American Head and Neck Society Endocrine Surgery Section and International Thyroid Oncology Group consensus statement on mutational testing in thyroid cancer: Defining advanced thyroid cancer and its targeted treatment

Abstract

Background:

The development of systemic treatment options leveraging the molecular landscape of advanced thyroid cancer is a burgeoning field. This is a multidisciplinary evidence-based statement on the definition of advanced thyroid cancer and its targeted systemic treatment.

Methods:

An expert panel was assembled, a literature review was conducted, and best practice statements were developed. The modified Delphi method was applied to assess the degree of consensus for the statements developed by the author panel.

Results:

A review of the current understanding of thyroid oncogenesis at a molecular level is presented and characteristics of advanced thyroid cancer are defined. Twenty statements in topics including the multidisciplinary management, molecular evaluation, and targeted systemic treatment of advanced thyroid cancer are provided.

Conclusions:

With the growth in targeted treatment options for thyroid cancer, a consensus definition of advanced disease and statements regarding the utility of molecular testing and available targeted systemic therapy is warranted.

1 ∣. INTRODUCTION

Thyroid cancer is a diverse group of malignancies, and together, is the most common endocrine malignancy with increasing incidence.¹ Although mortality due to differentiated thyroid cancer (DTC) is low, 15% of DTC is locally invasive and the 10-year overall survival is <50% in radioiodine refractory differentiated thyroid cancer (RAIR-DTC) patients with distant metastasis.^2,3 Thyroid cancers also include aggressive subtypes such as medullary thyroid carcinoma (MTC), poorly differentiated thyroid carcinoma (PDTC), and anaplastic thyroid carcinoma (ATC), each of which have high disease-specific mortality. Just a decade ago, treatments for recurrent and metastatic RAIR-DTC were limited. Studies with multikinase inhibitors (MKI), including randomized placebo-controlled phase 3 studies, have demonstrated efficacy in RAIR-DTC and MTC such that the evolution of genome-directed therapy has expanded the individualized therapeutic armamentarium significantly.^4-7 Even rare thyroid cancer subtypes, such as ATC, have evolving options.

The molecular genetics of both primary and advanced thyroid cancer, including ATC, have been investigated by multiple groups.^8-12 Many targetable alterations have been found with variable frequency in advanced thyroid cancer, including mutations in genes such as BRAF V600E, RET, PIK3CA, and PTEN, and fusions involving RET, NTRK, and ALK. Mutations such as TP53 and TERT promoter alterations may be associated with aggressive clinical behavior and poor prognosis.^11,13 In addition to these genetic alterations, microsatellite stability status, tumor mutational burden (TMB), and PD-L 1 status are being evaluated regarding utility for immunotherapy. The utility of genomic alterations and RNA sequencing for cytologically indeterminate thyroid nodules or advanced thyroid cancers for consideration of neoadjuvant therapy is also a burgeoning field.

Unfortunately, methodology to determine the somatic genetic landscape of a tumor is not uniform. There are many assays including polymerase chain reaction (PCR) and reverse transcription (RT)-PCR-based assays and assays based on next-generation sequencing (NGS) such as whole genome sequencing, as well as targeted NGS panels. In addition, in some alterations such as BRAF V600E or mismatch repair (MMR) deficiencies, immunohistochemistry can also be utilized. Recent publications have reported the use of liquid assays such as cell-free DNA or circulating tumor cells in advanced thyroid cancer.^14,15 Simultaneously, there is wide variability in practice regarding the timing of this testing during the disease course, as well as the extent of the testing with regards to the number of genes or markers to be tested. Finally, there are currently no guidelines or statements defining the subset of patients with thyroid cancer most likely to benefit from somatic mutational testing and targeted therapy. In fact, the very term “advanced thyroid cancer” has had many varied definitions in the otolaryngology (i.e., invasive, bulky, or non-resectable), endocrinology (i.e., RAI refractory) and medical oncology (i.e., distant metastasis) space.

The overarching goal of this consensus statement is to report consensus of an expert author panel with regard to the utility of somatic genetic alterations for prognosis and therapeutic intervention in advanced thyroid cancer. To achieve this goal, a multidisciplinary expert panel was convened under the umbrella of the American Head and Neck Society (AHNS) Endocrine Surgical Section in association with the International Thyroid Oncology Group (ITOG).

2 ∣. METHODS

In order to reflect the range of specialists involved in the care of patients with advanced thyroid cancer, the panel was comprised of 7 endocrinologists, 10 medical oncologists, 5 otolaryngologist-head and neck surgeons, 1 endocrine surgeon, 4 pathologists, and 1 molecular geneticist. This multidisciplinary author panel was assembled under the auspices of the American Head and Neck Society (AHNS) Endocrine Section and International Thyroid Oncology Group (ITOG). Once formed, the study group convened via telephone and web-based meetings at planned intervals, supplemented with additional electronic communication. To begin, the relevant published literature was queried through a key word search of “thyroid cancer” and “next-generation sequencing” (NGS). The English language literature from 2005 to 2019 using PubMed, Cochrane Library, Cochrane Central Register of Controlled Trials, and Medline was included. Additional literature was extracted from bibliographies of studies produced by the search. Finally, members of the panel were invited to supplement the search results with relevant, high-quality references. Available data were critically reviewed by the panel individually and discussed in meetings. This process resulted in a proposed set of 20 statements each generated by members of the author panel. The quality of evidence related to each proposed statement was categorized as “high” if there was supportive evidence from randomized clinical trials, “moderate” if the evidence was from prospective nonrandomized clinical studies, and “low” if the evidence was from retrospective analysis describing case series or retrospective single/multi-institutional data.

To quantitatively assess whether consensus was present for these statements, n = 27 panelists participated in a modified Delphi process. Within this process, each member independently rated each proposed statement on a 9-point Likert scale ranging from 1 “strongly disagree” to 9 “strongly agree.” A dedicated electronic interface was developed to present the statements securely and provide codified data for analysis. The system also allowed for additional free-text comments and feedback on the statements from panelists. Rating was performed in two rounds. In both rounds, statements were defined as meeting consensus if there was a mean score of at least 7.0 (“agree”) and 2 or fewer outlier responses or 3 outliers with no outlier response less than 5 (“neutral”). Outlier responses were defined as 2 Likert points away from the mean. After the initial round, responses were analyzed as above and descriptive feedback was reviewed by the panel and methodology leadership (D.S., G.W.R., B.S., J.J.S.). In follow-up, three additional statements were presented to all panelists in a second iteration of assessments. The proposed statements and results of these two iterative rounds of rating are summarized in Table 1. One author did not participate due to project timeline constraints. After the author panel provided its approval, the consensus statement was approved by the AHNS and ITOG and thereby has full organizational support as a statement of the AHNS and ITOG.

TABLE 1.

