It was not until the 19th century that Billroth, Kocher, Halsted, and others refined the thyroidectomy operation into a standard treatment for thyroid cancer with advancements in anti-septic technique, anesthesia, recurrent laryngeal nerve protection, and parathyroid preservation.^1,2 In the first half of the 20th century, oncologic resection for papillary thyroid cancer (PTC) commonly incorporated a “block dissection,” which sacrificed the sternocleidomastoid muscle, spinal accessory nerve, and marginal mandibular branch of the facial nerve resulting in significant deformity. George Crile Jr. heralded a more limited dissection with successful oncologic outcomes, which sparked the on-going debates regarding extent of dissection, implications of neck metastases, and prognostic factors for risk stratification.³ As early detection of PTC increased by the 1980s with the widespread use of diagnostic ultrasound and fine needle aspiration biopsy, controversy regarding the management of smaller tumors grew. Furthermore, with the development of radioactive iodine as a successful adjuvant therapy, there has been considerable debate regarding specific indications for administration.^4,5 Nevertheless, with these advancements, thyroid cancer has become a treatable disease and the stage has been set for development of modern-era treatments, particularly molecular-based targeted therapeutics.

Thyroid cancer is the most common endocrine malignancy accounting for over 60,000 new cases (3.6% of all cancer cases) diagnosed in 2013 in the United States.⁶ Its incidence has increased from 4.6 per 100,000 in 1975 to 11.6 per 100,000 in 2009, designating it as the fastest-growing cancer in recent years. However, the mortality rate has remained constant at approximately 0.5 per 100,000 with an overall 5-year survival rate of 94.2%.⁷

The increase in thyroid cancer is nearly completely attributable to papillary thyroid cancer (PTC). There has been a 2.9-fold increase of PTC over the past three decades, without a significant change in the incidence of follicular, medullary, or anaplastic tumors.⁸ PTC rates were also noted to increase most rapidly among females, rising nearly 100% in both the White non-Hispanic and Black populations.⁹ Furthermore, the highest rate of increase was for tumors smaller than 1 cm, accounting for 49% of the increased incidence of tumors between 1998 and 2002 with an annual percent change of 8.6% for men and 9.9% for women.^8,10 These overall trends, along with the unchanged PTC mortality rate, may suggest there has been increased detection of smaller tumors.

The growth in incidence is largely attributable to increased utilization and improvement of highly sensitive imaging modalities. Accordingly, the use of thyroid fine needle aspiration biopsy (FNA) has more than doubled from 2006 to 2011.¹¹ Considering that approximately 6% to 36% of the population have occult thyroid tumors upon autopsy and ultrasounds are able to detect lesions as small as 2 mm, it is reasonable to assume that the increased detection of occult tumors has led to the increased incidence.^12–14 However, while there has been a substantial increase in the number of sub-centimeter tumors, there has also been a smaller but significant increase in tumors of all sizes, including those greater than 4 cm and those with regional and distant metastasis.¹⁰ This suggests that other unknown factors such as genetic, dietary, and environmental influences may play a role in the increased incidence. Nevertheless, the number of thyroid operations for nodules has increased by 31% from 2006 to 2011, including a greater proportion of total thyroidectomy compared to lobectomy.¹¹ As more tumors are being diagnosed, it is important to identify and understand which tumors need more aggressive intervention.

Demographics and Clinical History

There exist patient characteristics and conditions that are correlated with an increased incidence of PTC. Females have a threefold higher incidence of thyroid cancer compared to men, and their age-specific rates peak earlier at the age of 47 compared to 53 in men. The effect of estrogen has been suggested as a causative factor, as studies have shown a positive correlation between parity and incidence of thyroid cancer.^15,16 However, oral contraceptives do not appear to significantly increase lifetime risk.^17,18 Taller stature and obesity have also been correlated with thyroid cancer in large observational studies.^19–22 The relative risk of PTC for patients with BMI >30 and height >180 cm (>170 cm in women) has been reported between 1.1 and 1.37.¹⁹ Additionally, childhood obesity and height one standard deviation above normal have been associated with a 1.2 increased lifetime risk of developing thyroid cancer.²³

