Novel Targeted Therapies for Metastatic Thyroid Cancer—A Comprehensive Review

The knowledge on thyroid cancer biology has grown over the past decade. Thus, diagnostic and therapeutic strategies to manage thyroid cancer are rapidly evolving. With new insights into tumor biology and cancer genetics, several novel therapies have been approved for the treatment of thyroid cancer. Tyrosine kinase inhibitors (TKIs), such as lenvatinib and sorafenib, have been successfully utilized for the treatment of radioactive iodine (RAI)-refractory metastatic differentiated thyroid cancer (DTC). In addition, pretreatment with mitogen-activated protein kinase (MAPK) inhibitors (trametinib and selumetinib) has been shown to restore RAI avidity in previously RAI-refractory DTCs. Local therapies, such as external beam radiation and radiofrequency/ethanol ablation, have also been employed for treatment of DTC. Vandetanib and cabozantinib are the two TKIs currently approved by the Food and Drug Administration (FDA) for the treatment of medullary thyroid cancer (MTC). Other novel therapies, such as peptide receptor radionuclide therapy and carcinoembryonic antigen (CEA) vaccine, have also been utilized in treating MTC. Ongoing trials on selective rearranged-during-transfection (RET) protooncogene inhibitors, such as LOXO-292 and BLU-667, have demonstrated promising results in the treatment of metastatic MTC resistant to non-selective TKIs. The FDA-approved BRAF/MEK inhibitor combination of dabrafenib and trametinib has revolutionized treatment of BRAFV600E mutation positive anaplastic thyroid cancer. Several other emerging classes of medications, such as gene fusion inhibitors and immune checkpoint inhibitors, are being actively investigated in several clinical trials. In this review, we describe the molecular landscape of thyroid cancer and novel targeted therapies and treatment combinations available for the treatment of metastatic thyroid cancer.


Introduction
Thyroid cancer is predicted to affect 52,890 new patients in the USA in 2020, with the incidence being three times higher in women as compared with men [1]. Until recently, thyroid cancer was the most rapidly increasing cancer in the USA, largely, but not solely, due to increasing use of sensitive diagnostic procedures. However, the increase of about 7% per year during the 2000s has slowed to 2% per year in men, and rates have stabilized in women from the period of 2012 to 2016, as a result of the more conservative diagnostic approach that has been recently implemented [1]. Differentiated thyroid cancer (DTC) that arises from follicular cells is further subclassified into papillary thyroid cancer (PTC), which is the most common histological type, follicular thyroid cancer (FTC), and Hürthle cell cancer (HTC). De-differentiated thyroid cancer is classified as poorly differentiated thyroid cancer (PDTC)
Mutations in mitochondrial genes, epigenetic modifiers, and components of DNA damage and repair pathway; chromosome 5 and 7 duplications; loss of heterozygosity; and in-frame gene fusions Nature of mutations remain to be explored HTC [10] Biological consequence of these mutations in thyroid cancer remain to be explored.  Members of the PI3K pathway that are mutated in thyroid cancer include PTEN, AKT, and PIK3CA. However, the frequency of their presence is low [41] (Figure 1 and Table 1). PTEN is a tumor suppressor gene, and mutations, deletions, and epigenetic changes leading to its loss of expression have been observed in DTC, PDTC, ATC, and HTC [7,10,11,42]. Somatic single nucleotide variants (SSNVs) of AKT1, AKT2, and AKT3 have been detected in thyroid cancer at a very low frequency [11,43]. An AKT1 (G49A) mutation has been seen in thyroid cancer metastasis, suggesting that this mutation arises late during progression [44]. Copy number gain of AKT1 has also been reported in ATC and FTC [21]. The genetic alterations reported in the PIK3CA gene in thyroid cancer include gene amplification and, less commonly, mutations [45]. As in AKT1, genetic alterations in PIK3CA appear late during malignant progression of thyroid cancer and are more common in ATC than DTC [44,45] (Figure 1 and Table 1).
In MTCs, the most common genetic alteration related to MAPK/PI3K signaling pathways is RAS mutation. Mutations in KRAS and HRAS genes have been identified in RET wild-type sporadic MTCs and are mutually exclusive events in MTCs [46].