Summary of statements and results of panel ratings

	Statement	Consensusa	Outliers	Quality of evidence
Statement 1	Optimal care of patients with advanced thyroid cancer is best delivered by a multidisciplinary team of endocrinologists, medical oncologists, surgeons, pathologists, radiologists and radiation oncologists	Yes (8.9)	0	Low
Statement 2	Molecular testing should be performed in Clinical Laboratory Improvement Amendments (CLIA)-accredited laboratories (or their international equivalent), on appropriate specimens, using clinically-validated procedures, which may include laboratory-developed tests or FDA-approved commercial assays	Yes (8.7)	1	High
Statement 3	When somatic mutational testing is performed for thyroid cancer, multiplexed next-generation sequencing (NGS)-based panels are superior to multiple single-gene tests	Yes (8.3)	1	High
Statement 4	Rapid testing to assess BRAF V600E mutational status should be obtained for patients with anaplastic thyroid cancer	Yes (8.7)	0	Low
Statement 5	NGS panels that include assays for gene fusions are preferred given the ability to detect multiple mutations and fusions in one assay thereby conserving tissue and limiting expense. When not available, a multistep approach may be considered	Yes (8.2)	2	Low
Statement 6	For differentiated thyroid carcinoma (DTC), poorly differentiated thyroid carcinoma (PDTC), and medullary thyroid carcinoma (MTC), the following conditions in the right clinical setting may define advanced disease:
	Bulky, invasive, or inoperable primary neck/locoregional disease	Yes (8.5)	1	Low
	Anatomically detectable clinically recurrent disease	No (6.9)	4	Low
	Distant metastatic disease including mediastinal disease	Yes (8.2)	2	Low
	Biochemical or structural volumetric doubling time <6 months	No (6.7)	2	Low
	Gross residual neck disease when further surgical resection is not feasible or should be delayed	No (8.0)	3	Moderate
	Poorly differentiated/other aggressive histology components	No (6.9)	2	Low
	RAI nonavid/unresponsive disease	No (6.9)	4	High
	Unresponsiveness to TSH suppressive dosage levothyroxine (target TSH <0.1 mIU/L)	No (5.9)	5	Moderate
	Anticipated imminent threat imposed by tumor extent	Yes (8.3)	1	Low
	Other features that portend aggressive behavior at the discretion of the treating physician	No (7.1)	6	Low
Statement 7	All patients with anaplastic thyroid cancer are classified as having stage IV disease and should be considered advanced, even when completely resected and incidentally identified pathologically	Yes (8.3)	2	High
Statement 8	Somatic mutational testing for RET, NTRK1, NTRK3, and ALK gene fusions should be performed on patients with advanced DTC that tested negative for BRAF V600E to search for therapeutic targets. This can optimally be achieved using NGS of tumor RNA or DNA, or other validated approaches including fluorescence in situ hybridization (FISH)	Yes (8.6)	1	Moderate
Statement 9	BRAF V600E mutational status and microsatellite instability (MSI) status should be evaluated in patients with advanced PTC/PDTC to search for additional therapeutic options	No (7.6)	4	Moderate
Statement 10	Tumor mutational burden (TMB) should be analyzed prior to initiation of immunotherapy as it may inform response to anti-programmed cell death 1 (anti-PD-1) therapy	No (6.5)	3	Low
Statement 11	Multikinase inhibitors such as lenvatinib should be offered for the treatment of patients with progressive or symptomatic advanced differentiated thyroid carcinoma (DTC) that has failed conventional therapeutic options or as neoadjuvant therapy in the setting of preoperative advanced local disease or as part of a clinical trial	No (7.2)	5	High
Statement 12	The use of agents targeting BRAF mutations should be considered in patients with advanced differentiated thyroid carcinoma (DTC) harboring BRAF V600E mutations	Yes (8.0)	1	Moderate
Statement 13	Treatment with RET inhibitors should be considered for patients with advanced, progressive and/or symptomatic RET-altered differentiated thyroid carcinoma (DTC)	Yes (8.5)	1	High
Statement 14	NTRK inhibitor therapy should be considered in NTRK-altered advanced differentiated thyroid carcinoma (DTC)	Yes (8.4)	1	Moderate
Statement 15	All index patients with newly diagnosed medullary thyroid carcinoma (MTC) regardless of disease stage, personal history of other endocrinologic disorder or family history should have genetic counseling and be tested for germline RET mutations	Yes (9.0)	0	High
Statement 16	Patients with advanced, sporadic medullary thyroid cancer (MTC) without germline mutation, should be offered somatic RET mutational testing, as RET status can impact systemic treatment options.	Yes (8.4)	1	High
Statement 17	Patients with locally advanced MTC posing significant challenge to upfront surgical management may be offered somatic RET mutational testing if the patient is to be considered for neoadjuvant systemic therapy or as part of a clinical trial	Yes (8.1)	2	Low
Statement 18	The somatic mutational testing for advanced medullary thyroid cancer (MTC) should also include HRAS, KRAS, and NRAS mutations	No (6.7)	1	Low
Statement 19	Comprehensive somatic mutational testing of all anaplastic thyroid carcinoma (ATC) should be mandatory at diagnosis	Yes (7.8)	2	Moderate
Statement 20	Dabrafenib and trametinib combination therapy may be initiated for patients with resectable anaplastic thyroid carcinoma (ATC) and should be promptly initiated for patients with unresectable ATC upon identifying BRAF V600E mutation by IHC, PCR or tissue/blood-based NGS	Yes (7.9)	2	Moderate

^aConsensus was defined as a mean modified Delphi score ≥7.0 (“agree”) and 2 or fewer outliers or 3 or fewer outliers if all outlier scores were 5 (“neutral”) or above.

3 ∣. TESTING

Given the aggressive nature of advanced thyroid cancer, treatment considerations should include current standard therapies and clinical trials. In patients with advanced thyroid cancer in need of systemic therapy, molecular testing is indicated in most cases to detect actionable mutations and identify optimal treatment options for individual patients as recommended in the current guidelines for thyroid carcinoma.¹⁶

Statement 1: Optimal care of patients with advanced thyroid cancer is best delivered by a multidisciplinary team of endocrinologists, medical oncologists, surgeons, pathologists, radiologists, and radiation oncologists.

In tertiary care settings, this team of physicians is typically organized as a tumor board and patients' care is discussed in a multidisciplinary fashion. This allows for delivery of a coordinated, evidenced-based, and individualized treatment plan to the patient. If a hospital is unable to provide coordinated multidisciplinary care, patients may benefit from referral to a tertiary care setting where additional resources are available.

Statement 2: Molecular testing should be performed in Clinical Laboratory Improvement Amendments (CLIA)-accredited laboratories (or their international equivalent), on appropriate specimens, using clinically validated procedures, which may include laboratory-developed tests or FDA-approved commercial assays.¹⁷

Testing may be performed on formalin-fixed, paraffin-embedded (FFPE) tissue, frozen samples or cytological specimens. Ideally, molecular analysis should be run on the most recent tumor sample obtained from the patient, and the specimen should undergo evaluation by a pathologist, to ensure adequate tumor cellularity and proper tissue preservation.

Statement 3: When somatic mutational testing is performed for thyroid cancer, multiplexed NGS-based panels are superior to multiple single-gene tests (see Appendix for additional information).

Statement 4: Rapid testing to assess BRAF V600E mutational status should be obtained for patients with anaplastic thyroid cancer (see Appendix for additional information).

Statement 5: Next-generation sequencing (NGS) panels that include assays for gene fusions are preferred given the ability to detect multiple mutations and fusions in one assay thereby conserving tissue and limiting expense. When not available, a multistep approach may be considered.

Although comprehensive DNA- or RNA-based NGS fusion assays may be ideal, testing all advanced tumors using this approach may not be feasible (Figure 1). In some settings, multistep testing algorithms combined with fast, less expensive fusion screening methods, to enrich for patients with fusion-positive tumors, may provide an attractive alternative to widespread NGS-based fusion analysis but their utility is currently under evaluation. Testing for gene fusions is more challenging than DNA mutational analysis due to the diversity of potential fusion partners and the multitude of possible genetic breakpoints and exon combinations. Different methodologies have different limitations, and there are currently no FDA-approved tests for the detection of RET or NTRK fusions to facilitate patient selection for treatment with TKIs.¹⁸

FIGURE 1.

Somatic mutational testing of advanced DTC

Fluorescence in situ hybridization (FISH) assays using dual-color break-apart probes may be used to identify RET, NTRK1, and NTRK3 gene rearrangements. FISH tests are typically reimbursable and widely available in clinical laboratories but may fail to detect intrachromosomal rearrangements¹⁸ (see Appendix for additional information).

4 ∣. DEFINITION OF ADVANCED DISEASE

Consideration of somatic genomic interrogation should balance threats and burdens posed by a particular patient's cancer against the opportunities and burdens posed by a particular treatment. If cancer burden and threat is low, there is no justification for somatic genomic interrogation. While the features of advanced thyroid cancer may be obvious to those physicians who treat this condition regularly, there is currently no widely accepted definition of advanced thyroid cancer in the literature. We propose a series of considerations that should define advanced thyroid cancer and thereby contribute to individualized shared decision-making regarding somatic genomic interrogation and targeted systemic treatment (Table 2).

TABLE 2.

