Molecular Markers in the Diagnosis of Thyroid Cancer in Indeterminate Thyroid Nodules

Abstract

The Bethesda System for the Reporting of Thyroid Cytology recognises six diagnostic categories of thyroid nodule cytology with an incremental risk of malignancy. Although the Bethesda system created a much-needed handhold by standardising the cytological diagnosis and management of thyroid nodules worldwide, the system does not provide a clear answer to the heterogeneous group of nodules with indeterminate cytology. Improvement in the assessment of indeterminate fine-needle aspiration (FNA) results with molecular testing allows better risk stratification and reduces the need for diagnostic thyroid surgery. The molecular markers are classified as a “rule out” test, which has a high negative predictive value and helpful in cases with a low pre-test probability of cancer to rule out thyroid cancer. The “rule in” test has a high positive predictive value and helps in confirming malignancy in those with a high pre-test probability of cancer. This review summarises the commonly used molecular studies in thyroid FNAC aspirates and their current role in clinical practice.

Introduction

The Bethesda System for the Reporting of Thyroid Cytology recognises six diagnostic categories of thyroid nodule cytology with an incremental risk of malignancy (Table 1) [1]. Although the Bethesda system created a much-needed handhold by standardising the cytological diagnosis and management of thyroid nodules worldwide, the system does not provide a clear answer to the heterogeneous group of nodules with indeterminate cytology [2]. The Bethesda category III of “Atypia of Undetermined Significance (AUS) or Follicular Lesion of Undetermined Significance (FLUS)” could be either “architectural atypia” or “nuclear atypia” or “preparatory artefacts related atypia.” In the Bethesda category IV (follicular neoplasm (FN)) cytology does not show the capsular and/or vascular invasion that distinguishes an FTC from a benign FA. It is the duty of the clinician to discuss with the pathologist to ascertain the reason for the categorisation of the cytology in the indeterminate category. Depending on the sonographic stratification as well as pre-test probability of thyroid cancer, the clinician can either decide to get a second opinion from an expert thyroid cytopatholgist, re-do the FNAC after a period of 3–6 months, send the patient for a diagnostic hemithyroidectomy, or use one of the molecular marker panel for aiding decision making [3]. The category of indeterminate nodule largely corresponds to histopathologically follicular-patterned lesions, both benign and malignant, including follicular adenoma (FA), non-invasive follicular thyroid neoplasm with papillary-like nuclear features (NIFTP), (encapsulated) follicular variant of papillary thyroid carcinoma (FVPTC or EFVPTC), and follicular thyroid carcinoma (FTC). These neoplasms are particularly difficult to differentiate on fine-needle aspiration cytology (FNAC). In the case of FTC, cytology does not show the capsular and/or vascular invasion that distinguishes an FTC from a benign FA. In FVPTC, the growth pattern is follicular but nuclear features of PTC cannot be identified cytologically [4]. The risk of malignancy with an indeterminate thyroid cytology ranges from 14 to 48% [5, 6]. However in tertiary care cancer centres the risk of cancer can reach up to 52% [7]. Majority of patients with a cytologic result of FN or FLUS/AUS confirmed on repeat aspiration undergo diagnostic thyroidectomies. If the final pathology was a benign nodule, then the patient had undergone an unnecessary surgery, and if it was thyroid cancer, then the patient may need to undergo a completion thyroidectomy. An additional preoperative test for thyroid nodules with indeterminate cytology could prevent diagnostic hemi-thyroidectomies for benign nodules and limit the number of two-stage surgeries for thyroid malignancies. With the doubling of incidence of thyroid cancer in the past two decades and the increasing detection of thyroid nodules on medical imaging, the need for a more accurate diagnostic procedure has grown [8, 9]. Improvement in the assessment of indeterminate fine-needle aspiration (FNA) results with molecular testing allows better risk stratification and reduces the need for diagnostic thyroid surgery. The molecular markers are classified as a “rule out” test, which has a high negative predictive value and helpful in cases with a low pre-test probability of cancer to rule out thyroid cancer. The “rule in” test has a high positive predictive value and helps in confirming malignancy in those with a high pre-test probability of cancer. The predictive capacity of these tests is largely based on the prevalence of thyroid cancer in one’s practice and the referral base of their institution.