Molecular Alterations in RTK Signaling Pathway
RTKs play a crucial role in the regulation of processes such as proliferation, differentiation, and metabolism, and genetic alterations in RTKs have been reported frequently in cancer [33]. In thyroid cancer, the RET proto-oncogene is one of the most common RTKs to be altered. The most prevalent genetic alterations of RET in PTC are RET-PTC fusions. The RET/PTC rearrangements are the predominant genetic alterations in sporadic childhood DTCs and in radiation-related DTCs [47]. The RET/PTC1 rearrangement was a common oncogenic alteration to be observed in childhood thyroid cancer cases after the Chernobyl nuclear power plant accident, which resulted in radiation exposures [47]. Moreover, RET fusions have been observed in almost 50% of atomic bomb survivors with PTCs who had high radiation exposure [48].
Various RET fusions have been reported with genes such as CCCD6 (Coiled Coil Domain  [7,11]. The genetic alterations in RET can lead to constitutive RET activation and stimulation of the MAPK and PI3K signaling pathways [7]. Neurotrophic receptor tyrosine kinase (NTRK) rearrangements have been reported in PTCs.
Mutations and gene rearrangements have also been reported in anaplastic lymphoma kinase (ALK) gene in thyroid cancer. The ALK gene fusions are mainly observed in ATC and PDTC, but are also present in PTC, at low frequencies [7,51]. Two ALK mutations (C3592T and G3602A) have been identified in ATC. These mutations lead to amino acid changes (L1198F and G1201E) within the ALK tyrosine kinase domain, resulting in increased tyrosine kinase activities [31]. The ALK-associated genetic alterations activate the downstream MAPK and PI3K signaling pathways [7,31] (Figure 1 and Table 1).
In MTCs, the most common RTK alteration is the RET gain of function point mutations in the RET oncogene. MTC can be sporadic (70-80% cases) or hereditary (20-30% cases), and about 40-70% of the sporadic and 95% of the hereditary cases harbor mutations in the RET oncogene [28,29]. The RET mutations associated with MTC occur at the cysteine-rich or tyrosine kinase domains located within the seven exons (8, 10, 11, 13, 14, 15, and 16) [52]. These mutations can lead to one of the three clinically distinctive subtypes of hereditary MTC-multiple endocrine neoplasia (MEN) type 2A (MEN2A), and MEN2B, and familial MTC [53]. The most frequent RET point mutation to be detected in 85-90% of MEN2A patients and a small subset of sporadic MTC patients occurs at codon 634 [28,52]. On the contrary, the most commonly observed RET mutation in 95% of the MEN2B and 75-95% of the sporadic cases is M918T [54]. To this date, several RET mutations have been identified in sporadic and hereditary MTC. It has been shown that different RET mutations can produce distinct phenotypes that differ in terms of the aggressiveness of MTC, age of onset of MTC, and the presence/absence of other endocrine malignancies [52]. Based on the genotype-phenotype associations, the RET mutations have been stratified into three risk groups (highest risk, high risk, and moderate risk) that reflects the aggressiveness of MTC [55]. The highest risk RET mutation is M918T mutation, and the patients with this mutation develop the aggressive form of MTC and metastasis at a very young age [55]. The RET mutations that fall in the high-risk group include mutations in codons 634 and 883, and the patients with these mutations are at risk of developing an aggressive form of MTC at an early age [55]. The moderate risk RET mutations occur at codons 533, 609, 611, 618, 620, 790, 804, and 891, and the patients with these mutations typically show a less aggressive form of MTC in comparison with the high-risk and highest-risk groups and later onset [55]. Moreover, RET and ALK gene fusions have also been observed in MTC at low frequencies [51,56].
Receptor tyrosine kinases (RTKs) are upstream of the MAPK and PI3K pathways. The mutations in the RTKs and RTK-gene fusions can alter the downstream RAS/MAPK and PI3K/AKT pathways, leading to oncogenic transformation/progression. Similarly, the mutations in the components of the MAPK and PI3K pathways can promote tumorigenesis. Genes that have shown to be mutated in thyroid cancer are shown in red with an asterisk.

Other Molecular Alterations
PAX8-PPARγ (Paired Box 8-Peroxisome Proliferator Activated Receptor Gamma) rearrangement is present in 30-35% of FTC and 2-13% of follicular adenomas (FA) [36,57]. Besides this, PAX8-PPARγ gene fusion also appears with variable frequency in the FVPTC [36,57]. This gene fusion can act as a dominant-negative activator of PPAR-γ and can act as a transcriptional activator of a subset of PPARγ-inducible genes leading to its oncogenic effects [36] (Figure 2).

Figure 2.
Genetic alterations affecting nuclear processes in thyroid cancer. The figure shows mutations/gene fusions affecting nuclear processes such as transcription, DNA repair, and epigenetic regulation of genes to promote oncogenesis. The genes that have been observed to be mutated in thyroid cancer are shown in red with an asterisk.
A comprehensive analysis of the genomic landscape of HTC was recently performed which revealed unique genetic alterations in addition to the common genetic alterations that are seen in other types of thyroid cancer [10]. The common genetic alterations in HTC include mutations in mitochondrial genes, epigenetic modifiers, and components of DNA damage and repair pathway; chromosome 5 and 7 duplications; loss of heterozygosity; and in-frame gene fusions [10].
The molecular landscape of thyroid cancer forms a basis for the implementation of targeted therapies.

Mechanisms of RAI-Refractoriness in DTC
RAI-refractoriness is the loss of ability of the DTC cells to take up and concentrate radioactive 131 I, which is necessary to eliminate foci of persistent/recurrent disease [58]. Approximately 5-15% of DTC patients become refractory to RAI by losing the expression of sodium iodide symporter (NIS), which occurs as a result of the loss of thyroid differentiation features [58,59]. There are multiple mechanisms responsible for the loss of NIS expression on the cell membrane of the thyroid cancer cells. In the thyroid, TSH regulates NIS expression through stimulation of TSHR, which activates adenylyl cyclase and results in the accumulation of cyclic AMP (cAMP) within thyroid cells ( Figure   Figure 2. Genetic alterations affecting nuclear processes in thyroid cancer. The figure shows mutations/gene fusions affecting nuclear processes such as transcription, DNA repair, and epigenetic regulation of genes to promote oncogenesis. The genes that have been observed to be mutated in thyroid cancer are shown in red with an asterisk. TP53 mutations leading to the inactivation of this tumor suppressor gene are common in ATC (50-80%) and PDTC (10-35%), but have been detected with variable frequency in FTC and PTC as well [7,35] (Figure 2, Table 1). These mutations usually occur between exons 5-8 and promote tumor development as well as progression [35]. Telomerase reverse transcriptase (TERT) promoter mutations are also detected in ATC (40-70%), PDTC (40%), an aggressive form of HTC (32%), PTC (10%) and FTC (20%) [7,10,37], ( Figure 2, Table 1). The other common mutations in ATC and PDTC include mutations of the components of the Wnt pathway [34]. This includes the gain of function mutations of CTNNB1 (β-catenin gene) and loss of function mutations of the Axin1 gene ( Figure 2 and Table 1). The mutations in β-catenin promote its nuclear localization and may increase transcription [34]. The other genetic alterations to be predominantly observed in less differentiated thyroid cancer include mutations in the members of the DNA Mismatch Repair pathway (MSH2-MutS Homolog 2, MSH6-MutS Homolog 6, and MLH1-MutL Homolog 1); members of the SWI-SNF chromatin remodeling complex; eukaryotic translation initiation factor 1A (EIF1AX); histone methyl-transferases (HMTs); and isocitrate dehydrogenase 1 (IDH1) [7] (Figure 2 and Table 1).
A comprehensive analysis of the genomic landscape of HTC was recently performed which revealed unique genetic alterations in addition to the common genetic alterations that are seen in other types of thyroid cancer [10]. The common genetic alterations in HTC include mutations in mitochondrial genes, epigenetic modifiers, and components of DNA damage and repair pathway; chromosome 5 and 7 duplications; loss of heterozygosity; and in-frame gene fusions [10].
The molecular landscape of thyroid cancer forms a basis for the implementation of targeted therapies.

