Review Article

Molecular Testing for Thyroid Nodules:

Review and Current State
Mara Y. Roth, MD1; Robert L. Witt, MD2,3; and David L. Steward, MD 4

Thyroid nodules affect nearly two-thirds of the world population. Fine-needle biopsy with cytologic evaluation remains the diagnostic
test of choice to distinguish benign from malignant thyroid nodules yet fails to discriminate as benign or malignant in up to one-third of
cases. This review discusses the limitation of current cytopathologic evaluation, the development of thyroid molecular testing, and the
strengths and limitations of commercially available tests. Initial cytomolecular testing sought to identify specific gene mutations associ-
ated with thyroid cancer. Although the presence of a mutation was strongly associated with cancer, the likelihood of identifying a muta-
tion was low; therefore, the test had low sensitivity. Subsequent tests developed have sought to improve the accuracy of cytomolecular
testing for thyroid fine-needle aspirations, both to reassure patients and providers when malignancy may be absent and to confirm the
malignancy when present. The development of cytomolecular testing for thyroid nodules has informed and improved current under-
standing of thyroid nodule formation and progression. When used appropriately and with clear understanding of the advantages and
disadvantages, cytomolecular testing has the potential to improve patient care in the setting of indeterminate thyroid nodules by help-
ing to guide both the need for and the extent of thyroid surgery. Cancer 2018;124:888-98. V C 2017 American Cancer Society.

KEYWORDS: cytology, diagnosis, molecular testing, thyroid cancer, thyroid nodules.

INTRODUCTION
Thyroid nodules are a common clinical problem worldwide, with studies suggesting that nearly two-thirds of the popula-
tion harbor thyroid nodules when evaluated by ultrasound.1 Although fine-needle aspiration (FNA) remains the diagnos-
tic test of choice to distinguish benign from malignant thyroid nodules, cytology alone fails to classify thyroid nodules in
15% to 30% of cases.2 Molecular testing for thyroid nodules has evolved rapidly over the past decade both to help improve
the diagnostic accuracy of thyroid cytology for indeterminate cases and potentially to guide the extent of surgery as initial
therapy for suspected thyroid malignancies. This report briefly reviews the development and current landscape of thyroid
molecular testing in the management of thyroid nodules.
The objective, as thyroid molecular testing initially developed, was to reduce unnecessary diagnostic thyroid surgery
for indeterminate thyroid nodules. The potential future direction is to help define prognosis and determine whether cyto-
molecular testing can reduce overtreatment of patients with low-risk malignancy while informing effective, tailored treat-
ment for those with higher risk thyroid malignancies. This prognostic information has the potential to inform both
surgical decisions and ongoing therapeutic decisions regarding the role of radioactive iodine therapy, and even the selec-
tion of targeted chemotherapy when necessary.
Although cytology is 1 of many clinical factors that inform decision making with regard to the management of thy-
roid nodules, additional risk factors to consider include: family history of thyroid malignancy, history of radiation expo-
sure, nodule size, sonographic risk assessment, patient symptoms, and thyroid function.3 Diagnostic test accuracy must
also be considered, with evaluation of test sensitivity, specificity, and the underlying disease prevalence in the population.
Bayes’ theorem informs us that disease prevalence impacts the predictive value of a test (Fig. 1). On the basis of given sensi-
tivity and specificity, when the prevalence of disease increases, the positive predictive value (PPV) goes up, whereas nega-
tive predictive value (NPV) goes down, and vice versa. Diagnostic tests with a high sensitivity and high NPV are
inherently good tests to “rule out” the presence of disease, depending on the disease prevalence within the population, sug-
gesting that a negative test result has high accuracy (approximately 95%) to reassure patients and providers that cancer is
not present in the thyroid nodule evaluated. Diagnostic tests with a high specificity and high PPV are good to “rule in”

Corresponding author: David L. Steward, MD, University of Cincinnati College of Medicine, 231 Albert Sabin Way, ML 0528, Cincinnati, OH 45267-0528; Fax:
(513) 558-5203; david.steward@uc.edu
1
Division of Metabolism, Endocrinology, and Nutrition, Department of Medicine, University of Washington, Seattle, Washington; 2Department of Otolaryngology,
Thomas Jefferson University, Philadelphia, Pennsylvania; 3Multidisciplinary Head and Neck Clinic, Helen F. Graham Cancer Center, Newark, Delaware; 4Department
of Otolaryngology, University of Cincinnati, Cincinnati, Ohio.

