Introduction

Thyroid cancer is the ninth most common cancer worldwide, with an estimated incidence of 586,202 new cases in 2020 [1]. Medullary thyroid carcinoma (MTC) accounts for only 1–2% of all thyroid cancers and it is associated with 8% of thyroid cancer-related deaths in the USA [2, 3]. However, the incidence of MTC in Asians is less than 1% of all thyroid cancers [4, 5]. MTC is a neuroendocrine neoplasm, arising from the C-cells of the thyroid. Its histological features differ from those of follicular cell-derived differentiated thyroid carcinomas, and are reflective of a more aggressive clinical outcome [2].

Advances in molecular biology over the past decades have led us to understand better how MTC develops and progresses by genetic and epigenetic alterations. Although contemporary tumor classification and risk stratification increasingly rely on molecular biomarkers that reflect genetic and epigenetic changes, histopathology retains its inherent value in the 2022 World Health Organization (WHO) classification of thyroid neoplasms [6]. The 2022 WHO classification adopted the prognostic International Medullary Thyroid Carcinoma Grading System (IMTCGS) among recently introduced histological grading systems which has been shown to facilitate the risk stratification and clinical management of patients with MTC [7].

MTC is primarily treated surgically. However, in case of recurrence or distant metastases, molecular testing is crucial to select patients for RET-targeted tyrosine kinase inhibitors and other targeted therapies [8, 9].

This review provides an overview and recent updates on the molecular genetics of MTC, disease risk stratification based on molecular profile and histopathology, and targeted molecular therapies in patients with advanced MTC.

Developmental Origin of C-Cells and Their Neoplasms

MTC is derived from parafollicular C-cells of the thyroid gland, which secrete calcitonin, a calcium-lowering hormone. The thyroid gland parenchyma is composed of endoderm-derived follicular cells and endodermal cells that migrate and/or associate with ultimobranchial/pharyngeal bodies originated parafollicular C-cells. The latter component comprises less than 0.1% of the thyroid epithelial cells [10, 11]. During the embryonic development of vertebrates, the ultimobranchial body is technically derived from the fifth pharyngeal pouch. However, its origin is practically the ventral recess of the fourth pharyngeal pouch because the fifth is rudimentary and merges with the fourth [12]. The fourth pharyngeal pouch gives rise to the superior parathyroid glands, the apex of the pyriform fossa, and laryngeal cartilages and muscles, as well as the ultimobranchial body [12]. The ultimobranchial body migrates caudally, invades the develo** thyroid gland, fuses to the thyroid primordium, and eventually gives rise to the parafollicular C-cells in mammals [10]. However, the ultimobranchial anlage of nonmammalian vertebrates and monotremes does not merge into thyroid primordium, instead remains as a separated organ, the ultimobranchial gland [12]. This ultimobranchial body origin concept supports that parafollicular C-cells are located mainly in the middle of the thyroid lobes and are very rare in the peripheral and isthmic areas [11]. Parafollicular C-cells cluster around the solid cell nests, which are considered to represent ultimobranchial body remnants [11]. Recent data on embryonic development proved that C-cells are of no neural crest origin and are of endodermal cells which migrate and/or associate with ultimobranchial/pharyngeal bodies [13, 14]. A minority of parafollicular C-cells may also be derived directly from endodermal follicular primordium progenitor cells [11].

MTC can present as a mixed tumor composed of MTC and follicular cell-derived thyroid carcinoma, including papillary thyroid carcinoma, follicular thyroid carcinoma, oncocytic thyroid carcinoma, poorly differentiated thyroid carcinoma, or anaplastic thyroid carcinoma [6]. The endodermal origin of parafollicular C-cells may explain the rare occurrence of follicular and C-cell-derived tumors within the same lesion [11]. The stem cell properties of ultimobranchial cells to differentiate into both C-cells and follicular cells contribute to the development of the mixed thyroid carcinomas [11]. Many clinical case reports support the common origin theory from an ultimobranchial stem cell. However, unlike the MTC component, the non-MTC tumor component shows co-expression of thyroglobulin and monoclonal PAX8 (especially C-terminus specific antibodies) and lacks expression for calcitonin, calcitonin gene-related peptide, monoclonal CEA, chromogranin-A, synaptophysin, and INSM1 [15,16,17,18]. The mixed tumors can develop lymph node metastases having both the cell types. Mixed carcinomas composed of MTC and papillary thyroid carcinoma have been reported in patients with pathogenic germline RET variants [19].

MTC and follicular cell-derived thyroid carcinoma components may indeed have different clonality. For example, BRAF p.V600E can be found in the papillary carcinoma component of a mixed MTC and papillary thyroid carcinoma (Fig. 1). The MTC component also harbors a similar molecular profile as that of the pure MTC.

Fig. 1
figure 1

Mixed medullary and papillary thyroid carcinoma. A The papillary carcinoma component is intimately admixed with the medullary carcinoma and shows follicular growth with rudimentary papillae (arrow). B VE1 immunostaining highlights the classic papillary carcinoma component harboring the BRAF p.V600E variant

Amphicrine carcinoma expressing MTC and exocrine (mucin production) features in the same tumor cells has been reported in less than 10 cases in the thyroid gland [20]. However, evidence of mixed endocrine-exocrine differentiation in other thyroid neoplasms is largely lacking, with the exception of metaplastic trans-differentiation that occurs in salivary gland-type mucoepidermoid carcinoma of the thyroid [21].

