The thyroid gland gives rise to the most common endocrine tumors, and therefore, it represents the largest chapter in the new 5th edition of the WHO Classification of Endocrine and Neuroendocrine Tumors. The approach taken in this edition is somewhat different from its prior iterations; there is a focus on taxonomy following the approach of Carl Linnaeus [1, 2], and cytogenesis forms the basis of framework for this new classification, with histology and molecular features defining tumor types and subtypes.

Over the last 15 years, the importance of molecular biology in thyroid pathology has both revolutionized the discipline and, at the same time, proven the inherent value of classical histopathology. It is a field in which pathologists have long recognized patterns that reflect specific molecular alterations, but the addition of molecular tools to the pathologists’ armamentarium has enhanced our ability to prognosticate and predict the efficacy of targeted therapies.

This review will focus on the most important and novel changes in the new WHO thyroid tumor classification scheme (Table 1) employing a question–answer framework. We encourage the reader to use this as a framework to understand the new classification, but it is not a substitute for the actual text that provides detailed descriptions, illustrative figures, and a highly structured approach to assist in identifying the complexities of diagnosis and differential diagnosis of thyroid neoplasms.

Table 1 WHO classification scheme of thyroid neoplasms, 5th edition

Question 1: What Are the New Categories of Benign Follicular Cell Thyroid Lesions and Why Are They Included in the WHO 5th Edition?

The 4th edition of the WHO classification of endocrine tumors [3] included a single benign lesion: follicular adenoma (FA). While FAs are well-recognized tumors, they occur in many different scenarios with distinct clinical, radiological, biochemical, and morphological features. In this edition, the important variants are described.

The clinical entity known as multinodular goiter has been used for pathology diagnosis but this is inappropriate, since many lesions, including thyroiditis, hyperplasias, and neoplasms, can give rise to a clinically enlarged, multinodular thyroid gland. The entity that is most commonly associated with this clinical scenario is a disorder characterized by multiple thyroid lesions composed of follicular epithelial cells that have highly variable architecture; they can be very small or very large, they range from colloid-rich macrofollicular nodules to cellular microfollicular nodules, and they can be poorly delineated or well circumscribed with absent, well-defined or incomplete capsules (Fig. 1). These lesions have not generally been classified as neoplasms. Pathologists have used many different names for this enigmatic entity; they have been called “colloid nodules,” but most common diagnostic verbiages include the term “hyperplasia” as well as “adenomatous” and “adenomatoid,” reflecting the fact that the nodules in this disorder may morphologically mimic adenomas. Interestingly, multiple studies have shown that these nodules are frequently but not always clonal [4,5,6,7,8]; therefore, some are indeed adenomas, while others are hyperplastic. The clonality of these lesions explains why foci of malignant transformation can occur within the nodules of multinodular goiter. An alternative terminology proposed to address this enigma is “thyroid follicular nodular disease,” a term that avoids defining a lesion as hyperplastic, neoplastic, or the contradictory “adenomatous hyperplasia” [9]. This term achieved consensus support from the WHO editorial board.

Fig. 1
figure 1

Thyroid follicular nodular disease. Grossly, the thyroid gland is enlarged with variably sized multiple nodules (A). By light microscopy, this disorder manifests with a spectrum of morphologies from small colloid–rich nodules with Sanderson’s polsters (B) to large poorly defined colloid-rich macrofollicular nodules (C). They are usually multiple and generally show highly variable delineation and encapsulation (D). Some are more well-defined and microfollicular, resembling adenomas (E); however, while clonality studies have shown that some are monoclonal while others are polyclonal, there is no good correlation between clonality and morphology. Large lesions can have central degeneration with fibrosis and calcification (F)

An unusual but clinically important tumor is follicular adenoma with papillary architecture. This is a benign non-invasive encapsulated follicular cell–derived neoplasm characterized by a distinct “centripetal” intrafollicular papillary architecture that is more organized than papillary thyroid carcinoma (PTC), lacks nuclear features of PTC, and is often associated with autonomous hyperfunction (Fig. 2). Unlike follicular adenomas that harbor RAS mutations, these tumors are often associated with activating TSHR mutations (in up to 70% of cases) or GNAS mutations (in a small subset) [10,11,12] and/or EZH1 mutations [13, 14]. These molecular alterations result in activation of adenylyl cyclase, increased intracellular cyclic AMP, and unrestrained stimulation of function and proliferation [15]. These tumors are features of McCune-Albright syndrome due to germline mosaic GNAS mutations, and Carney complex, due to germline inactivating mutations in PRKAR1A that also cause constitutive activation of the cAMP-protein kinase A (PKA) pathway [16]. These clinical associations as well as the more common sporadic tumors that cause clinical or subclinical hyperthyroidism are important for clinic-pathological correlation, recognizing the radiological correlates of hot nodules.

Fig. 2
figure 2

Papillary adenoma — non-invasive encapsulated neoplasm characterized by a distinct “centripetal” intrafollicular papillary architecture lacking nuclear features of PTC

The importance of oncocytic change in the thyroid cannot be overemphasized therefore oncocytic follicular adenomas now hold their own special place in the classification. The term “Hürthle cell” is discouraged; it is actually a misnomer since Hürthle described the C cells of the thyroid gland. These tumors have distinct genomic alterations in the mitochondrial genome (mtDNA) [17,18,19,20] or in the related GRIM19 (NDUFA13) gene [21], and more than one-third have copy number variations [22]. It is well known that follicular adenomas can have focal oncocytic change; the definition of > 75% oncocytic cytology is used in this classification, but this remains to be proven as a valid criterion.

Question 2: What Distinguishes the Various Low-Risk Follicular Thyroid Neoplasms?

The 2022 WHO classification of endocrine tumors has organized follicular cell–derived neoplasms into three categories: benign neoplasms, low-risk neoplasms, and malignant neoplasms. The low-risk neoplasms are borderline tumors that are morphologically and clinically intermediate between benign and malignant tumors (Table 2). These neoplasms have the potential to develop metastasis, but the incidence of metastasis is extremely low. Histologically, they are classified into three types including non-invasive follicular thyroid neoplasm with papillary-like nuclear features (NIFTP), thyroid tumors of uncertain malignant potential (UMP), and hyalinizing trabecular tumor (HTT). The term “tumor” was intended to reduce the risk of overtreatment for these low-risk neoplasms. These terms are the same as those in the previous edition. In the 2017 WHO classification, HTT was described in a different chapter from that of NIFTP and UMP tumors, but in the new edition, they were all combined into one category of low-risk follicular cell–derived neoplasms [3].

Table 2 Pathological and molecular correlates of NIFTP and tumors of uncertain malignant potential, compared to other encapsulated follicular-patterned tumors

The diagnosis of NIFTP requires assessment of strict diagnostic criteria in a surgical resection specimen and requires meticulous microscopic examination of the entire tumor capsule/periphery to rule out invasive growth (Fig. 3A, B). The NIFTP terminology was proposed in 2016 and included as a new entity in the 2017 WHO classification [3]. Since then, there has been debate about the criterion allowing less than 1% of true papillae because some studies reported BRAF V600E and lymph node metastasis in a subset of NIFTPs with < 1% papillae [23,24,25]. To avoid misdiagnosing these malignant tumors as NIFTP, the NIFTP consensus group changed the diagnostic criterion of < 1% papillae to no well-formed papillae in 2018 [26]. However, in subsequent studies using the original criteria of < 1% true papillae, no adverse events were found in NIFTP patients [27,28,29,30,31,32,33]. In a study performed at Memorial Sloan Kettering Cancer Center, lymph node metastasis or tumor recurrence was not found in non-invasive encapsulated PTC even if the criterion for the percentage of papillae extended to 10% [29]. Therefore, in the absence of BRAF V600E mutation, the original criterion [27] allowing less than 1% true papillae remains unchanged in the 2022 WHO classification. It is important to distinguish between true papillae and pseudopapillary structures seen in NIFTP. While true papillae have a fibrovascular core lined by tumor cells with well-formed nuclear features of PTC, pseudopapillary structures are abortive (rudimentary) papillary formations without fibrovascular cores or hyperplastic-type structures called Sanderson polsters [27, 34].

Fig. 3
figure 3

A, B Representative histologic features of non-invasive follicular thyroid neoplasm with papillary-like nuclear features (NIFTP). The NIFTP is well circumscribed with a thin fibrous capsule without invasion (A). The tumor is composed of small- to normal-sized follicles and has nuclear features of papillary thyroid carcinoma showing enlarged nuclei, irregular nuclear membranes, and chromatin clearing (B)

The original study by the NIFTP consensus group did not include tumors ≤ 1 cm in size and oncocytic tumors fulfilling the histologic criteria of NIFTP. Therefore, these tumors were diagnosed as subtypes of PTC rather than NIFTP. However, these tumors are considered as subtypes of NIFTP in the 2022 WHO classification because they have been shown to behave like NIFTP with negligible risk of lymph node metastasis and tumor recurrence [28, 32]. Oncocytic NIFTPs are composed of at least 75% oncocytic cells [30]. Although the term “subcentimeter NIFTP” can be applied to any tumor < 1 cm, the diagnosis is usually unattainable in tumors ≤ 2 mm because it is difficult to be sure that the tumor is non-invasive and has < 1% true papillae [32]; however, even invasive tumors of this size do not warrant aggressive management.