DOI: 10.1002/cncr.30708, Received: October 24, 2016; Revised: March 2, 2017; Accepted: March 6, 2017, Published online December 26, 2017 in Wiley Online
Library (wileyonlinelibrary.com)

888 Cancer March 1, 2018


Figure 1. Positive predictive value (PPV) and negative predictive value (NPV) are impacted by the prevalence (Prev) of cancer.
Sens indicates sensitivity; Spec, specificity.
disease, suggesting that a positive test result has high accu-
racy to confirm that a nodule is indeed cancerous, similar
to the risk of malignancy on histopathology observed on
cytology results that are consistent with malignancy.
However, the accuracy of the test must be interpreted
with regard to the disease prevalence within the popula-
tion evaluated. An ideal test would provide high sensitiv-
ity and high specificity, indicating high accuracy at
distinguishing benign from malignant disease; however,
as a general rule, increasing sensitivity tends to reduce
specificity, and vice versa.
LIMITATIONS OF THYROID CYTOLOGIC
AND HISTOPATHOLOGIC EVALUATION
Ultrasound and ultrasound-guided FNA has long been the
primary method of evaluating thyroid nodules for malig-
nancy.4 Although it is highly accurate for the diagnosis of
benign colloid nodules (the most common finding) and
classic papillary thyroid carcinoma (PTC) (the most com-
mon thyroid malignancy), thyroid FNA is much less accu-
rate for the diagnosis of follicular variant of PTC.
Furthermore, the ability to distinguish between follicular
(or Hurthle cell) adenomas and carcinomas requires histo-
logic assessment for capsule or vascular invasion and thus
cannot be diagnosed with cytology alone. In addition, sev-
eral benign conditions pose cytologic challenges, such as
hyperplastic adenomatoid nodules, and inflammatory
conditions, such as lymphocytic (Hashimoto) or granulo-
matous thyroiditis. In many cases, thyroid FNA results
Cancer March 1, 2018
Molecular Testing for Thyroid Nodules/Roth et al
are cytologically “indeterminate” and historically have
required diagnostic thyroid surgery, primarily hemithyroi-
dectomy/lobectomy, for a definitive diagnosis.
To address high variability in the reporting and clas-
sification of cytology findings, the Bethesda System for
Reporting Thyroid Cytopathology was developed and
published in 2009 (Fig. 2).5,6 Although the Bethesda sys-
tem has been widely adopted and clearly has improved the
consistency of terminology and communication regarding
the potential risk of malignancy of indeterminate cyto-
logic findings, the rates of malignancy within the subcate-
gories can vary widely by institution. Furthermore,
significant interobserver and intraobserver variability exist
in interpretation, with concordance rates of 65% to 75%
for thyroid cytology and 90% for histopathology.7 Cyto-
molecular testing of thyroid nodules has the potential to
distinguish indeterminate as either benign or neoplastic
(including both premalignant and malignant pathology)
and thus to improve the quality of care and guide surgical
decision making. However, the incorporation of molecu-
lar testing into the options for diagnosis in thyroid cancer
may actually contribute to changes in cytopathology inter-
pretation and changes in the rates of indeterminate nod-
ules.8 Even final histologic assessment of benign versus
malignant disease is in transition, as indicated by the
recent change in the nomenclature when discussing
noninvasive, encapsulated follicular variant of PTC,
which was recently changed to be considered nonmalig-
nant and redefined as a noninvasive follicular tumor with
889
Review Article

Figure 2. The Bethesda system has improved the consistency of terminology and communication regarding the potential risk of
malignancy of indeterminate cytologic findings. Adapted from: Baloch ZW, LiVolsi VA, Asa SL, et al. Diagnostic terminology and
morphologic criteria for cytologic diagnosis of thyroid lesions: a synopsis of the National Cancer Institute Thyroid Fine-Needle
Aspiration State of the Science Conference. Diagn Cytopathol. 2008;36:425-437.6

papillary-like features (NIFTP). However, diagnostic
lobectomy is still required for both pathologic diagnosis
and clinical treatment of NIFTP.9,10 Similar to distin-
guishing between follicular adenoma and follicular carci-
noma, histopathologic evaluation is necessary to exclude
capsular or vascular invasion in the diagnosis of NIFTP.
THE DEVELOPMENT OF THYROID
MOLECULAR TESTING
The initial development of immunohistochemical evalua-
tion with Galectin-3 and other biomarkers demonstrated
some potential to improve the diagnostic accuracy of thy-
roid cytology, but it lacked sufficient reproducibility and
sensitivity to reliably exclude malignancy.11 Subsequently,
new investigations sought to identify specific gene muta-
tions and gene rearrangements that were pathogenic in
thyroid cancer. The feasibility of cytomolecular testing for
known DNA and RNA mutations, which accounted for
70% of thyroid carcinomas (Harvey rat sarcoma viral
oncogene homolog [H-Ras], neuroblastoma rat sarcoma
viral oncogene homolog [N-Ras], Kirsten rat sarcoma
viral oncogene homolog [K-Ras], rearranged during trans-
fection/PTC 1 [Ret/PTC1], Ret/PTC2, Ret/PTC3, v-Raf
murine sarcoma viral oncogene homolog B1 [BRAF], and
paired box gene 8-peroxisome proliferator-activator
receptor c [Pax8-PPARc]) was demonstrated, which
improved the diagnostic accuracy of needle biopsy when
added to cytologic evaluation (Fig. 3). The most common
mutations identified in thyroid cancer improved the spe-
cificity of thyroid cytology, but it was still not a sensitive
enough detection method to rule out cancer for cases in
which no mutation was identified.12
Approaching the question of benign versus malignant
cytopathology from a different perspective, a gene expres-
sion classifier (GEC) was developed through an iterative
process designed a priori to have a high sensitivity and high
NPV, similar to the NPV of thyroid nodules diagnosed as
benign on cytology.13,14 Designed specifically to help rule
out malignancy, the Afirma GEC (Veracyte, South San
Francisco, Calif) was developed to potentially decrease the
rate of diagnostic surgeries in the case of indeterminate
nodules.
In addition, as molecular testing was beginning to
develop for clinical use to improve the diagnostic accuracy
of thyroid FNA, The Cancer Genome Atlas (TCGA) pro-
ject mapped the mutations attributed to both classical and
follicular variants of PTC, significantly reducing the num-
ber of unknown specific mutations causing differentiated
thyroid cancer.15 Incorporating the new information
obtained through TCGA, an updated gene panel using
next-generation sequencing (NGS) was developed for
clinical use (ThyroSeq NGS, v2.0 [and now v2.1];
CBLPath, Orlando, Fla).16-18 This test was developed to
improve the overall accuracy of molecular testing to