Pathogenesis and Molecular Genetics

Up to 75% of all MTC cases develop sporadically. Sporadic MTCs are diagnosed at an average age of 45–55 years [2, 22]. In addition, 20–25% of cases of MTC are hereditary and associated with multiple endocrine neoplasia type 2 (MEN2) syndrome; occurring due to germline alterations in the RET proto-oncogene [22]. MTC occurs as a dominant presentation of the autosomal dominant tumor syndrome. It develops at an earlier age (10–20 years) in MEN2B (formerly known as MEN3) than in MEN2A (25–35 years), and at a variable age in isolated familial MTC [3, 23].

The RET proto-oncogene on chromosome 10q11.2 encodes a RET receptor tyrosine kinase that comprises an extracellular domain, a transmembrane domain, and an intracellular domain with tyrosine kinase activity (Fig. 2) [23]. Activating RET variants, usually point mutations or gene fusions, leads to constitutive activation of tyrosine kinase (Fig. 2). These genetic events are well-known oncogenic drivers in the development of thyroid cancer. RET point mutations are the most frequent driver events in MTC, whereas RET fusions are found in papillary thyroid carcinoma and poorly differentiated thyroid carcinoma [3, 23,24,25].

Fig. 2
figure 2

The RET receptor tyrosine kinase signaling pathway and oncogenic signal activation. The extracellular domain of the RET receptor contains four cadherin-like domains and a cysteine-rich region. In a normal RET activation pathway, each of glial cell line-derived neurotrophic factors (GDNFs) interacts with a co-receptor of GDNF family receptor-α (GFRα) family ligands. The complex of GDNF-family ligands and co-receptors GFRα1-4 recruits two RET receptors closely and leads to the dimerization of RET receptors and autophosphorylation of tyrosine residues. However, oncogenic RET alterations have different activation mechanisms. Pathogenic RET variants involving the cysteine residues (indicated by asterisks) of the extracellular domain constitutively activate RET through the formation of intermolecular cysteine disulfide bridges leading to the dimerization of altered RET. Germline RET alterations at the cysteine residues cause multiple endocrine neoplasia type 2A (MEN2A) and familial medullary thyroid carcinoma (FMTC). RET variants affecting the tyrosine kinase residues (indicated by asterisks) of intracellular domains lead to a conformation change in the ATP binding pocket of RET and constitutively activate the monomeric form of RET. Germline pathogenic RET variants involving the intracellular tyrosine kinase domains lead to multiple endocrine neoplasia type 2B (MEN2B) or FMTC. Additionally, most RET fusion proteins are cytosolic without transmembrane domain and have constitutively active tyrosine kinase domains of RET receptors. RET gene fusions occur in papillary thyroid carcinoma (PTC) and poorly differentiated thyroid carcinoma (PDTC) but in MTC. Activation of normal RET or oncogenic RET signaling pathway promotes activation of downstream signaling pathways, including MAPK, PI3K/AKT, JAK2/STAT3, or PLCγ pathway

Germline Pathogenic RET Variants

Germline RET alterations are found in almost all patients with MEN2 and have a well-defined molecular genotype and clinical phenotype correlations [26]. Of the patients with MEN2A syndrome, about 95% have pathogenic germline RET alterations in codons 609, 611, 618, and 620 at exon 10 or 630 and 634 at exon 11 within the extracellular cysteine-rich domains (Fig. 3) [23]. Germline RET p.C634A is the most common genetic alteration in MEN2A patients [23, 27]. These gain-of-function variants lead to constitutive RET kinase dimerization by forming disulfide-bounded homodimers, resulting in ligand-independent activation of the RET tyrosine kinase receptor [26].

Fig. 3
figure 3

Relatively common RET gene alterations and their associated RET proteins in hereditary and sporadic medullary thyroid carcinomas (MTCs). Common RET variants are shown in different colors according to multiple endocrine neoplasia (MEN) subtypes, familial medullary thyroid carcinoma (FMTC), and sporadic MTC. Pathogenic variants affecting the extracellular cysteine-rich domain codons account for most cases of MEN2A and FMTC. Mutations of codons at exons 13–16 activate the RET receptor tyrosine kinase. Two ATP-competitive small-molecule inhibitors of RET, selpercatinib and pralsetinib, selectively inhibit the tyrosine kinase activity of wild-type and altered RET. They were approved by the Food and Drug Administration for treating patients with RET-altered MTC and RET fusion-positive follicular cell-derived thyroid cancers, and non-small cell lung cancers

Isolated familial MTC, a subtype of MEN2A, accounts for about 15% of all hereditary MTCs [23]. Most germline RET variants associated with isolated familial MTC affect extracellular codons other than codon 634 in the cysteine-rich domain located on exons 10 and 11 or intracellular non-cysteine codons 768, 790, and 791 at exon 13, codon 804 at exon 14, or codon 891 at exon 15 (Fig. 3) [23].

