stains.60 In a benign follicular nodule, finding a minority of the follicular cells arranged in microfollicles or crowded groups is acceptable.
Figure 10.4 Benign follicular nodule.
A, The follicular cells comprising this unfolded macrofollicle are evenly spaced. Pale green–blue colloid has escaped from the follicle (Papanicolaou stain). B, Normal follicular cells have coarsely granular chromatin, and cytoplasm can be scant or moderately abundant (Papanicolaou stain).
Colloid is usually abundant and can take the appearance of dense blobs with a hyaline texture, hard edges, and cracking artifact (Fig. 10.5). These thick colloid chunks have a similar appearance on smears and liquid-based preparations. More often, colloid is mostly thin and watery, covering large areas of the smear as a translucent film with bubbles, folds, cracks, and a circular “chicken wire” artifact (Fig. 10.6A). On liquid-based preparations, this watery colloid has a characteristic “folded tissue paper” appearance26 (Fig. 10.6B). Colloid stains pale pink, pale green, or orange with the Papanicolaou stain, and magenta with Romanowsky-type stains.
Figure 10.5 Colloid—benign follicular nodule.
Colloid can congeal into opaque, irregularly shaped hyaline chunks with hard edges that are clearly visible on smears and liquid-based preparations (Papanicolaou stain).
Figure 10.6 Colloid—benign follicular nodule.
A, On smears, watery colloid often covers the entire slide as a thin, translucent film (pink in this image) and is often admixed with blood (Papanicolaou stain). B, On liquid-based preparations, watery colloid has a transparent, tissue-paper appearance with numerous linear folds (Papanicolaou stain).
Hürthle cells are seen in up to 50% of cases of MNG60 (Fig. 10.7). Macrophages with foamy or hemosiderin-laden cytoplasm, including multinucleated giant cells, are often present and are a nonspecific finding indicating cystic degeneration60 (see Fig. 10.2). A minor population of elongated, large “cyst lining cells” with pale and grooved nuclei are sometimes present. Because of their streaming (“tissue culture”) arrangement, they resemble reparative epithelial changes and likely represent altered follicular cells lining areas of cystic degeneration61 (Fig. 10.8).
Figure 10.7 Focal Hürthle cell metaplasia, multinodular goiter (MNG).
Hürthle cells are a common finding in some cases of MNG. The cells are large and polygonal and have abundant, finely granular cytoplasm. Variation in nuclear size, as seen here, is common in non-neoplastic Hürthle cell proliferations and does not signify neoplasia (Papanicolaou stain).
Figure 10.8 Cyst lining cells.
Reactive follicular cell changes are often seen adjacent to areas of cystic degeneration. They typically have a pulled-out appearance that mimics reparative epithelium (Papanicolaou stain).
The diagnosis of a benign follicular nodule can be made confidently when there is a mostly orderly macrofollicular architecture. Follicular carcinomas are virtually never composed of orderly macrofollicles, but instead have a preponderance of microfollicles and/or trabeculae of crowded, overlapping cells.