Tumors of UMP are defined as “well-differentiated thyroid tumors with follicular architecture that are encapsulated or unencapsulated but well-circumscribed, in which invasion remains questionable after thorough sampling and exhaustive examination” in the 2022 WHO classification. The definition is the same as the previous edition. Thyroid tumors of UMP are divided into two subtypes according to their nuclear alterations: follicular tumor of uncertain malignant potential (FT-UMP) that lacks PTC-like nuclear features (nuclear score of 0–1) and well-differentiated tumor of uncertain malignant potential (WDT-UMP) that has more or less pronounced nuclear features of PTC (nuclear scores of 2–3). The term “atypical adenoma” is not recommended. Some of these tumors can have oncocytic features [35,36,37] and may display clear cells [38] or glomeruloid features and mucinous stromal changes [39]. Tumors of UMP are distinguished from follicular adenoma and NIFTP by the presence of questionable capsular or vascular invasion.

Encapsulated or well-demarcated follicular-patterned thyroid neoplasms are characterized by the high prevalence of RAS-like molecular alterations and the lack of BRAF V600E. Although molecular profiles of NIFTP are different from those of multinodular goiter, non-follicular PTC subtypes, and high-grade follicular cell–derived carcinomas, NIFTPs share molecular characteristics with tumors of UMP, follicular adenoma/carcinoma, and invasive encapsulated follicular variant PTC (Table 2). The overlap makes it difficult to distinguish NIFTP from other follicular-patterned tumors by molecular testing on pre-operative cytology specimens.

Among retrospectively reviewed PTC cases, the prevalence of NIFTP is lower in Asian countries (0.5–5%) compared to Western countries (15–20%) [33, 40,41,42]. The prevalence of tumors of UMP depends on whether the term is used in routine diagnostic practice. In institutions where UMP is diagnosed in daily practice, the incidence is 0.5–3% of all thyroidectomies [35, 43]. The prevalence of HTT is estimated to be less than 1% of thyroid neoplasms [44].

Thyroid lobectomy with clinical and radiologic surveillance is the treatment of choice for NIFTP and HTT. Radioiodine therapy after complete surgery should be avoided because these tumors almost always follow a benign course. UMP tumors require close follow-up since their biologic potential is not certain.

HTTs are well demarcated nodules with PTC-like nuclear changes, trabecular architecture, and a peculiar prominent intratrabecular hyaline material not seen in other thyroid neoplasms that has accumulated as the result of secretion of an active basal membrane type of protein (Fig. 4). The relationship between HTT and PTC was initially suggested by the detection of RET::CCDC6 rearrangements [45,46,47]; however, these findings were not confirmed in follow-up studies [48, 49]. Recent molecular studies have demonstrated that HTT is a distinct thyroid neoplasm with a specific molecular alteration. GLIS gene rearrangements define HTT and have not been identified in other thyroid tumors [49, 50]. Moreover, HTTs lack BRAF and RAS mutations [47]. The two most common rearrangement types are PAX8::GLIS3 (the most frequent type) and PAX8::GLIS1, which are formed by fusions of exon 3 of GLIS3 gene and exon 2 of GLIS1 gene to exon 2 of PAX8 gene, respectively [50]. These fusion gene transcripts lead to overexpression of the 3′ portions of the GLIS genes, which induces upregulation of extracellular matrix-related genes including collagen genes [50].

Fig. 4
figure 4

AC Hyalinizing trabecular tumor (HTT). The HTT is solid and well delineated from surrounding thyroid tissue (C). Tumor cells are arranged in trabecular architecture. The cells are elongated or polygonal and oriented perpendicular to axis of trabeculae. The cytoplasm is abundant and eosinophilic with hyaline material (D). The MIB1 immunostain shows diagnostic membranous staining (E)

Fig. 5
figure 5

A, B Classic PTC subtype showing complex papillary growth pattern (A) lined by cells with nuclear features of papillary thyroid carcinoma (B)

Fig. 6
figure 6

A, B Tall cell PTC subtype showing tightly packed follicles and papillae (A), tumor cell height at least 3 × the width, eosinophilic cytoplasm, and easily identifiable nuclear features of PTC (B)

Fig. 7
figure 7

A, B Columnar cell PTC subtype consisting of columnar cells with pale to eosinophilic cytoplasm and prominent pseudostratification (A). CDX2 expression is noted in around 50% of these tumors (B)

Fig. 8
figure 8

Hobnail PTC subtype showing papillary growth and tumor cells with oncocytic cytoplasm and enlarged nuclei, bulging from the apical surface

Fig. 9
figure 9

A, B Solid PTC subtype showing solid and trabecular growth pattern (A) with variable nuclear features of papillary thyroid carcinoma. There is no tumor necrosis. The mitotic activity is less than 3 per 2 mm2

Fig. 10
figure 10

A, B Diffuse sclerosing PTC subtype. An infiltrative tumor (A) with numerous psammoma bodies and associated chronic lymphocytic thyroiditis.Tumor cells arranged in solid nests and papillary formations with squamous metaplasia (B)

Fig. 11
figure 11

A, B Warthin-like PTC subtype showing papillary growth (A) with core of papillae containing lymphoplasmacytic infiltrate (B)

An HTT diagnosis can also be confirmed by a peculiar membrane staining of MIB1 antibody tested at room temperature, in conjunction with a positive expression of follicular markers (thyroglobulin, TTF1, and PAX8). The specific GLIS fusion product can be identified by molecular techniques and also by immunohistochemistry. Tumor cells with PAX8::GLIS3 fusion express strong nuclear and cytoplasmic immunostaining using antibody to the C-terminus region of GLIS3 [49, 50]. While this biomarker is largely unavailable in most diagnostic immunohistochemistry laboratories, the detection of a GLIS rearrangement or GLIS protein expression enables a pre-operative diagnosis of HTT in cytology specimens [51]. The intratrabecular eosinophilic hyaline material resembles amyloid but is negative with the Congo red stain. The hyaline material is diastase-resistant material that stains with the periodic acid-Schiff stain and is immunoreactive for collagen IV.

While thyroid lobectomy alone is curative and no metastases are found in almost all patients with HTT [44], lymph node or distant metastases have been reported in exceptionally rare cases [44, 52,53,54]. The malignant counterpart of HTT has tumor capsular or vascular invasion. No patients with GLIS-rearranged thyroid neoplasms have developed tumor recurrence or any other adverse events to date [50, 51]. However, it is unclear whether a metastatic thyroid tumor with a GLIS rearrangement has existed or could occur.

Question 3: Why Is Follicular Variant of PTC Separated from the Other Subtypes of PTC? What Distinguishes It from Follicular Thyroid Carcinoma? What Are Histological Subtypes of Follicular Thyroid Carcinoma and Follicular Variant Papillary Thyroid Carcinoma?

Molecular studies have taught us that the morphologic features of differentiated thyroid carcinomas correlate with two major classes of mutations found in these tumors. RAS-like mutations result in tumors that have an expansile pattern of growth and subtle/less florid nuclear atypia, whereas BRAF-like mutations give rise to infiltrative tumors with florid nuclear atypia. The history of classification of thyroid carcinoma, which initially relied on architecture, became convoluted when classification became based more on nuclear features, the origin of “follicular variant of PTC (FVPTC).”

FVPTC has two distinct variants including the infiltrative and encapsulated forms with invasion of blood vessels or of the tumor capsule. The former subtype is an infiltrative malignancy with all the features of classical PTC except papillae. It has florid nuclear atypia, psammoma bodies, and fibrous stroma and often exhibits perineural and lymphatic invasion. In contrast, the encapsulated form resembles follicular thyroid carcinoma, growing as an expanding lesion with a well-defined border that may or may not incite a fibrous reaction to create a tumor capsule; it then invades locally into the capsule or adjacent tissue (if there is no capsule), and when there is involvement of vessels, they are frequently the blood vessels of the tumor capsule rather than lymphatics. Molecular studies have shown that infiltrative FVPTC is a BRAF-like tumor, a member of the PTC family, whereas encapsulated FVPTC is a RAS-like neoplasm; placing it closer to follicular thyroid carcinoma (FTC) than to PTC. The infiltrative FVPTC category should only be used in the absence of true papillae, and there is still a debate among experts whether infiltrative FVPTC with a BRAF V600E mutation may represent a classic PTC with predominant follicular growth and subtle papillae which may be identified in subsequent levels.

The distinction of encapsulated FVPTC from FTC is based entirely on nuclear morphology. The definition of FTC includes the statement that it lacks nuclear features of PTC. However, it is well known that the diagnosis of nuclear features of PTC suffers from tremendous interobserver variability [55, 56]. When subtle nuclear atypia due to convolution of nuclear membranes is included [57], virtually all FTCs have nuclear atypia, making the distinction almost academic.

Both FTC and encapsulated FVPTC have cytologic variations including oncocytic and clear cell change; however, these are not clinically relevant and therefore are not classified as subtypes in the new WHO classification. FTCs and FVPTCs that are composed of > 75% oncocytic cells with no high-grade features qualify as oncocytic carcinomas and FVPTCs [58, 59].

Both tumor types are subtyped based on the type and the form and/or degree of invasion that are essential in determining dynamic-risk stratification and clinical management. These tumors may be only minimally invasive (tumor capsular invasion only), may invade into blood vessels (angioinvasive FTC or FVPTC), or may be widely invasive and each type has a different prognosis. The 40-month disease-free survival has been reported as 97% for minimally invasive, 81% for angioinvasive, and 45% for widely invasive FTC [60], and similar clinical behaviors have been reported for encapsulated FVPTC [61].

There is still controversy about the definition of angioinvasion and whether the criteria used are more important than the number of vessels involved, but despite this, the presence of a single focus of angioinvasion (vascular invasion) in the absence of widely invasive growth is diagnostic of angioinvasive FTC or FVPTC. Minimally invasive tumors are uniformly considered to be low risk and can be treated with local resection alone. In contrast, those that are widely invasive into surrounding parenchyma or angioinvasive tumors may require completion thyroidectomy and adjuvant therapy to prevent loco-regional recurrence and/or distant metastasis, based on clinical dynamic risk assessment.