The MEN2B syndrome accounts for about 5% of all hereditary MTCs and is mainly associated with the germline pathogenic variant, RET p.M918T involving exon 16 within the intracellular tyrosine kinase domain [23]. The genetic variant at codon 918 leads to autophosphorylation of the kinase domain and activates the RET tyrosine kinase receptor as a dimerization-independent monomeric form (Fig. 3) [23].

Germline RET variants have been found in almost all MEN2 families. However, according to the three European studies, 1.1% (6/532) of patients with hereditary MTC had no alteration in the RET gene [27]. Moreover, not all patients with germline RET variants develop the MEN2 phenotype. For example, 1.2% of Italian MEN2 families with germline RET mutation did not clinically develop MTC [27]. In addition, rare examples of inherited MTCs that lack pathogenic germline RET variants were reported to harbor pathogenic germline MET variants [28]; therefore, evidence suggests that whole exome sequencing may identify additional genetic alterations involved in the pathogenesis of hereditary MTC.

Risk Stratification for Germline Pathogenic RET Variant Carriers

Specific germline pathogenic variants involving RET gene confer different cancer risks and aggressiveness. Therefore, current American Thyroid Association (ATA) guidelines stratify patients with germline RET variants into the highest, high, and moderate risk levels of aggressive MTC and recommend the appropriate timing of prophylactic thyroidectomy based on the specific RET alteration [2]. The highest risk category includes patients with MEN2B and RET p.M918T (Fig. 4). The high-risk category includes patients with the RET variants involving codon 634 and RET p.A883F. Patients with hereditary MTC and other germline pathogenic RET variants belong to the moderate-risk category.

Fig. 4
figure 4

Risk stratification of hereditary and sporadic medullary thyroid carcinoma (MTC) according to the molecular profiles. Current American Thyroid Association guidelines classify patients with pathogenic germline RET variants into 3 groups based on the aggressiveness of the MTC: highest risk includes patients with RET p.M918T variant, high risk includes patients with RET p.C634F/G/R/S/W/Y variants or p.A883F alteration, moderate risk includes patients with other pathogenic RET variants. Germline RET p.M918T and p.A883F variants mainly occur in multiple endocrine neoplasia type 2B (MEN2B), whereas germline RET alterations at codon C634 occur in MEN2A. In sporadic MTCs, patients with somatic RET p.M918T variant have the worst prognosis, whereas those with somatic RAS mutations have the best prognosis

Somatic Pathogenic RET Variants

Somatic pathogenic RET variants are found in about half of all patients with sporadic MTC (Fig. 4) [23, 26, 29]. The variant prevalence is lower in small MTCs and higher in patients with distant metastases [23, 30] and cervical lymph node metastases [Medullary Microcarcinoma

Medullary microcarcinoma is a subtype of MTC measuring 10 mm or less in diameter [70]. There is no minimum size criterion. While some authors have shown tumors measuring 5 mm or less to have a better outcome compared to those with larger tumors [71], others documented smaller tumors to also have a variable risk of lymph node metastases and post-operative hypercalcitonemia [72, 73]. Kazaure et al. analyzed 310 patients with medullary microcarcinoma in the SEER registry from 1988 to 2007 [72]. Medullary microcarcinoma ≤ 5 mm had about 23% risk of lymph node metastasis, whereas the risk increased from 28% to 37% as tumor size increased from 6 mm to 10 mm [72]. In addition, Machens et al. found that with the increase in tumor size, more patients with medullary microcarcinoma had lymph node metastases from 13% in ≤ 2 mm tumors to 43% in 9– 10 mm [73]. A meta-analysis study showed that medullary microcarcinomas are associated with longer disease-free survival and a lower rate of extrathyroidal extension and lymph node metastases [74]. The prevalence of somatic RET p.M918T variant is also lower in microcarcinoma than in larger MTC [75]. However, there is no significant difference in the frequency of multifocality and distant metastases between microcarcinomas and larger tumors [74].

Hereditary Versus Sporadic MTC

Patients having hereditary MTC diagnosed by screening have overall survival rates similar to the general population [63]. The excellent prognosis is attributed to an early diagnosis when the tumor is confined to the thyroid gland [2, 63]. Compared to hereditary MTCs, sporadic MTCs present at an older age and have higher lymph node recurrence [62]. However, whether MTC is sporadic or hereditary is not an independent prognostic factor [62, 76]. DSS and clinical outcomes are similar between hereditary MTCs and sporadic MTCs after adjusting for the tumor stage [76].

Circulating Serum Tumor Markers

Calcitonin and carcinoembryonic antigen (CEA), two important hormones secreted by the C-cells, act as tumor markers for MTC. While calcitonin has diagnostic relevance, both are used in follow-up and have a role in predicting disease progression [2]. Niederle et al. demonstrated that basal serum calcitonin levels of > 85 pg/mL for females and > 100 pg/mL for males had 100% sensitivity for diagnosing lateral neck lymph node metastasis [77]. However, calcium-stimulated calcitonin levels did not improve preoperative diagnosis [77]. Basal serum calcitonin levels correlate with the extent of the disease progression. Machens et al. showed that thresholds of 20, 50, 200, and 500 pg/mL indicated lymph node metastasis at ipsilateral central and lateral neck, contralateral central neck, contralateral lateral neck, and upper mediastinum, respectively [78]. Through this study, the authors also deduced that bilateral compartment-oriented neck surgery could achieve biochemical cure in at least half of the patients with preoperative basal calcitonin levels of < 1000 pg/mL but in none of the patients with levels > 10,000 pg/mL [78]. Preoperative basal serum calcitonin level > 1000 pg/mL predicts failure to reach biochemical cure [79]. Preoperative serum CEA levels also correlate with the extent of lymph node involvement [80].