890 Cancer March 1, 2018

 Molecular Testing for Thyroid Nodules/Roth et al

Figure 3. The mitogen-activated protein kinase (MAPK) pathway is the principal pathway leading to differentiated thyroid cancer.
The phosphatidylinositol-4,5-bisphosphate 3-kinase–protein kinase B (PI3K-AKT) pathway is an important pathway leading to fol-
licular thyroid cancer and undifferentiated thyroid cancer. a, b, and g indicate rearranged during transfection (RET) receptors a,
b, and g, respectively; BRAF, v-Raf murine sarcoma viral oncogene homolog B1; ERK, extracellular signal-regulated kinase; GF,
growth factor; GPT, glutamic-pyruvic transaminase; H-Ras, Harvey rat sarcoma viral oncogene homolog; K-Ras, Kirsten rat sar-
coma viral oncogene homolog; MEK, mitogen activated protein extracellular signal-regulated kinase; mTOR, mammalian target of
rapamycin; N, nucleus; N-Ras, neuroblastoma rat sarcoma viral oncogene homolog; Pax8/PPARc, paired box gene 8-peroxisome
proliferator-activator receptor g; PTEN, phosphatase and tensin homolog; RAF, RAF proto-oncogene serine/threonine-protein
kinase; RAS, a family of rat sarcoma (ras) small guanosine triphosphate hydrolases (GTPases); RET, rearranged during transfec-
tion; Ret/PTC, rearranged during transfection/papillary thyroid cancer 1; RTK, receptor tyrosine kinase.

identify both benign and malignant thyroid cytopathol-
ogy by improving sensitivity and NPV with only a mild
reduction in specificity and PPV.
Additional approaches to molecular testing for
thyroid cytopathology include the development of micro-
RNA (miRNA) testing. MiRNAs act as negative regula-
tors of gene expression and may impact cellular processes
that lead to carcinogenesis. Because miRNA expression
may be dysregulated in thyroid cancer, this represents an
additional approach to molecular test development with
high sensitivity and high specificity.19-21
The ultimate goal of all the described approaches to
thyroid molecular testing is to improve the clinical man-
agement of patients with thyroid nodules. Reducing the
number of unnecessary diagnostic thyroid surgeries has
huge implications for overall patient quality of life and
cost of health care. Therefore, as the number of patients
with thyroid nodules continues to rise, the goal is to
decrease potential morbidity for benign thyroid disease
and to optimize initial treatment of thyroid cancer.
A “Rule-In” Test: Gene Mutation Panel
In the setting of a suspicious thyroid nodule with a clear
gene mutation that identifies malignancy, thyroid cancer
could be diagnosed before surgery and effectively decrease
the number of patients who require a 2-part procedure for
definitive primary treatment of thyroid cancer. The 7-
gene specific mutation panel has been validated as an
effective “rule-in” cytomolecular test for thyroid malig-
nancy based on the high specificity and PPV (Table 1),
but it has not been considered an effective “rule-out” test
because of low sensitivity (Table 2). The largest
prospective, single-institution study included 247
Bethesda III thyroid nodules and 214 Bethesda IV thyroid

Cancer March 1, 2018 891

Review Article

TABLE 1. “Rule-In” Testing

Test

ThyroSeq v2.1 Seven-Gene Rosetta GX

ThyGenX/ThyraMIR (Nikiforov 201416, Afirma Mutation Panel Reveal
Result (Labourier 201521) Nikiforov 201517) (Alexander 201214) (Nikiforov 201122) (Lithwick-Yanai 201623)

AUS/FLUS, %
Prevalence 32 23 24 14 21
Specificity 80 92 52 99 74
PPV 68 77 38 88 43
FN/SFN. %
Prevalence 32 27 25 27 21
Specificity 91 93 52 97 74
PPV 82 83 37 87 43

Abbreviations: AUS, atypia of undetermined significance; FLUS, follicular lesion of undetermined significance; FN, follicular neoplasm; PPV, positive predictive
value; SFN, suspicious for follicular neoplasm.

