Advances and practical use of the molecular markers for thyroid cancer

Abstract

Background

Thyroid nodules are very common; however, only 5–15% of nodules are malignant. Although most of the malignant thyroid cancers can be identified pre-operatively by cytological analysis, approximately 20–30% of these nodules are classified indeterminate. Subsequently, repeated Fine Needle Aspiration (FNA) or thyroid surgery may be required for a more definitive diagnosis. In the United States, the incidence of thyroid cancer has tripled from 1975 with an incidence rate of 14.3 per 100,000 individuals, primarily due to an increased diagnosis of papillary thyroid cancer which increased by 9.1 per 100,000 (RR 3.7, 95% CI (3.4–4.0)); however, mortality has remained stable. Many patients with cytologically indeterminate nodules undergo unnecessary diagnostic surgeries, placing them at risk of potential surgical complications and need for lifelong levothyroxine replacement. The advent of molecular gene testing has drawn significant attention toward improved stratification of these indeterminate lesions into benign or malignant. Detection of any pertinant molecular mutations in the indeterminate categories- atypia of undetermined significance (AUS)/ follicular lesion of undetermined significance (FLUS), follicular neoplasm (FN)/suspicious for a follicular neoplasm (SFN), and suspicious for malignant cells (SMC) has been shown to confer the risk of histologic malignancy of 88, 87 and 95% respectively, while a negative test for the mutation was reassuring with a very high negative predictive value over 95%.

Purpose

The purpose of this article is to review the current recommendations to implement molecular genetic testing for thyroid nodules. We review the most common gene mutations harbored in the various thyroid cancers, including BRAF, RAS, and RET/PTC. We will also review the more recently discovered gene mutations for thyroid cancer, including TERT mutations. Finally, we will discuss the practice guidelines to implement molecular gene testing for the indeterminate cytology of thyroid nodules as well as the new category of non-invasive follicular thyroid neoplasm with papillary-like nuclear features as an intermediate pathology which will not require aggressive therapy after diagnostic lobectomy.

In Context

Thyroid nodules are very common; however, rate of malignancy is low, with only 5–15% of nodules identified as malignant. Fine needle aspiration (FNA) may in 20–30% cases identify an “indeterminate” cytology which leads to repeat FNA or thyroid surgery for a more definitive diagnosis. In many cases, these diagnostic surgeries prove to be unnecessary placing patients at risk of surgical complications and need for lifelong levothyroxine replacement. Molecular testing helps to stratify these indeterminate lesions into benign or malignant. Detection of specific molecular mutations in the indeterminate categories is used as a “rule-out” or “rule-in” for thyroid malignancy. This article summarizes the most common gene mutations for thyroid cancer. We also review the current recommendations for implementing molecular gene testing for thyroid nodules particularly in the indeterminate category and summarize the methods, strengths, limitations, and pre-and post-test probabilities of each test. We also review the newest classification of thyroid neoplasm called non-invasive follicular thyroid neoplasm with papillary-like nuclear features (NIFTP) which was intended to reduce overdiagnosis and treatment of thyroid cancer.

• Thyroid cancer
• Thyroid nodule
• Molecular markers
• fine needle aspiration
• indeterminate cytology
• mutation analysis
• genetic testing
• Non invasive follicular thyroid neoplasm with papillary-like nuclear features (NIFTP)