Doubling times of calcitonin and CEA levels, calculated by measuring them over multiple follow-up intervals, help in determining the tumor growth rate and are predictive of disease progression [81, 82]. The ATA recommends estimating doubling times of calcitonin and CEA in patients with detectable serum levels of calcitonin and CEA post-thyroidectomy [2]. The ATA advises against systemic therapy in patients with stable low-volume metastatic disease and serum calcitonin and CEA doubling times > 2 years. Systemic therapy should also not be administered to those without documented metastatic disease, even with increasing serum calcitonin and CEA levels [2].

Postoperative levels of calcitonin and CEA can predict tumor dedifferentiation. An increasing CEA level associated with a stable or declining calcitonin level is indicative of dedifferentiation and a worse prognosis [83]. MTCs with advanced disease and dedifferentiation often show normal or low calcitonin levels [84]. However, CEA levels may still be elevated, making CEA a more reliable biomarker in dedifferentiated MTCs.

Procalcitonin produced by C-cells and MTC cells is very stable, and its concentration is unaffected by blood sampling and storage [85]. A recent meta-analysis suggested that owing to its high accuracy and better analytical characteristics, procalcitonin should replace calcitonin in the diagnosis and follow-up of MTC patients [86]. A high procalcitonin to calcitonin ratio predicts an increased risk of progressive disease and a shortened progression-free survival [87]. The high procalcitonin level is associated with larger primary tumor size, more lymph nodes involved, distant metastases, and lower biochemical cure rates [88]. The biochemical cure rates declined to no more than 71%, 36%, 23%, and 10%, respectively, when the procalcitonin levels crossed thresholds of 1, 5, 10, and 50 ng/mL [88]. The serum levels also correlated with the lymph node compartments involved, which is essential for deciding the extent of required surgery [88].

Limited evidence suggests the utility of serum carbohydrate antigen 19–9 (CA19-9) as a marker of poor prognosis in MTC. Elisei et al. found elevated levels of CA19-9 in 16 of the 100 patients with advanced structural recurrent or persistent disease but in none of the 100 patients with biochemical cure [89]. Elevated serum CA19.9 was also a predictor of mortality [89]. In a more recent study, the same research group showed that serum CA19.9 positivity and doubling times of < 6 months and < 1 year were associated with mortality but not a progressive structural disease [90]. Importantly, the authors highlighted the cases of three patients having CA19.9 doubling times < 6 months. They died of disease after 6, 5, and 3 months, respectively, although there was no evidence of progressive disease [90].

Histologic Features

Many studies have investigated different pathological parameters for their potential utility in the prognostication of MTC. However, none of the histological features, such as the predominant histologic pattern (solid, trabecular, or acinar), the predominant cytological appearance (plasmacytoid, fusocellular, small cells, oncocytic, clear cells), amount of amyloid deposited, presence or absence of calcification, C-cell hyperplasia, and multifocality, was prognostically relevant [68, 91].

C-cell hyperplasia is found in almost all prophylactic thyroidectomies for MTC and is considered the precursor of hereditary MTCs [92, 93]. Morphological features associated with germline RET alterations are C-cell hyperplasia in association with at least a single focus of MTC or medullary microcarcinoma and/or multifocal C-cell neoplasia (Fig. 5) [93]. These C-cells are neoplastic and have mild-to-moderate nuclear atypia and abundant cytoplasm that are identified in H&E-stained slides and confirmed by calcitonin immunostaining (Fig. 5). The criteria of C-cell hyperplasia remain not fully defined and problematic since the number of C-cells normally found in humans varies from 4–6 cells around a follicle to 50 cells per low-power (100 ×) field [70]. While C-cell hyperplasia may occur as a reactive (secondary) phenomenon (e.g., underlying follicular nodular disease, hyperparathyroidism, aging, medications), bilateral and multifocal C-cell hyperplasia in association with medullary (micro)carcinoma or multifocal medullary (micro)carcinoma should prompt genetic testing to rule out germline pathogenic RET variants. The distinction of medullary microcarcinoma from a nodular C-cell hyperplasia requires the demonstration of tumor invasion. Infiltrating tumor cells tend to result in desmoplastic tissue reaction and are often associated with irregular outlines at the periphery of the lesion, whereas C-cell hyperplasia lacks these histologic features [70].

Fig. 5
figure 5

C-cell hyperplasia in a patient with multiple endocrine neoplasia type 2A. A The C-cells have round nuclei and clear or granular cytoplasm (arrows). B Calcitonin-positive C-cells show focal segmental proliferation within follicles

The prognostic relevance of amphicrine phenotype remains uninvestigated primarily due to its rarity. Following a review of the limited literature available as case reports, a recent publication has suggested that the amphicrine morphology may imply a poor prognosis in MTC [20].