Benign thyroid conditions (such as goiter and adenomas), hyperthyroidism, asthma, and benign breast disease have been linked to a higher risk for PTC.^24,25 However, the risk of thyroid cancer in patients with multinodular goiter (MNG) is unclear: while there appears to be a wide range of thyroid cancer incidence (4% to 18%) in patients with MNG across several studies,^26–28 others have reported a lower risk for thyroid cancer when compared to those with a solitary nodule. ^29–31 Nonetheless, up to 60% of tumors are detected in a nondominant nodules upon surgical pathology review, thus patients with MNG warrant close observation with consideration for FNA biopsy of up to four nodules if they are >1 cm or suspicious on ultrasound. ^28,32

In hypothyroid patients, elevated TSH, even within the normal range, increases the risk of PTC in patients with nodules.³³ Moreover, studies have shown that thyroid autonomy is a protective factor against the development of PTC.³⁴ It is important to note, however, that hypothyroidism in the absence of nodularity is not an observed risk factor.²⁵ Given these overall findings, accurately titrating levothyroxine dosage is of utmost importance in patients with thyroid disorders.

Autoimmune Thyroid Disease

Autoimmune inflammatory states generally predispose to tumorigenesis; however, there is considerable debate whether Hashimoto’s thyroiditis (HT) is a risk factor for the development of thyroid cancer. Some groups have shown no correlation, while others have demonstrated over a sixfold increase in relative risk of developing PTC.^35–39 Several studies have shown that patients with PTC in the setting of HT present with a better clinical stage and improved prognosis.^40–42 Importantly, however, selection bias plays a pivotal role in patients with HT, as they typically undergo more frequent thyroid surveillance with diagnostic ultrasound, and some speculate that this leads to the increased detection rate. Since most of the literature regarding this topic is based on retrospective data, it is difficult to draw concrete conclusions about the risk of a chronic autoimmune inflammatory state and the development of PTC.

Graves’ disease is another autoimmune disease whose correlation with thyroid cancer is controversial. Retrospective studies have found a low incidence of thyroid cancer in Graves’ patients.^43,44 However, several cohorts of variable sizes have found up to a 10-fold increase in thyroid cancer in patients with Graves’ disease, particularly within 3 years of diagnosis.^45–47 Similar to Hashimoto’s thyroiditis, there may be a selection bias by increased surveillance of patients with Graves’ disease, and these results must be interpreted cautiously. Further studies need to be performed to determine the true correlation between autoimmune disease and thyroid cancer.

Iodine and Nutrition

Iodine is necessary for the production of thyroid hormone, and its deficiency results in a rise in TSH and a subsequent thyrotropic state. Accordingly, there is a higher prevalence of thyroid nodules in iodine-deficient areas. While the prevalence of thyroid cancer is actually lower in these areas, the majority appear to be follicular and anaplastic carcinomas.^48,49 Iodine’s association with PTC is less clear: areas of the world with iodine supplementation have an increased incidence of thyroid cancer, with an increased ratio of papillary:follicular carcinomas ranging from 1.8:1 to 6.5:1.^50–52 According to one study analyzing the histologic subtypes of thyroid cancer before and after iodine prophylaxis in Argentina, the incidence of PTC doubled after iodine prophylaxis, while follicular and medullary carcinomas remained nearly equal.⁵³ This increase in PTC is likely multifactorial, influenced not necessarily by iodine supplementation alone, but by additional causes such as improved detection methods.