Mechanisms of RAI-Refractoriness in DTC
RAI-refractoriness is the loss of ability of the DTC cells to take up and concentrate radioactive 131 I, which is necessary to eliminate foci of persistent/recurrent disease [58]. Approximately 5-15% of DTC patients become refractory to RAI by losing the expression of sodium iodide symporter (NIS), which occurs as a result of the loss of thyroid differentiation features [58,59]. There are multiple mechanisms responsible for the loss of NIS expression on the cell membrane of the thyroid cancer cells. In the thyroid, TSH regulates NIS expression through stimulation of TSHR, which activates adenylyl cyclase and results in the accumulation of cyclic AMP (cAMP) within thyroid cells ( Figure 3) [60]. This increase in cAMP induces NIS transcription by stimulating thyroid-specific transcriptional factors (TTFs), like paired box 8 (PAX8) [59,60]. Studies have suggested that aberrant activation of MAPK/ERK and PI3K/AKT pathways are responsible for the TTF repression and loss of NIS expression ( Figure 3) [59]. The presence of BRAF V600E mutation has been strongly correlated to the loss of NIS expression and RAI refractoriness in PTC patients [61]. BRAF activation has been shown to repress PAX8 binding to the NIS promoter by activating TGFβ (transforming growth factor β)/Smad3 signaling [62,63]. In addition, activated BRAF can also promote NIS silencing by driving histone deacetylation of the NIS promoter [63,64]. Targeted treatment with BRAF and MAPK inhibitors have been shown to re-sensitize thyroid cancer cells/tumors to RAI therapy by improving NIS expression ( Figure 3) [65][66][67].
Cancers 2020, 12, x FOR PEER REVIEW 9 of 37 3) [60]. This increase in cAMP induces NIS transcription by stimulating thyroid-specific transcriptional factors (TTFs), like paired box 8 (PAX8) [59,60]. Studies have suggested that aberrant activation of MAPK/ERK and PI3K/AKT pathways are responsible for the TTF repression and loss of NIS expression (Figure 3) [59]. The presence of BRAFV600E mutation has been strongly correlated to the loss of NIS expression and RAI refractoriness in PTC patients [61]. BRAF activation has been shown to repress PAX8 binding to the NIS promoter by activating TGFβ (transforming growth factor β)/Smad3 signaling [62,63]. In addition, activated BRAF can also promote NIS silencing by driving histone deacetylation of the NIS promoter [63,64]. Targeted treatment with BRAF and MAPK inhibitors have been shown to re-sensitize thyroid cancer cells/tumors to RAI therapy by improving NIS expression ( Figure 3) [65][66][67]. Activation of the PI3K-AKT pathway has also been correlated to loss of NIS expression. Various growth factors, such as insulin, insulin-like growth factor 1 (IGF-1), hepatocyte growth factor (HGF), or epidermal growth factor (EGF), leading to PI3K activation and AKT phosphorylation, downregulated NIS expression in thyrocytes [68]. Conversely, inhibition of the PI3K pathway promoted NIS expression in rat thyroid cells and human PTC cells [69]. Activation of mTOR-a downstream target of PI3K/AKT pathway-also resulted in decreased NIS expression in thyrocytes, thus suggesting the involvement of the PI3K-AKT pathway in controlling NIS expression in thyroid cells ( Figure 3) [68]. Similarly, RET/PTC rearrangement has also been reported to reduce NIS expression in thyroid cells [69]. In addition to this, the presence of TERT promoter mutations has been correlated to radioiodine refractoriness in distant metastatic DTC patients [70]. In thyroid tumor cells with hyperactive MAPK and PI3K/AKT signaling, the NIS transcription is repressed, resulting in loss of its cell surface expression and RAI resistance. (C) Targeted treatment with inhibitors that block the MAPK/PI3K signaling pathways improves NIS expression and RAI avidity. * Trastuzumab has been tested in breast cancer model to target HER2 signaling but not in thyroid cancer model. Activation of the PI3K-AKT pathway has also been correlated to loss of NIS expression. Various growth factors, such as insulin, insulin-like growth factor 1 (IGF-1), hepatocyte growth factor (HGF), or epidermal growth factor (EGF), leading to PI3K activation and AKT phosphorylation, downregulated NIS expression in thyrocytes [68]. Conversely, inhibition of the PI3K pathway promoted NIS expression in rat thyroid cells and human PTC cells [69]. Activation of mTOR-a downstream target of PI3K/AKT pathway-also resulted in decreased NIS expression in thyrocytes, thus suggesting the involvement of the PI3K-AKT pathway in controlling NIS expression in thyroid cells ( Figure 3) [68]. Similarly, RET/PTC rearrangement has also been reported to reduce NIS expression in thyroid cells [69]. In addition to this, the presence of TERT promoter mutations has been correlated to radioiodine refractoriness in distant metastatic DTC patients [70].