Table 1.

Bethesda system diagnostic categories for reporting thyroid cytopathology

The Bethesda System for Reporting Thyroid Cytopathology: Implied risk of malignancy and recommended Clinical management
Diagnostic category	Risk of malignancy (%)	Usual management
Non-diagnostic category		Repeat FNA with ultrasound guidance
Benign	0–3	Clinical follow-up
Atypia of Undetermined Significance or follicular lesion of undetermined significance (AUS/FLUS)	~~ 5–15	Repeat FNA
Follicular neoplasm or suspicious for a follicular neoplasm (specify if Hürthle type or oncocytic)	15–30	Surgical lobectomy
Suspicious for malignancy	60–75	Near-total thyroidectomy or surgical lobectomy
Malignant	97–99	Near-total thyroidectomy

Summary of Molecular Markers

There has been considerable progress in the understanding of thyroid carcinogenesis and the emergence of numerous molecular markers in the recent years with potential to be used in the diagnostic algorithm of these nodules [10, 11]. The molecular markers on the FNAC aspirates use one of the techniques below:

1. Identification of a particular mutation/rearrangement/gene fusion or a panel of commonly occurring variations in thyroid cancer

2. Use of high-density genomic data for molecular classification (a genomic sequencing classifier)

3. Use of micro-RNA (miRNA) classifier

Mutational Analysis

The most common molecular markers studied are the somatic BRAF and RAS point mutations, and RET/PTC and PAX8-PPRɣ rearrangement, predominantly involving the mitogen-activated protein kinase (MAPK) signalling pathway or the PI3-AKT pathway [12].

• BRAF mutation is one of the most extensively studied mutation in thyroid cancer [13]. A point mutation (T1799A) causes V600E amino acid substitution in the BRAF protein, which is one of the common mutations that constitutively activate serine–threonine kinase. The frequency of this point mutation can be as high as 90% but is seen on an average in 45% of papillary thyroid cancer (PTC) [12]. BRAF mutation is not seen in benign thyroid nodules. BRAF mutation has been associated with increased aggressiveness of the thyroid cancer and nodal metastasis [14, 15].

• RAS mutations are the second most common mutations in DTCs [12]. RAS in active state is bound to GTP and has intrinsic GTPase activity that converts GTP to GDP thus inactivating it. Mutation results in loss of this GTPase activity leading to its constitutive activation. There are 3 isoforms—HRAS, KRAS, and NRAS; the most common mutations in thyroid cancers are in NRAS. The RAS mutations have a predominant role in follicular patterned neoplasm including follicular thyroid carcinoma (FTC), wherein it promotes invasiveness and metastases. However, this appears to be an early mutational event having been identified in a portion of follicular adenomas also [16].

• The RET proto-oncogene encodes a cell membrane receptor tyrosine kinase. RET is highly expressed in para-follicular C cells. It is usually not expressed in follicular cells, but it can be activated by chromosomal rearrangement: the RET/PTC translocation. This occurs due to genetic recombination between 3′ tyrosine kinase of RET and 5′ portion of a partner gene. The translocation constitutively activates tyrosine kinase activity of RET. RET/PTC activates both the MAPK and PI3-AKT pathways [12]. There are more than 10 types of this translocation, and the most common are RET/PTC1 and RET/PTC3 [17, 18]. This translocation is seen in PTC related to radiation exposure.