Dietary influences that may increase risk of thyroid cancer are high consumption of starch, butter, and cheese, with odds ratios ranging from 1.4 to 2.1 in some studies.^54–56 Conversely, consumption of raw fish, fruits, and vegetables may have a protective effect, with odds ratios ranging from 0.6 to 0.7.^54,57,58 Interestingly, according to a large pooled analysis of five prospective studies, current cigarette smoking and alcohol consumption greater than 7 drinks/week are associated with reduced risk of PTC, with hazard ratios of 0.68 and 0.72, respectively.⁵⁹

Radiation

Radiation exposure during childhood, either environmental or for medical treatment, is a major risk factor for the development of PTC.⁶⁰ The pediatric thyroid gland is very sensitive to the tumorigenic effects of external radiation, and post-radiation tumors can arise anywhere from 5 to 40 years after exposure. However, the risk of thyroid cancer is inversely correlated with age at exposure, with a high likelihood of tumorigenesis from childhood exposure and, although controversial, minimal risk from exposure after age 20.^61,62 Many studies evaluating the carcinogenic effects of ionizing radiation originated from the Chernobyl disaster in 1986. These studies demonstrate a linear dose-dependent risk of thyroid cancer that plateaus above 2 Gy of exposure in children who were irradiated at 15 years of age or younger.⁶³ Specifically, there is an estimated fivefold increased risk of PTC in patients with a childhood exposure of greater than 1.0 Gy of radiation.^63,64 As demonstrated by the Chernobyl disaster, irradiation results in the development of PTC tumors that are often poorly differentiated with aggressive features and have a strong association with RET/PTC rearrangements.

The incidence of thyroid cancer in patients with adult exposure is lower and more controversial.^64,65 Some groups have found a positive correlation between thyroid cancer and adult radiation exposure, but other studies suggest there is no association.⁶⁴ Moreover, adult patients who are exposed to iatrogenic radiation for medical diagnostics or therapeutics do not appear to have significantly increased risk for PTC. While patients who undergo radiation for Hodgkin’s disease have a 1.7% overall risk of developing thyroid cancer (ranging 9 to 15 times that of the normal population), these patients are usually irradiated either during childhood or early adulthood.^66,67 Furthermore, there does not appear to be increased risk for thyroid cancer after radiation for breast cancer treatment.^68–72

Familial Associations

Having at least one first-degree relative with thyroid cancer increases a patient’s risk of thyroid cancer up to 10-fold compared to the normal population.⁷³ Most cases of PTC arise sporadically, however, about 5% of cases are attributed to hereditary non-medullary thyroid cancer (HNMTC). These patients most commonly present with isolated primary thyroid tumors, which are classified as familial nonmedullary thyroid cancer (FNMTC); the remainder of HNMTC patients have a known Mendelian cancer syndrome, such as familial adenomatous polyposis (FAP), Gardner syndrome, Carney’s complex type 1, or Cowden disease.⁷³

FNMTC is defined by the presence of a primary well-differentiated thyroid cancer of follicular cell origin in two or more first-degree relatives in the absence of another predisposing hereditary syndrome. These tumors tend to present in younger patients as multifocal, bilateral, and more aggressive lesions with higher recurrence rates compared to sporadic PTC (16% vs. 9.6%).^74–76 Six potential chromosomal regions are implicated in FNMTC: MNG1 (14q32), TCO (19p13.2), fPTC/PRN (1q21), NMTC1 (2q21), FTEN (8p23.1–p22), and the telomere-telomerase complex. Genes commonly associated with sporadic PTC such as RET, TRK, MET, APC, PTEN, and TSHR have been excluded.⁷⁷ No difference in overall survival has been demonstrated in patients with FNMTC tumors compared to sporadic tumors, but disease-free survival has been shown to be lower in FNMTC.^78,79 As such, most endocrine surgeons recommend more aggressive treatment, including total thyroidectomy with prophylactic central neck lymph node dissection (CND) and adjuvant radioactive iodine ablation (RAI) in these patients.^78,80