Progression from DTC to Advanced Thyroid Carcinomas
The integrative analysis of the genomic landscape of well-differentiated thyroid tumors revealed the presence of a low frequency of somatic alterations, with the majority of those tumors harboring mutually exclusive mutations in BRAF and RAS genes, as well as gene fusions that primarily involve RTKs [38]. The progression of DTC to advanced thyroid carcinomas involves the acquisition of mutational hits in either oncogenes or tumor suppressor genes. The molecular profiling of advanced thyroid carcinomas showed that the presence of mutations is a much more common event in comparison to well-differentiated thyroid tumors [71]. The predominant mutations in the advanced thyroid tumors were in the components of the MAPK and PI3K/AKT pathway [71]. BRAF V600E mutation was prevalent in advanced PTC, PDTC, and ATC, whereas NRAS mutations were common in advanced FTC and PDTC. The predominant tumor suppressor mutation occurred in TP53 and was prevalent in HTC, ATC, and PDTC [71]. The presence of co-mutations is also a frequent event in advanced thyroid cancers. Concomitant mutations in RET/PTC, RAS, and BRAF have been observed frequently in advanced PTC and were associated with poor prognosis [72]. In another study, the integrative analysis of aggressive thyroid carcinomas, including ATC and advanced DTC, revealed the frequent presence of TERT, AKT1, PIK3CA, and EIF1AX mutations in combination with BRAF V600E and RAS mutations in these advanced forms of thyroid cancers [73]. The most common oncogenic duet that has been investigated for its association with aggressive thyroid cancer phenotype is BRAF V600E and mutated TERT promoter. The presence of TERT promoter mutations in synergy with BRAF mutations has been associated with aggressive thyroid tumor characteristics, which include lymph node metastasis, distant metastasis, tumor recurrence, multifocality, extrathyroidal extension, and patient mortality [74][75][76][77]. The two parts of this oncogenic duet (BRAF V600E and mutated TERT promoter) have been shown to work cooperatively together to promote oncogenesis, leading to aggressive thyroid cancer phenotype [78]. The underlying mechanism involves increased expression of an E26 transformation-specific (ETS) transcription factor, GA-binding protein subunit beta (GABPβ) through BRAF V600E driven MAPK pathway leading to the formation of GABPα-GABPβ complex, which binds to the mutant TERT promoter and increases TERT expression [78]. Besides BRAF V600E , TERT promoter mutations have also been observed to coexist with RAS mutations in aggressive thyroid tumors; however, the clinical significance of this oncogenic duet is still not fully understood [37]. It is likely that RAS mutations work cooperatively together with TERT promoter mutations through the activated PI3K pathway, to promote aggressive tumor characteristics [37], but this needs to be investigated further.

DTC
In November 2013, sorafenib was the first tyrosine kinase inhibitor (TKI) to be approved by the Food and Drug Administration (FDA), for the treatment of progressive metastatic DTC refractory to RAI treatment. Sorafenib targets VEGFR 1-3, PDGFR, RET, FLT, and c-kit. The approval was based on the results of the DECISION trial, which was a phase III, multicenter, double-blind, placebo-controlled trial conducted in 417 patients with progressive DTC which failed standard treatment. Sorafenib was associated with a significantly longer median progression-free survival (PFS) (10.8 months) when compared to the placebo (5.8 months). The most common side effects encountered in patients treated with sorafenib were hand-foot skin reaction (76.3%), diarrhea (68.6%), alopecia (67.1%), and rash or desquamation (50.2%) [79,80].
In February 2015, a second TKI lenvatinib, which targets VEGFR2, VEGFR3, EGFR, PDGFR, KIT, and RET, was approved for the treatment of progressive DTC refractory to RAI, based on the SELECT trial-A phase III, randomized, double-blind, multicenter study involving 261 patients with progressive DTC. Lenvatinib was associated with a longer median PFS of 18.3 months vs. 3.6 months in the placebo group. Of the 20 deaths that occurred on lenvatinib, six were attributed to the treatment itself, as TKIs lead to QT interval prolongation and fatal tachyarrhythmias [81]. Adverse reactions led to dose reductions in 68% of patients receiving lenvatinib and discontinuation in 18% of patients [82]. Given lack of permanent complete remissions after therapy with the FDA-approved agents, the use of alternative TKIs, as well as combination therapies involving TKIs and mTOR inhibitors or immune checkpoint inhibitors, is being evaluated in several ongoing clinical trials [83][84][85][86][87][88][89][90][91][92][93][94][95][96][97][98][99][100][101][102] (Table 2).      Since BRAF V600E is one of the most common mutations in PTC, BRAF inhibitors such as vemurafenib or dabrafenib have been implemented in the management of thyroid cancer. Interestingly, a basket trial of vemurafenib in BRAF V600E mutation-positive tumors showed significant responses in patients with ATC [105]. However, thyroid cancers exhibit primary resistance to RAF kinase inhibition due to reduced negative feedback, and hence a combination of RAF and MEK kinase inhibitors is necessary to effect MAPK pathway inhibition [115] (Figure 1). An open-label phase II study of 16 patients with BRAF V600E -mutant ATC treated with dabrafenib in combination with the MEK inhibitor trametinib showed a remarkable overall response rate of 69% in patients with ATC, with one patient achieving complete response [106]. The 12-month duration of response, PFS, and overall survival were estimated at 90%, 79%, and 80%, respectively, a phenomenon not previously seen in ATC, which is usually characterized by overall survival of 3-6 months post-diagnosis [116]. This observation has revolutionized the management of this very aggressive cancer and led to the FDA approval of combination therapy with dabrafenib and trametinib in BRAF V600E -mutant ATC in May 2018 [106].