Thyroid carcinoma: Definition of advanced disease

Category	Characteristic of advanced disease	Quality of evidence	Reference
1. Structural/Surgical	Bulky, invasive, or inoperable primary neck/locoregional disease	Low	19,20
	Recurrent disease	Low	23
	Distant metastatic disease including mediastinal disease	Low	24,25
	Gross residual neck disease when further surgery is not an option	Low	N/A
	Rapid radiographic progression or doubling time	Low	27,144
	Anticipated imminent threat imposed by tumor extent	Low	N/A
2. Biochemical	RAI nonavid/unresponsive	High	5,32
	Unresponsive to TSH suppression	Moderate	40
	Rapid calcitonin, CEA, or thyroglobulin doubling times	Low	28-30
3. Histologic/molecular	Poorly differentiated or other aggressive histology components	Low	33,34
	High Ki 67 labeling index	Low	36,37
	High mitotic count/ tumor necrosis	Low	35
	All anaplastic thyroid carcinoma	High	41
4. Clinician prerogative	Features that portend aggressive behavior at the discretion of the treating physician	Low	N/A

Statement 6: For differentiated thyroid carcinoma (DTC) (papillary, follicular, Hurthle cell thyroid cancer [HCC], poorly differentiated thyroid carcinoma [PDTC]), and medullary thyroid carcinoma (MTC), the following conditions in the right clinical setting may define advanced disease:

1. Bulky, invasive, or inoperable primary neck/locoregional disease

Invasive disease and bulky cervical nodal disease are associated with worse overall survival and increased risk of locoregional recurrence, respectively.^19,20

Every effort should be made to treat locoregionally threatening disease with locoregional approaches before considering systemic therapy, as any systemic therapy will be expected to offer limited duration of benefit and will require enhancement regardless of initial benefit attained.²¹ To a large extent “resectability” is surgeon-specific and not an absolute, making expert (high volume, surgical oncology training, and/or experience) surgical input paramount. Impacts of locoregional therapies must be considered carefully; for example, if laryngectomy is required to resect bulky neck disease, is this justified by expected benefits and acceptable within the context of the patient's goals of care.²² There is currently an ongoing effort to define the role for neoadjuvant (before surgical intervention) targeted therapy and these approaches are currently evolving.¹⁵

2. Anatomically detectable clinically recurrent disease

Clinically recurrent disease can be generally categorized as Type 1—easily treated or observed and Type 2—advanced, not easily treated. Somatic testing most likely benefits only the latter. Patients who undergo surgical treatment of regional clinically recurrent disease have a high risk of disease persistence or future recurrence.²³

3. Distant metastatic disease including mediastinal disease

While patients with metastatic DTC and MTC can achieve favorable outcomes, metastatic disease (including mediastinal nodal disease) has a negative impact on overall survival.^24,25 The impact of metastatic disease on survival is even more profound in patients with MTC and PDTC.²⁶

4. Biochemical or structural volumetric doubling time <6 months

If tumor growth rate raises concern wherein systemic therapy is expected justified, somatic genomic interrogation is justified.^27-30

5. Gross residual neck disease when further surgical resection is not feasible or should be delayed

Somatic testing could also be considered in patients with gross residual or recurrent disease locoregionally where further surgical resection is not feasible due to anatomical constraints or when medical comorbidities preclude safe anesthesia administration/surgical resection. The determination of whether disease can be resected or not should be made by an expert surgeon. External beam radiation therapy can be considered to improve locoregional control but does not improve overall survival or distant metastasis failure-free survival and is associated with additional morbidity.³¹ Systemic therapy may be favored over radiation therapy in this situation.

6. Poorly differentiated/other aggressive histology components

Certain histologic tumor variants (columnar cell variant, hobnail variant, tall cell variant [TCV] of papillary thyroid carcinoma [PTC]), characteristics (tumor necrosis, high mitotic activity), or subtypes such as PDTC are associated with more aggressive clinical behavior.^32-35 This also includes patients with a high Ki 67 labeling index.^36,37 It should be noted however that some of the tumors with the above histologic features can be small and minimally invasive and therefore do not necessarily require somatic molecular testing.

7. RAI nonavid/unresponsive disease

RAI insensitivity is a reason to consider targeted therapy initiation and/or for somatic genomic interrogation in DTC and PDTC.⁵ While FDG avidity often accompanies iodine nonavidity, it is not an absolute predictor of RAI insensitivity.³⁸ In cases where RAI resensitization therapy is considered somatic genomic testing should be undertaken so as to best inform the selection of candidate RAI sensitizing agents.³⁹

8. Unresponsiveness to TSH suppressive dosage levothyroxine (target TSH <0.1 mIU/L)

Some DTC are exquisitely responsive and well controlled in response to TSH suppressive dosage levothyroxine, making unresponsiveness to TSH suppression below 0.1 mIU/L an indication for somatic genomic testing and/or escalation to targeted therapy.⁴⁰

9. Anticipated imminent threat imposed by tumor extent

Examples of imminent threat include disease whose presence implies impending neural or respiratory decompensation. If the extent and locations of metastatic deposits pose anticipated imminent structural or functional threat, somatic genomic testing and escalation of targeted therapy are merited.

10. Other features that portend aggressive behavior at the discretion of the treating physician

Some characteristics of thyroid cancer not outlined above may represent aggressive disease. For example, symptomatic disease progression with potential for therapeutic palliation should prompt consideration of both systemic therapy and somatic genomic interrogation, as symptomatic benefit might result. Additionally, unpredictable behavior requiring episodic salvage therapies may indicate aggressive disease. Some patients, even if imaged every 3 months, develop unpredicted/unexpected disease progression requiring urgent palliation despite otherwise manifesting no cancer-related symptoms and having otherwise indolent cancer growth and doubling time. Features such as these can provide rationale for systemic therapy initiation and for somatic genomic interrogation at the discretion of the multidisciplinary care team.

Statement 7: All patients with anaplastic thyroid cancer are classified as having stage IV disease and should be considered advanced, even when completely resected and incidentally identified pathologically.

Patients with ATC often are in immediate need of systemic therapy to control existent metastatic disease—or will be expected to be so shortly. Consequently, there is rationale for somatic mutational testing in all patients with ATC, as they have potential to harbor productively targetable somatic alterations. Moreover, for ATC with a somatic BRAF V600E mutation, paired therapy (dabrafenib + trametinib) is FDA approved in this context and brisk initial responses have been described. Thus, in all patients with ATC, rapid examination for BRAF V600E mutation should be performed; more comprehensive analysis should be limited to ATC without a BRAF mutation or to those that have a mutation but do not respond, or lose response to targeted therapy.⁴¹

5 ∣. DIFFERENTIATED THYROID CANCER

5.1 ∣. Molecular pathways

Over the last decade, technological advancements including NGS have significantly expanded knowledge of the molecular pathways pertinent to thyroid cancer pathogenesis. Collaborative efforts across many institutions, including The Cancer Genome Atlas (TCGA), have successfully utilized NGS to uncover the key mutations and structural alterations in PTC. The comprehensive molecular catalog that has emerged from these studies has directly impacted the diagnosis, prognosis, and treatment of this disease.^42,43 The following section will focus on the genetic basis of the various subtypes of DTC and thereby provide a molecular framework for approaching these tumors.

5.1.1 ∣. Papillary thyroid carcinoma

Molecularly, well-differentiated PTCis usually characterized by a low tumor mutational burden and recurrent alterations involving the MAPK signaling pathway.⁴³ Although altered MAPK signaling is usually the main driver event in PTC, changes in PI3K, WNT, and epigenetic pathways, and mutations in the TERT promoter are additional genomic events that when concurrent with a MAPK-activation have been associated with more aggressive behavior. The genetic signatures of PTC correlate with the three dominant subtypes that have been described based upon histological findings. Classical or conventional PTC is frequently characterized by activating BRAF V600E point mutations or, to a lower extent, by rearrangements involving RET, BRAF, or NTRK1/3 fusion products.^44,45 TCV, which is a more aggressive type of PTC, is usually driven by BRAF V600E point mutations and is associated with a higher mutational density than other PTC.^46,47 Finally, follicular variant PTC is predominantly characterized by RAS family point mutations (most commonly NRAS) and, less frequently, by BRAF V600E mutations.^43,48,49 Copy number variations (CNV) are also more frequent in follicular variant PTC.⁴⁸ The PAX8-PPARG rearrangement can be seen in follicular variant PTC, FTC, and PDTC.^43,50 Of these genetic aberrations, RAS mutations and PAX8-PPARG rearrangements also occur in benign thyroid follicular adenomas (FA), thus they are not carcinoma-specific.