Thyroid cell growth and oncogenesis is complex but appears to be largely tied to a few key molecular pathways: TSH working through cAMP, and molecular signals working through MAP kinase (MAPK), phosphatidylinositol-3-kinase (PI3 K), and RASSF1 cascades. Activating mutations of the TSH receptor and adenylyl cyclase are associated with benign functioning adenomas while constitutive activation of the MAP kinase pathway contributes to greater than 70% of PTC cases.⁸¹ The MAP kinase pathway includes three pivotal kinases, namely RAF, MEK (MAPK kinase), and ERK (MAPK), which ultimately regulate cell differentiation, proliferation, and survival.⁸² The PI3K-AKT pathway also regulates cell growth and is more commonly activated in FTC and poorly-differentiated tumors, but it also has interactions with the MAPK and RASSF1 cascades. The RASSF1 cascade is a pro-apoptotic cascade that becomes inhibited in thyroid tumors through RAF and AKT interactions. (Fig. 33-1)

These pathways play a central role in cell growth, proliferation, and apoptosis, as well as expression of proteins essential for thyroid function, such as the sodium-iodide symporter and thyroid peroxidase. Accordingly, dysregulation of these central pathways is implicated in tumorigenesis, particularly in radioactive iodine refractory (RAIR) tumors. It is also important to understand that while aberrancies in certain pathways may preferentially lead to a specific tumor type, there is a high degree of interaction between signal cascades. Identifying oncogenes and tumor suppressors, as well as their pathway interactions, provides guidance for risk stratification and development of potential targeted molecular therapeutics.

BRAF

BRAF serine-threonine kinase is a member of the RAF protein family that acts on the MAPK cascade. In its wild-type dephosphorylated state, it is activated by RAS binding at the cell membrane, which then phosphorylates MEK, leading to the activation of ERK and downstream signaling and transcription. A point mutation at nucleotide 1799 results in a valine-to-glutamate substitution at protein residue 600 (V600E): this altered protein causes continuous phosphorylation of MEK and constitutive activation of the signal cascade. It also has an inhibitory effect on the RASSF1-MST1-FOXO3 tumor suppressor pathway and a stimulatory effect on the PI3K-AKT pathway (Fig. 33-1). This mutation is present in melanomas and colorectal tumors, but is also noted as the most common genetic alteration in PTC. It is found in 36% to 70% of PTC cases, particularly in adults with classic or tall-cell variants.^83,84

Clinicopathologically, it has been shown to be associated with extra-thyroidal extension (ETE) and cervical lymph node metastasis; however, there is conflicting evidence regarding aggressive behavior in papillary microcarcinomas.^85–91 Furthermore, it has been associated with poorer prognosis, as patients are more likely to present with stage III and IV disease and are more likely to have recurrence even with early stage presentation.^92–95 Interestingly, BRAF mutations have also been found to impair tumor cell ability to express and target the sodium-iodide symporter (NIS) to the cell membrane. Thus, these tumors are unable to trap and retain radioactive iodine, which may contribute to their poorer prognosis.^96,97 Furthermore, silencing the activating BRAF mutation in BRAF-mutant PTC cell lines increases sodium-iodine symporter gene expression as well as ¹³¹I uptake in vitro.⁹⁸

Despite these findings, however, the utility of BRAF mutation positivity as a prognostic indicator has recently become a matter of debate. In a multicenter study evaluating the rate of central lymph node metastasis in PTC patients who underwent prophylactic central node disection (CND), BRAF mutation was initially found to be an independent predictor of central lymph node metastasis in the overall cohort; however, importantly, this association lost significance when follicular and tall-cell variants were excluded and the analysis only included classic variant PTC. Additionally, in a separate study evaluating 310 indeterminate FNA biopsies, preoperative BRAF mutation screening only had a 15% sensitivity of detecting malignancy in indeterminate nodules. Moreover, mutation positivity did not impact initial surgical management: 12 out of 13 patients with BRAF mutation in an indeterminate nodule initially underwent total thyroidectomy due to worrisome cytologic features and, thus, only one patient’s surgical management would have been altered.⁹⁹ Given these data, BRAF mutation screening alone may not be a robust independent preoperative diagnostic test.