Angioinvasion

Angioinvasion is a reliable predictor of disease recurrence and DSS. By using pre-defined strict criteria for angioinvasion (tumor cells invading through a vessel wall and forming tumor-fibrin complexes or intravascular tumor cells admixed with fibrin/thrombus) (Fig. 6) [94], Erovic et al. found that 60% of the MTCs with angioinvasion developed metastases and 100% of them developed locoregional and/or distant metastasis [95]. Other studies also revealed the detrimental effect of angioinvasion on DSS in MTC patients [91].

Fig. 6
figure 6

Medullary thyroid carcinoma with angioinvasion (A) and lymphatic invasion (B) in surrounding thyroid parenchyma. This tumor metastasized to the cervical lymph nodes and lung. A Angioinvasion is characterized by thick-wall vascular spaces and intravascular tumor cells admixed with fibrin and thrombus. B Lymphatic tumor emboli are shown within the clear lumens of thin-wall lymphatic vessels lacking a muscular layer

Lymph Node Metastasis

At the diagnosis of MTC, lymph node metastases are already present in 47–75% of patients [96]. In a recent study on 152 MTC patients from 2000 to 2020, lateral lymph node metastases and lymph node ratio (positive lymph nodes/retrieved lymph nodes) > 1/3 were independently associated with progressive disease [97]. The presence of contralateral nodal disease is also an indicator of poor outcomes [97]. A novel indicator is the log of the ratio between the number of positive nodes and the number of negative nodes [98]. The log odds of positive lymph nodes have been demonstrated to have better prognostic efficacy than lymph node status and the number of positive lymph nodes [98]. Multivariate analysis has confirmed the absence of regional metastases to predict long-term biochemical cure [63]. Kuo et al. found that older age, tumor size > 2 cm, regional spread (extrathyroidal extension or lymph node metastasis), and distant metastatic disease were independent risk factors for DSS [99]. Lymph node metastasis increased the risk of re-surgery. Central and lateral lymph node dissection at the time of initial surgery was associated with decreased recurrences [99].

Under the 8th edition of the AJCC staging system, MTCs with T1–3, central lymph node metastasis (N1a), and no distant metastasis (M0) are classified in stage III [69]. MTCs with lateral lymph node metastases (N1b) and without distant metastasis (M0) are classified in stage IV.

Proliferative Activity, Tumor Necrosis, and Grade

The National Cancer Institute defines “histologic grade” as “a description of a tumor based on how abnormal the cancer cells and tissue look under a microscope and how quickly the cancer cells are likely to grow and spread” (https://www.cancer.gov/publications/dictionaries/cancer-terms/def/histologic-grade). The grading systems differ for each type of cancer. The commonly used parameters for grading cancers include proliferative activity (e.g., mitotic count and Ki67 index) and coagulative necrosis.

The proliferative activity gives information regarding how rapidly the tumor cells are multiplying. This can be assessed by more than one method, the most common being calculation of mitoses. It has the advantage of easy applicability using routinely available hematoxylin and eosin (H&E)-stained sections. Traditionally, the number of mitotic figures is expressed per high power field (HPF). However, the method has its limitations; the major being differing fields of view across different microscopes, even at the same magnification. The problem has been further compounded by the increasing shift to digital pathology systems, where the fields are rectangular and the displayed area variable depending on the image viewer system. This would, beyond doubt, affect grading, especially in cases with borderline scores. To circumvent this, the fifth edition of the WHO classification of tumors recommends using standardized international (SI) units, first adopted in 2021 [100]. Mitoses now need to be counted per mm2 [101]; most grading schemes require documentation of mitotic figures per 2 mm2 based on a formal count from 10 mm2. Therefore, the area to be evaluated and the method used (“hotspot” or “average”; “conventional or phosphohistone-H3” assisted) also need mentioning [100].

The mitotic activity in a tumor can be expressed as mitotic count, mitotic index, or mitotic rate [100]. The mitotic count is the number of mitoses per mm2. It does not consider the size of neoplastic cells, the section thickness, and the tumor cell density; the latter being affected by the presence of intra-tumoral stroma like fibrosis and areas of necrosis.

In comparison, the mitotic index is calculated by dividing the number of tumor cells in mitosis by the number of tumor cells counted [100]. This overcomes the concerns of cell size, size of the field of view, and intra-tumoral stroma, but is labor-intensive for implementation in routine clinical practice. However, the mitotic index can be assessed easily using digital pathology.

The mitotic rate is the rate at which cells are entering the mitotic phase [100]. It is expressed as a percentage of the M phase entering cells counted per hour. It needs metaphase arrest or labeling of viable cells and is not feasible in routine practice.

As there is interobserver variability in the identification of mitoses, immunohistochemistry for proteins expressed during mitosis has been suggested as an alternative to visual counting [101]. Ki67 immunohistochemistry is a method of assessing the proliferative activity of a tumor. It labels the nuclei in G1, G2, or S phase [100,99,102]. Immunohistochemistry for phosphoHistone-H3 antibody stains the phosphorylated nuclear histone protein, formed during late G2 to late anaphase, and has been shown to have better accuracy than the visual counting of mitosis and the Ki67 labeling index [100, 101, 103, 104].