MTC
Two additional TKIs-vandetanib, which targets VEGFR2, VEGFR3, EGFR, KIT, and RET, and cabozantinib, which targets VEGFR2, RET, MET, FLT3, and AXL (AXL Receptor Tyrosine Kinase)-have been FDA-approved for the treatment of metastatic progressive MTC. The treatment with the first one was associated with a longer median PFS of 22.6 compared with 16.4 months in individuals exposed to placebo, while the latter led to a PFS of 11.2 months in the treated group vs. four months in the placebo arm in multicenter, randomized, double-blind phase III clinical trials [103,104,117,118]. Again, the side effect profile was similar to other TKIs and included QT prolongation and arrhythmias, severe hypertension leading to reversible posterior leukoencephalopathy syndrome, headaches, hand-foot syndrome, fistula formation, profound fatigue, decreased appetite, nausea, diarrhea, and abdominal pain [104]. In addition, interstitial lung disease and Stevens Johnson syndrome have been associated with use of vandetanib [104].
The large number of severe side effects of TKIs result from its ability to target multiple kinases (Figure 4) which play a role not only in cancer progression but also in many important physiological processes. Therefore, more targeted therapeutic strategies have been recently developed and include inhibition of RET-only in RET-mutated MTCs and PTCs and inhibition of BRAF-only in BRAF V600E -mutated tumors [106,115]. Since BRAFV600E is one of the most common mutations in PTC, BRAF inhibitors such as vemurafenib or dabrafenib have been implemented in the management of thyroid cancer. Interestingly, a basket trial of vemurafenib in BRAFV600E mutation-positive tumors showed significant responses in patients with ATC [105]. However, thyroid cancers exhibit primary resistance to RAF kinase inhibition due to reduced negative feedback, and hence a combination of RAF and MEK kinase inhibitors is necessary to effect MAPK pathway inhibition [115] (Figure 1). An openlabel phase II study of 16 patients with BRAFV600E-mutant ATC treated with dabrafenib in combination with the MEK inhibitor trametinib showed a remarkable overall response rate of 69% in patients with ATC, with one patient achieving complete response [106]. The 12-month duration of response, PFS, and overall survival were estimated at 90%, 79%, and 80%, respectively, a phenomenon not previously seen in ATC, which is usually characterized by overall survival of 3-6 months post-diagnosis [116]. This observation has revolutionized the management of this very aggressive cancer and led to the FDA approval of combination therapy with dabrafenib and trametinib in BRAFV600E-mutant ATC in May 2018 [106].

MTC
Two additional TKIs-vandetanib, which targets VEGFR2, VEGFR3, EGFR, KIT, and RET, and cabozantinib, which targets VEGFR2, RET, MET, FLT3, and AXL (AXL Receptor Tyrosine Kinase)have been FDA-approved for the treatment of metastatic progressive MTC. The treatment with the first one was associated with a longer median PFS of 22.6 compared with 16.4 months in individuals exposed to placebo, while the latter led to a PFS of 11.2 months in the treated group vs. four months in the placebo arm in multicenter, randomized, double-blind phase III clinical trials [103,104,117,118]. Again, the side effect profile was similar to other TKIs and included QT prolongation and arrhythmias, severe hypertension leading to reversible posterior leukoencephalopathy syndrome, headaches, hand-foot syndrome, fistula formation, profound fatigue, decreased appetite, nausea, diarrhea, and abdominal pain [104]. In addition, interstitial lung disease and Stevens Johnson syndrome have been associated with use of vandetanib [104].
The large number of severe side effects of TKIs result from its ability to target multiple kinases (Figure 4) which play a role not only in cancer progression but also in many important physiological processes. Therefore, more targeted therapeutic strategies have been recently developed and include inhibition of RET-only in RET-mutated MTCs and PTCs and inhibition of BRAF-only in BRAFV600Emutated tumors [106,115].  There are two RET inhibitors currently implemented in the management of RET mutation-positive tumors in the clinical trial setting-BLU-667 and selpercatinib (LOXO-292) ( Table 2) [119][120][121][122]. BLU-667 inhibits the protein product of RETM918T, as well as RETV804L/M gatekeeper mutations conferring resistance to TKIs, while LOXO-292 is a highly selective RET kinase inhibitor with nanomolar potency against the canonical RET MTC drivers, RET gatekeeper mutations, and RET fusions [115]. Preliminary data from the ongoing clinical trials documented minimal side effects, excellent tolerability, and high efficacy, with objective response rates (complete and partial responses) ranging from 47% to 62% (Table 2) [115,[119][120][121][122].

Therapies Targeting Gene Fusions
An additional development in the management of thyroid cancer which has been proven highly effective in preliminary clinical trials is the use of therapy aimed at targeting NTRK and ALK gene fusions ( Table 2 and Figure 4). Larotrectinib is a selective inhibitor of tropomyosin receptor kinase (TRKA), TRKB, and TRKC which has been approved by the FDA for treatment of solid tumors with NTRK fusion [107,123]. The objective response rate in phase I and II clinical trials was remarkably high (80%) with 63% of patients with NTRK-positive tumors experiencing partial response and 13% complete response [107,123]. Most of the adverse events reported in the primary analysis were grade 1 and grade 2, with the most common being elevated ALT and AST (42%), fatigue (36%), vomiting (33%), dizziness (31%), nausea (31%), diarrhea (29%), and anemia (29%). (Table 2 and Figure 4).
Unfortunately, at the moment, there are no clinical trials targeting THADA or other described above fusions in thyroid cancer, such as PAX8/PPARγ, apart from a one case report describing application of pioglitazone in PPARγ fusion protein positive metastatic HTC and resulting in the reduction of tumor size and improved pain control [124].