A minor population of crowded follicular cells and/or microfollicles is often present in benign follicular nodules and does not alter the benign interpretation. If macrofollicles (intact and/or fragmented) outnumber the crowded, microfollicular/trabecular groups, the findings are consistent with a benign nodule. Follicular cells trapped in fibrin clots are often disrupted and architecturally distorted, conferring a crowded appearance that may potentially lead to an erroneous suspicious interpretation. An interpretation should not be rendered on the basis of the presence of clotted follicular cells.62
It is important to note that not all MNGs and follicular adenomas are predominantly macrofollicular. Some nodules in MNG and some follicular adenomas are comprised predominantly of microfollicles and/or trabeculae and are inevitably interpreted as suspicious by FNA. This is a well-known limitation. Of note, morphometry and image analysis have not demonstrated sufficient reliability for classifying individual cases.63–66 Although some separation of MNG, follicular adenoma, and carcinoma is possible,67 these methods are not sufficiently accurate to be useful for clinical diagnosis.68 Neither is DNA quantitation by flow cytometry useful, because 25% of follicular adenomas are aneuploid.69–72 Immunohistochemistry for MIB-1 is of limited value because there is significant overlap in the proliferative activity between benign and malignant follicular nodules.73 Early reports on the application of immunohistochemistry for galectin-3 appeared promising,74 but subsequent studies have not been able to demonstrate reliable results for FNA. An understanding of the underlying genetics of these diseases may eventually provide the key to more accurate distinction, and encouraging inroads have been made. The translocation t(2;3) (q13;p25), which results in a PAX8-PPARγ gene fusion, is a relatively specific marker of follicular carcinoma that is virtually never present in follicular adenoma, MNG, or papillary carcinoma, but it has low sensitivity (26% by fluorescence in situ hybridization [FISH]).75
Focal cytologic atypia is present in some benign nodules and may present diagnostic difficulties. Hürthle cell (oncocytic) metaplasia is common in MNG, and the Hürthle cells sometimes show marked nuclear atypia. Such changes raise the possibility of a Hürthle cell neoplasm. If Hürthle cells are admixed with ordinary follicular cells arranged in macrofollicles, particularly if there is abundant colloid, it is appropriate to interpret the findings as benign. By contrast, Hürthle cell neoplasms yield an exclusive population of oncocytic cells. A suspicious interpretation should be rendered only when the sample is composed exclusively (or virtually exclusively) of Hürthle cells, particularly when many isolated cells are noted.76
The flat sheets of fragmented macrofollicles are reminiscent of the flat sheets of papillary thyroid carcinoma. The distinction between a benign nodule and papillary carcinoma depends primarily on careful examination of nuclear features. The distinction can be difficult in some cases, particularly with the macrofollicular variant of papillary carcinoma, in which the nuclear changes are often subtle and focal.77,78 Besides differences in the nuclei, the malignant cells of papillary carcinoma are arranged more haphazardly, with more crowding, than the follicular cells of benign nodules. In rare cases, macrophages in MNG can aggregate, and their large pale nuclei might be mistaken for papillary carcinoma.79
The large, pale nuclei of cyst lining cells bear a resemblance to the cells of papillary carcinoma. When the nuclear changes are mild, and the great majority of the sample looks benign, the benign, reactive nature of these cells is easily recognized. In some cases, nuclear atypia is marked and papillary carcinoma cannot be excluded. Depending on the degree of atypia, such cases are reported as AUS or even “suspicious for PTC.”61
Chronic Lymphocytic (Hashimoto) Thyroiditis
Chronic lymphocytic (Hashimoto) thyroiditis (HT) is a common autoimmune disease in which thyroid follicles are destroyed by a marked lymphoid infiltrate. Almost 95% of HT occurs in women, with a peak incidence between the ages of 40 and 60 years. The disease results in a diffuse, painless goiter with or without nodularity. Ultrasonography reveals a gland with a heterogeneous appearance. Most patients are hypothyroid. Histologic examination reveals an intense, patchy lymphoid infiltrate with germinal centers, follicular atrophy, and Hürthle-cell metaplasia. Fibrosis is prominent in a minority of cases. The diagnosis is established by correlating clinical findings with serologic test results. One or more of a variety of circulating autoantibodies are identified in almost all patients. The most common are anti thyroglobulin and anti thyroid peroxidase (TPO). (TPO was originally called the microsomal antigen.) In a patient with HT, FNA is performed only if there is a suspicious nodule that raises the possibility of a coexisting malignancy.