RET/PTC

RET is a cell membrane tyrosine kinase receptor that is not expressed in normal thyroid follicular cells. However, it can be constitutively activated by a chromosomal rearrangement called RET/PTC, which results in the fusion of the 3’ portion of RET to the 5’ portion of a variety of extraneous genes. There are many RET/PTC rearrangements, but the most common forms are RET/PTC1 and RET/PTC3, caused by the fusion of RET to the H4 (D10S170) gene or NCOA4 (ELE1) gene, respectively.^100,101 The resultant fusion protein activates the RAS-RAF-MAPK and PI3K-AKT cascades and has been shown to promote tumorigenesis in vivo.^102–104 RET/PTC rearrangement is the second most common type of genetic alteration in PTC: its prevalence is approximately 20% in sporadic cases, being more frequent in young patients with prior irradiation.^81,105 Clinicopathologically, patients with RET/PTC rearrangements demonstrate more nuclear grooves, more psammoma bodies, and a higher rate of lymph node metastasis; however, 94% of these cases present with either stage I or II disease which portends a more favorable prognosis.^106,107 Notably, up to 13% of benign thyroid nodules have the RET/PTC rearrangement, which may predispose the nodule to faster growth.^108–110

RAS

RAS genes encode G-proteins located on the inner surface of the cell membrane and are key members of signal transduction from tyrosine kinase and G-protein coupled receptors. In its wild-type state, RAS has intrinsic GTP-ase activity that hydrolyzes activated GTP into GDP and transforms RAS back to its inactive state, resulting in cessation of downstream signaling. However, RAS point mutations inactivate the intrinsic GTP-ase function, which leads to a constitutively active GTP-bound state and subsequent phosphorylation of the MAPK and PI3K-AKT pathways, with preferential activation of the PI3K-AKT pathway.^107,111 This mutation accounts for less than 10% of genetic alterations in classic PTC, and are instead more commonly found in up to 50% of follicular thyroid carcinomas as well as follicular-variant PTC.^112–116 When found in PTC, these mutations are associated with lower risk of lymph node metastasis, but a higher risk of distant metastasis and mortality.^116,117 The RAS mutation is also present in up to 18% of benign nodules, commonly follicular adenomas, and has been implicated as a potential contributor to thyroid tumorigenesis.^114,115

NF- B

The NF- B family of transcription factors has an important role regulating inflammation and apoptosis whose amplified activation has been recently shown to be implicated in thyroid cancer tumorigenesis.^118,119 This pathway involves the upregulation of oncogenic proteins and appears to be intertwined with the PI3K-AKT and MAPK cascades.¹²⁰ RET/PTC, RAS, and BRAF genetic alterations can activate NF- B, suggesting a dual coupling of these oncogenes to the MAPK and NF- B pathways.^119,121 Accordingly, the NF- B pathway may have an important synergistic role with the MAPK cascade specifically through BRAF, as treatment of BRAF-mutated thyroid cancer cell lines with NF-B inhibitors potentiates the anti-tumor effects of MEK inhibitors.¹²²

RASSF1-MST1-FOXO3 Signaling Pathway

Activation of the RASSF1-MST1-FOXO3 cascade promotes pro-apoptotic activity, and its suppression is implicated in thyroid tumorigenesis. RASSF1 is a member of the RAS family, and when activated it phosphorylates MST1, which promotes translocation of transcription factor FOXO3 to the nucleus for transcription of pro-apoptotic genes.¹²³ The BRAF-V600E mutation has been shown to inhibit the ability of MST1 to phosphorylate FOXO3, thereby decreasing its translocation to the nucleus for transcription.¹²⁴ Furthermore, phosphorylated AKT can inhibit FOXO3 transcription by promoting translocation of activated FOXO3 from the nucleus back into the cytoplasm.^125,126 The FOXO3 pathway is thus implicated as a tumor suppressor pathway that becomes inhibited at least in part by the PI3K-AKT pathway and BRAF-mutated protein.