Restoration of RAI Uptake via MEK and BRAF Inhibition
Another important strategy implemented in the management of de-differentiated DTC and PDTC relies on upregulation of sodium iodide symporter NIS via inhibition of its negative regulators MEK and RAF signaling. This strategy re-enables incorporation of iodine within the cancer cell and thus consists of the pretreatment with MEK and BRAF inhibitors, followed by RAI therapy [125,126].
The landmark study by Ho et al. has proven the principle that the inhibition of MEK1 and MEK2 by selumetinib induces RAI uptake in RAI-non-avid DTC (Figure 3) [109]. Furthermore, an individualized approach consisting of evaluation of the amount of restored RAI uptake, as measured by 124 I positron emission tomography/computed tomography (PET/CT)-based tumor dosimetry, has identified patients meeting the tumor accumulation threshold warranting RAI therapy. Among eight patients subsequently treated with RAI, five had a partial response and three had stable disease [109]. Interestingly, these preliminary data suggest a particularly high efficacy of this therapeutic approach in NRAS-mutated tumors. Given these promising results, there are several ongoing clinical trials utilizing selumetinib as an adjunct to RAI therapy in larger patients populations (Table 2) [127]. The second-generation MEK inhibitor trametinib is being evaluated in an ongoing phase II trial implementing pretreatment with trametinib, followed by 124I PET/CT-based lesional dosimetry and RAI therapy in patients with sufficiently restored RAI uptake (clinicaltrials.gov identifier NCT02152995) ( Table 2).
Additional clinical trials utilize individualized approaches to patients with RAS-mutated and BRAF-mutated tumors ( Table 2). Patients with tumors harboring an NRAS mutations are treated with the MEK inhibitor trametinib, while those with tumors characterized by BRAF V600E mutation are treated with combination therapies, including BRAF inhibitors such as dabrafenib and vemurafenib and MEK inhibitors ( Figure 3). Preliminary data suggest high efficacy of such an approach with partial response to RAI therapy at three months' follow-up observed in all patients who achieved sufficient restoration of RAI uptake [110]. A similar concept is being utilized in another ongoing multicentric prospective non-randomized phase II trial with two independent arms studying the use of trametinib for NRAS mutation-harboring tumors and dabrafenib for BRAF V600E -harboring tumors (clinicaltrials.gov identifier NCT03244956) ( Table 2).

Peptide Receptor Radionuclide Therapy in Thyroid Cancer
Another therapeutic concept utilizing radiolabeled agents for treatment of metastatic thyroid cancer had been based on a subset of thyroid cancer expresses somatostatin receptors (SSTR) that could be targeted with peptide receptor radionuclide therapy (PRRT) [128,129].
Radiolabeled somatostatin receptor (SSTR) analogs, such as 68 Ga-DOTATATE, which recognize predominantly SSTR type 2, are utilized for positron emission tomography/computed tomography (PET/CT) imaging. Of note, 68 Ga-DOTATATE was FDA approved in June 2016 for clinical use for imaging of patients with neuroendocrine tumors, including MTC. Identification of patients with RAI-non-avid DTC and metastatic MTC characterized by positive 68 Ga-DOTATATE uptake enables identification of individuals who may benefit from PRRT. There were small pilot studies performed in Europe that documented the utility of treatment of progressive RAI-non-avid metastatic DTC and MTC with SSTR agonists radiolabeled with 177 Lutetium or 86 Ytrium. The overall response rate was similar to current standard-of-care therapy with TKIs, while the quality of life was better and complication rate was lower after the therapy with PRRT (Table 2) [111][112][113][130][131][132].

DTC
The tumor microenvironment has been acknowledged as the major player in cancer progression and response to therapy. Therefore, the immune landscape of the thyroid cancer has emerged as a potential therapeutic target. Immune checkpoints, such as programmed cell death protein 1 (PD1) and its ligand-PDL-1, as well as cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) inhibitors, exhibit antitumor effects by altering the interaction between the immune system cells and tumor cells [133]. The expression of PD1/PDL1 in thyroid cancer has been extensively studied for both diagnostic and prognostic purposes [134]. In one meta-analysis by Aghajani et al. the expression of PDL1 in thyroid cancer was associated with tumor recurrence and poor survival [135]. Based upon analysis of the TCGA database, a higher level of PD-L1 mRNA was associated with lymph node metastasis, extrathyroidal invasion, and shorter disease-specific survival [136]. The latter was further supported by two independent cohort studies [89,137]. Nevertheless, DTC is thought to be poorly immunogenic due to a relatively low mutation burden. Consistently, published studies have shown that thyroid cancer has relatively poor response to immunotherapy with checkpoint inhibitors [114,138] ( Table 2). In order to enhance the efficacy of immunotherapy, combination treatments are being tested in several clinical trials in solid tumors, including thyroid cancer ( Table 2). The efficacy of these therapies might be better particularly in de-differentiated tumors, such as widely invasive HTC and ATC, as they are characterized by high mutation burden and higher likelihood of introducing immunogenicity [10,139].

MTC
Another interesting concept utilizing the ability of MTC cells to produce carcinoembryonic antigen (CEA) consisted of the introduction of dendritic cell vaccination with anti-CEA vaccine [140][141][142]. Yeast-CEA (GI-6207) is a therapeutic cancer vaccine genetically modified to express recombinant carcinoembryonic antigen (CEA) protein. In a phase 1 trial involving 25 patients with CEA expressing cancers, the treatment with Yeast-CEA showed stabilization of disease and its biochemical biomarkers. It was also well tolerated, with the most common adverse effect being grade 1/2 injection-site reaction [142]. A phase II trial was designed at the National Cancer Institute to evaluate the use of GI-6207 in recurrent medullary thyroid cancer and documented biochemical response without a significant tumor response [143].