TABLE 2. “Rule-Out” Testing

Test

ThyroSeq v2.1 Seven-Gene Rosetta GX

ThyGenX/ThyraMIR (Nikiforov 201416, Afirma Mutation Panel Reveal
Result (Labourier 201521) Nikiforov 201517) (Alexander 201214) (Nikiforov 201122) (Lithwick-Yanai 201623)

AUS/FLUS, %
Prevalence 32 23 24 14 21
Sensitivity 94 90 92 63 74
NPV 97 97 95 94 92
FN/SFN, %
Prevalence 32 27 25 27 21
Sensitivity 82 90 92 57 74
NPV 91 96 94 86 92

Abbreviations: AUS, atypia of undetermined significance; FLUS, follicular lesion of undetermined significance; FN, follicular neoplasm; NPV, negative predictive
value; SFN, suspicious for follicular neoplasm.

nodules with both cytology and histologic follow-up. The
investigators reported 63% sensitivity and 99% specificity
for Bethesda III nodules and 57% sensitivity and 97%
specificity for Bethesda IV nodules. A positive test result
(the PPV) increased the cancer risk to 88% for Bethesda
III nodules and 87% for Bethesda IV nodules, whereas
the absence of a mutation or fusion was associated with a
cancer risk of 6% (NPV, 94%) and 14% (NPV, 86%),
respectively.22
Although the early adoption of the 7-gene mutation
panel to guide surgery was very positive,22 with a signifi-
cant decrease in the number of patients requiring a com-
pletion thyroidectomy, several changes in the overall
approach to thyroid cancer treatment may limit this early
promise. A more judicious approach to thyroid surgery
has been recommended, as described in the 2015 Ameri-
can Thyroid Association guidelines, recognizing that a
total thyroidectomy may not be the recommended initial
surgery even for confirmed thyroid malignancies that are
clinically considered low risk.3 Early studies with the 7-
gene mutation panel indicated that rat sarcoma (RAS) (a
family of small guanosine triphosphate hydrolases) muta-
tions were most common but were not highly specific for
thyroid cancer, thereby limiting the PPV of the test.
Recent studies have demonstrated that RAS mutations
also can be identified in up to 48% of benign nodules;
therefore, the type of RAS mutation and the degree of alle-
lic frequency can be used to guide decision-making.24
Similarly, many thyroid nodules with RAS and PAX8/
PPARc mutations are the histologically noninvasive
encapsulated follicular variant of PTC,10 now reclassified
as NIFTP.9 The definition of NIFTP was only recently
published and reclassified to highlight the limited malig-
nant potential of these precancerous lesions. However,