Other somatic events found in PTC include mutations of E1F1AX and the TERT promoter, as well as alterations in DNA repair and tumor suppressor genes, such as CHEK2, ATM, ERCC5, TP53, and NF1/2. The TERT promoter variants in the presence of BRAF mutations are prognostic of more aggressive disease as they are associated with increased recurrence rates and decreased survival.⁴³ The aforementioned BRAF mutations and RTK gene rearrangements (including RET and NTRK1/3 fusion products) are well-characterized oncogenic drivers that may have significant implications when considering targeted treatment options for advanced DTC.^51-53

5.1.2 ∣. Follicular thyroid carcinoma

Follicular thyroid carcinoma (FTC) is most frequently found to harbor mutations in the RAS family of proto-oncogenes, reported in 22%–49% of tumors, with mutations in NRAS being more commonly observed than in KRAS or HRAS.^54,55 The PAX8-PPARG fusion gene product, reported in 3%–36% of FTC, is mutually exclusive of RAS mutations and further defines a distinct class of FTC.^54,56,57 PTEN tumor suppressor mutations have also been well-described in FTC, implicating activated PI3K signaling in tumor formation.⁵⁸ All of these genomic changes occur also in FA, and germline mutations in PTEN also confers increased risk for FA, FVPTC, and FTC (i.e., Cowden syndrome) suggesting a common genomic lineage for these tumor-types involving PI3K signaling. Loss of heterozygosity (LOH) and DNA copy number alterations are more common in FTC than in PTC and likely play a role in tumorigenesis.^59,60 Finally, TERT promoter mutations have been reported in 17% of FTC and, like PTC, are observed in older patients and associated with an increased risk for distant metastatic disease.⁶¹

5.1.3 ∣. Hürthle cell thyroid carcinoma

Hürthle cell thyroid carcinoma (HCC) is a unique form of thyroid cancer with a remarkable abundance of mitochondria that was once considered a subtype of FTC.⁶² Recent genetic analyses clearly define HCC to be a separate entity.⁶³ The most distinctive molecular features of HCC are recurrent mutations in mitochondrial DNA (mtDNA), extensive LOH across multiple chromosomes, and activation of mTOR.^64,65 The mtDNA mutations are highly pervasive, occurring in ~60% of HCC, and enriched for disruptive changes in genes encoding subunits of complex I (CI) of the electron transport chain.^64-66 Widespread LOH events either lead to near haploid genomes or whole genome duplication events.^64,65,67 TERT promoter mutations and TP53 mutations are more common in widely invasive HCC, while RAS and BRAF mutations are found less frequently than in FTC and PTC, respectively.^65,68-70

5.2 ∣. Indicated molecular testing

If systemic therapy with targeted inhibitors is being considered for a patient with progressive DTC that is not responsive to conventional approaches, the following somatic genetic testing is recommended prior to initiating therapy whenever possible (Figure 1 and Table 3).

TABLE 3.

Genetic abnormality and corresponding targeted systemic therapy option for patients with advanced thyroid cancer being considered for treatment with selective inhibitors

Genetic abnormality	Molecular testing recommendation	Drug (mechanism: targets)	FDA status (indication)	No. of mutation+/fusion+ patients enrolled in CT (objective response rate)	Reference
BRAF V600E	DTC: strong, moderate MTC: weak, low ATC: strong, high	Dabrafenib (STKI: BRAF V600E/K) single agent	Approved (BRAF V600E mutant melanoma)	DTC: 10 (3/10) DTC: 22 (50%)	76,145
		Dabrafenib (STKI: BRAF V600E/K) with Trametinib (STKI: MEK1/2)	Approved (BRAF V600E/K mutant melanoma; BRAF V600E mutant ATC and NSCLC)	ATC: 16 (69%) DTC: 24 (54%)	41,76
		Vemurafenib (STKI: BRAF V600E)	Approved (BRAF mutant melanoma)	ATC: 7 (2/7) DTC: 48 (39%, no previous TKI therapy; 27%, prior VEGFR TKI)	75,146
NTRK fusions (NTRK1, NTRK3)	DTC: strong, moderate MTC: weak, low ATC: weak, low	Entrectinib (TRKI: NTRK fusions)	Approved (NTRK fusion+ solid tumors, ROS1 fusion+ NSCLC)	n/a	129
		Larotrectinib (TRKI: NTRK fusions)	Approved (NTRK fusion+ solid tumors)	Histology not reported: 24 (79%)	53 a
RET fusions (DTC, ATC)	DTC: strong, high MTC: strong, high	Selpercatinib (TKI: RET inhibitor)	Approved (RET-mutant MTC, RET fusion+ thyroid cancer, RET fusion+ NSCLC)	MTC: 55 (69%) DTC: 26 (79%)	72,147
RET mutations (MTC)	ATC: weak, low
		Pralsetinib (TKI: RET inhibitor)	Approved (RET-mutant MTC, RET fusion+ thyroid cancer, RET fusion+ NSCLC)	MTC: 79 (65%) DTC: 12 (75%)	73,99,148
ALK fusions	DTC: weak, low MTC: weak, low ATC: weak, low	Crizotinib (TKI: ALK/ROS1 fusions)	Approved (ALK fusion+ or ROS1 fusion+ NSCLC)	ATC: 1 (1/1, case report) MTC: 1 (1/1) MTC: 1 (1/1, case report)	141,149,150
		Alectinib (TKI: ALK fusions)	Approved (ALK fusion+ NSCLC)	MTC: 1 (1/1, case report)	150

Abbreviations: STKI, serine/threonine kinase inhibitor; TKI, tyrosine kinase inhibitor; TRKI, tropomyosin receptor kinase inhibitor.

^aPooled data.

Statement 8: Somatic mutational testing for RET, NTRK1, NTRK3, and ALK gene fusions should be performed on patients with advanced DTC that tested negative for BRAF V600E to search for therapeutic targets. This can optimally be achieved using NGS of tumor RNA or DNA, or other validated approaches including fluorescence in situ hybridization (FISH).

RET gene fusions, detected in ~10% of PTC, are oncogenic and predict response to targeted inhibitors, including the FDA-approved drugs selpercatinib and pralsetinib.^43,71-73 Genetic rearrangements involving NTRK1/3 are relatively rare, reported in only ~2% of PTC.⁴³ However, NTRK fusion-positive tumors, including thyroid cancer, can respond to treatment with TRK inhibitors, including FDA-approved larotrectinib and entrectinib.^51-53 ALK fusions are reported in <1% of PTC but are more frequent in PDTC. There are currently no FDA-approved ALK inhibitors for the treatment of DTC; however, they are FDA-approved for solid tumors that harbor ALK fusions and a few patients with thyroid cancer are included in the reported clinical trials and/or case reports (Table 3). Thus, testing for ALK rearrangements is indicated in the context of either “off-label” treatment or clinical trial enrollment. It is appropriate to include ALK fusion testing in larger, multigene testing panels.

Statement 9: BRAF V600E mutational status and microsatellite instability (MSI) status should be evaluated in patients with advanced PTC/PDTC to search for additional therapeutic options.

Despite encouraging clinical data, there are currently no BRAF-inhibitor therapies approved by the FDA for the treatment of DTC, although vemurafenib and dabrafenib are FDA-approved for BRAF V600E-mutated melanomas.^41,74-76 Therefore, BRAF testing is recommended in the context of either “off-label” long term or redifferentiation/I-131 therapy or clinical trial enrollment and are part of the NCCN guidelines for the treatment of BRAF V600E-mutated DTC.¹⁶ Additionally, though the estimated overall prevalence of MSI-high cancer in DTC is low, such patients are eligible for treatment with the programmed death-1 (PD-1) inhibitor pembrolizumab given the drug's tissue agnostic approval for MSI-high cancers.^77,78

Statement 10: Tumor mutational burden (TMB) should be analyzed prior to initiation of immunotherapy as it may inform response to anti-programmed cell death 1 (anti-PD-1) therapy.

While TMB is typically low in DTC and limited data are available regarding thyroid cancers specifically, studies of solid tumors including thyroid cancer have demonstrated that high TMB is associated with response to pembrolizumab.^79,80 Limited information is available on utility of PD-L 1 IHC in thyroid cancers being considered for immunotherapy.

5.3 ∣. Systemic therapy

The importance of molecular alterations and signaling pathways in thyroid cancer has led to the development of new targeted therapeutic drugs within the last decade, affording much needed treatment options for patients with advanced DTC.

Statement 11: Multikinase inhibitors such as lenvatinib should be offered for the treatment of patients with progressive or symptomatic advanced differentiated thyroid carcinoma (DTC) that has failed conventional therapeutic options or as neoadjuvant therapy in the setting of preoperative advanced local disease or as part of a clinical trial.