Thyroid cancer is the most common type of cancer among endocrine tumors. There is an estimated incidence of 14.3 per 100,000 individuals per year in the United States (1). The incidence of thyroid cancer, both in the United States and worldwide, has increased over the last 4 decades. Diagnosis of thyroid cancer is usually obtained through ultrasound examination and fine-needle aspiration biopsy of suspicious nodules. American Association of Clinical Endocrinologists (AACE) and American Thyroid Association (ATA) have created guidelines to guide clinicians in choosing which nodules require biopsy (2, 3). Currently, cytological examination of cells collected by FNA is the most reliable method for evaluating thyroid nodules (3, 4). In 2009, the Bethesda reporting system for classifying thyroid cytology was proposed by the National Cancer Institute in Bethesda, MD, USA and provides diagnostic categories with accompanying risk stratification (4). According to the data from the National Cancer Institute in Bethesda, MD, USA, thyroid FNAs in the benign category have a 0–3% risk of malignancy, whereas thyroid FNAs in the malignant category have a 97–99% risk of malignancy. However, 20–30% of thyroid FNA specimens fall into the category of indeterminate. The indeterminate category can be divided into three subcategories: 1) Atypia of undetermined significance/follicular lesion of undetermined significance (AUS/FLUS, category III), 2) follicular neoplasm/suspicious for follicular neoplasm (FN/SFN, category IV), and 3) suspicious for malignancy (SMC, category V) (4). The predicted probability of malignancy in each of these categories is 5–15%, 15–30%, and 60–75%, respectively (4). Of the indeterminate thyroid nodules that are surgically resected, 15–30% nodules are confirmed to be malignant, although the incidence varies at different institutions. Subsequently, the majority of diagnostic surgeries are performed on benign nodules. The cost benefit of surgery versus molecular testing needs to be assessed. However, molecular testing is advantageous compared with unnecessary surgery in terms of morbidity because any surgical intervention is associated with risks, even in the hands of the most experienced thyroid surgeons.

Recommendations for implementation of molecular testing for evaluation of thyroid nodules

When assessing the clinical usefulness of molecular testing, it is important to recognize that even the gold standard of thyroid cytologic analysis is limited by variations in interpretation and hence there is a role for molecular markers to increase the diagnostic sensitivity for malignancy (5). The physician and patient ought to be aware that because molecular testing is in its infancy, continued research is required to validate all of these tests (5). Nonetheless, several large, prospective, multicenter studies have independently validated the utility of testing for markers of thyroid cancer with reproducible results (6–9). The current molecular markers that have received the most attention in thyroid cancer include BRAF, NRAS, KRAS, HRAS point mutations, and PAX8/PPARG and RET/PTC rearrangements, with the aim to identify (rule-in) thyroid cancer (10). These studies have found that a positive result carries a 77% positive predictive value (PPV) and 97% negative predictive value (NPV) for predicting thyroid cancer, although clinical validation has not been performed for nodules that are less than 1 cm (6, 11) (12). More recent studies have focused on telomerase reverse transcriptase promoter mutations on chromosome five (TERT mutations) which is seen at a higher prevalence in aggressive thyroid cancers (13).

Presently, the role of molecular testing for evaluation of thyroid cancer is mainly reserved for indeterminate nodules to help stratify these nodules and prevent unnecessary surgery or to prevent unnecessary repeated FNA if surgery is warranted.

Review of molecular mutations in thyroid cancer

The BRAF V600E point mutation is the most common genetic mutation detected in patients with papillary thyroid cancer (PTC) and occurs in approximately 45–60% of patients (14, 15). The BRAF V600E mutation results in an amino acid substitution at position 600 in BRAF, from a valine to a glutamic acid, with an increase in kinase activity (16). BRAF is located on chromosome 7 and is the most potent activator of the mitogen-activated protein kinases (MAPK) pathway among the three forms of the RAF kinases (16). BRAF mutations seem to be involved in tumor initiation as they are found in microscopic PTC (10). Prevalence of BRAF V600E mutation is much higher in classical PTC (45%) than follicular variant of PTC (20%) and is rarely found in follicular thyroid cancer (1.4%) (17). For tall cell variant of PTC, it has been reported as 50% (17). The BRAF V600E mutation is associated with an aggressive tumor phenotype and higher risk of recurrent and persistent disease in patients with conventional PTC (14, 15). Coexisting BRAF V600E and TERT promoter mutations have a marked synergistic effect on the aggressiveness of PTC, including a sharply increased tumor recurrence and patient mortality (13, 18) (19).