Question 4: What Are the Diagnostic Criteria and Clinical Significance of the Subtypes of Papillary Thyroid Carcinoma?

Papillary thyroid carcinoma (PTC) is the most common malignancy of follicular cell derivation in both adult and pediatric populations [62]. PTC commonly occurs as a sporadic tumor; however, familial forms are being increasingly appreciated [63]. Until the 2017 WHO classification of thyroid tumors [3], PTC was exclusively diagnosed based on characteristic nuclear cytology regardless of growth pattern and invasive features [64, 65]. This changed with the introduction of the diagnostic term NIFTP [27]; similar to the previous edition of the WHO [3], either papillary growth or invasion was added to the definition of PTC in the 5th edition of WHO classification of thyroid tumors. Molecular studies have shown that encapsulated purely follicular-patterned lesions are RAS-like and more closely resemble follicular thyroid carcinomas. Thus, invasive encapsulated follicular variant lesions are not classified in the same group as PTC that is a BRAF-like family of malignancies [66]

The incidence of PTC has been gradually increasing worldwide; most agree that this is a consequence of current trends in screening and diagnostic practices [67,68,69]. Even though a majority of diagnosed PTCs measure ≤ 1.0 cm, several authors have also reported a rise in the number of large PTCs, possibly as a result of evolving trends to diagnose PTC based on nuclear alterations [70,71,72].

Based on molecular data, PTC can be considered as a “less well-differentiated” form of carcinoma than follicular carcinoma or encapsulated follicular variant papillary carcinoma [9, 73]. The frequently encountered molecular events in PTC are either point mutations or gene rearrangements involving the MAPK pathway [73,74,75,76,77]. BRAF V600E is the most common molecular alteration in classic PTC and its subtypes with papillary growth pattern and infiltrative tumors with follicular architecture. These BRAF-like tumors show focal to diffuse papillary growth and readily identifiable characteristic nuclear features; they are most often infiltrative, but can be localized with expansile growth or pushing borders, or confined to a cyst [66, 78]. Telomerase reverse transcriptase (TERT) promoter mutations, as a secondary pathogenic event, are encountered in 10% of PTCs and are usually associated with an aggressive clinical course [73, 79, 80]. RET gene rearrangements (CCDC6::RET and NCOA4::RET) are found in classic PTC and other subtypes [81, 82]. A strong association is reported between RET rearrangements and radiation-induced PTC [81]. Other less common molecular variations in PTC include gene fusions in NTRK [83] and other genes, mutations in other genes, copy number variations, alterations in gene expression, and altered microRNA expression [73, 74, 84, 85].

In the new WHO classification of thyroid tumors, the term “variant” has been replaced by “subtype” to allow for consistency with other WHO tumor classification schemes and avoid confusion with the molecular diagnostic term “genetic variant(s).” The list of PTC subtypes with their key histologic and molecular features is depicted in Table 2.

Traditionally, PTCs measuring ≤ 1.0 cm have been called papillary microcarcinoma, papillary microtumor, occult, and incidental and occult sclerosing PTC. It is well documented in the literature that most cases of these small PTCs, when identified as incidental findings, carry an excellent prognosis, similar to other incidental neoplasms throughout the body [86, 87]. However, there does exist a group of these tumors that display aggressive pathologic features and clinical behaviors, including regional and distant metastasis and structural recurrence after surgery [88,89,90,91]. Rare cases have even been fatal to the patient, with progression to high-grade carcinoma usually present in metastatic lymph nodes [92, 93].

Therefore, it is not farfetched to conclude that all so-called incidental and occult PTCs are not created equal and in fact the terminologies used may be misleading for patients and treating clinicians. Therefore, in the 5th edition of the WHO classification of thyroid neoplasms, it is recommended that “PTC-microcarcinoma” should not be considered as a distinct subtype. This is also in alignment with clinical management guidelines which rely on multiple pathologic features rather than solely on size to develop personalized risk stratification protocols for patients diagnosed with PTC [94,95,96].

Papillary Thyroid Carcinoma Subtypes (Table 3) (Figs. 5, 6, 7, 8, 9, 10, 11)

Table 3 Key histopathologic criteria and molecular profiles of subtypes of papillary thyroid carcinoma (Figs. 410)

Classic PTC is the paradigm for all PTC subtypes; it is defined by well-formed papillae lined by tumor cells with nuclei showing a specific set of nuclear features: enlargement, peripheral margination of chromatin, clearing of nucleoplasm, irregular contours forming grooves, and resulting in cytoplasmic pseudoinclusions. Lymphatic permeation by PTC is the cause of a high rate of regional lymph node metastasis. Vascular invasion is less common [65, 97]. The encapsulated subtype is completely encased by a thick fibrous capsule which can be either intact or infiltrated partially or in its entire thickness by tumor. The encapsulated classic PTC lacking invasive features is associated with excellent clinical prognosis [65, 98,99,100,101].

Infiltrative follicular variant PTC is a BRAF-like lesion that has the infiltrative growth pattern of classic PTC but lacks prominent papillae; it has predominant follicular architecture but florid nuclear atypia, prominent psammoma bodies, and stromal fibrosis, and careful examination usually (but not always) identifies focal small papillary structures.

Among PTC subtypes, tall cell (TC), columnar cell (CC), and hobnail (HN) subtypes are of undisputable clinical significance due to their aggressive clinicopathologic features as compared to classic PTC [102,103,104,105]. The risk stratification scheme developed by the American Thyroid Association identifies these as having intermediate risk of structural recurrence [67]. The aggressive histologic subtypes of PTC can also be fully encapsulated and/or present as clinically low stage tumors lacking pathologic features such as extrathyroidal extension, lymphatic and vascular invasion, and lymph node metastasis [102, 106,107,108,109].

BRAF V600E mutations are most common in TC-PTC (approximately 90% of cases) as compared to other subtypes; TERT promoter mutations have also been reported in some cases. The other molecular events seen in TC-PTC include loss of heterozygosity for chromosome 1 and TP53 mutations [110]. The CC-PTC is also associated with BRAF V600E mutation; less common are BRAF fusions, RAS mutations, TERT promoter mutations, and loss of CDKN2A and TP53 mutations [111]. Most HN-PTC cases harbor BRAF V600E mutations, typically associated with mutations in TP53, TERT promoter, and PIK3CA [112,113,114].

Other subtypes of PTC are highlighted in the WHO 5th edition of thyroid neoplasms. The diffuse sclerosing (DS)-PTC is characterized by diffuse unilateral or bilateral involvement of the thyroid gland with extensive lymphatic infiltration, dense sclerosis, numerous psammoma bodies, and associated chronic lymphocytic thyroiditis [115, 116]. The solid/trabecular subtype has solid, trabecular, or nested growth pattern that can mimic poorly differentiated carcinoma but lacks necrosis and prominent mitoses. An aggressive clinical course can also be encountered in diffuse sclerosing and solid subtypes of PTC [104, 105, 117]. Other subtypes without known impact on prognosis include the following: the oncocytic classic PTC with well-developed papillae and tumor cells with oncocytic cytoplasm (> 75%) and PTC nuclei [58, 59, 118, 119]; the Warthin-like PTC essentially an oncocytic PTC with papillary growth and heavy lymphoplasmacytic infiltrate in its stroma that bears morphologic similarities to Warthin tumor of salivary glands; and the rare clear cell subtype [120, 121].

Other less common PTC subtypes discussed in the 5th edition of WHO classification of thyroid neoplasms include spindle cell PTC and PTC with fibromatosis/fasciitis-like/desmoid-type stroma [122,123,124]. The former can be difficult to distinguish from other neoplasms without the use of proper ancillary tools. The latter is an unusual tumor that has two discrete components: a BRAF-mutated PTC embedded within a fibromatosis that has CTNNB1 mutation and nuclear localization of beta-catenin.

Question 5: What Are the Subtypes and Features of “High-Grade” Thyroid Carcinoma?

The search for the criteria to identify and diagnose the group of thyroid carcinomas with a prognosis intermediate between the favorable outcome of differentiated follicular cell–derived thyroid carcinomas (papillary and follicular carcinoma) and the very poor outcome of anaplastic carcinoma has been a topic debated for decades [125]. Poorly differentiated architecture — solid, trabecular — was proposed in the mid 1980s as a criterion [126]. Insular carcinoma [127] — which combined poorly differentiated “insular” architecture with high proliferative grade (mitotic activity, tumor necrosis) — was regarded as the prototype of this tumor group [128]. The Turin consensus criteria [129] — endorsed by the 2017 [3] as well as the current WHO classification of tumors of endocrine organs — clarified the histologic criteria to diagnose a poorly differentiated thyroid carcinoma and validated its prognosis as intermediate between well and undifferentiated (anaplastic) carcinomas. As early as 2000 [130], it was proposed to grade PTC based on tumor necrosis and high mitotic rate. It has subsequently become clear that proliferative grading — defined on the basis of high mitotic activity and tumor necrosis — also identifies tumors of intermediate prognosis, regardless of histologic differentiation in terms of papillae, follicles, or solid/trabecular/insular growth patterns [125, 131,132,133]. Thus, the new WHO classification recognizes two groups of high-grade non-anaplastic follicular cell–derived carcinomas that have intermediate prognostic risk (Table 4):

  1. A.

    Poorly differentiated thyroid carcinoma (PDTC): These are invasive, high-grade follicular cell–derived carcinomas that are histologically poorly differentiated because of their solid, trabecular, and insular growth patterns (or combinations of these) [129].