Some of the other methods available for assessing tumor proliferative activity include immunohistochemistry for proliferating cell nuclear antigen, topoisomerase IIa, cyclin A, geminin, and minichromosome maintenance; flow cytometric evaluation for the S phase fraction; and incorporation methods like thymidine labeling index and bromodeoxyuridine labeling index. All have their advantages but are of limited clinical use [105].

Histologic grade integrating necrosis and proliferation index has been well-established for long as an important prognostic factor in most neuroendocrine neoplasms, particularly in the gastrointestinal tract and pancreas [106]. However, this concept has only recently been adapted to MTC. Since 2020, three grading systems have been proposed for MTC with prognostic implications. An initial proposal by a group from the Memorial Sloan Kettering Cancer Center (MSKCC, New York, NY, USA) introduced a two-tier grading system [107]. The authors retrospectively reviewed 144 MTCs for multiple clinicopathologic parameters and identified tumor necrosis and increased mitotic activity were the only independent predictors for DSS in multivariate analysis. Therefore, they defined low-grade MTCs as cases with a mitotic count < 5 mitoses per 10 HPFs (2 mm2) and no tumor necrosis. Cases with increased proliferation of ≥ 5 mitoses per 2 mm2 and/or presence of tumor necrosis were classified as high-grade (Fig. 7). Based on this grading system, 113 (78%) MTCs were classified as low-grade whereas the remaining 31 (22%) were considered as high-grade. Univariate survival analysis demonstrated that the MSKCC grading system predicted DSS well. The second system was proposed by a group from the Royal North Shore Hospital (Sydney, NSW, Australia) [108]. The authors evaluated multiple clinicopathologic parameters in 76 patients with MTCs and found that advanced age, absence of MEN2 syndrome, high Ki67 proliferative index, high mitotic count, and coagulative necrosis were significantly associated with worse overall survival in univariate analyses. A three-tier grading system was developed using a combination of Ki67 proliferative index, mitotic count, and the presence or absence of coagulative tumor necrosis (Figs. 7 and 8). Low-grade MTCs were defined as cases with low proliferative activity (< 3 mitoses per 2 mm2 and a Ki67 labeling index < 3%) and no tumor necrosis; intermediate-grade MTCs as cases with low proliferative activity but with tumor necrosis, or with intermediate proliferative activity (3–20 mitoses per 2 mm2 and/or 3–20% Ki67) and no tumor necrosis; and high-grade MTCs as cases with intermediate proliferative activity but with tumor necrosis, or with high proliferative activity (> 20 mitoses per 2 mm2 and/or > 20% Ki67) regardless of tumor necrosis status. Based on the Sydney grading system, 62 (81.5%) cases were low-grade, 9 (12%) were intermediate-grade, and 5 (6.5%) were high-grade. The survival analysis demonstrated low-grade MTCs with a mean overall survival of 195 months, followed by the intermediate-grade MTCs with a mean overall survival of 137 months, and then the high-grade MTCs with a mean overall survival of only 45 months.

Fig. 7
figure 7

Comparison of histological grading methods in medullary thyroid carcinoma (MTC). Memorial Sloan Kettering Cancer Center (MSKCC) grading system is a two-tier system to differentiate high-grade MTC based on mitotic count and tumor necrosis [107]. Sydney grading system uses a three-tier grading system based on the mitotic count, tumor necrosis, and Ki67 index [108]. The international MTC grading system (IMTCGS) is a two-tier grading system based on the mitotic count, tumor necrosis, and Ki67 index [7]. The 2022 WHO classification adopted the IMTCGS as a prognostic variable. The area of 10 high-power fields (HPFs) is equivalent to 2 mm2 in most microscopes

Fig. 8
figure 8

High-grade medullary thyroid carcinoma showing tumor necrosis (A), mitotic count > 5 per 2 mm2 (B), and Ki67 proliferative index > 5% (C). Arrow indicates a mitotic figure

Although both grading systems have been validated to predict progression-free survival and DSS in an independent cohort of 44 sporadic MTCs from the Brigham and Women’s Hospital (Boston, MA, USA) [109], it was problematic to have two different grading systems for the clinical management. Therefore, a joint attempt has been made by both groups to create a third grading system (the international MTC grading system, IMTCGS) to unify the parameters that have been validated with consensus cutoffs [7]. In this multi-institutional study including a total of 327 MTC patients from five medical centers, a retrospective review of mitotic activity, Ki67 proliferative index, and necrosis using uniform criteria was conducted and a two-tier consensus grading system was developed. High-grade MTCs were defined as MTCs with at least one of the following features: mitotic index ≥ 5 per 2 mm2, Ki67 proliferative index ≥ 5%, or presence of tumor necrosis (Fig. 7). Based on this grading system, 246 (75%) MTCs were low-grade and 81 (25%) MTCs were high-grade. The IMTCGS was predictive of overall survival, DSS, distant metastasis-free survival, and locoregional recurrence-free survival.