The aspirate is usually very cellular, with numerous lymphoid cells, including small lymphocytes, centrocytes, centroblasts, and dendritic cells (Fig. 10.9A and B). With liquid-based preparations, the dispersed lymphoid component is admixed with the white blood cells from contaminating blood. Dendritic-lymphocytic aggregates (germinal center fragments containing dendritic cells) are also present, as are tingible-body macrophages. Colloid is usually absent or scant. Normal follicular cells are infrequent or absent altogether; instead, there may be occasional clusters of Hürthle cells (Fig. 10.9C). The Hürthle cells and any residual follicular cells occasionally display focal chromatin clearing and nuclear grooves. Multinucleated giant cells can be present but are usually not as numerous as in subacute thyroiditis.
Figure 10.9 Chronic lymphocytic (Hashimoto) thyroiditis (HT).
A, Lymphoid cells are the predominant feature. Most are small, mature lymphocytes (Papanicolaou stain). B, The heterogeneity of the lymphoid cells is better appreciated on air-dried preparations (Romanowsky stain). C, Hürthle cells with abundant cytoplasm are usually identified in clusters in HT. Note the scattered lymphoid cells (Romanowsky stain).
Primary thyroid lymphoma is an uncommon malignancy that must be considered in the differential diagnosis of HT. Almost all cases of thyroid lymphoma arise in patients with longstanding HT, and patients with HT have a significantly higher relative risk for developing thyroid lymphoma compared with the general population.80 Histologically, primary thyroid lymphomas are either extranodal marginal zone B-cell lymphomas, diffuse large B-cell lymphomas, or a mixture of the two.59 Cytologically, thyroid lymphomas are composed of a uniform population of either small or large lymphoid cells. Because lymphoma is uncommon, immunophenotyping is not performed routinely on cases of typical HT and might actually be counterproductive. Some HT specimens contain clonal B-cell populations (demonstrated by a restricted κ/λ light chain ratio by flow cytometry or an IgH gene rearrangement by the polymerase chain reaction) that do not equate to malignancy and thus can be misleading.81,82 For this reason, immunophenotyping and/or molecular genetics are recommended only if the clinical presentation (rapid growth, large nodule) or cytomorphology (monomorphous and/or atypical lymphoid population) suggest lymphoma.
Subacute (de Quervain) Thyroiditis
Subacute (de Quervain) thyroiditis (ST) is a rare, painful enlargement of the thyroid gland in which thyroid follicles are damaged by a chronic granulomatous reaction, possibly due to a viral infection. Early, transient hypothyroidism is very common, but permanent hypothyroidism occurs only in a minority of patients. ST usually lasts several months and is self-limited, but it can recur in about 4% to 9% of patients.83,84 Patients are treated with nonsteroidal anti-inflammatory medications and, in more severe cases, with corticosteroids. In most cases, FNA plays no role in the diagnosis, which is usually established clinically. Because the gland can appear inhomogeneous, however, an FNA is sometimes performed to rule out a malignant nodule.
Aspirates can be sparsely, moderately, or highly cellular.85 Multinucleated giant cells are usually a prominent, striking finding. They are numerous and very large, with bizarre, angular shapes (Fig. 10.10A). In ST, they tend to have densely granular but not vacuolated cytoplasm, in contrast with the frothy cytoplasm typical of the multinucleated cells in MNG nodules with cystic degeneration.86 Some multinucleated cells have ingested colloid droplets.86 Granulomas (Fig. 10.10B) are the hallmark of the disease but are not always present. Granulomas are crowded aggregates of epithelioid histiocytes that have an oval, spindle-shaped, kidney bean-shaped, or curved (“boomerang”) nucleus and abundant granular or vacuolated cytoplasm. Cell membranes are poorly defined, and as a result, granulomas have a syncytium-like appearance. Granulomas, if present, are rare and need to be hunted for in most cases.85 Scant unremarkable macrofollicle fragments and colloid are usually present.85
Figure 10.10 Subacute thyroiditis.
A, The most conspicuous finding is an abundance of multinucleated giant cells with bizarre shapes (Papanicolaou stain). B, Granulomas are often few in number. The nuclei of epithelioid histiocytes have a variety of elongated and curved shapes. Cytoplasm is abundant, clear, and poorly demarcated (Papanicolaou stain).