892 Cancer March 1, 2018


treatment by lobectomy is still recommended.9,10 The
2015 American Thyroid Association guidelines do not
mandate a total thyroidectomy for differentiated thyroid
cancer <4 cm with any of the mutations identified in the
7-gene mutation panel, further limiting its utility in either
decreasing unnecessary surgery or guiding the extent of
surgery for patients with indeterminate nodules.13 Reduc-
ing staged completion thyroidectomy can only be
achieved in the more limited circumstance for the patient
with a highly specific mutation for cancer (BRAF and
RET/PTC) coupled with a surgeon and patient favoring
total thyroidectomy over lobectomy.25
A “Rule-Out” Test: Afirma GEC
With the development and adoption of the Bethesda clas-
sification system for cytopathology,2 the estimated rates
of malignancy in indeterminate nodules, ranging from
5% to 15% in Bethesda III nodules (atypia of undeter-
mined significance/follicular lesion of undetermined sig-
nificance) to 15% to 30% in Bethesda IV nodules
(suspicious for follicular neoplasm/follicular neoplasm),
remains too high to consider ongoing clinical surveillance
(Tables 1 and 2). According to the guidelines proposed by
the National Cancer Institute to define a good diagnostic
test, a “good” test should rule out cancer in 95% of cases.
Because a benign cytologic result on a thyroid FNA is
thought to rule out malignancy in 97% of cases, a molecu-
lar test would ideally have similar accuracy to confirm
benign disease. A GEC was developed with the objective
of having high sensitivity and NPV, similar to those of
thyroid nodules diagnosed as benign on cytology.13,14
The result of this process is a commercially available GEC
(Veracyte), which uses an algorithm to evaluate the molec-
ular expression of 167 genes to further classify cytologi-
cally indeterminate thyroid nodules into either benign or
suspicious categories.
The initial prospective, blinded, multi-institutional
validation study involving 577 indeterminate thyroid
FNA samples, 265 of which were used for the validation
analysis, reported 92% overall sensitivity and 52% specif-
icity for indeterminate nodules classified as Bethesda III
(atypia of undetermined significance/follicular lesion of
undetermined significance), IV (follicular neoplasm/sus-
picious for follicular neoplasm), or V (suspicious for
malignancy). The NPVs for a benign GEC result were
95%, 94%, and 85% for the 3 indeterminate Bethesda cat-
egories, respectively, with rates of malignancy among the 3
categories of 24%, 25%, and 62%, respectively.14 Of the 7
false-negative results, 6 samples had insufficient follicular
material for evaluation, but the seventh was an adequate
Cancer March 1, 2018
Molecular Testing for Thyroid Nodules/Roth et al
sample in which the GEC failed to identify an invasive
oncocytic carcinoma. Although the GEC test was not
designed to provide high specificity or PPV, it is impor-
tant to note that the PPV reported in the initial validation
study for Bethesda III and IV nodules was 38%.14 More
recently, ongoing studies evaluating the accuracy of the
Afirma GEC to effectively rule out malignancy have iden-
tified additional cases of malignancy with benign GEC
results.25
It is important for the patient and clinician to recog-
nize that a suspicious classification by GEC is not a diag-
nosis of malignancy, nor is it equivalent to the Bethesda V
suspicious for malignancy category. The notable excep-
tion is in medullary thyroid carcinoma, in which the PPV
is 98% for the GEC.26,27
Designed to lower patient morbidity and the cost of
care by decreasing unnecessary surgery for benign thyroid
nodules, the majority of studies have demonstrated the
utility of the Afirma GEC, whereas others have ques-
tioned the actual cost-benefit analysis of reflexively incor-
porating it into clinical care.8,28-44 Behind some of the
various interpretations of the performance of the GEC are
the inherent differences in agreement among thyroid cyto-
pathologists.7 However, another concern is the lack of a
true “gold standard” in the subsequent validation studies,
for which very few of the GEC benign cases have histo-
logic correlation. This limitation prevents any further vali-
dation of the sensitivity and specificity of the GEC test
and raises the concern of possible cases that have been mis-
classified as benign because of the limited length of
follow-up in subsequent clinical validation studies.
Evaluation of the GEC test among unique cytopa-
thologic classifications leads to additional questions about
the utility of the GEC test in subsequently decreasing the
number of patients sent for diagnostic surgery. Recent
evaluation of which malignancies are detected by the
GEC test reveal a high prevalence of the follicular-variant
of PTC, many of which have been reclassified as
NIFTP.45 This suggests a growing role for limited surgery
by hemithyroidectomy or lobectomy to treat low-risk
malignancies or precancerous nodules with a “suspicious”
GEC result.46 In the setting of Hurthle cell lesions, recent
data have indicated much higher rates of “suspicious”
GEC results, ranging from 70% to 90%, but with much
lower rates of malignancy, suggesting limited clinical util-
ity in this specific population.47-50 Attempts to increase
the PPV of a GEC “suspicious” result by adding the
BRAF gene mutation test did not increase the PPV of the
test result because of a low prevalence of classical PTC in
the Bethesda III and IV categories.
893
Review Article
Although it may be debatable whether the incorpo-
ration of cytomolecular testing into clinical care has
improved clinical outcomes or influenced the overall cost
of care, the addition of molecular testing in thyroid nod-
ules has informed the overall discussion of thyroid nod-
ules and the malignant potential of specific thyroid
cytology results. The utility of the GEC for cytologically
indeterminate nodules, as supported by the research and
multiple validation studies published, should focus on the
NPV of the test, which performs best when the estimated
likelihood of malignancy based on clinical, sonographic,
and cytologic evaluation is less than or equal to approxi-
mately 25% and patient preference is toward avoiding
surgery if the GEC test is benign.
COMPREHENSIVE TESTS TO IMPROVE
OVERALL ACCURACY
ThyroSeq NGS assay
The limitations of both the “rule-out” and “rule-in”
approaches have been partially addressed with advances
derived from TCGA that have been harnessed with NGS,
a high-throughput sequencing analysis of large areas of
the human genome (Tables 1 and 2).18 An NGS-based
assay (ThyroSeq v2.0) is reported with 90% sensitivity,
93% specificity, an 83% PPV, and a 96% NPV, with
prevalence of cancer at 25% for indeterminate thyroid
nodules in a recent single-center study of 143 consecutive
Bethesda IV nodules.16 Subsequently evaluated for
Bethesda III nodules, the commercially available version,
ThyroSeq v2.1 (CBLPath), identifies 15 point mutations
and 42 gene fusions with 90.9% sensitivity, 92.1% specif-
icity, a 76.9% PPV, and a 97.2% NPV17. On the basis of
both high sensitivity and high NPV, ThyroSeq v2.1 is
promoted as a comprehensive test to both “rule in” and
“rule out” thyroid malignancy.
A test with increasing sensitivity will inherently
cause reduced specificity to some degree, as discussed
above. The improvement in sensitivity and NPV of Thy-
roSeq v2.1 over the 7-gene mutation panel is achieved by
testing for more pathogenic mutations. Although the
NPV with ThyroSeq v2.1 is able to rule out malignancy
with greater than 95% accuracy for both Bethesda III and
Bethesda IV nodules while maintaining a relatively high
PPV (approximately 75%), it is similar to the PPV for
nodules deemed cytologically suspicious for malignancy
but not as high as that for nodules cytologically diagnosed
as malignant (>90%). In addition, ongoing research into
the correlation of specific mutations with specific histo-
pathologic correlations may help further guide surgical
894
planning when a preoperative diagnosis of an indolent
versus aggressive tumor mutation is identified.51-53
ThyroSeq v2.1 has added potential benefit for
expanding presurgical risk stratification by providing spe-
cific results for identified mutations.29,30 Prognosis and
staging-based treatment decisions have been extensively
studied for the BRAF mutation. Multivariate analysis in
some but not all studies54-60 have demonstrated that
BRAF is an independent predictor of recurrence. How-
ever, BRAF mutation does not appear to improve progno-
sis or direct treatment beyond traditional TNM
classifications, and the independent use of BRAF would
incorrectly upstage the majority of indolent behaving,
lower stage classical PTCs. Furthermore, it has been dem-
onstrated that certain RAS mutations have a higher likeli-
hood of distant metastasis and increased mortality
compared with BRAF mutations, but RAS mutations can
also be identified in benign follicular adenomas and in the
noninvasive encapsulated follicular variant of PTCs
(NIFTP), now reclassified as a premalignant tumor.24
Two molecular markers that appear to confer an increased
risk of tumor recurrence and tumor-related mortality are
tumor protein 53 (TP53) and telomerase reverse tran-
scriptase (TERT) mutations. TP53 mutations are late
events in tumor clone progression and have been known
to occur mostly in poorly differentiated and anaplastic
thyroid cancers, often deriving from well differentiated
PTCs.40 TERT mutation is an independent predictor of
disease-free survival.61 The combination of a TERT
mutation and a BRAF or Ras mutation within the same
tumor is associated with a high risk of structural disease
recurrence and mortality.39,60-63 However, these are rare
mutations and often clinically present with advanced and
aggressive disease; thus, the current state of thyroid molec-
ular testing does not add significantly to prognosis and
treatment decisions beyond traditional pathologic and
summary TNM staging.51 Although the role of using
molecular testing to guide treatment decisions beyond the
primary management of thyroid nodules remains
unknown, ongoing research is working to expand our
understanding and ability to further risk stratify thyroid
cancer and potentially guide early treatment decisions,
including the extent of surgery.
ThyGenX/ThyraMIR
ThyGenX/ThyraMIR (Interpace Diagnostics, Parsip-
pany, NJ) is also promoted as a comprehensive test with
both high PPV and high NPV, based on very limited pub-
lished data. This 2-part molecular test incorporates the
original 7-gene mutation panel along with the
Cancer March 1, 2018