Gene expression profiling of PTC with RET/PTC, RAS, and BRAF mutations accurately classifies the mutational status of the cancers, further supporting that each of these oncogenes induces specific phenotypic expression (16, 20). Tumors associated with RET/PTC1 rearrangements have the conventional PTC histology. RET/PTC3 rearrangements are seen with solid variant sporadic PTC (21). Follicular variant PTCs (FVPTCs) harbor RAS mutations or PAX8/PPARG rearrangements and less commonly distinct BRAF mutations and have comparatively low frequency of lymph node metastases (10, 16) (22). Most PTCs with BRAF mutation have a classic histology, approximately 80% of the tall cell variant PTCs have the BRAF T1799A mutation; this histologic subtype is believed to be present more often in advanced disease (14, 23). BRAF mutations also have a higher recurrence rate, and metastatic recurrences have diminished radio-iodine avidity (10, 15). Thus, alternate therapies would have to be considered for aggressive metastatic PTCs. The fact that PTC is associated with alterations to MAPK pathway offers medical therapy that can be targeted to this pathway (10).

Overview of commercially available genetic tests for thyroid cancer

Afirma gene testing

Afirma gene expression classifier is based on 167 gene expression profiles of surgically proven benign and malignant thyroid nodules and evaluates for the presence of a benign gene expression profile (6, 24; Table 1). Based on validation studies demonstrating high NPV among nodules with cytology of AUS/FLUS or FN/SFN, it has been designated as a ‘rule-out’ test to identify nodules that are benign. For nodules in the AUS/FLUS and FN/SFN categories, Afirma demonstrated a high NPV of 95 and 94%, respectively (6, 24).

Table 1. Overview of four commercially available genetic tests for indeterminate thyroid cytology

Test	Afirma	ThyroSeq verion 2 next generation sequencing	ThyGenX/ThyraMIR	Rosetta GXReveal
Company	Veracyte	University of Pittsburgh Medical Center through CBLPath	Interspace Diagnostics	Rosetta Genomics
Methods	Based on mRNA (gene expression); classified as benign or malignant	Next-generation sequencing which detects 14 gene mutations (AKT1, BRAF, CTNNB1,GNAS, HRAS, KRAS, NRAS,PIK3CA, PTEN, RET, TP53, SHR,TERT, and EIF1AX, and >42 gene rearrangements (RET-PTC, PAX-8/PPARG, etc.)	ThyGenX: Uses PCR for detection of 7 common gene mutations (BRAF, RAS, NRAS, and KRAS) and rearrangements (RET-PTC1, RET-PTC3, and PAX8-PPARG) first with sequence-specific probes. If initial mutation panel is negative then further testing for 10-gene microRNA classifier (ThyraMIR: expression analysis) done Classified as ‘or ‘positive’	microRNA-based diagnostic assay First thyroid test that works on stained FNA smears containing at least 60 cells for analysis
Strengths	High NPV: AUS/FLUS− 95% FN/SFN− 94% ‘Rule-out’ test Identifies MTC with MTC gene classifier	High PPV 68–72% ‘Rule-in test’ High NPV 96% ‘Rule-out’ assay Risk stratification based on mutation and prognostic value	High NPV 94% High PPV 74%	High NPV 91% Rule-out test
Limitations	Low PPV AUS/FLUS− 38% FN/SFN− 37% High false-positive results	Newer assay with limited experience	New assay with limited experience	Low PPV 59%

PPV, positive predictive value; NPV, negative predictive value.

However, it has a low PPV of only 38% for AUS/FLUS and 37% for FN/SFN which renders high false-positive results, and its usefulness would depend on local prevalence of the thyroid cancer (24). It also carries high false-positive results for Hurthle cell lesions. Afirma test can also be used to identify medullary thyroid cancer by using MTC gene expression classifier (24).

ThyroSeq verion 2 next generation sequencing

Assessment of 14 gene mutations and over 42 gene rearrangements has recently been available commercially as ThyroSeq v2 which can detect BRAF and RAS point mutations as well as common rearrangements of RET/PTC and PAX8/PPARG (25). It is estimated that one of these mutations is present in approximately 70% of well differentiated thyroid cancers. ThyroSeq v2 oncogene panel assay is considered a ‘rule-in’ assay since nodules harboring these mutations have a high likelihood of cancer given the test's high PPV of 68–87% (24). It is also useful as a ‘rule-out’ assay since the negative mutation result carries a very high NPV of 96% (24). ThyroSeq v2 also includes six gene expressions for medullary thyroid cancer and parathyroid tissues.