  2. B.

    Differentiated high-grade thyroid carcinoma (DHGTC): These are invasive high-grade follicular cell–derived carcinomas that are still differentiated since they retain the distinctive architectural and/or cytologic properties of well-differentiated histotypes of carcinoma of follicular cell derivation, such as nuclear features and/or architecture of papillary carcinoma and follicular growth pattern of follicular carcinoma [132, 133].

Table 4 Prognostically relevant classification of follicular cell derived carcinomas of the thyroid

This approach is justified by the recognition that approximately 50% of high-grade non-anaplastic thyroid carcinomas will not take up radioactive iodine [134], and new treatment modalities, in particular systemic therapies focusing on the particular molecular signature of the tumors, may be needed to treat these patients [135]

From clinical and epidemiologic viewpoints, high-grade non-anaplastic carcinomas of follicular cells, both PDTC and DHGTC, share certain characteristics. They are rare, ranging from less than 1 to 6.7% of all thyroid carcinomas. Higher frequencies are seen in Europe, and South America, while lower prevalence rates are reported in North America and Japan. This suggests the possibility of either ethnic or dietary (iodide) factors in the development of these tumors. Clinically, they occur in adults usually over age 50 and develop as rapidly growing masses and with a slight female preponderance [136]. Surgically, the tumors are often large (4 cm or more) and extend beyond the thyroid, with gross vascular invasion, infiltration of perithyroidal soft tissues and skeletal muscle, and perineural attachment or invasion. About 30 to 50% are associated with easily identifiable lymph node metastases [127, 131, 136, 137].

Pathologically, these tumors are grossly widely invasive, but may rarely appear partially encapsulated. Foci of hemorrhage and tumor necrosis may be seen macroscopically [127, 131, 136, 137].

The differences between PDTC and DHGTC are found at histologic examination (Table 5) (Figs. 1213). PDTC has solid trabecular or insular growth. In some instances, tumor cells have small dark nuclei with a convoluted “raisin-like” appearance, reminiscent of papillary carcinoma nuclei [127, 129]. Rarely, large pleomorphic nuclei may be identified. The hallmark of PDTC is the presence of tumor necrosis which may occur as small areas or large swaths of necrotic material with ghost outlines of tumor cells and nuclear dust. If tumor necrosis is absent, mitotic count should reach at least 3 mitoses per 10 high-power fields/ ~ 2 mm2 for the carcinoma to qualify as poorly differentiated. Mitotically active oncocytic carcinomas often have necrosis: since they typically have solid or trabecular growth, they usually fulfill the criteria for PDTC [138, 139]. Rare PDTCs are composed of clear cells [140].

Table 5 Diagnostic criteria for high-grade follicular cell–derived thyroid carcinomas
Fig. 12
figure 12

Poorly differentiated thyroid carcinoma showing solid nests, absence of nuclear features of papillary carcinoma, and tumor necrosis

Fig. 13
figure 13

High-grade papillary thyroid carcinoma. This photomicrograph illustrates tumor necrosis in a BRAF-like high-grade papillary thyroid carcinoma

On the other hand, DHGTC shows a growth pattern similar to well differentiated tumors and this is papillary in the vast majority of cases. Nuclear features characteristic of papillary carcinoma may be present throughout, although some areas of the tumor may show nuclear enlargement and pleomorphism. The characteristic histologic feature to confirm the diagnosis is necrosis and/or excess mitotic activity (5 mitoses per 10 high-power fields/ ~ 2 mm2, 400 ×) [131]. Vascular, lymphatic, perineural and extrathyroidal invasions are commonly found [132, 133].

The use of high-power fields is no longer supported by the new WHO classification schemes, since these vary depending on the microscope and ocular used; the WHO encourages the use of mm2 that is a standard measure irrespective of the microscope used and applicable to digital whole slide images [141]. However, since the majority of studies in this field have used high-power field measures, these are retained in the description until more accurate data can be obtained.

Immunohistochemical staining in both PDTC and DHGTC indicates that they are positive for TTF1, PAX8, cytokeratins (usually cytokeratin 7), and thyroglobulin [3]. Thyroglobulin tends to be weak and focal with dot-like reactivity. The Ki67 proliferation index is high, usually in the range of 10 to 30% [142]. Ancillary immunostaining is indicated to exclude other tumor types, such as medullary carcinoma, parathyroid carcinoma, and metastases to the thyroid gland and may be used to confirm vascular invasion. Of particular concern are aggressive medullary thyroid carcinomas with mitotic activity and tumor necrosis [143,144,145]; these often feature solid and trabecular growth: immunostaining for calcitonin, chromogranin, and monoclonal CEA may be necessary to distinguish them from follicular cell–derived poorly differentiated tumors.

From a molecular biology standpoint, PDTC and DHGTC harbor driver mutations in BRAF (BRAF V600E), RAS, or — much less frequently — gene fusions (usually RET or NTRK3). In addition, they also carry aggressive secondary mutations, most frequently of the TERT promoter and in some cases of PIK3CA and TP53 [73, 135, 146]. Poorly differentiated thyroid carcinomas are enriched in RAS mutations, a consequence of their strict definition that requires the absence of papillary carcinoma nuclear features [133, 146]. In contrast, the vast majority of DHGTC are BRAF V600E-driven since most display the cytoarchitectural features of papillary carcinoma [132, 133, 146]. This likely explains the higher propensity for cervical lymph node metastases in DHGTC [133].

In large studies, PDTCs (based on the Turin proposal) have a 10-year overall survival of 46% [147] and a 60% disease-specific survival at 10 years [133]. High-grade non-anaplastic follicular cell–derived carcinomas that do not fit the Turin proposal (i.e., DHGTC) have an approximately similar disease-specific survival (56% at 10 years) [133], although disease-free survival may be worse for high-grade papillary thyroid carcinoma compared with PDTC [132]. The prognosis of poorly differentiated oncocytic thyroid carcinoma is apparently similar to that of its non-oncocytic poorly differentiated counterpart [59, 138, 139].

Rare forms of high-grade non-anaplastic follicular cell–derived carcinomas are encapsulated and confined within the thyroid [148]. Although some of these tumors are angioinvasive, once the tumor is resected, patients have median disease-specific survival of 8 to 10 years [148]. Poorly differentiated carcinoma is a rare finding in teenagers. A series of six cases culled from three institutions showed that these tumors harbor somatic mutations in DICER1; two of the six patients also carried germline mutations and thus had DICER1 syndrome. A 30% mortality rate due to tumor was noted [149].

Question 6: What Is New in the Understanding of Anaplastic Thyroid Carcinoma and Why Is “Squamous Cell Carcinoma” Included in This Section?

In the previous WHO classification [3], squamous cell carcinoma of the thyroid (i.e., carcinoma composed almost entirely of squamous cells without a differentiated carcinoma component) was considered a separate entity from anaplastic thyroid carcinoma. Several lines of evidence point towards this tumor being a morphologic pattern of anaplastic thyroid carcinoma. In a recent multi-institutional study, pure squamous cell carcinoma with (Fig. 14) or without a differentiated thyroid carcinoma component displayed BRAF V600E mutations in 87% of cases and had an outcome similar to anaplastic thyroid carcinoma in general [150]. In addition, these squamous cell carcinomas express PAX8 and TTF1 in 91% and 38% of cases, respectively, confirming their follicular cell origin [150]. They harbor a differentiated thyroid carcinoma in three quarter of cases (almost all being papillary carcinoma). Furthermore, pure squamous cell carcinoma without any differentiated thyroid carcinoma component (i.e., fulfilling the 2017 WHO definition of squamous cell carcinoma) carries BRAF V600E mutations in 60% of cases and has the same prognosis as anaplastic carcinoma in general [151]. For the above reasons, squamous cell carcinoma of the thyroid is now classified as a morphologic pattern of anaplastic thyroid carcinoma. Another important addition to the anaplastic carcinoma section is the emphasis on rapid and prompt testing of all anaplastic carcinomas for the presence of BRAF V600E mutation. This testing is mandatory since the combination of BRAF and MEK inhibitors was found to be active against BRAF V600E–mutated anaplastic carcinoma [152]. This testing can be performed using immunostaining against the mutated protein or through genotyping.

Fig. 14
figure 14

Anaplastic thyroid carcinoma with prominent squamous differentiation. The tumor is arising in association with a papillary thyroid carcinoma, hobnail subtype (A). Anaplastic thyroid carcinoma showing spindle cell morphology (B)

Question 7: What Are the Clinicopathological Correlates of Oncocytic Thyroid Carcinomas?

Oncocytic follicular cell–derived thyroid carcinomas as a group can include many different entities: oncocytic PTC, oncocytic encapsulated follicular subtype of PTC, oncocytic poorly differentiated carcinoma, and oncocytic medullary thyroid carcinoma [58, 153]. However, the term “oncocytic carcinoma of the thyroid” is used in the new WHO to refer to invasive malignant follicular cell neoplasms composed of at least 75% oncocytic cells in which the nuclear features of PTC and high-grade features are absent. This term replaces Hürthle cell carcinoma, a misnomer given that Hürthle actually described parafollicular C cells. Oncocytic cells have abundant granular eosinophilic cytoplasm secondary to a marked accumulation of dysfunctional mitochondria. Oncocytic carcinoma of the thyroid (OCA) represents the malignant counterpart of oncocytic adenoma [153]. Although there are no major substantive changes to the description of these tumors in the 2022 WHO, it is helpful to review the distinct clinicopathologic correlates of OCA.