phosphatidylinositol-4,5-bisphosphate 3-kinase cata-
lytic subunit a (PIK3CA) mutation, labeled as Thy-
GenX. PIK3CA mutation is associated with follicular
carcinoma and undifferentiated carcinoma.21 If the
ThyGenX panel has no mutation, then the ThyraMIR
test is reflexively offered. ThyraMIR is a 10-miRNA
GEC. This test was validated in a cross-sectional
cohort analysis of 109 samples of thyroid FNA mate-
rial from Bethesda III and IV nodules with paired his-
tologic specimens, which were gathered across 12
clinical institutions with cytology and pathology speci-
mens evaluated locally, without central adjudication
of cytology or pathology results. In that small study,
only samples with paired histopathology were
included, and only 35 of the enrolled samples had
malignant pathology. When evaluating the characteris-
tics of this test for Bethesda III and IV nodules com-
bined, the sensitivity of the ThyGenX/ThyraMIR test
was 89%, with 85% specificity, a 68% PPV, and a
97% NPV for Bethesda III nodules (91% for
Bethesda IV nodules), with an overall cancer preva-
lence rate of 32%.21 Additional clinical validation of
this test is warranted to determine how effectively and
accurately the ThyGenX/ThyraMIR platform can help
distinguish between benign and neoplastic thyroid
nodules.
Rosetta GX Reveal
MiRNAs are an important class of noncoding RNAs
implicated in gene expression regulation, and their expres-
sion profiles have been associated with the pathophysiol-
ogy of cancer, including thyroid cancer. Rosetta GX
Reveal (Rosetta Genomics Ltd, Rehovot, Israel)20 is a
miRNA-based assay for diagnosing indeterminate thyroid
FNAs using stained FNA smears prepared for initial cyto-
pathologic evaluation. The assay classifier measures 24
miRNAs. In this test, expression levels of a set of miRNA
biomarkers are measured and used to classify thyroid nod-
ules as benign or suspicious, and its analytical validity has
been reported, supporting this novel approach to molecu-
lar testing using thyroid FNA smears prepared from rou-
tine cytology slides.20 This approach to molecular testing
may prove to be less invasive for patients by allowing the
use of cytology slides for the acquisition of cellular mate-
rial rather than requiring a repeat thyroid FNA or dedi-
cated collection of sample material at the time of original
the FNA.
A blinded, retrospective, multicenter validation of
the novel Rosetta GX Reveal assay was recently published
evaluating its performance on 189 FNA samples
Cancer March 1, 2018
Molecular Testing for Thyroid Nodules/Roth et al
(Bethesda III-V nodules) with matched surgical speci-
mens.63 The study outlined 3 phases leading to the dis-
covery of specific miRNAs for inclusion in the assay, the
training of the assay classifier, and the final validation set,
including 189 FNA samples. Overall performance charac-
teristics revealed 85% sensitivity, 72% specificity, a 59%
PPV, and a 91% NPV. Several post hoc analyses were
included in the published results, separating the validation
set into an “agreement” set, in which pathology was con-
cordant among 3 pathologists. This improved the perfor-
mance characteristics, suggesting that the assay
performance may vary based on the accuracy of the histo-
pathology results as well. Nine malignant samples were
misclassified as benign, 5 of which were the encapsulated
follicular variant of PTC, 2 were classic PTC, 1 was a non-
encapsulated variant of PTC, and 1 was a follicular carci-
noma with minimal capsular invasion.
The Rosetta GX Reveal assay adds a unique
approach to molecular testing for thyroid FNA samples
by using a cytology slide, thereby minimizing the impact
to the patient at the time of FNA. This assay does not
require a specialized sample-collection kit and thus may
provide an option for molecular testing when needed
rather than requiring specific sample collection at the time
of the initial procedure. The validation data currently
available highlight the challenges of using the miRNA
platform for distinguishing between benign and neoplas-
tic thyroid nodules but point to areas for further growth
and development of this assay. Although additional vali-
dation will help further define the utility of this test specif-
ically in Bethesda III versus Bethesda IV nodules, the
recent publication adds to the growing options for molec-
ular testing in clinical practice.
CONCLUSIONS
The science of molecular testing has progressed rapidly,
but caution must be used when incorporating molecular
testing for thyroid nodules into clinical practice. A clear
understanding of the goals and limitations of molecular
testing must be incorporated into how physicians use and
explain molecular testing to patients.
Molecular tests can rule in cancer for indeterminate
thyroid nodules with highly specific mutations for cancer,
such as BRAF and RET/PTC, and can reduce completion
thyroidectomy rates for surgeons recommending total
thyroidectomy. The PPV of malignant cytology (Bethesda
VI) is 98%3; and, although molecular testing improves
specificity analysis and PPV, it falls short of this ideal for
other mutations. Molecular testing can rule out thyroid
cancer for the indeterminate thyroid nodule in selected
895
Review Article
cases. Indications favoring “rule-out” testing include a
low institutional prevalence of cancer for indeterminate
thyroid nodules and patients with no high-risk history,
physical features, or ultrasound features. The NPV of
benign cytology is 96%.3 Prospective studies suggest that
this ideal has been met with several molecular tests, but
careful surveillance is recommended given the lack of
long-term outcomes. In addition, many of the molecular
platforms described herein have limited testing against a