Several subsequent studies have validated the applications of the IMTCGS. Vissio et al. compared these three different grading systems using 151 MTC cases from two hospitals in Italy [110]. The results showed that although all three grading systems were significantly associated with disease recurrence, only the IMTCGS demonstrated a significant impact on DSS.

In a recent reproducibility study, Williams et al. demonstrated that the IMTCGS had high interobserver agreement with an overall kappa value of 0.87 [111]. In addition, Nigam et al. found that IMTCGS high-grade patients demonstrated significantly faster calcitonin doubling times and majority of them were less than 2 years [112]. Although the IMTCGS is not a WHO-grading system, it has been endorsed by the 2022 WHO classification as a prognostic parameter [6, 113]. The pathologists are still required to document clearly the numeric value for mitotic count per 2 mm2 and the Ki67 labeling index irrespective of the tumor grade. Further prospective studies on this grading system are required for real-world application.

Progress in Systemic Therapy Options in MTCs

MTC is resistant to cytotoxic chemotherapy and does not respond to radioactive iodine and thyroid-stimulating hormone suppression therapy that is used for treating differentiated thyroid cancer. The evolving role of radiolabelled somatostatin analogs (e.g., 177Lu-DOTATATE) in the peptide receptor radionuclide therapy (PRRT) has been a focus of interest in somatostatin receptor (SSTR)-avid neuroendocrine neoplasms [114]. There is also emerging evidence that patients with metastatic SSTR-avid MTCs may benefit from PRRT [115]; however, access and indications vary in clinical practices.

Recently, a promising treatment strategy has been to inhibit RET kinase, which is highly expressed in MTC. Multi-targeted tyrosine kinase inhibitors were used for treating patients with advanced MTC before selective RET inhibitors became available (Table 1). However, multikinase inhibitors insufficiently inhibited RET and showed more off-target toxicities than selective RET inhibitors. Unlike the non-selective inhibitors, pathologists are required to confirm the RET mutation status for the use of selective RET inhibitors.

Table 1 FDA-approved tyrosine kinase inhibitors for the treatment of medullary thyroid carcinoma

Non-selective RET Inhibitors

Two multikinase inhibitors, vandetanib and cabozantinib, have been approved by United States Food and Drug Administration (FDA) and European Medicines Agency (EMA) for the treatment of patients with a progressive metastatic MTC [9]. Vandetanib inhibits the activity of multiple receptor tyrosine kinases including vascular endothelial growth factor receptor (VEGFR)1/2/3, RET, and epidermal growth factor receptor (EGFR). The ZETA clinical trial defined that the use of vandetanib showed, compared to the control group, increased progression-free survival (30.5 vs. 19.3 months) and objective imaging response (45% vs. 13%), irrespective of RET mutation status, progression rate, disease location, and tumor extent at baseline [116].

Cabozantinib inhibits multiple tyrosine kinases including RET, VEGFR1/2/3, TIE-2, MET, FLT-3, and KIT [117]. Its efficiency was documented in the EXAM clinical trial, which reported an improved progression-free survival (11.2 vs. 4.0 months) and an objective imaging response (28% vs. 0%) were observed compared to placebo, irrespective of age, tumor burden, tumor location, progression rate, or prior tyro kinase inhibitor treatment [117]. Both drugs have been associated with disease stabilization in 30% and partial regression in 35% of cases; severe toxicities may occur, although manageable in most cases. However, the estimation of the benefit of these agents on overall survival is difficult as most patients have a long-life expectancy.

Selective RET Inhibitors

Recently, two novel selective RET inhibitors, selpercatinib (LOXO-292) and pralsetinib (BLU-667), have been approved for the same indications in patients with RET-aberrant thyroid cancer (RET-altered MTC and differentiated thyroid cancer with RET fusion) and non-small cell lung cancer [9]. Unlike previously targeted small molecules (vandetanib and cabozantinib) used in differentiated thyroid cancer and MTC, both selpercatinib and pralsetinib are highly selective inhibitors for RET alterations, including point mutations and gene fusions [118, 119]. Furthermore, as these small molecules are not primarily based on the inhibition of VEGFRs, the representative primary molecular target for anti-cancer therapy by small molecules, we can expect excellent drug durability and continuation of medications without severe adverse events [120].

Selpercatinib was rapidly approved in May 2020 by FDA for RET mutated MTCs or RET fusion-positive differentiated thyroid cancers, based on the results of the phase 1–2 clinical trial (LIBRETTO-001) [121]. The study enrolled RET-altered MTC cases with or without previous vandetanib or cabozantinib treatment. The reported results were an objective response rate (complete or partial response) of 73% and 1-year progression-free survival of 92% from previously untreated MTC patients (n = 88). In previously treated MTC patients, objective response rate was 69%, and 1-year progression-free survival was 82%.

Pralsetinib secured FDA approval in patients with advanced or metastatic RET-altered MTC and RET fusion-positive follicular cell-derived thyroid cancer in December 2020, based on a phase 1–2 clinical trial (ARROW) [9, 122]. The clinical trial reported an objective response rate of 71% and an estimated 1-year progression-free survival of 81% in treatment-naïve MTC RET-altered patients (n = 21) [122]. In previously treated RET-altered MTC patients (n = 55) with vandetanib or cabozantinib, an objective response rate of 60% and an estimated 1-year progression-free survival of 75% were obtained from the study results. In previously treated RET fusion-positive follicular cell-derived thyroid cancers, objective response rate of pralsetinib was 89%.