OCA, which accounts for approaching 5% of differentiated thyroid carcinomas in the USA [154], can occur anywhere in the thyroid and usually presents as a slowly enlarging painless solitary thyroid nodule. Thyroid ultrasound cannot distinguish between oncocytic adenoma and OCA, though larger tumors have a higher rate of malignancy [155]. There are no known risk factors for developing OCA. The mean age at diagnosis is approaching 60 years, which is roughly 10 years later than the mean age of diagnosis for patient with follicular thyroid carcinoma [154, 156, 157]. OCA, although more common in women (with a 1.6 to 1 female-to-male ratio), has a lower female-to-male ratio than is seen with follicular thyroid carcinoma [157]. Histologically, OCAs are encapsulated tumors with capsular and/or vascular invasion and at least 75% oncocytic cells (Fig. 15). OCAs are subclassified into minimally invasive (those with capsular invasion only), encapsulated angioinvasive, and widely invasive (those with gross invasion through the gland) tumors due to differences in clinical outcome. When evaluating OCA, it is important not only to document extent of invasion, but also to evaluate for progression to oncocytic poorly differentiated thyroid carcinoma; thus, all tumors should be assessed for increased mitotic activity (3 or more mitoses per 10 high-power fields/ ~ 2 mm2) and tumor necrosis. OCA can metastasize to lymph nodes [158]; however, some authors have shown that most of the so-called lymph nodes metastasis of OCA represent tumor plugs in veins in the neck and not lymph nodes involved by tumor [158]. The important clue is the almost perfect roundness of these tumor plugs compared to oval or somewhat irregular outline for nodal metastases [158]. OCA (like follicular thyroid carcinoma) usually spreads to distant sites via blood vessels. Distant metastasis at presentation are seen in 15–27% of patients with OCA and in up to 40% of tumors with extensive vascular invasion [159].

Fig. 15
figure 15

Oncocytic carcinoma of the thyroid. A Invasive growth through the capsule is evident at low power. B At high power, the cells have abundant granular cytoplasm and prominent nucleoli

Prognostic parameters for OCA include patient age, tumor size, vascular invasion, extrathyroidal extension, and the presence of distant metastases [154, 160]. Distant metastases at diagnosis are the most important prognostic factor for OCA [154, 160]. For OCA, the 5-year overall survival has been reported to be 85%, but only 24% among patients with distant metastases at diagnosis compared to 91% for patients with M0 disease at diagnosis [160]. Although it is not clear that OCA is more aggressive than follicular thyroid carcinoma after adjusting for variables such as patient age, gender, and tumor stage [156, 157], due to decreased efficacy of radioactive iodine with OCA compared to follicular thyroid carcinoma, treating OCA is currently more difficult once there is disease recurrence.

Benign and malignant oncocytic thyroid tumors have both been shown to harbor homoplasmic or highly heteroplasmic (> 70%) mitochondrial DNA mutations in complex I subunit genes of the electron transport chain [18, 161, 162]. Additionally, OCAs demonstrate widespread chromosome losses that result in near-genome-wide haploidization with or without subsequent genome endoreduplication [161]. Chromosomal changes have been found to be associated with extent of invasion: most OCAs with capsular invasion only or focal vascular invasion have been shown to be diploid, whereas tumors with extensive vascular invasion and widely invasive tumors are usually polysomic and nearly always demonstrate chromosome 7 amplification [161]. Additionally, the near-haploid state has been shown to be maintained in metastases, implying selection during tumor evolution [162]. OCAs have also been shown to have recurrent DNA mutations, including RAS mutations (though at a lower rate than is seen with follicular thyroid carcinoma), EIF1AX, TERT, TP53, NF1, and CDKN1A, among others [161,162,163].

Question 8: What Are the Molecular Biomarkers of Tumor Progression or Adverse Biology in Follicular Cell–Derived Thyroid Carcinomas?

While ancillary molecular testing is not required for the diagnostic workup of follicular cell–derived thyroid carcinoma, next-generation sequencing techniques are gaining ground as reliable analyses complementing gold-standard cytological and histological examinations. For example, molecular panels for pre- and post-operative analyses are routinely used by some academic centers, in which pre-operative testing of fine-needle aspiration biopsies (FNABs) may guide the patient to the right surgical procedure, while post-operative testing is focused on risk stratification and the identification of potentially actionable genetic events in case of therapy-resistant disease progression [164,165,166]. Indeed, follicular cell–derived thyroid carcinomas often harbor mutations or fusions in clinically targetable genes, thus motivating why clinicians should be acquainted and up-to-date with these mechanisms.

PTCs are in general driven by very few somatic mutations or mutually exclusive fusions involving genes regulating the mitogen-activated protein kinase (MAPK) signaling cascade (Fig. 16). PTCs exhibit among the lowest mutational burden of all human cancers and are highly stable from a pan-genomic perspective with few gross alterations observed, further emphasizing the overall placid genetic background of these lesions. The recurrent p. V600E missense mutation of the BRAF proto-oncogene is the most common genetic alteration in PTCs and is particularly enriched in specific, high-risk PTC subtypes (such as the classical, tall cell, and hobnail variants) [167, 168]. The mutation causes activation of the MAPK signaling pathway, which in turn will stimulate extracellular signal–regulated kinases (ERK) transcriptional programs, leading to increased proliferation, angiogenesis, and invasiveness, as well as down-regulating transcription of genes responsible for thyroid differentiation (Fig. 16). While the p. V600E BRAF mutation may be associated to a poorer patient outcome in some series, the results are conflicting — and the mutation is also prevalent in papillary thyroid microcarcinomas, which are known to have an excellent long-term prognosis.

Fig. 16
figure 16

MAPK and AKT pathways in thyroid cancer. Schematic overview of the mitogen-activated protein kinase (MAPK) and AKT pathways in the physiological (left) and neoplastic state (right). Upon activation of various report tyrosine kinases (RTKs; including RET and TRK1/3) via extracellular ligands, the RTKs dimerize and activate the MAPK pathway via activation of RAS proteins and also stimulate the PI3K-AKT cascade via PIK3. Both pathways stimulate proliferation, angiogenesis, and migration of cells. In thyroid cancer (right), various fusion genes or mutations simulate physiological activation of the RTKs, thereby leading to constitutively active MAPK and PI3K-AKT signaling, even in the absence of extracellular ligands. RET and NTRK1/3 fusions (yellow stars), activating PIK3CA, AKT, RAS, and BRAF mutations (green star) as well as deleterious PTEN mutations (red star), are highlighted. Created with

Additional factors besides BRAF mutations are thought to influence the aggressive clinical phenotype observed in subsets of PTC patients with this genetic alteration. Indeed, telomerase reverse transcriptase (TERT) promoter mutations have been found to predict worse clinical outcome in PTC patients with synchronous BRAF mutations, suggesting a synergic effect [169, 170]. Occurring in approximately 10–15% of unselected PTC cases, the recurrent TERT promoter mutations (C228T and C250T) are thought to enhance the TERT gene output, which in turn could lead to immortalization through the activation of telomerase (Fig. 17). Telomerase has the ability to maintain the length of telomeres by addition of telomeric repetitive sequences, avoiding critical shortening of the chromosomal ends after a specific number of cell divisions and thereby ensuring limitless replicative potential (Fig. 16). TERT promoter mutations are predominantly found in older patients (> 55 years) with large tumors that often exhibit extensive local infiltration [171, 172]. Although we lack data suggesting that the finding of a TERT promoter mutation in a PTC should alter the clinical management of the individual patient, the occurrence of a TERT promoter mutation should possibly alert the clinical team, as these mutations are highly associated with radioiodine refractory disease, distant metastases, and thyroid cancer dedifferentiation (Figs. 1718) [135, 172,173,174,175]. Moreover, apart from promoter mutations, gain of the TERT gene locus at chromosome 5p15.33, aberrant TERT promoter methylation patterns, and TERT mRNA overexpression are features coupled to adverse prognosis in well-differentiated follicular cell–derived thyroid carcinomas, of which some may be of clinical utility [176,177,178]. Indeed, the THY1 signature platform includes chromosome 5p status as one of the parameters tested [177]. Moreover, while PTC patients < 55 years rarely exhibit TERT promoter mutations, the presence of detectable TERT promoter mRNA is associated to worse outcomes and may thus be a promising marker for future studies [178].

Fig. 17
figure 17

TERT functions in thyroid cancer. Left: The telomerase reverse transcriptase (TERT) gene located on chromosome 5p15 encodes the catalytic subunit of the human telomerase complex, which is imperative for immortalization of stem cells and cancer cells alike. In the normal follicular cell, the protective, repetitive sequences at the end of chromosomes (telomeres) are shortened at the 3′ end with each replicative cycle. Upon reaching a critical limit, the cell is forced into cell cycle arrest. Right: Two hotspot mutations in the promoter sequence (denoted C228T and C250T) have been established as an event signifying poor prognosis in thyroid cancer. The mutations lead to an augmented recruitment of various transcription factors and thereby increased TERT gene expression. TERT/telomerase will extend the telomeric repeats at the chromosomal ends, preventing telomeric shortening, and thereby contributing to immortalization. Created with

Fig. 18
figure 18

Schematic overview of the multistep process of thyroid cancer development and progression. Via modern genetic analyses, thyroid cancer can be divided into BRAF V600E–like and RAS-like tumors based on mutational and transcriptomic profiles. BRAF V600E–like alterations typically include gene fusions in ALK, BRAF, RET, NTRK1/3, and MET, while RAS-like alterations include BRAF K601E, DICER1, EZH1, EIF1AX, and PTEN mutations and gene fusions in PPARG and THADA. While many PTC subtypes adhere to the BRAF-like cluster, IEFVPTC and FTC in general cluster together as both entities harbor RAS-like aberrations. Not shown here is the third cluster composed of non-BRAF-/non-RAS-like tumors. *Note that while PTEN mutations are considered an early event in thyroid carcinoma with RAS-like alterations, additional PTEN pathway alterations are recurrently noted in more advanced thyroid carcinomas (DHGTC, PDTC, and ATC). PTC, papillary thyroid carcinoma; IEFVPTC, invasive encapsulated follicular variant papillary carcinoma; FTC, follicular thyroid carcinoma; DHGTC, differentiated high-grade thyroid carcinoma; PDTC, poorly differentiated thyroid carcinoma; ATC, anaplastic thyroid carcinoma