Two multikinase inhibitors have been approved by the FDA for the treatment of DTC: lenvatinib and sorafenib. Both of these compounds are multikinase inhibitors (MKIs) that block activation of a number of key receptors that regulate thyroid cancer progression. Lenvatinib therapy is associated with longer median progression-free survival in patients with DTC compared to placebo (18.3 vs. 3.6 months) with a superior hazard ratio for progression or death (0.21, 99% CI 0.14–0.31) and response rate (64.8% vs. 1.5%) in a phase III study.⁴ Sorafenib was shown in a phase III trial with 417 patients to be superior to placebo for treatment of RAIR DTC with significantly longer median progression-free survival (10.8 vs. 5.8 months, HR 0.59, 95% CI 0.45–0.76).⁵ Given these agents' multiple targets, both drugs are associated with considerable toxicity profiles, which often limits their clinical utility. In addition, complete remissions have not been described and the duration of response in in the 12–24 month range, thus their use in patients with stable, asymptomatic disease is discouraged. A head-to-head comparison study of the two drugs has not been performed. Clinical trials evaluating the neoadjuvant use of multikinase inhibitors are in progress.

Statement 12: The use of agents targeting BRAF mutations should be considered in patients with advanced differentiated thyroid carcinoma (DTC) harboring BRAF V600E mutations.

The use of agents targeting BRAF V600E, commonly found in PTC, has been described. Vemurafenib, approved for the treatment BRAF V600E-mutated unresectable or metastatic melanoma, has demonstrated clinical response and was studied in a Phase II study where objective responses were observed in 38.5% of patients.^74,75 The combination of dabrafenib (BRAF inhibitor) and trametinib (MEK inhibitor), approved in BRAF V600E mutated melanoma and ATC, also has been studied in BRAF-mutated PTC.⁷⁶ One study compared single-agent dabrafenib with the combination dabrafenib/trametinib in which both groups demonstrated high response rates (50% single-agent dabrafenib vs. 54% combination, modified RECIST criteria) and median PFS 11.4 vs. 15.1 months.^41,76 These responses were durable with median duration of response not yet reached for both treatment groups.

Statement 13: Treatment with RET inhibitors should be considered for patients with advanced, progressive and/or symptomatic RET-altered differentiated thyroid carcinoma (DTC).

Currently, very highly selective targeted therapies have been shown not only to offer continued high response rates and clinical benefit but to avoid the extensive side effects seen with approved multikinase inhibitors. Selpercatinib (LOXO-292) is a highly selective RET inhibitor that was recently approved by the FDA for the treatment of advanced or metastatic RET fusion-positive RAI-refractory thyroid carcinoma and RET-mutated MTC. Drug efficacy was investigated in the LIBRETTO-001 phase 1/2 clinical trial of selpercatinib, which included 26 RET-fusion positive PTC⁷¹ with a reported ORR of 62% in the most updated results.⁷² To date, patients included on this study have exhibited few adverse events, making this a favorable treatment option with excellent disease control rates and a tolerable side effect profile. Similarly, pralsetinib is another highly selective inhibitor for RET-altered cancers, including PTC, with similar efficacy and tolerable toxicity, and is approved for RET-altered thyroid cancer.⁷³

Statement 14: NTRK inhibitor therapy should be considered in NTRK-altered advanced differentiated thyroid carcinoma (DTC).

The NTRK inhibitor, larotrectinib, was evaluated in a series of clinical trials in which thyroid cancer comprised a significant proportion of the NTRK fusion-positive tumors included on the initial study and the pooled analysis.^51,53 The study reported an ORR of 75% with most responses sustained >1 year and an absence of grade 3 or 4 adverse events. A recent study focused on thyroid cancer showed an ORR of 90% making larotrectinib a prototypic targeted therapy.⁸¹ A recent study Entrectinib, another highly selective NTRK inhibitor, also demonstrates excellent disease control, and both drugs are approved for optimal frontline systemic treatments for thyroid cancers harboring a NTRK-fusion alteration.⁵²

Despite recent advancements with remarkable results of targeted therapies, further investigation is warranted to provide more clinically relevant treatment options for patients with advanced thyroid cancer. Now that the molecular landscape of DTC has become better elucidated, drug development and subsequent clinical trial enrollment can be geared towards affording many more precision medicine options for such patients in the future.

5.4 ∣. Redifferentiation therapy

Recent insights into the role of MAPK pathway oncogenic gene mutations in suppressing the expression of follicular cell genes responsible for radioiodine (RAI) uptake and retention has led to a new field of clinical investigation, whereby MAPK targeted therapies are used to restore RAI avidity in RAIR patients,^82,83 a therapeutic strategy termed “redifferentiation.” A potential advantage of restoring RAI efficacy in these patients is creating an effective systemic therapy option that avoids the toxicities associated with continuous dosing required for standard RAIR drug therapies. A pilot clinical trial utilizing investigational ¹²⁴I PET/CT lesional dosimetry demonstrated that in a subset of RAIR patients a 4-week administration of the allosteric MEK 1/2 inhibitor selumetinib (Astra-Zeneca) enhanced RAI uptake in 8 of the 20 patients treated, translating to tumor regressions in response to ¹³¹I therapy.⁸⁴ MEK inhibition was less effective for patients whose tumors harbored BRAF V600E mutations (1/9 patients achieved redifferentiation) compared to those with NRAS mutations (5/5 patients achieved redifferentiation). Two subsequent studies evaluating the small molecule BRAF inhibitors vemurafenib⁸⁵ and dabrafenib⁸⁶ showed higher rates of redifferentiation (40%–60%) over what was previously observed with selumetinib (11%) in BRAF V600E mutant RAIR metastatic thyroid cancers, supporting the hypothesis that tailoring therapy to the tumor gene mutation to optimize pathway inhibition is a more effective strategy for restoring clinically relevant RAI responses. Case series describing off label experiences with RAF kinase inhibitors for redifferentiation therapy in RAIR patients have been reported by several groups.^87-89 There has also been an effort to evaluate how redifferentiation might be applied in other clinical settings, including augmenting the efficacy of postoperative adjuvant RAI for patients at high risk of disease persistence or recurrence. The ASTRA study (Adjuvant Selumetinib for differentiated Thyroid cancer Remission After RAI) was an international trial that compared the 18-month biochemical and structural complete remission rate in patients randomized 2:1 to selumetinib or placebo in combination with standard RAI after thyroidectomy.⁹⁰ The study did not demonstrate an advantage with selumetinib in this setting, though subset analyses suggested that future strategies designed to encourage treatment compliance, the use of more potent kinase inhibitors, and tailoring drug therapy to tumor genetics may be critical for improving complete remission rates with redifferentiation.

Redifferentiation is currently being investigated in several ongoing clinical trials internationally to identify the most effective targeted therapy regimens for each genomic subset and provide more data on the clinical utility of this approach. Clinical and molecular biomarkers correlating to redifferentiation efficacy promise to guide patient selection to enrich for clinical benefit elicited with this therapeutic approach.⁹¹

6 ∣. MEDULLARY THYROID CANCER

6.1 ∣. Molecular pathways

Medullary thyroid cancer (MTC) is a subtype of thyroid cancer which arises from the neuroendocrine parafollicular C cells of the thyroid gland and accounts for 1%–2% of thyroid cancer in the United States.⁹²

Approximately 75%–80% of MTC occur sporadically, and the remaining 20%–25% occur as an inherited tumor syndrome, either multiple endocrine neoplasia (MEN) 2A or 2B. Hereditary MTC are associated with genomic alterations of the RET proto-oncogene and inherited in an autosomal dominant fashion. Hereditary MTC display a distinct genotype-phenotype correlation, with age of MTC onset and co-occurring conditions, such as pheochromocytoma and hyperparathyroidism varying depending in part on the specific underlying RET alteration.

In MEN2A, the majority of RET mutations occur in the extracellular cysteine rich coding domain, the most common of which are at codon C634 (e.g., C634R, C634W). Patients harboring these mutations are at high risk for the onset of MTC in childhood, and thus prophylactic thyroidectomy should be considered in carriers with age recommendations ranging from early childhood (<5 year of age) to young adulthood depending on the family history and the particular gene mutation, clinical evaluation and family/patient preferences. Nearly all patients with MEN2B harbor germline RET M918T mutations in the protein's tyrosine kinase domain. MEN2B patients have the highest risk for early onset MTC requiring consideration of thyroidectomy in the first year of life.