p53

The p53 tumor suppressor is mutated in up to 50% of all human cancers and is associated with poorly differentiated thyroid cancer.^127,128 The main modes of genetic alteration are somatic mutations, inherited mutations (e.g., Li-Fraumeni), and polymorphisms resulting in malignant phenotypes. Usually these genetic alterations are missense mutations caused by single amino-acid changes which can occur at a variety of different sites.¹²⁷ Genetic and immunohistochemical studies have identified p53 gene mutations and mutant P53 protein expression in thyroid cancer, usually as a marker of tumor progression to poorly differentiated or anaplastic thyroid carcinomas.^129,130 Furthermore, anaplastic tumors harboring PTC foci are found to have both p53 and BRAF-mutations, but not RET/PTC rearrangements, suggesting that some anaplastic tumors may be derived from BRAF-mutated PTC with an additional p53 mutation driving progression to a poorly differentiated state.¹³¹ Interestingly, some studies have identified expression of the mutated P53 protein without a genetic mutation in 13% to 27% of primary PTC tumors, for unclear reasons.^132,133 While the exact role and mechanism of p53 in PTC tumors is ambiguous, it is clear that p53 is a marker of aggressive disease and progression to a poorly differentiated phenotype. Clinicopathologically, PTC tumors with p53 mutations and P53 protein expression have aggressive features and are correlated with increased tumor size, lymph node metastasis, and advanced stage.^134,135

Non-PTC Genetic Alterations

Additional genetic alterations include RET point mutations (associated with medullary carcinoma and MEN2A/2B), PAX8-PPARg rearrangements (associated with follicular carcinomas), PI3K-AKT pathway gene amplifications (associated with anaplastic carcinoma), and tumor suppressor PTEN deletions (associated with follicular carcinoma in Cowden syndrome). These mutations are not prevalent in PTC.^107,111

Macroscopically, PTC lesions have variable gross features. Most tumors are located in a background of normal thyroid tissue and are commonly firm, white, and calcified with an average diameter of 2 to 3 cm. These lesions can also have variable solid, cystic, or sclerosing components, which may confound accurate diagnosis.¹³⁶ Extensive necrosis is not a typical feature and may be indicative of a higher-grade or aggressive tumor.¹³⁷ Microscopically, neoplastic papillae exist within a core of fibrovascular tissue and are lined by multiple layers of cells with crowded oval nuclei. These papillae can undergo progressive infarction, forming densely calcified “Psammoma bodies,” which are found in up to 50% of PTC cases; if they are seen in normal thyroid tissue or cervical lymph nodes, then there is strong evidence for PTC in the gland.^136,138,139

Most tumors will have papillary architecture focally or diffusely, and may contain follicles as well. The tumor cells themselves typically have a columnar or cuboidal morphology with eosinophilic cytoplasm and nuclei displaced from their normally-basal polarity.¹³⁶ Non-neoplastic papillae (as found in benign hyperplastic nodules) appear within a similar fibrovascular core and may be misdiagnosed as PTC; however, these cells’ nuclei are normal-appearing, basally-oriented, and round with even distribution, thus lacking the nuclear changes associated with PTC.¹⁴⁰ Neoplastic nuclei, on the other hand, are overlapping and ovoid in shape, usually displaying “grooves” created by folds in the nuclear membrane, nuclear clearing with hypodense chromatin (“Orphan Annie-eyed” nuclei), or intranuclear inclusions created by cytoplasmic invaginations of the nuclear membrane.^136,139,140 One must be meticulous and look for several prevalent pathologic features for accurate diagnosis of PTC because “Orphan Annie-eyed” nuclei can be seen in Hashimoto’s thyroiditis and nuclear grooves can be seen in Hashimoto’s thyroiditis, adenomatous hyperplasia, follicular adenomas, and diffuse hyperplasia.¹³⁶