ThyGenX 7-gene mutation analysis panel and 10-gene miRInform classifier

This test can help reduce the cost by half by first testing for the common seven gene mutations, and if the initial mutation panel is negative, further testing for 10-gene miRInform classifier (24). It is useful for the combined AUS/FLUS and FN/SFN lesions with high NPV of 94% and PPV of 74%.

Rosetta GXReveal miRNA

This new molecular test has the benefit of avoiding an extra pass or repeat FNA biopsy as the extraction can be performed by using the existing FNA slides containing at least 60 cells for the analysis. It has high NPV of 91% but low PPV of 59%, and therefore can be used as a rule-out test (26).

Clinical practice guidelines (ATA vs. AACE)

For a nodule with indeterminate cytology (AUS/FLUS) result, ATA recommends molecular testing (strong recommendation, low-quality evidence) after counseling the patient regarding benefits and limitations of testing given there are no long-term outcomes data on the clinical and therapeutic implications of results (2).

The recently updated AACE 2016 guidelines recommend molecular testing to complement, not replace cytologic evaluation (Grade A) (3). Molecular testing may be considered for cytologically indeterminate nodules. Testing for detection of BRAF, RET/PTC, PAX8/PPRG, and RAS mutations (Grade B) are recommended. Due to insufficient evidence and limited follow-up studies, use of gene expression classifiers for cytologically indeterminate nodules (Grade B) are neither recommended nor rejected.

Non-invasive follicular thyroid neoplasm with papillary-like nuclear features

The two common histological subtypes of PTC, namely classic PTC and Follicular Variant PTC (FVPTC) are well known. Among FVPTC, two major subtypes are described: infiltrative and encapsulated. Infiltrative FVPTC resembles classic PTC, while encapsulated FVPTC (EFVPTC) when invasive behaves like follicular carcinoma (27).

There is a non-invasive variant of EFVPTC which lacked capsular/vascular invasion but due to nuclear features was previously diagnosed as a type of cancer. A multi-disciplinary team reviewed clinical and cytopathologic data of 268 tumors categorized as EFVPTC. This was a retrospective study of patients with thyroid nodules diagnosed as EFVPTC, including 109 patients with non-invasive EFVPTC observed for 10–26 years and 101 patients with invasive EFVPTC observed for 1–18 years (27). At median follow-up of 13 years, clinical outcomes showed no evidence of disease in the group with non-invasive EFVPTC and molecular analysis showed they did not harbor BRAF mutations seen in classic PTC (27, 28). This data led to a reclassification of this subset of tumors as NIFTP. This reclassification was intended to reduce overtreatment and to ameliorate psychological impact of a cancer diagnosis (27, 28).

This classification will substantially reduce thyroid cancer prevalence and reduce overtreatment of indolent tumors. With change in terminology and not considering NIFTP as a cancer, the rate of malignancy in Bethesda diagnostic categories would decrease substantially. The expected rate of decrease differs among authors, with Strickland et al. (29) predicting in AUS/FLUS, FN/SN, and SM 17, 8, and 41% reductions in risk of malignancy while Faquin et al. (30) predicts 13, 15, and 23% reductions in the same categories. Follow-up studies will be required to re-assess pretest and posttest probabilities of molecular testing with incorporation of this new classification.

Conclusion

Thyroid molecular testing can be a useful tool in conjunction with clinical and ultrasound risk assessment of thyroid nodules, especially for patients with indeterminate thyroid cytology (Bethesda category III, IV, and V) to reduce the need for diagnostic surgery and guide further management of the nodules. However, there is no need for this costly testing if the ultrasound and/or FNA showed entirely benign or highly suspicious cytology for malignancy.

Among differentiated thyroid cancer, identification of aggressive mutation phenotypes, for example, coexistence of BRAF or RAS with the TERT mutations would warrant more aggressive treatment and close surveillance due to very poor prognosis.

Future classification of thyroid cancer will likely incorporate the genetic mutation information to help clarify the subtypes and stage of the tumor, that is, benign, intermediate category such as NIFTP, or cancer, and also to inform us about the cancer risk and prognosis of the thyroid nodules.

Conflict of interest and funding

The authors have not received any funding or benefits from industry or elsewhere to conduct this study.