In patients with sporadic MTC, somatic RET mutations are found in approximately 50% of cases, with the M918T mutation being the most common, followed by C634R. About 6% of patients with clinically apparent sporadic MTC without family history or other endocrinologic disorder are found to harbor a germline RET mutation. In patients with sporadic RET-wildtype MTC, somatic mutations in HRAS (~25%) or KRAS, and rarely NRAS genes have been found, which are mutually exclusive with RET mutations.⁹³ About 20% of sporadic MTC harbor neither RET nor RAS gene alterations.⁹⁴

RET encodes a single-pass transmembrane tyrosine kinase receptor protein. Dysregulation of RET signaling is associated with several other malignancies. While in MTC RET activation is caused primarily by point mutations, in most other malignancies (e.g., lung adenocarcinomas, papillary thyroid cancers, and chronic myelomonocytic leukemia) RET is activated via chromosomal rearrangements.^95,96 Whether RET activation is by point mutation or gene fusion, these alterations all result in ligand-independent constitutive activation of the RET signaling pathway and oncogenesis.

6.2 ∣. Indicated molecular testing

6.2.1 ∣. Germline RET mutation

Statement 15: All index patients with newly diagnosed medullary thyroid carcinoma (MTC) regardless of disease stage, personal history of other endocrinologic disorder or family history should have genetic counseling and be tested for germline RET mutations.

If a germline RET mutation is confirmed, genetic germline testing and cascade testing of the family, starting with first-degree relatives, should also be offered. This is critically important in order to identify carriers for whom thyroidectomy can be performed prior to or early in the development of MTC.

RET mutation testing can be performed either by NGS assay of the entire RET exomic sequence or in a tiered manner in which common hotspots of recurrent mutations in exons 10, 11, and 13–16 are analyzed first. In this case, the sequencing of additional exons may be considered in patients with strong clinical features or family history highly suggestive of hereditary MTC syndromes, and with no mutations found in exons 10, 11, or 13–16 (a tiered approach). The family members of MTC patients with a known germline mutation may have a targeted test for that specific RET mutation.

6.2.2 ∣. Somatic genetic testing

Statement 16: Patients with advanced, sporadic medullary thyroid cancer (MTC) without germline mutation, should be offered somatic RET mutational testing, as RET status can impact systemic treatment options.

Somatic RET mutations lead to more aggressive MTC, including higher T- and N-stage, and increase the rate of distant metastasis. Thus, while approximately 60% of sporadic MTCPT harbor somatic RET mutations, patients with recurrent/metastatic MTC are even more likely to harbor somatic RET mutations.

Statement 17: Patients with locally advanced MTC posing significant challenge to upfront surgical management may be offered somatic RET mutational testing if the patient is to be considered for neoadjuvant systemic therapy or as part of a clinical trial.

Statement 18: The somatic mutational testing for advanced medullary thyroid cancer (MTC) should also include HRAS, KRAS, and NRAS mutations.

6.3 ∣. Systemic therapy

Two multikinase inhibitors (MKIs), vandetanib and cabozantinib, are approved by the U.S. Food and Drug Administration (FDA) for the systemic treatment of MTC. Both MKIs have moderate activity against RET and strong activity against VEGFR kinases. Vandetanib has inhibitory activity against RET, VEGFR2, 3, and EGFR and showed an improvement in progression-free survival (PFS) compared to placebo in a pivotal randomized phase III study.⁷ Cabozantinib targets VEGR2, c-MET, AXL, and RET, and was approved for the treatment for advanced MTC based on a PFS benefit compared to placebo in a randomized phase III study.⁶ However, because of their activity against multiple kinases, both MKIs have a narrow therapeutic window and the effective dose for sufficient RET pathway suppression is limited by off-target toxicities driven primarily by VEGFR2 inhibition. In addition, these MKIs lack efficacy in tumors with acquired gatekeeper resistance mutations at RET codon V804.

Recently, highly selective and potent RET inhibitors have been developed and show promising efficacy with a favorable toxicity profile in tumors with various RET gene alterations.⁹⁷ Selpercatinib (LOXO-292) is a highly selective RET kinase inhibitor with nanomolar potency (IC50 0.2–12.5 nM) against a variety of RET alterations, including gene fusions, oncogenic mutations, even the V804 gatekeeper mutation. In an early presentation of data from the phase I/II study of selpercatinib (LIBRETTO-001), 56% of patients (n = 55) with RET-mutant MTC who were previously treated with vandetanib and/or cabozantinib achieved objective responses.⁷² The adverse events were mostly grade 1 or 2. Based on the results of this trial, selpercatinib was recently approved by the FDA for the treatment of adult and pediatric patients ≥12 years of age with advanced RET-mutant MTC that requires systemic therapy. An ongoing randomized study is evaluating selpercatinib compared with standard MKI therapy in treatment-naïve patients with RET-mutant MTC. Finally, neoadjuvant Selpercatinib prior to surgery is a novel treatment strategy which has been reported for locoregionally advanced and distant metastatic RET-mutated MTC.⁹⁸

Pralsetinib (BLU-667) is another highly selective and potent RET inhibitor which has been recently approved by the FDA for the treatment of patients with advanced or metastatic RET-mutant MTC (IC50 0.3–5 nM). An interim analysis of a RET-altered thyroid cancer cohort in the phase I/II study (ARROW) of pralsetinib showed 65% of objective response rated in patients with RET-mutant MTC (n = 71), including patients with MKI resistant tumors and with known gatekeeper mutations. The drug was well tolerated with most treatment related adverse events being low grade and reversible.^73,99

For patients with advanced RET-mutant MTC who require systemic therapy, these two selective RET inhibitors build upon the known biology of the disease and are expected to offer patients new treatment alternatives that are highly efficacious and have a relatively favorable side effect profile.

7 ∣. ANAPLASTIC THYROID CANCER

7.1 ∣. Introduction

Anaplastic thyroid cancer (ATC), one of the most aggressive malignancies in humans, accounts for approximately 1% of all thyroid cancers.¹⁰⁰ The majority present with distant metastases and loco-regional extension beyond the thyroid.¹⁰¹ Prognosis has historically been dismal with median survival of approximately 5 months and 1-year survival rate of <20%.¹⁰² With evolving treatments for molecular subsets of ATC, it is imperative that diagnosis and treatment be expedited. These guidelines will provide a necessary platform for uniformity across institutions.

7.2 ∣. Molecular pathways

ATC is an undifferentiated follicular cell derived carcinoma broadly composed of three patterns (spindle, giant cell and epithelial) which can be found singly or in combination. Other patterns can also occur such as squamoid and rhabdoid.⁶³ By immunostaining, these tumors can be positive for PAX8, are often negative for TTF-1 and almost always lack thyroglobulin.¹⁰³ Most studies suggest progressive dedifferentiation of DTC to PDTC and then to ATC due to progressive accumulation of mutations. Many studies have looked at genomic profiles of ATC and systematic reviews have cataloged these findings.^{12,101,104-108} Table 4 shows common molecular aberrations.

TABLE 4.