As opposed to follicular carcinomas that spread hematogenously, PTC tumors more commonly invade the lymphatic system, which account for its high rate of multifocality (up to 78% overall, 61% bilateral) and lymph node metastasis (50%).^136,141 While there is evidence that multifocal PTC arise as monoclonal proliferations secondary to intra-thyroidal lymphatic metastasis of the primary tumor,^142,143 there is also evidence reporting each focus arising as an independent primary lesion.¹³⁹ Both mechanisms are plausible, but the prevalence and impact on prognosis of each remain to be elucidated.

Variants

There are numerous subtypes of PTC based on histopathologic variations. Classic PTC comprises up to 80% of all PTC cases, follicular variant PTC comprises 10% to 15%, and the other non-classic variants are much more rare.^144,145 We will discuss the more prevalent and aggressive variants including: follicular, tall cell, columnar cell, diffuse sclerosis, and solid (Fig. 33-2). While these subtypes have variable aggressive features, survival is largely dependent on tumor grade, size, and stage, not necessarily the histologic subtype itself.¹³⁷

The follicular variant resembles a follicular neoplasm upon low-power magnification: there are extensive follicular cells without papillary architecture, usually in the presence of a deeply eosinophilic, “scalloping” colloid.¹⁴⁶ However, its diagnosis hinges on cytologic features that resemble classic PTC: presence of psammoma bodies, nuclear grooves, powdery chromatin, and intra-nuclear inclusions.^147,148 More recently, further classification of fvPTC into encapsulated, partially encapsulated, nonencapsulated, and infiltrative subtypes have been described.¹⁴⁹ While exact clinicopathologic differences between each subtype are controversial, the infiltrative subtype appears to have a greater propensity for more aggressive cytologic features and lymph node metastasis without significant difference in survival or re currence.^149–152

The tall cell variant accounts for approximately 4% to 5% of PTC cases.¹⁴⁵ By definition, these tumor cells are two to three times taller than wide, and must comprise at least 50% of the cell population—though there is recent evidence that a lower percentage of tall cells may also confer the same clinicopathologic features.¹⁵³ There is also a predominant papillary morphology often with a lymphocytic infiltration.^136,139 Aggressive clinicopathologic features, such as extrathyroidal extension, necrosis, numerous mitoses, and distant metastases, are more prevalent with corresponding lower disease-specific and recurrence-free survival rates compared to classic PTC.^137,154

The columnar cell variant is a rare subtype of PTC, but it displays particularly aggressive features. They often present as large tumors (>6 cm) with papillary morphology, greater than 30% tall columnar cells, and scant clear cytoplasm. Nuclear features are atypical compared to classic PTC (and tall cell variant): they are hyperchromatic with dense punctate chromatin and nuclear pseudo-stratification.¹³⁹ Aggressive clinicopathologic features, such as extrathyroidal extension, numerous mitoses, and distant metastases, are more prevalent especially if there is extra-capsular invasion. This variant should be treated aggressively as 3-year overall survival is approximately 70%, with some case series reporting even worse survival rates.¹⁵⁵

The diffuse sclerosis variant accounts for 3% of all PTC cases, often presenting at a younger age. Diffuse tumor papillae, dense sclerosis, patchy lymphocytic infiltrate, and numerous psammoma bodies infiltrate throughout the thyroid.¹⁴⁶ Uniquely, serum antimicrosomal and antithyroglobulin antibodies may be elevated.¹⁵⁶ Aggressive clinicopathologic features, such as extrathyroidal extension, cervical lymph node metastasis, and distant metastasis, are highly prevalent; however, overall prognosis and survival rates are controversial.^136,155