• PAX8/PPARγ rearrangement is caused by (2;3)(q13;p25) translocation that leads to fusion between the PAX8 gene and the peroxisome proliferator-activated receptor-γ (PPARγ) gene [19]. PAX8/PPARγ has an inactivating effect on the wild-type tumour suppressor PPARγ and also transactivates certain PAX8 responsive genes. This translocation occurs in about 30–60% of FTC [20, 21] and also in 38% of follicular variant of papillary thyroid cancer (FVPTC) [21].

The early mutational analysis assays tested for common gene mutations associated with thyroid cancer, BRAF, RAS, RET/PTC, and PAX8/PPAR gamma missed malignancy in 14% of cases under the FN category [22]. The ThyroSeq v2, with additional point mutations and gene fusions had a negative predictive value for malignancy of 96–97% and a positive predictive value of 77–83% for cytology showing AUS/FLUS or FN [23]. The third version of this assay, ThyroSeq v3, tests for 112 thyroid cancer-related genes and has an improved sensitivity of 91–97%, specificity of 75–85%, negative predictive value of 97–98%, and positive predictive value 64–68% for FLUS/AUS and FN cases [24].

We prospectively studied FNAC samples from 69 patients who presented with palpable thyroid nodules and used a next-generation sequencing (NGS) panel to query multiple variants in the DNA and RNA. The panel detected point mutations in codons 12, 13, 61, and 146 of HRAS, KRAS, NRAS, and codon 600 in BRAF in the DNA fraction, and RET/PTC1, RET/PTC2, and RET/PTC3 and PAX8-PPAR훾 rearrangements were detected from the RNA fraction. The panel showed an overall sensitivity of 81.25% and a specificity of 100%, while in the case of the categories of Bethesda III, IV, and V the sensitivity was higher (87.5%), and the specificity was established at 100% [25]. Studies of molecular markers in both FNAC aspirate and surgical specimens from India are summarised in Table 2 [26–35].

Table 2.

Molecular markers studied in thyroid FNAC aspirates and thyroid cancer specimens from India

Authors	Number of samples	Molecular marker studied	Test performance
Aron et al., Delhi [26]	70	Galectin-3	The smears stained positive for Galectin-3 in 80% of PTCs, 37.5% of FNs, and in 60% of benign nodules.
Choudhury et al., Delhi [27]	25	p53 and Ki-67	Ki-67 scores ranged from 2.2 to 13 in the DTCs while ranged 0.5–2 in benign nodules
Mehrotra et al., Lucknow [28]	123	Ki-67 and AgNORs	Mean AgNOR counts and Ki-67 labelling index were consistently higher in Follicular cancers compared to adenomas and hyper plastic nodules
Hemalatha et al., Vellore [29]	277	BRAF V600E	27.2% of FNA samples were positive for mutations; only 1 sample of those labelled as AUS was positive for BRAF; BRAF positive PTCs had higher rates of lymph node metastasis
George N et al., Lucknow [30]	109* PTC	BRAF V600E, RAS, RET/PTC	BRAF V600E noted in 51.38%, NRAS in 7.34% of all PTCs. No RET/PTC rearrangements were observed
Ahmad F et al., Mumbai [31]	95* Thyroid tumours	BRAF V600E	38% were BRAF positive
George N et al., Lucknow [32]	30* FVPTC	BRAF and RAS and NIS expression	BRAF mutation was observed in 62.5%; No NRAS mutation was found. Sodium iodide symporter (NIS) expressions were down-regulated in invasive and infiltrative/diffuse FVPTC but not in NIFTP.
Krishnamurthy A et al., Chennai [33]	79* DTC	BRAF V600E	31% of the DTCs were positive for BRAF
Nair CG et al., Kochi [34]	59* PTC	BRAF V600E	51% harboured BRAF V600E mutation, but the mutation status was not associated with aggressive tumour factors and adverse outcome.
Chakraborty A et al., Mumbai [35]	140* Thyroid tumours	BRAF V600E	BRAF V600E in 53.4% of PTC. Classic PTC (61%) FVPTC (11.7%). Significant correlation between BRAF mutation status and extra-thyroidal invasion, lymph node metastasis, and tumour stage.