true gold standard; because benign nodules, as determined
by molecular tests, rarely undergo definitive histologic
evaluation. Similarly, the true specificity of molecular tests
also must be evaluated in benign nodules to further char-
acterize the accuracy of molecular testing results.3
Although there is hope that cytomolecular testing
can potentially inform treatment decisions regarding the
clinically low-risk PTCs that may be indolent; currently,
molecular tests fail to improve the predictive ability
beyond the TNM classification system.51 The develop-
ment of the American Thyroid Association risk-
stratification system remains the best way to predict future
disease recurrence after initial therapy.3 However, the
future of molecular testing in thyroid pathology holds
great promise and may inform not only nodule manage-
ment but also the role for radioactive iodine therapy and
potentially systemic therapy as well. Further understand-
ing of the molecular mutations causing thyroid cancer
may allow for new drug development for high-risk disease,
help identify and change treatment approaches for
patients with specific mutations, and allow more directed
risk-stratification and prognosis discussions. Early data
have already begun to identify specific molecular markers
in thyroid cancer associated with negative outcomes and
decreased survival.63 Further research into the implica-
tions of molecular testing on treatment decisions and dis-
ease prognosis is warranted.
Whereas each test described has its own strengths
and limitations, no data compare the performance charac-
teristics of each commercially available test among a given
collection of samples. Ideally, future research would
include head-to-head comparisons of molecular tests,
long-term prospective data, and cost-effectiveness studies;
however, currently, clinicians and patients who use the
tests must understand and weigh the advantages and
limitations.
FUNDING SUPPORT
No specific funding was disclosed.
896
CONFLICT OF INTEREST DISCLOSURES
Mary Y. Roth reports grants from Rosetta Genomics outside the
submitted work. David L. Steward reports clinical trial support
through the University of Pittsburgh Cancer Institute, from Rosetta
Genomics, Veracyte Inc, and AstraZeneca during the conduct of
the study. Robert L. Witt made no disclosures.
AUTHOR CONTRIBUTIONS
David L. Steward: Conception and design of work; acquisition,
analysis, and interpretation of data; writing–initial draft; writing–
review and revisions; final approval of the version to be published;
and agreement to be accountable for all aspects of the work. Mara
Y. Roth: Acquisition, analysis, and interpretation of data; writing–
review and revisions; final approval of the version to be published;
and agreement to be accountable for all aspects of the work. Robert
L. Witt: Acquisition, analysis, and interpretation of data; writing–
review and revisions; final approval of the version to be published;
and agreement to be accountable for all aspects of the work.
REFERENCES
1. Guth S, Theune U, Aberle J, Galach A, Bamberger CM. Very high
prevalence of thyroid nodules detected by high frequency (13 MHz)
ultrasound examination. Eur J Clin Invest. 2009;39:699-706.
2. Cibas ES, Ali SZ. The Bethesda System for reporting thyroid cyto-
pathology. Thyroid. 2009;19:1159-1165.
3. Haugen BR, Alexander EK, Bible KC, et al. 2015 American Thyroid
Association management guidelines for adult patients with thyroid
nodules and differentiated thyroid cancer: the American Thyroid
Association Guidelines task force on thyroid nodules and differenti-
ated thyroid cancer. Thyroid. 2016;26:1-133.
4. Cooper DS, Doherty GM, Haugen BR, et al. Management guide-
lines for patients with thyroid nodules and differentiated thyroid
cancer. Thyroid. 2006;16:109-142.
5. Baloch ZW, LiVolsi VA, Asa SL, et al. Diagnostic terminology and
morphologic criteria for cytologic diagnosis of thyroid lesions: a syn-
opsis of the National Cancer Institute Thyroid Fine-Needle Aspira-
tion State of the Science Conference. Diagn Cytopathol. 2008;36:
425-437.
6. Crippa S, Mazzucchelli L, Cibas ES, Ali AZ. The Bethesda System
for reporting thyroid fine-needle aspiration specimens [author reply].
Am J Clin Pathol. 2010;134:345.
7. Cibas ES, Baloch ZW, Fellegara G, et al. A prospective assessment
defining the limitations of thyroid nodule pathologic evaluation.
Ann Intern Med. 2013;159:325-332.
8. Sacks WL, Bose S, Zumsteg ZS, et al. Impact of Afirma gene expres-
sion classifier on cytopathology diagnosis and rate of thyroidectomy.
Cancer Cytopathol. 2016;124:722-728.
9. Nikiforov YE, Seethala RR, Tallini G, et al. Nomenclature revision
for encapsulated follicular variant of papillary thyroid carcinoma: a
paradigm shift to reduce overtreatment of indolent tumors. JAMA
Oncol. 2016;2:1023-1029.
10. Armstrong MJ, Yang H, Yip L, et al. PAX8/PPARg rearrangement
in thyroid nodules predicts follicular-pattern carcinomas, in particu-
lar the encapsulated follicular variant of papillary carcinoma. Thy-
roid. 2014;24:1369-1374.
11. Bartolazzi A, Orlandi F, Saggiorato E, et al. Galectin-3-expression
analysis in the surgical selection of follicular thyroid nodules with
indeterminate fine-needle aspiration cytology: a prospective multi-
centre study. Lancet Oncol. 2008;9:543-549.
12. Nikiforov YE, Steward DL, Robinson-Smith TM, et al. Molecular
testing for mutations in improving the fine-needle aspiration diagno-
sis of thyroid nodules. J Clin Endocrinol Metab. 2009;94:2092-2098.
13. Walsh PS, Wilde JI, Tom EY, et al. Analytical performance verifica-
tion of a molecular diagnostic for cytology-indeterminate thyroid
nodules. J Clin Endocrinol Metab. 2012;97:E2297-E2306.
Cancer March 1, 2018