Common genetic alterations in ATC

Genetic alteration	Frequency	Treatment based on mutation	Level of evidence (NCCN)
BRAF mutation/fusion	28%–45%11,12,151	FDA approval of BRAF inhibitor’ dabrafenib in combination with MEK inhibitor, trametinib for BRAF V600E positive locally advanced, or metastatic ATC with no locoregional treatment options in 2018119	Category 2A
NTRK1/3 fusions/mutations	152	Tumor-agnostic FDA approval of larotrectinib in 2018 for patients with metastatic unresectable solid tumors that harbor neurotrophic receptor tyrosine kinase (NTRK) gene fusion127	Category 2A
Mismatch repair gene mutation	0%–12%11,12,151	Tumor-agnostic approval of pembrolizumab in 2017 for treatment of patients with advanced unresectable or metastatic microsatellite instability-high (MSI-H) or mismatch repair deficient (dMMR) solid tumors including thyroid cancer130	Category 2A
Anaplastic Leukemia Kinase (ALK) fusions	Anaplastic Leukemia Kinase (ALK) fusions139-141	Preclinical data and case reports of ALK kinase inhibitors activity in ALK-fusion positive ATC140,141	N/A
RAS mutation Mutations in NRAS, HRAS, and KRAS isoforms reported in ATC	23%–29%11,12,151	N/A	N/A
PI3CA mutation	10%–18%11,12,151	MTOR inhibitors such as everolimus may have some efficacy in patients with tumors harboring this mutation if there are no other co-occurring targetable mutations142,153	N/A
PTEN mutation	9%–15%11,12,151	N/A	N/A
TERT promotor mutation	6%–73%11,12,151	N/A	N/A
TP53 mutation	46%–73%11,12,151	N/A	N/A
RB1 mutation	4%–9%11,12,151	N/A	N/A
NF1/2 mutation	10%–21%11,12,151	N/A	N/A
CDKN1/2 mutation (CCNE1 amplification)	4%–8%12,151	N/A	N/A
STK11 mutation	0%–6%11,12	N/A	N/A
SWI/SNF chromatin remodeling complex mutations	18%–36%11,12	N/A	N/A
ATM	4%–9%11,12,151	N/A	N/A
RET	Rare but noted in some series72	RET inhibitors Selpercatinib and Pralsetinib	N/A

7.3 ∣. Indicated molecular testing

Statement 19: Comprehensive somatic mutational testing of all anaplastic thyroid carcinoma (ATC) should be mandatory at diagnosis.

Given the aggressive course and resistance to chemotherapy, radiotherapy and radioactive iodine, and the potential to identify a targetable mutation, all patients with ATC must undergo expeditious staging, histological confirmation and broad-spectrum NGS testing (Figure 2). The turn-around time for tissue-based NGS ranges 10–28 days depending on methodology, number of genes tested and the laboratory. Several approaches have been taken to more rapidly determine if a tumor has a BRAF V600E mutation to enable earlier initiation of targeted therapy if positive. For example, some centers employ a rapid PCR assay to detect BRAF V600E in DNA isolated from paraffin blocks with 48–72 h turn-around. A second option is the use of peripheral blood NGS (cell-free DNA) which has sensitivity of 75%–90% and turnaround time of 3–7 days. This avoids the delay in procuring tissue-based NGS results, enabling early initiation of appropriate therapies.^109-112 Mutation-specific IHC for BRAF V600E is reported to be useful in expeditiously identifying patients who might benefit from approved targeted therapy.^113-115 IHC for BRAF V600E should not be performed on FNA smears, as this may result in false positive readings. Core biopsy, FNA cell block, or surgical specimens are required for this test.¹¹⁵

FIGURE 2.

Somatic mutational testing of ATC

7.4 ∣. Systemic therapy

R0/R1 total thyroidectomy with lymph node dissection followed by radiotherapy +/− radiosensitizing chemotherapy for resectable (stage IVA) ATC is recommended.¹¹⁶ For patients with unresectable disease without distant metastatic disease (stage IVB), external beam radiation with radiosensitizing chemotherapy has historically been the preferred approach. This has changed with the availability of effective therapy for BRAF V600E mutated ATC. In patients with stage IVB and IVC, unresectable BRAF mutated ATC, neoadjuvant BRAF-directed therapy followed by surgical resection +/− radiation is an alternative to upfront radiation.¹¹⁶ For patients without an actionable mutation such as BRAF V600E and unresectable and metastatic disease (stage IVC), palliative radiation (in select cases) followed by systemic therapy is recommended. Chemotherapeutic drugs like paclitaxel, doxorubicin, and platinum, as single agents or combinations, have minimal benefit with most patients experiencing rapid progression amidst significant toxicity.^117,118 These can, however, be utilized as bridging chemotherapy until molecular testing results are available. Targeted therapy is recommended in patients whose tumor has an actionable target.

7.4.1 ∣. Systemic therapy for BRAF-mutated ATC

The FDA approved the combination of BRAF and MEK inhibitors, dabrafenib and trametinib, for BRAF V600E mutated locally advanced/metastatic ATC in May 2018.¹¹⁹ This was based on interim results of a multicenter phase II trial for BRAF V600E positive solid tumors that included an ATC cohort.⁴¹ Updated results presented at ESMO 2019 reported ORR of 67% (95% CI, 46%–84%).¹²⁰ PFS of 60 weeks (95% CI, 20 weeks-not estimable [NE]) and a median OS of 86 weeks (95% CI, 35 weeks to NE). The combination was well tolerated with common adverse events being fatigue (38%), fever (37%), and nausea (35%).¹²⁰ Since this initial study, further studies including some in the neoadjuvant setting have documented significant and rapid responses of BRAF-mutated ATC to BRAF/MEK inhibitors, such that BRAF/MEK inhibitor therapy is now considered an acceptable approach for patients with BRAF-mutated ATC.^15,121

7.4.2 ∣. Anti-angiogenic drugs

Lenvatinib, a multikinase inhibitor, is approved in Japan but not in the United States for ATC based on encouraging phase 2 results.^122,123 A larger study in the United States and Europe was closed early for futility (NCT02657369).¹²⁴ Lenvatinib should be used with caution as there have been reports of fatal bleeding and fistula formation.^125,126 Due to the invasiveness of ATC tumors, this risk is likely even higher than reported.

7.4.3 ∣. NTRK fusions

The US FDA approved larotrectinib in November 2018 for patients with metastatic/unresectable solid tumors harboring NTRK gene fusions identified by NGS or FISH based on three multicenter, open-label, singlearm studies which collectively included 55 patients across 12 cancer types.¹²⁷ ORR was 75% (95% CI: 61%, 85%), inclusive of 22% CR and 53% PR. Median DoR had not been reached at the time of data lock. However, response duration was ≥6 months for 73%, and ≥12 months for 39% of patients. An analysis of the patients with thyroid cancer included in these studies showed an ORR of 29% in ATC (n =7) with a median OS of 14 months. This ORR was far lower than DTC where the ORR was 90% and median OS was not reached.¹²⁸ Larotrectinib was tolerated well with fatigue and gastrointestinal issues as most common side effects. Entrectinib is another agent that has good tolerability profile and remarkable activity in NTRK fusion-positive solid tumors.¹²⁹ The efficacy in ATC of NTRK directed therapy is unclear at this time.

7.4.4 ∣. Tumor-agnostic FDA-approved targeted therapies for genetically altered solid tumors with applicability in ATC

In May 2017, US FDA approved pembrolizumab for unresectable or metastatic microsatellite instability-high (MSI-H) or mismatch repair deficient (dMMR) solid tumors.¹³⁰ Approval was based on results from five trials which enrolled 149 patients across 15 cancer types. Even though most common cancers were gastrointestinal and endometrial, there was 1 thyroid cancer (unspecified histology). ORR was 39.6% with majority maintaining response for ≥6 months. With multiple studies suggesting that some ATC are MMR deficient, it would be prudent to test all tumors for this biomarker.¹² However, in a pilot study evaluating tremelimumab and durvalumab with SBRT in metastatic ATC which included two patients with MSI-H tumors, no responses were observed.¹³¹ Pembrolizumab is now also approved for unresectable/metastatic solid tumors of any tissue origin with high tissue tumor-mutational burden (TMB-H; ≥10 mutations/megabase). This was based on promising results of Keynote-158 trial which was an open-label study that included several types of tumors including DTC.¹³² ATC however rarely has high TMB and this approval may not be of high relevance in this disease.¹³³ Efficacy and safety of checkpoint inhibitors (CPI) as monotherapy specifically for ATC was evaluated in a single arm clinical trial involving an anti-PD1 agent, spartalizumab.¹³⁴ Forty-two patients were enrolled in the trial. ORR was 19%, which included three patients with CR and five with PR. Response rates were higher in PD-L1-positive (8/28; 29%) versus PD-L1-negative (0/12; 0%) patients. The highest rate of response was observed in the patients with tumor PD-L1 ≥ 50% (6/17; 35%). Responses were seen in both BRAF-wild-type and BRAF-mutant patients though responses in BRAF-mutant population was much lower at 8% compared to 25% in the nonmutant group. Median PFS and OS amongst all patients was only 1.7 and 5.9 months, respectively. However, OS for PD-L1 positive patients was not reached at the time of reporting and 1-year survival was 52.1% in this population. Combining checkpoint inhibitors is currently being explored in some trials. Results of an investigator-initiated Phase II trial of ipilimumab and nivolumab in thyroid cancer was presented at ASCO 2020 by Lorch et al.¹³⁵ The trial included an exploratory ATC cohort of 10 patients, 3 of who had profound PR (30% RR, 95 CI: 7%−65%). Two of these patients had a sustained continued response at 26 and 13 months. Immune related side effects were similar to what has been reported with this combination in other diseases except for an unexpected number of treatment-related adrenal insufficiency (8%). Retrospective studies have demonstrated potential benefit of immunotherapy in combination with BRAF/MEK inhibitor.^121,136,137 Preliminary results of a prospective clinical trial with of the combination of PD-L1 inhibitor, atezolizumab with vemurafenib/cobimetinib for BRAF-mutated or cobimetinib only in those with NF1/2 or RAS-mutated ATC showed these regimens are effective with some patients having a dramatic response enabling subsequent complete tumor resections.¹³⁸ Prospective trials combing CPI with other agents are ongoing (NCT04238624; NCT03181100).