Limitations of Targeted Therapies
The major limitation in applying targeted therapies is its side-effects profile, as well as transient efficacy due to the development of escape mechanisms by the tumors. The first limitation could be overcome by dose de-escalation and adjunct therapy with agents targeting adverse symptoms. The latter requires a thorough investigation of the mechanisms of tumor resistance and analysis of acquired second hit mutations, in order to tailor therapy against changing over time the molecular landscape of the tumor.

Tyrosine Kinase Inhibitors
Systemic treatment with TKIs often leads to side effects affecting different organs and systems. Fatigue is a common side effects that is reported with different agents ( Table 2). It can be severe enough (grade 3 and grade 4) to significantly affect the quality of life, leading to interruption or discontinuation of treatment [88,103]. Another common side effect is diarrhea, which can be mild, moderate, or severe. It is thought that it results from the effect of the agent on VEGF and EGF receptors in the bowel [144]. When diarrhea is of grade 1 or grade 2, the treatment is conservative, with antimotility agents such as loperamide and/or diphenoxylate/atropine, as well as cholestyramine and tincture of opium [144][145][146].
The most common cardiovascular adverse event associated with TKIs is elevated blood pressure, which is mediated by the anti-angiogenic effect on VEGF receptors, leading to reduction in endothelial nitric oxide, causing increased vascular tone [147,148]. Given the associated morbidity and mortality from hypertension in these patients, blood pressure should be frequently monitored with a target of <140/90 [148]. If blood pressure is above the target, a treatment with any of the first line anti-hypertensive agent should be initiated [148]. If hypertension is of grade 3 (SBP > 160 and/or DBP > 100) despite optimal treatment with anti-hypertensive agents, then treatment with TKI should be transiently discontinued till blood pressure is at target [148]. Grade 4 hypertension is an indication to stop treatment and consider other treatment options [148]. Other encountered cardiovascular adverse event of TKIs, particularly vandetanib, include QT prolongation, which can cause torsades de pointes ventricular tachyarrhythmia and subsequent cardiac arrest [104]. Therefore, ECG should be routinely done, to monitor QT interval, and QT interval prolongation to >500 milliseconds is an indication to stop treatment [149].
Another organ that can be affected by TKIs is the skin, as muco-cutaneous adverse reactions such as dryness, rash, hyperkeratosis, hand-foot syndrome, alopecia and mucositis occur frequently [150]. Hand-foot syndrome can manifest as hyperkeratosis, skin ulceration, or vesicles and present with pain, tingling, or paresthesia in affected skin areas which, if severe enough, may lead to interruption discontinuation of treatment [151,152]. The management of mucocutaneous side effects relies on preventive measures such as good skin hydration, urea creams, symptomatic treatment such as targeting neuropathic pain with gabapentin, dose reduction, or discontinuation of treatment with TKI [145,150,151].
Some adverse effects are more specific to certain TKIs; for example, pazopanib causes skin and hair hypopigmentation, while sorafenib causes skin hyperpigmentation [79,90]. Treatment with BRAF inhibitors dabrafenib and vemurafenib can be associated with the abnormal skin growth, forming papilloma, keratoacanthoma, nevi, or even skin cancers such as squamous cell carcinoma and melanoma [99,153].
Other encountered adverse effects of antiangiogenic targeted therapies are impaired wound healing, hemorrhage (within the tumor), fistula formation, or viscus perforation (in GI or respiratory tract) [144,154]. Therefore, it is important not to start these agents directly after the surgery for thyroid cancer.

Therapies Targeting Gene Fusion-Tropomyosin Receptor Kinase Inhibitor
The most common side effect of larotrectinib is abnormalities in liver function test [107]. However, transaminitis was rarely severe (grade > 3 or more) and required dose de-escalation only in four patients out 55 involved in the clinical trial targeting NTRK-positive tumors [107]. Other side effects that were less frequent and led to dose de-escalation were neutrophilia and dizziness [107]. Of note, none of the patients who received larotrectinib discontinued treatment due to adverse events [107].
The most frequent adverse events seen with entrectinib were dysgeusia, which was mild and reversible [108]. Other side effects, such as central nervous system related events, weight gain, congestive heart failure anemia, elevated creatinine level, and fatigue, were more severe, requiring dose reduction, treatment interruption, or even discontinuation [108].

Peptide Receptor Radionuclide Treatment
The treatment with (90)Yttrium and (177)Lutetium labeled somatostatin analogs in one study was associated with minor hematological abnormalities, such as leukopenia, erthryocytopenia, and/or decreased hemoglobin level [111]. Most of the hematological abnormalities seen with PRRT are self-limiting, but some patients may develop chronic complications such as myelodysplastic syndrome or leukemia [155]. Elevated liver enzymes, including ALT, AST, GGT, and ALP abnormalities, were also observed [111]. PRRT-related renal toxicity can lead to chronic renal disease/failure especially in patients with comorbidities that affect kidney function, such as hypertension and diabetes [156,157]. This complication is more likely to be observed with 90Ytrium-labeled somatostatin analogs rather than 177 Lutetium radio-labeled. In addition, amino acids infusion before the therapy has reno-protective properties. Unfortunately, amino acids infusion is often associated with nausea and vomiting, which are usually transient [112].

mTOR Inhibitors
A very common side effect of mTOR inhibitor everolimus is stomatitis, which presents as mouth pain, difficulty swallowing, or loss of taste [158]. When severe enough, stomatitis may require stopping oral intake and initiation of parenteral nutrition [158]. Patients receiving stomatitis should be counseled regarding this side effect and encouraged to keep good oral hygiene, to prevent stomatitis [159]. Treatment options include use of mouthwash, topical analgesics, and/or topical corticosteroids [158]. When symptoms of stomatitis are severe and affect the patient, oral-intake treatment should be interrupted till symptoms resolve [158]. In life-threatening situations, everolimus should be permanently discontinued [158].
Another common side effect of everolimus is skin rash, which usually presents as macules and/or papules. This is often associated with pruritis [160]. The mainstay for prevention of skin rash is the use of moisturizers and avoidance of sun exposure [161]. For acne-like rash topical treatment is the first option, but there may be a need to oral antibiotics in severe cases [161].
Other side effects seen with mTOR inhibitors include metabolic derangements such as hyperglycemia and hyperlipidemia, which mandate close monitoring of blood glucose and lipid profile routinely while on everolimus and utilization of anti-hyperglycemic agents and statins for the biochemical control [162].