In terms of genomic rearrangements, PTCs may harbor fusions involving a wide variety of cancer-related genes, of which RET, NTRK1-3, BRAF, and ALK are all clinically important to recognize due to the therapeutic value of these alterations, as they may be targeted by specific drugs. RET fusion partners in thyroid cancer encompass 27 different genes, of which 24 are reported in PTCs [179]. Of these, CCDC6 is the most common partner gene, followed by NCOA4 [179]. RET fusions are the most common genetic aberrancy in PTC developing after radiation exposure and are the main culprit events in pediatric PTCs overall [180]. Irrespectively of fusion partner, the RET kinase is activated due to the switch of the RET promoter with that of the fusion partner; activated RET receptors stimulate the MAPK and PI3K-AKT signaling pathways and thus promote tumorigenesis (Fig. 16). While not as common as RET fusions, NTRK fusions (NTRK1 and NTRK3, encoding the receptor tyrosine kinases Trk-A and Trk-C, respectively, fusing with 5′ sequences of various activating genes) are detected in 3–5% of PTCs commonly seen in pediatric and adolescent patients with BRAF wild-type PTCs. These tumors display a predominant follicular growth pattern, often displaying a non-infiltrative tumor border, clear cell change, and less pronounced nuclear atypia as compared to BRAF-mutated cases [181,182,183,184,185]. The fusion products lead to a constitutively active receptor complex, in turn activating the MAPK and PI3K-AKT pathways.

Apart from the abovementioned genetic alterations, PLEKHS1 gene aberrations have been established in PTCs, including promoter mutations, aberrant promoter methylation, and gene overexpression leading to AKT pathway activation and poorer patient outcomes [173, 186]. In a recent study, PLEKHS1 promoter mutations were present in 13% of PTCs presenting with distant metastases and associated to radioiodine (RAI) refractory disease [173], and the authors suggested that the presence of a mutation in either the TERT promoter, the PLEKHS1 promoter, or the TP53 gene (in association with either BRAF or RAS mutations) will predict worse clinical outcome in well-differentiated thyroid carcinoma [173] (Fig. 19). The authors also suggested that losses of 9q and 11q were associated with cancer-specific mortality. Another group investigated the genomic landscape of fatal differentiated and poorly differentiated thyroid carcinomas (PDTCs) and found that an overrepresentation of chromosome 1q gain, as well as mutations in the mRNA processors MED12 and RBM10, encodes proteins involved in the regulation of RNA polymerase II–driven transcription and mRNA splicing functions, respectively [187]. Mutations in these genes were foremost observed in PDTCs and subsets of PTCs. Additional epigenetic aberrations include global DNA hypomethylation, which is coupled to an increased risk of distant metastases in differentiated thyroid cancer [188, 189] (Fig. 19).

Fig. 19
figure 19

Adverse markers in follicular cell–derived thyroid carcinoma (FCDTC). In papillary thyroid carcinoma (PTC), the synergistic effect of BRAF V600E and TERT promoter mutations indicates an increased risk of distant metastases and worse overall survival. Moreover, gain of chromosome 1q as well as mutations in the PLEKHS1 promoter (PLEKSH1p) may be observed in PTCs with poorer outcomes. In follicular thyroid carcinoma (FTC), TERT promoter mutations are associated to distant metastases and poor patient outcomes. Moreover, TP53 gene mutations and high tumor mutational burden (TMB) have been demonstrated in subsets of cases and may predict worse prognosis. In both PTCs and FTCs, global DNA hypomethylation may signify poorer patient outcomes. Oncocytic thyroid carcinomas are usually associated to mitochondrial DNA mutations and various copy number alterations, of which intensified gross chromosomal alterations (including whole-chromosomal duplications of chromosomes 5 and 7 and near-haploid genotypes) are coupled to worse outcomes. Finally, in differentiated high-grade thyroid carcinoma (DHGTC) and poorly differentiated thyroid carcinoma (PDTC), TERT promoter mutations may signify an increased risk of distant metastases. Due to their overall dismal prognosis, anaplastic thyroid carcinoma (ATC) is excluded from this algorithm. Created with

On the microRNA (miRNA) level, several gene products have been associated to worse outcome in PTCs, including overexpression of miR-146a/b, miR-221, and miR-222 [190].

Follicular thyroid carcinoma (FTC) and encapsulated follicular variant papillary thyroid carcinoma (EFVPTC) harbor somatic mutations of RAS gene family members NRAS, HRAS, and KRAS that are recurrently found in follicular patterned tumors (Fig. 16). Of these, NRAS mutations are the most prevalent, most often represented by recurrent codon 61 mutations, and less often of codon 12/13 alterations. The mutations cause activation of the MAPK pathway, but not to the same extent as mutant BRAF — as ERK may dampen the effect of mutant RAS through a negative feedback loop inhibiting various RAF proteins which is not possible when BRAF itself is mutated [191]. Therefore, the MAPK signaling output is lower, which could be a plausible explanation to the improved prognosis in RAS mutated tumors compared to BRAF-mutated cases [191].

While RET and TRK fusions do not seem to play a role in the development of follicular patterned lesions, PAX8::PPARG rearrangements are frequently reported (10–40% of cases) [192]. This genetic aberrancy is also reported in subsets of follicular thyroid adenomas, thereby preventing it from having diagnostic properties on the pre-operative level. The juxtapositioning causes a constitutively active transcription factor (PAX8) to overexpress the PPARG gene, which encodes a nuclear receptor with the ability to interact with cancer-related pathways [192]. PAX8::PPARG fusions and RAS mutations are mutually exclusive, and the rearrangement is predominantly observed in younger patients with smaller primary tumors [193, 194]. Moreover, FTCs often display alterations in genes related to the AKT signaling pathway, such as deleterious PTEN mutations, activating PIK3CA mutations, and PIK3CA copy number gain [195, 196] (Fig. 14). As PTEN negatively regulates PI3K mediated activation of the oncogenic PI3K-AKT pathway, these alterations lead to augmented AKT signaling and tumorigenesis (Fig. 16).

Apart from MAPK and PI3K-AKT signaling pathway alterations, a subset of FTCs may harbor mutations in miRNA processor genes DICER1 and DGCR8 [82, 197,198,199]. The mutations are believed to perturb the maturation of miRNAs, causing dysregulation of various target genes. Interestingly, a genotype–phenotype correlation may be at play for follicular-patterned differentiated thyroid carcinoma, as these cases are overrepresented in terms of DICER1 mutations [82, 200,201,202]. As in PTCs, subset of FTCs (15–20%) and IEFVPTC harbors TERT promoter mutations, and FTCs with this alteration exhibit a specific transcriptome and miRNA landscape compared to wild type cases [198, 203]. TERT promoter mutations are particularly amassed in cases with distant metastases and poorer patient outcomes [37, 174, 204,205,206] (Fig. 17). Moreover, tumor mutational burden has also been implicated as a prognostic tool in FTC, and this factor was more confident in predicting patient outcome than conventional histological subtyping [199] (Fig. 19).

Oncocytic carcinomas of the thyroid are enriched for deleterious mitochondrial DNA (mtDNA) mutations, more specifically genes encoding for complex I electron transport chain proteins [18, 162, 207]. These mutations cause a reduction of ATP generation and lead to an increase in reactive oxygen species, possibly driving tumor formation [17, 162]. Another distinctive feature of oncocytic tumors is the occurrence of genome haploidization leading to uniparental disomy, which usually does not involve chromosomes 5 and 7 (155) (156) [208]. The variable involvement of different chromosomes in individual tumors by this process results in the gross chromosomal alterations frequently reported in oncocytic tumors. These include loss of chromosomes 2, 8, and 22 and gains of chromosomes 7, 12, and 17 [198, 209]. Interestingly, oncocytic thyroid carcinomas with widespread chromosomal aberrations in general may be associated with poorer patient outcomes [161, 210] (Fig. 18).

Differentiated high-grade follicular cell–derived thyroid carcinoma (DHGTC) and poorly differentiated thyroid carcinoma (PDTC) usually develop through a genetic multistep fashion from a pre-existing well-differentiated thyroid carcinoma and are therefore in large parts driven by established mutations in MAPK or PI3K-AKT signaling pathways associated with the preceding PTC or FTC [132, 211, 212] (Fig. 17). Therefore, DHGTCs and PDTCs exhibit BRAF- or RAS-like signatures depending on the early driver gene event [73, 146]. As the majority of DHGTCs develop from PTCs, the BRAF V600E mutation and PTC-related gene fusions are the most common “early-type” driver gene events in these lesions, while PDTCs are enriched for RAS mutations, indicating a relationship to FTCs or EFVPTC [132, 146] (Fig. 19). Of course, DHGTCs and PDTCs acquire numerous additional genetic changes along their path of progression, including TP53 and TERT promoter mutations — of which the latter are associated with higher risk of distant metastases [146, 213] (Fig. 18). In adolescent patients with PDTC, DICER1 mutations seem common and may reflect a coupling between high-grade features and an underlying disturbance of miRNA regulation [149].