Most of the patients experienced several adverse events. However, most of them were less intense than the two non-selective (multikinase) inhibitors. The rates of discontinuation of selpercatinib and pralsetinib due to adverse events were 2% and 4% in the clinical trials, respectively. Therefore, selpercatinib and pralsetinib showed more potent efficacy and good safety profiles than multikinase inhibitors.

Although the survival benefits of selpercatinib and pralsetinib for advanced and metastatic RET-altered MTC have not been proved yet, ongoing phase 3 trial may have a high probability of a better outcome than non-selective RET inhibitors, cabozantinib or vandetanib.

A Case Report of Selective RET Inhibitor for Refractory MTC

A 63-year-old male patient presented with a 4.5-cm-sized nodule involving the left thyroid with multiple bilateral enlarged cervical lymph nodes. Fine needle aspiration was suspicious for MTC. The serum calcitonin level was 1017.0 pg/mL (normal value, < 10 pg/mL). He underwent total thyroidectomy with bilateral lymph node dissection. The pathology confirmed high-grade MTC with extrathyroidal extension and extensive bilateral lymph node metastasis. Subsequent lung computed tomography scan showed small multiple lung metastases. Germline variant study using blood showed c.2692G > T (p.D898Y) in RET exon15. The patient and his family have not demonstrated clinical features characteristic of MEN-related diseases. As multiple lung metastases were progressive with a maximal diameter of 13 mm and the doubling time of serum calcitonin levels showed 5.75 months, the patient was started on vandetanib, an approved first-line multikinase inhibitor, with a total dose of 300 mg per day. During the 2-year vandetanib treatment, the patient showed stable disease of multiple lung metastases. After that, however, the treatment dose was lowered to 100 mg daily due to frequent adverse events, such as QT prolongation on the electrocardiogram, gum bleeding, and moderate to severe hypertension or headache. Finally, he stopped the treatment because of a progressive disease of lung metastasis.

Without further treatment for metastatic lesions, the patient developed multiple spine metastasis and rapid progression of metastatic diseases at lung, parahilar, mediastinal, and lateral neck lymph nodes. Due to intolerable neck pain, immediate neck area radiation therapy was performed to relieve the pain. As there was no approved second-line tyrosine kinase inhibitor in MTC, potential candidate tyrosine kinase inhibitors were being searched as an early access through a Named Patient Program (NPP) and then selpercatinib was administered to the patient with dose of 120 mg two times per day.

Although the patient experienced 3 cm sized intracranial hemorrhage on right frontal cerebrum 1 year before selpercatinib treatment, he did not show any severe adverse events associated with selpercatinib, except one-time hypercalcemia and mild dry mouth, and showed partial response in nearly all metastatic lesions following 10-month treatment of selpercatinib (Fig. 9).

Fig. 9
figure 9

A case of treatment with selpercatinib for refractory medullary thyroid carcinoma. Computerized tomography scans before and after 10 months of selpercatinib treatment (120 mg twice per day) show regression of left supraclavicular lymph node metastasis (A, B), subcarinal lymph node metastasis (C, D), and multiple lung metastases (E, F). Arrows indicate metastatic tumors

Conclusion

MTC is not the only neuroendocrine neoplasm in the thyroid gland. Other neuroendocrine neoplasms include intrathyroidal thymic neuroendocrine neoplasms, intrathyroidal parathyroid neoplasms, and thyroid paragangliomas as well as metastatic neuroendocrine neoplasms [18, 70]. The demonstration of diffuse CEA expression often helps distinguish various other calcitonin and/or calcitonin gene-related protein (CGRP)-producing neuroendocrine neoplasms from MTCs. Therefore, the first responsibility of a pathologist is to distinguish MTC from other mimics using appropriate immunohistochemical biomarkers. The second responsibility includes meticulous assessment of the status of angioinvasion, tumor necrosis, proliferative rate (mitotic count and Ki67 labeling index), and histologic grade (low- or high-grade) along with the tumor stage and the resection margins. Given the morphologic and proliferative heterogeneity in these neoplasms, an exhaustive sampling is strongly recommended. Routine molecular testing for pathogenic germline RET variants is typically performed in all patients with a diagnosis of MTC; however, C-cell hyperplasia in association with at least a single focus of MTC (or microcarcinoma) and/or multifocal C-cell neoplasia is morphological harbingers of germline RET alterations [93]. The status of somatic RET alterations should also be determined in all advanced/progressive or metastatic MTC, especially when selective RET inhibitor therapy is considered. While the role of routine SSTR2/5 immunohistochemistry requires further clarification, evidence also suggests that patients with SSTR-avid metastatic disease may benefit from 177Lu-DOTATATE PRRT. Finally, the authors of this review also make a call to support the nomenclature change of MTC to C-cell neuroendocrine neoplasm to align this entity with the IARC/WHO taxonomy since MTCs represent epithelial neuroendocrine neoplasms of endoderm-derived C-cells.