The solid variant is a rare subtype that is more prevalent in women, children, and patients with prior radiation exposure. It displays the classic nuclear features of PTC, but morphologically consists of solid sheets of tumor cells occupying >50% of the lesion. Vascular invasion, extrathyroidal extension, and cervical lymph node metastasis appear to be more common; however, much like the other subtypes, the effect on overall survival and prognosis is controversial.^136,146,155

Adjuncts for Diagnosis

As several benign disease states can mimic PTC and the diagnosis of different variants may not be straightforward, several immunostains are available to help confirm a diagnosis of PTC. The most common immunostains include cytokeratin 19 (CK19), galectin-3, and HBME-1. Data suggest that HBME-1 is most specific for PTC, as some non-neoplastic lesions will stain for CK19 and galectin-3.^157,158 However, the most accurate test is to combine all three markers into one panel, which has a sensitivity of 85%, specificity of 97%, and diagnostic odds ratio of 95 for detecting PTC.¹⁵⁹ Other stains useful for questionable cases or confirmation of metastasis are CITED1, TTF-1, and thyroglobulin.^146,158,160

Patients typically present with an asymptomatic, palpable thyroid nodule discovered on routine physical exam or by the patient. Thirteen percent to 40% of patients may present with a nodule found incidentally on imaging performed for another reason.^161,162 Key elements of initial clinical evaluation include obtaining a history of rapid nodule growth, hoarseness, dysphagia, prior cervical or total-body medical irradiation, exposure to environmental radiation (such as Chernobyl), and a family history of thyroid tumors. Pain is an atypical symptom that may be worrisome for a more aggressive tumor.

On physical exam, the nodule should be palpated to assess for size, firmness, mobility, and tenderness (rare). Thorough evaluation of the cervical, submandibular, submental, peri-auricular, and supraclavicular regions is necessary to assess for lymphadenopathy, although submental and peri-auricular metastases are rare in PTC. While thyroid nodules are prevalent, 90% are benign and do not require intervention.¹⁶³

Seventeen different PTC staging methods exist, which reflects the level of disagreement as to which classification most accurately predicts overall outcomes. Each system attempts to incorporate either extent of disease or certain prognostic factors to determine outcome, but there are conflicting data as to which classification should be the standard. The most popular schemes are TNM staging (based on extent of disease), the AMES classification (based on age, metastasis, extent, and size), and the MACIS classification (metastasis, age, completeness of resection, local invasion, and size).^164,165 The most accurate staging system is controversial; however, several large retrospective studies have suggested that MACIS and TNM staging are superior.^166,167 Table 33-1 depicts the TNM stage along with each stage’s correlated survival risk. Approximately 75% of patients present with Stage I or II disease, which have greater than 90% 15-year cancer-specific survival.¹⁶⁶

These classification schemes can identify up to 85% of patients who have low-risk for mortality; however, they are unable to accurately predict long-term outcomes in the smaller proportion of high-risk patients. Additionally, these classifications only predict cancer-specific survival, not recurrence. In order to design a reliable staging scheme predictive for risk of recurrence, the American Thyroid Association (ATA) proposed the AJCC/International Union against Cancer (AJCC/UICC) classification that categorizes patients as either low risk, intermediate risk, or high risk (Table 33-2).¹⁶⁸ Furthermore, a subsequent validation study incorporated response-to-treatment in the classification with several interesting results.¹⁶⁸ First, they demonstrated that the likelihood of developing recurrent disease detected after a period of “no evidence of disease” (NED) was similar in all AJCC stages (1% to 2%). Second, they showed the new ATA risk stratification effectively defines short-term risk of recurrence and can be used to guide management during the first 2 years of follow-up: low-, intermediate-, and high-risk groups had a 3%, 21%, and 68% risk for recurrence, respectively. Lastly, their response-to-treatment analysis gives a valuable quantification of the risk for recurrence after either successful or unsuccessful therapy. Thus, in summary, the AJCC staging appears to be a reliable tool to predict survival, while the new ATA risk stratification appears to be a reliable tool to predict recurrence.