Anaplastic follicular cell–derived thyroid carcinoma (ATC) has been clarified by next-generation sequencing and clonality analyses that have increased our understanding of ATCs tremendously. Nowadays, it is acknowledged that the majority of ATCs develop through dedifferentiation from a preceding well-differentiated DHGTC or PDTC, and the driver events are therefore inherited from the more differentiated tumor type [135]. Thus, ATC developing from PTCs or DHGTCs often carry TERT promoter mutations with synchronous BRAF mutations, while ATCs arising from FTCs or EFVPTCs usually carry combinations of RAS and TERT promoter mutations [214]. Moreover, TP53 gene mutations are commonly reported, as are deletions of CDKN2A/B (encoding cell cycle regulators p16 and p14, respectively) [214]. Subsets of ATC with mismatch repair (MMR) gene mutations display a hypermutator phenotype compared to non-MMR gene mutated cases, but the clinical relevance of this observation is not known [150, 215, 216].

The use of molecular immunohistochemistry offers tools to enhance diagnosis. Even though pathology laboratories usually do not offer comprehensive genetic analyses in routine clinical practice, it should be stressed that immunohistochemistry can act as a screening tool for specific genetic events in PTCs, not least the mutation-specific BRAF antibody (clone VE1) to screen for V600E alterations, the pan-RAS Q61R (clone SP174) antibody that detects the most common HRAS/NRAS/KRAS Q61R mutations as well as pan-TRK staining for NTRK1/3 fusions and the 5A4 and D5F3 antibodies optimized for the detection of ALK fusions (Fig. 20). Notably, the VE1, SP174, and 5A4 antibodies have shown excellent concordance with molecular screening results [217,218,219]. However, the performance of pan-TRK immunohistochemistry suffers from a lower sensitivity and variable specificity [220]. In a recent study, pan-TRK staining was only positive in 3/7 PTC cases with an ETV6::NTRK3 fusion, of which two cases only stained focally, while 2/2 NTRK1 rearranged PTCs were both diffusely positive [183]. Thus, the staining pattern may vary in terms of which NTRK gene is rearranged, and staining heterogeneity may also confuse the interpretation.

Fig. 20
figure 20

Molecular immunohistochemistry in follicular cell–derived thyroid carcinoma (FCDTC). Immunohistochemistry may aid in identification of molecular aberrancies with prognostic and/or therapeutic importance. A Mutation-specific BRAF antibody (clone VE1) detecting a BRAF V600E mutation in a tall cell variant PTC. B The pan-RAS Q61R (clone SP174) antibody, indicating an underlying HRAS/NRAS/KRAS Q61R mutation. C Positive pan-TRK staining indicating an NTRK1/3 fusion. D The ALK 5A4 antibody optimized for the detection of therapeutically relevant ALK fusions. E PTEN global loss in a follicular patterned PTC is illustrated. This specimen had several PTEN-immunodeficient follicular patterned nodules. Subsequently, the patient was found to harbor germline pathogenic PTEN variant, consistent with PTEN-hamartoma tumor syndrome

PTEN immunohistochemistry may be helpful in cases in which Cowden syndrome may be suspected, as global absence of PTEN immunoreactivity in all benign and malignant follicular-patterned neoplasms is indicative of this syndrome [221, 222]. Moreover, loss of 5-hydroxymethylcytosine (5-hMC) seems to correlate with the presence of a TERT promoter mutation in thyroid cancer and may therefore be a promising immunohistochemical triaging marker — especially as TERT protein immunohistochemistry itself seems unreliable in this context [223, 224].

Question 9: What Is New in the Field of Medullary Thyroid Carcinoma?

The most important update for medullary thyroid carcinoma in this WHO edition is the introduction of a grading scheme. Since the definitive histologic description of medullary thyroid carcinoma in 1959 [225], there has been no widely recognized histopathological grading system for this entity. In 2020, two groups independently developed grading schemes for medullary thyroid carcinoma based on proliferative activity (mitotic count and Ki67 proliferative index) and tumor necrosis [143, 144]. The study of Al-Zumailli et al. is based on a two-tiered system with high-grade tumors defined by the presence of tumor necrosis and/or ≥ 5 mitoses per 10 high-power fields, 400 × (equivalent to approximately 2 mm2 in most microscopes) [143]. Fuchs et al. designed a three-tiered grading scheme based on mitotic count, Ki67 proliferative rate, and tumor necrosis with different cutoffs [144]. Both systems were shown to be independent predictors of outcome [143, 144]. The prognostic significance of these grading systems was also validated in an independent cohort [226]. Subsequently in 2021, an international study of 327 medullary thyroid carcinoma patients developed a two-tiered grading system named “The international medullary thyroid carcinoma grading scheme.” In this system, high-grade tumors are defined as having at least one of the three following features: tumor necrosis, mitotic count ≥ 5 per 2 mm2, and/or a Ki67 proliferation index ≥ 5% [226] (Fig. 21). Twenty-five percent of medullary thyroid carcinoma patients had high-grade tumors using this scheme. This two-tiered grading system was independent from AJCC 8th edition stage grouping, age, sex, tumor size, margin status, post-operative calcitonin and CEA serum levels in predicting locoregional recurrence, distant metastasis-free, disease-specific, and overall survival [226]. Although this grading scheme is independent from germline RET mutation status in predicting outcome, its prognostic value vis-à-vis the mutations found in sporadic medullary carcinoma is still not known. In a study of 44 sporadic medullary carcinomas graded on the basis of the original 2020 publications (based on mitotic count, tumor necrosis and Ki67 proliferative rate) [143, 144], there was no correlation between grade and RET or RAS mutation status [226]. However, larger studies are needed to investigate whether there is a correlation between genotype and grade in sporadic medullary thyroid carcinomas. Because tumor necrosis can be focal, it is essential to generously sample these tumors and grading of biopsies is thus not recommended. In addition to reporting histologic grade and tumor necrosis, the precise mitotic count and Ki67 index should be recorded (both should be based on the area of tumor with highest proliferative activity). In fact, mitotic count and Ki67 proliferative rate are continuous variables; as each variable increases, outcome worsens [145]. Because of its robust independent value, this novel histologic grade may benefit high-grade patients leading to closer follow-up, low thresholds for cross-sectional imaging, and careful monitoring for distant metastasis [145]. Additionally, this histologic grade can be used as a data point in clinical trials of adjuvant therapy since it is likely that adjuvant therapy will have greatest benefit in high-grade medullary carcinomas [145].

Fig. 21
figure 21

Example of a high-grade medullary thyroid carcinoma. Tumor necrosis is present (A). The mitotic count was 16 per 10 high-power fields (~ 2 mm2), with readily apparent mitoses present as seen in this field (B), and the Ki67 proliferative index was 35% (C)

Question 10: The Categories of “Salivary Gland–Type Neoplasms” and “Thymic Tumors Within the Thyroid” Represent a New Approach — What Information Prompted These Changes?

The 5th edition of the WHO classification of endocrine tumors has introduced minor changes and a new approach in the chapters on “salivary gland–type neoplasms” and “thymic tumors within the thyroid.” Awareness of these very rare primary tumors is relevant for appropriate diagnostic and therapeutic decisions. The group of salivary gland tumors includes mucoepidermoid carcinoma with its subtype “mucinous carcinoma,” and the secretory carcinoma (corresponding to former mammary analogue secretory carcinoma). These are very rare entities with 48 and 12 reported cases, respectively [227, 228]. Sclerosing mucoepidermoid carcinoma with eosinophilia shares some features with the above tumors, but has currently been classified within the group of uncertain histogenesis tumors. Intra- or perithyroidal thymic tumors are grouped under the umbrella of thymic tumors and thymoma and spindle epithelial tumor with thymus-like elements and thymic carcinoma.

Mucoepidermoid Carcinoma.

Mucoepidermoid carcinoma (MEC) has been defined as a malignant neoplasm of salivary gland type, characterized by mucinous, intermediate, and squamoid tumor cells, having a solid or cystic growth pattern (Fig. 22). Generally, epidermoid cells are cytologically bland and predominate over mucocytes in the dense fibrotic stroma of the tumor, although the proportion of each component can vary significantly. In fact, the very rare mucinous carcinoma of the thyroid has been incorporated in this entity, at the one extreme dominated by glandular differentiation, signet ring features [13], and accumulation of extracellular mucin among neoplastic cells [13].

Fig. 22
figure 22

Mucoepidermoid carcinoma of thyroid, an infiltrative tumor (A) showing mucinous, intermediate, and squamoid tumor cells, having a solid or cystic growth pattern

The histogenesis of MEC is controversial; a link to ectopic salivary gland tissue or solid cell nests has been suggested. Nevertheless, up to half cases are associated with a conventional PTC or, more rarely, a follicular or oncocytic carcinoma and to lymphocytic thyroiditis. Thus, squamous metaplasia is the currently favored precursor event [229]. Immunohistochemistry may help confirm this diagnosis with cytokeratin and p63 expression, in the absence of neuroendocrine markers and calcitonin. Conversely, follicular lineage markers are expressed in a fraction of cases only, including thyroglobulin, TTF1, and PAX8. The separation of this rare tumor type from conventional thyroid tumors seems therefore justified, also based on its indolent behavior and favorable outcome after surgery [229,230,231], with extremely rare aggressive cases reported, usually in association with anaplastic carcinoma transformation [232,233,234,235,236]

Secretory Carcinoma.