*Studies on thyroid cancer specimens

mRNA Genomic Sequencing Classifier

The first version of the gene expression Afirma classifier measured the activity level of 167 genes within the thyroid nodule, and this had a negative predictive value for malignancy of 95 and 94% for samples showing FLUS/AUS and FN, respectively, but the positive predictive value was only 38 and 37%, respectively [36–38]. However, using the test in patients with indeterminate thyroid cytology would save over 60% of patients from diagnostic thyroid surgery, which would result in an overall lower cost of care [39]. The second version of Afirma is a genomic sequencing classifier (GSC) that adds a cascade of classifiers utilising several thousand genes to better discriminate Hürthle cell neoplasms from non-neoplastic Hürthle cell lesions, as well as classifiers that identify medullary thyroid cancer, parathyroid lesions, and BRAF V600E mutations. Afirma has introduced an Xpression Atlas as an add-on test, available for GSC-suspicious and Bethesda V and VI nodules, which assesses 761 DNA variants and 130 RNA fusions in 500 genes.

MicroRNAs (miRNA) Gene Expression Combined with Mutational Analysis

The miRNA classifier is a multi-platform test based on the expression level of 10 miRNA genes and mutational analysis to detect the presence of eight oncogenes (ThyraMIR and ThyGenX). In a study of 109 indeterminate nodules, 54 of which were resected and 35 of which were malignant, the combined test had a negative predictive value of 97 and 91% for FLUS/AUS and FN, respectively, and a positive predictive value for malignancy of 68 and 82%, respectively [40].

Interpretation of these Tests and Clinical Decision Making

1. Benign molecular pattern—For patients with indeterminate FNA cytology who have a benign molecular pattern (no mutations on mutational analysis or a benign GSC or miRNA gene expression classifier result), observation rather than diagnostic lobectomy could be preferred. The negative predictive value for malignancy with these techniques (94 to 96%) seems to be sufficiently high to warrant observation over diagnostic surgery [36, 37]. Repeat ultrasound is performed in 12 to 24 months to assess stability. However, the decision to observe a patient with a benign profile should be reassessed as more data become available.

• 2.Suspicious pattern—Patients with point mutations in genes that are strongly associated with thyroid cancer (e.g., BRAF, TERT, and RET/PTC) require thyroid surgery. The choice of lobectomy versus total thyroidectomy may depend on tumour size, the specific mutation, patient preference, and the presence of nodes or contralateral nodules on ultrasound.

Limitations of Molecular Testing

It is important to keep in mind that the validation studies for all molecular methods involve very small numbers of malignant nodules [41] and they are quite expensive. Their precise role in routine clinical practice continues to be a contentious issue. Majority of the studies in this context are retrospective, and impact of these mutations is not independent of other prognostic factors making the interpretation difficult. Whether some of the mutational testing offer newer insights more than gleaned by the clinical or ultrasound characteristics is not clear and long-term data utilising molecular testing are still lacking. The biggest challenge with all the molecular marker testing is that their performance depends on the prevalence of cancer in the setting in which it is studied. The negative and positive predictive values noted above are highly dependent on the prevalence of thyroid cancer in any clinician’s cytologically indeterminate nodule population, with the negative predictive value decreasing as the cancer prevalence increases. Unless the prevalence of malignancy in indeterminate nodules is known for each individual institution or clinic, it is difficult to estimate the negative and positive predictive values of some of these molecular tests. Before utilising these expensive tests, one should also consider the sonographic characteristics and the size of the nodule, the degree of patient concern, and the availability of follow-up imaging, as well as the patient’s candidacy for surgery.

Conclusion

The role of molecular markers in the management of patients with thyroid nodules and cancer is evolving. While these molecular markers can guide clinical decision making they are unlikely to replace clinical judgement. We should stay tuned for further updates.