Immunotherapy
Immune checkpoint inhibitors are associated with immune related adverse effects involving different organs and systems in the human body. Skin is commonly affected by immune checkpoint inhibitors, and the presentation can vary from mild dry skin to life-threatening Steven Johnson Syndrome (SJS) or Toxic Epidermal Necrolysis (TEN) [163,164]. Management of skin manifestation depends on severity; for mild-skin-lesion treatment with antipruritic agents or topical steroids is indicated. Severe manifestations such as SJS require inpatient or even intensive critical care unit admission and treatment with systemic steroids. Grade 4 reactions mandate permanent discontinuation of treatment [164,165]. Another common side effect of immune checkpoint inhibitors is colitis, which often presents as diarrhea. Once other causes of diarrhea are ruled out and the diagnosis is confirmed, treatment can be initiated with oral steroids (budesonide) and antimotility agents [166]. In more severe/refractory cases, treatment interruption or discontinuation and the use systemic steroids or infliximab is indicated [167,168]. Immune checkpoint inhibitors can have different effects on the endocrine system. The effect of immune checkpoint inhibitors on the thyroid gland can lead to hypothyroidism and, less frequently, hyperthyroidism, which is usually not an important consideration in thyroidectomized patients with thyroid cancer [169]. Other less common endocrinological adverse manifestations of immune checkpoint inhibitors are hypophysitis, adrenalitis, and development of type 1 diabetes mellitus [169]. Inflammation of other organs has also been described and includes myocarditis, pneumonitis, hepatitis, nephritis, pancreatitis, and others [170][171][172][173].

Mechanisms of Tumor Escape from Targeted Therapies
The targeted therapies have shown promising results in the treatment of thyroid cancer, but their biggest limitation is that the tumor cells develop resistance against those therapies over a period of time. The tumor cells acquire resistance to the treatment by developing an escape mechanism against the targeting drugs [174]. Some tumor types have intrinsic resistance to certain drugs; for example, MTCs with V804M-RET and V804L-RET gatekeeper mutations do not respond to vandetanib treatment. These mutations can also be acquired, over time, upon exposure to vandetanib if they were not already present before treatment [175]. Most of the tumors develop resistance against targeted drugs by acquiring new mutations that result in overactivation of pathways or through induction of alternate pathways to bypass the action of the drug [174].
MAPK and PI3K pathways are the overlapping pathways, and the upregulation of either of these pathways can lead to the same outcome. One possible mechanism by which thyroid tumors overcome the action of multiple kinase inhibitors is through induction of alternate signaling pathways or by upregulating the expression of tyrosine kinase receptors on the cell surface [176]. For example, targeting RAS and RET/PTC oncogenes in DTC can lead to the induction of PI3K pathway and activation of downstream targets, such as mTOR, forkhead family of transcription factor (FoxO), and others, thus helping tumors escape the drug action [176].
Thyroid tumors acquire RAI resistance by multiple mechanisms, and not all of them are well understood. One of the mechanisms by which thyroid tumors acquire RAI resistance is through the upregulation of the human epidermal receptor (HER) family of receptor tyrosine kinases [176]. HER2 and HER3 are the members of the EGFR family and are upstream of both the MAPK and PI3K pathways [176]. Preclinical data have shown that BRAF-mutant cells escape from the vemurafenib (BRAF inhibitor) effects by overexpressing HER2 and HER3 receptors, leading to the activation of mTOR and MAPK signaling pathways [63,176]. The tumor escape can be overcome by the combination treatment of vemurafenib and HER2 inhibitor trastuzumab, to improve RAI sensitivity (Figure 3) [63].
ALK fusion proteins also help thyroid tumors to escape from the effects of TKIs. ALK fusions are primarily seen in aggressive thyroid-cancer types, and they are known to activate MAPK and PI3K signaling pathways [177]. EML4-ALK fusion and various other ALK translocations have been identified in RAI resistant DTC [63]. STRN-ALK fusions have been observed in up to 4% ATC and 9% PDTC cases [176]. Combining TKIs with ALK specific inhibitors such as crizotinib will be beneficial to patients with ALK-positive tumors [63].
Thyroid tumors can escape targeted therapies by multiple mechanisms, and to this date, we have very little understanding of those pathways. The biggest challenge in overcoming the tumor escape mechanisms is to identify the mutations or alterations that play a role in the development of resistance against targeting drugs. Identification of the tumor escape mechanism will be helpful in the development of alternative therapies that can be combined with standard therapies to overcome the development of resistance against targeted therapies.

Conclusions
Several novel targeted therapies have recently been approved by the FDA for use in advanced thyroid cancer. There are many ongoing national and multi-national studies implementing targeted therapies in thyroid cancer. There is a shift in the management paradigm, with the molecular landscape rather than histology/morphology driving an individualized treatment approach. Even amongst the same histological groups of thyroid cancer, the response to therapy is different based on genetic and immune signature of the tumor, further confirming that "one size does not fit all". The use of the therapeutic agents should be individualized and based on shared decision-making after informing the patient about the possible benefits and disadvantages of the therapy. Encouraged by the results of individualized therapies documenting even complete responses when targeting driver mutations, the aim of future research should focus on finding efficacious treatments that will have a long-lasting curative effect.