Although only 12 cases are on record in the English literature [227,228,229], secretory carcinoma (SC) of the thyroid gland (also known as mammary analog of secretory carcinoma) has been incorporated in the new WHO classification of thyroid tumors as part of the salivary gland–type neoplasms. It is both morphologically and genetically similar to its mammary and salivary gland counterparts and does not share the histological and immunophenotypical features of differentiated follicular cell–derived carcinomas (Fig. 23). Rather, it contains specific molecular alterations involving the ETV6 gene. Analogous to SC of other sites, the diagnostic hallmarks include a solid, papillary, tubular, or microcystic growth of eosinophilic cells with vacuolated, “bubbly” cytoplasm; rare cases may show high-grade features [237]. Nuclear atypia is bland, and occasional nuclear clearing and contour irregularities can be detected. Immunohistochemically, SC is diffusely reactive for GATA3, mammaglobin, and S100 [83] in the absence of thyroglobulin, TTF1, and PAX8 expression [238]. Cases associated with PTCs have been reported [239].

Fig. 23
figure 23

Secretory carcinoma of thyroid showing a solid and microcystic growth of eosinophilic cells with vacuolated, “bubbly” cytoplasm. Nuclear atypia is bland and occasional nuclear clearing and contour irregularities are seen

ETV6 translocations are the hallmark of SC, with fusion products with NTRK3. As a consequence, these carcinomas may respond to targeted therapy with TRK inhibitors [83, 227]. In a salivary gland SC series, a novel fusion of ETV6 with RET was detected [240], but this alteration has not been reported in the thyroid, yet. Apart from these novel treatments, primary thyroid SC follows a more aggressive course than that of other locations, with locoregional recurrences and distant spread in up to 30% of cases [227,228,229, 241].

Intrathyroid Thymic Tumors.

Compared to the WHO 2017 classification [3], no major changes occurred in the new WHO scheme: three types of thymic tumors in the thyroid region are described, with benign and malignant tumors sharing a thymic epithelial differentiation included under the term “thymoma family” and “thymic carcinoma family.” They are postulated to develop from either ectopic thymus (but the term “ectopic” has been discouraged) or from branchial pouch remnants differentiating along the thymic line. Hassall bodies may in fact be seen at the periphery of these tumors [13].

Thymomas develop within or attached to the thyroid, generally in the left lower lobe portion, from embryological remnants. Histologically, they resemble their mediastinal counterparts, all subtypes being represented, with solid or lobular growth of cuboidal, spindled, or squamoid cells in a loose stroma infiltrated by immature, TdT-positive, T lymphocytes [242,243,244,245,246]. Rare cases are invasive, but most are well circumscribed or encapsulated tumors, occasionally entrapping thyroid follicles, and after surgery follow an uneventful course [242, 247].

Intrathyroidal spindle epithelial tumor with thymus-like elements (SETTLE) is a rare malignant tumor generally arising in the pediatric age or in young male adults with less than 50 cases reported in the literature [248,249,250]. Histologically, a lobulated growth of two epithelial cell types with spindle or cuboidal shape (the latter arranged in tubules, papillae or glands) is observed [251,252,253,254]. Tumor cells have low-grade atypia and a low proliferative activity, which helps to distinguish this tumor from sarcomas, medullary carcinoma, and sarcomatoid anaplastic carcinoma. No recurrent molecular alterations have so far been identified [255]. SETTLE has a relatively good prognosis with over 80% 5-year survival rate, although cases with locoregional and distant metastases (mostly to the lung) are on record [256, 257].

Intrathyroid thymic carcinoma (ITC) is a malignant tumor with thymic epithelial differentiation that affects the adult population with a slight female preponderance and a high occurrence in Asians. It develops in intra- or perithyroidal regions, generally in the lower poles as a firm, solid mass of variable size [13, 258]. The old terminologies CASTLE and lymphoepithelioma-like carcinoma are no longer recommended in the new WHO classification. Histologically, ITC shares the orthotopic thymic carcinoma features, with a lobulated growth of squamous (keratinizing or basaloid) epithelial cells in a desmoplastic stroma infiltrated by lymphocytes and plasma cells [253, 259,260,261] (Fig. 24). Tumor cells have poorly defined cell borders, mild nuclear atypia, distinct nucleoli, and a low proliferative activity. The ITC immunophenotype includes expression of cytokeratins, CD5, p63, CD117, CEA, p53, and bcl-2, in the absence of thyroid follicular markers (thyroglobulin, TTF1) and of EBV-related markers (EBER) [258]. Recurrent TERT promoter mutations have been reported in ITC, but not in mediastinal thymic carcinomas [262]. ITC is a relatively indolent neoplasm associated with a median disease-free survival of 144 months (and median overall survival not reached), as reported in a pooled analysis of 132 cases [263].

Fig. 24
figure 24

Intrathyroidal thymic carcinoma — an infiltrative tumor with solid growth pattern composing of polygonal medium size lesional cells in a lymphocytic background (A). The majority of tumor cells express CD5 with variable intensities (B)

Molecular Profile and Ancillary Tests:

Regarding salivary gland–type carcinomas, the demonstration of MAML2 rearrangements in MEC and of ETV6::NTRK3 fusions in secretory carcinoma may be helpful to confirm the diagnosis of these cases [238, 240, 264]. Nevertheless, no molecular profiling is essential to reach a correct diagnosis. At this moment, no molecular tests are required for the diagnostic classification of thymomas, SETTLE, and intrathyroid thymic carcinomas.

Question 11: Why Were Sclerosing Mucoepidermoid Carcinoma with Eosinophilia and Cribriform-Morular Thyroid Carcinoma Reclassified as Tumors of Uncertain Histogenesis? What Are the Clinicopathological Characteristics of These Tumors?

The tumor previously known as the “cribriform-morular variant of papillary thyroid carcinoma” can be associated with familial adenomatous polyposis (FAP) or may occur as a sporadic form. It was originally classified as a subtype of papillary carcinoma because of the presence of papillae and in some instances diagnostic nuclear features (Fig. 25). Several studies have shown that this tumor has a molecular profile distinct from follicular cell–derived thyroid carcinoma. In contrast to the latter that harbor mutations in the MAPK pathway (e.g., BRAF, RAS), these tumors do not display BRAF V600E mutations [265] and only rarely have RAS or PIK3CA mutations [266, 267]. Almost all cribriform-morular tumors have genetic alterations in the Wnt/beta-catenin pathway [265] with APC mutations being the most common and found in both the familial and sporadic setting [266, 267]. Mutations in other genes involved in the Wnt/Beta-catenin pathway such as CTNNB1 have also been detected [266, 267]. By immunohistochemistry, these tumors show diffuse cytoplasmic and nuclear beta-catenin expression. In addition, a recent study from two institutions questioned the follicular cell derivation of this neoplasm since it often lacks PAX8 and thyroglobulin expression while retaining TTF1 protein only in the cribriform elements [268]. The cribriform areas also express estrogen and progesterone receptors. The morulae are positive for CD5, CK5, CDX2, and CK5, but lack TTF1 expression [268] (Fig. 25).

Fig. 25
figure 25

Cribriform-morular thyroid carcinoma. Tumor displays complex cribriform architecture with focal morulae (A; arrows indicate morulae). The hallmark of this tumor is the diffuse nuclear and cytoplasmic beta-catenin expression (B). Unlike follicular cell–derived thyroid carcinomas with differentiated architecture, these tumors are often negative for PAX8 (C) and typically negative for thyroglobulin (D). The cribriform component is diffusely positive for TTF1, whereas the morulae are negative for TTF1 (E; morular structure highlighted) and positive for CDX2 (F; morular structure highlighted). These tumors tend to show estrogen receptor expression (G). The morulae are also positive for CD5 (H; morular structure highlighted). Scattered intratumoral lymphoid cells also express CD5 (H)

For the above reasons, the tumor nomenclature was changed to “cribriform-morular thyroid carcinoma” in the new WHO classification scheme and the tumor is now classified under the umbrella of thyroid tumors of uncertain histogenesis.

Sclerosing mucoepidermoid carcinoma with eosinophilia (SMECE) is a rare thyroid tumor with less than 60 reported cases, characterized by a morphology partially overlapping with that of MEC, in association with a marked infiltration of lymphocytes and eosinophils in the fibrotic stroma [13, 228] (Fig. 26). A background of thyroiditis is usually observed. Association with papillary carcinoma is reported in 20% of cases. The immunoprofile resembles that of MEC with no expression of follicular markers, except for TTF1 in half of the cases [228, 245, 269,270,271,272] A few genotyped cases of SMECE lack the typical genetic alteration of mucoepidermoid carcinoma of salivary gland (i.e., MAML2 translocation) [270, 273]. Associations with anaplastic thyroid carcinoma [274] and with NUT carcinoma [275] have been reported.

Fig. 26
figure 26

Sclerosing mucoepidermoid carcinoma with eosinophilia of thyroid tumor cells with squamous differentiation arranged in cords, tubules, and nests with marked infiltration of lymphocytes and eosinophils in the fibrotic stroma

The histogenesis of this low-grade malignant tumor is debated and the new WHO classification included this entity among tumors of uncertain histogenesis. An origin from ultimobranchial cells is favored, but the genomic profile has not identified gene alterations associated with either MEC or PTC [245, 270, 275], with rare exceptions [274, 276]. Mutations typical of follicular cell–derived thyroid carcinomas such as BRAF V600E have also not been found [245]. While some have hypothesized a follicular cell origin through the mechanism of squamous metaplasia of the follicular epithelium [277], this view is not supported by the lack of PAX8 and thyroglobulin positivity [228]. Because of the presence of p63 in the tumor, some authors have favored an origin from solid cell nests [278] while other investigators acknowledge that the origin of this rare carcinoma is unknown [228].

In view of the above, a definite classification of this rare primary thyroid tumor is not yet possible. Thus, in the 5th edition of the WHO blue book, SMECE is classified under thyroid tumors of uncertain histogenesis.