7.4.5 ∣. Other potential targets

ALK fusions are seen in aggressive thyroid cancers of various histologies including ATC. The most common fusion partner is striatin gene (STRN) and the kinase activity of STRN-ALK responds to inhibition by ALK inhibitors in vitro.¹³⁹ Case reports of impressive response of ALK-rearranged ATC to ALK-inhibitors such as crizotinib, ceritinib, and brigatinib have been published.^140,141 P1K3CA mutations are sometimes noted in ATC. While they can co-exist with BRAF V600E, they frequently occur in tumors without other actionable mutations. Phase II trial and case series have shown that everolimus, a PIK3CA/mTOR inhibitor results in sustained response in these patients.¹⁴² There are reports of ATC patients with mutations in the TSC1/TSC2 genes experiencing impressive durable response to everolimus.¹⁴³

Statement 20: Dabrafenib and trametinib combination therapy may be initiated for patients with resectable anaplastic thyroid carcinoma (ATC) and should be promptly initiated for patients with unresectable ATC upon identifying BRAF V600E mutation by IHC, PCR or tissue/blood-based NGS.

In summary, ATC is a highly lethal disease for which patients have traditionally had poor outcomes with surgery, chemotherapy and radiotherapy. Approval of dabrafenib and trametinib for treatment of BRAF V600E mutated ATC is the beginning of a new era of targeted treatment options for ATC. At this time, testing for BRAF V600E mutation as part of comprehensive NGS or mutation-specific molecular assays or IHC should be considered mandatory at diagnosis. With discovery of many other molecular targets, and emerging literature showcasing promise of matched targeted therapies, we recommend that all patients with ATC have comprehensive genomic profiling on tumor tissue through NGS.

8 ∣. CONCLUSION

The incidence of thyroid cancer is increasing, and a subset of patients will be affected by advanced disease. This paper defines the characteristics of advanced thyroid carcinoma and thereby identifies the patients who would benefit from somatic mutational testing of their tumors. All patients with ATC and most patients with incurable RAI-refractory DTC should undergo somatic mutational testing. NGS can reveal targetable mutations and potentially give patients affected by advanced thyroid carcinoma systemic treatment options that can prolong survival. These new innovative approaches are changing the landscape of clinical care for patients with advanced thyroid cancer.

Abbreviations:

AJCC

American Joint Committee on Cancer

ALK

anaplastic lymphoma kinase

ATM

ataxia-telangiectasia mutated

TTF-1

thyroid transcription factor 1

BRAF

rapidly accelerated fibrosarcoma gene, homolog B

CCNE

cell cycle protein cyclin E

CDKN

cyclin-dependent kinase inhibitor

IHC

immunohistochemistry

CR

complete response

DoR

duration of response

ESMO

European Society for Medical Oncology

FISH

fluorescence in situ hybridization

MEK

mitogen-activated protein kinase

mTOR

mammalian target of rapamycin

NCCN

National Comprehensive Cancer Network

NF

neurofibroma gene

NTRK

neurotrophic tropomyosin receptor kinase

ORR

objective response rates

OS

overall survival

PI3CA

phosphatidyl inositol 3-kinase catalytic subunit alpha

PFS

progression-free survival

PR

partial response

PTEN

Phosphatase and tensin

RAS

rat sarcoma virus oncogene

RB

retinoblastoma gene

RET

rearranged during transfection

SD

stable disease

STK11

serine/threonine kinase 11

SWI/SNF

switch/sucrose nonfermentable

PD

programmed death

TERT

telomerase reverse transcriptase

TKI

tyrosine kinase inhibitors

TP53

tumor protein 53 gene

TSC1/TSC2

tuberous sclerosis 1/tuberous sclerosis 2

US FDA

United States Food and Drug Administration

APPENDIX A: ADDITIONAL INFORMATION ON TESTING

A.1 ∣. NGS sequencing

Highly sensitive, NGS technologies enable high-throughput screening of numerous gene loci, using nano-quantities of tumor nucleic acid, and are being adopted into clinical practice to examine FFPE surgical samples, core biopsies, and even fine needle aspirates (FNA). Various commercially available genetic profiling platforms have been developed for the detection of mutations and/or gene rearrangements. The FDA has approved several NGS-based solid tumor profiling platforms, as well as liquid-based testing panels, including: FoundationOne^® CDx, FoundationOne^® Liquid CDx, Guardant360^® CDx, MSK-IMPACT, and Oncomine Dx Target Test. FoundationOne^® CDx is FDA-approved for testing all solid tumors and encompasses 324 genes. Tempus∣xT^® is another option that includes over 600 genes. Both companies also offer testing options for circulating tumor DNA (FoundationOne^®Liquid; Tempus∣xF^®), which are less comprehensive but still include the most common mutations associated with advanced thyroid carcinoma. ThyroSeq^® is another test developed specifically for evaluating thyroid cancer, and its most recent version ThyroSeq v3, released in 2017, is a 112-gene panel capable of probing >12 000 mutation hotspots and >150 gene fusions involved in thyroid cancer.¹⁵⁴

A.2 ∣. BRAF mutational testing

NGS-based genetic platforms provide comprehensive tumor information but require elaborate analyses that may result in treatment delays. Stand-alone BRAF testing may be performed using any clinically-validated method (direct sequencing with Sanger sequencing or pyrosequencing; mass-spectometry approaches; targeted polymerase chain reaction (PCR)-based methods such as real-time PCR, SNaPshot, peptide nucleic acid (PNA)-clamp amplification, etc). The cobas^® 4800 BRAF V600 Mutation Test (Roche), is a commercially available real-time PCR assay, approved by the FDA for the detection somatic BRAFV600E mutations, to help select melanoma patients for treatment with vemurafenib.

Immunohistochemistry (IHC) using mutation-specific antibodies against BRAF V600E (VE1 clone, Spring Bioscience, Pleasanton, CA) provides an alternative, inexpensive method for the rapid identification of BRAF V600E mutation-positive thyroid tumors.^155,156 This approach is reliable in detecting BRAF V600E in FFPE tumor tissue and in FNA cell blocks, but it is not recommended for FNA smears.¹¹⁵

A.3 ∣. Multi-step testing

Given the relatively low prevalence of RET and NTRK1/3 fusions in DTC, PDTC, and ATC, some laboratories may opt to only test tumors that are negative for mutations in more common (and mutually exclusive) oncogenic drivers, such as BRAF and RAS genes, with which RET and NTRK1/3 fusions are mutually exclusive.

RET IHC is not a viable pre-screening option.¹⁵⁷ Consensus guidelines by an international expert panel do not recommend IHC for NTRK fusion confirmation, but suggest that, in some cases, IHC may be used as a preliminary screening assay.¹⁵⁸ In a recent study evaluating NTRK fusion detection methods across a variety of solid tumors, DNA-based and RNA-based NGS assays were compared to IHC with a pan-Trk rabbit monoclonal antibody (clone EPR17341, Abcam, Cambridge, MA) that recognizes a conserved C-terminal epitope in the tyrosine kinase domains of TrkA, TrkB, and TrkC (the proteins encoded by the NTRK1, NTRK2, and NTRK3 genes, respectively).¹⁵⁹ Pan-Trk IHC showed 82% sensitivity and 100% specificity for thyroid carcinoma samples (13 NTRK fusion-positive and 27 NTRK fusion-negative tumors).¹⁵⁹ These results suggest that IHC may be an effective initial screening tool for NTRK fusions in advanced thyroid cancer.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.