14. Alexander EK, Kennedy GC, Baloch ZW, et al. Preoperative diagno-
sis of benign thyroid nodules with indeterminate cytology. N Engl J
Med. 2012 23;367:705-715.
15. Cancer Genome Atlas Research Network. Integrated genomic charac-
terization of papillary thyroid carcinoma. Cell. 2014;159:676-690.
16. Nikiforov YE, Carty SE, Chiosea SI, et al. Highly accurate diagnosis
of cancer in thyroid nodules with follicular neoplasm/suspicious for
a follicular neoplasm cytology by ThyroSeq v2 next-generation
sequencing assay. Cancer. 2014;120:3627-3634.
17. Nikiforov YE, Carty SE, Chiosea SI, et al. Impact of the multi-gene
ThyroSeq next-generation sequencing assay on cancer diagnosis in thy-
roid nodules with atypia of undetermined significance/follicular lesion
of undetermined significance cytology. Thyroid. 2015;25:1217-1223.
18. Nikiforova MN, Wald AI, Roy S, et al. Targeted next-generation
sequencing panel (ThyroSeq) for detection of mutations in thyroid
cancer. J Clin Endocrinol Metab. 2013;98:E1852-E1860.
19. Nikiforova MN, Chiosea SI, Nikiforov YE. MicroRNA expression
profiles in thyroid tumors. Endocr Pathol. 2009;20:85-91.
20. Benjamin H, Schnitzer-Perlman T, Shtabsky A, et al. Analytical
validity of a microRNA-based assay for diagnosing indeterminate
thyroid FNA smears from routinely prepared cytology slides. Cancer
Cytopathol. 2016;124:711-721.
21. Labourier E, Shifrin A, Busseniers AE, et al. Molecular testing for
miRNA, mRNA, and DNA on fine-needle aspiration improves the
preoperative diagnosis of thyroid nodules with indeterminate cytol-
ogy. J Clin Endocrinol Metab. 2015;100:2743-2750.
22. Nikiforov YE, Ohori NP, Hodak SP, et al. Impact of mutational
testing on the diagnosis and management of patients with cytologi-
cally indeterminate thyroid nodules: a prospective analysis of 1056
FNA samples. J Clin Endocrinol Metab. 2011;96:3390-3397.
23. Lithwick-Yanai G, Dromi N, Shtabsky A, et al. Multicentre valida-
tion of a microRNA-based assay for diagnosing indeterminate thy-
roid nodules utilising fine needle aspirate smears [published online
head of print October 26, 2016]. J Clin Pathol. doi: 10.1135/jclin-
path-2016-204089.
24. Najafian A, Noureldine S, Azar F, et al. RAS mutations and RET/
PTC and PAX8/PPAR-gamma chromosomal rearrangements are also
prevalent in benign thyroid lesions: implications thereof and a sys-
tematic review. Thyroid. 2017;27:39-48.
25. Al-Quarayshi Z, Deniwar A, Thethi T, et al. Association of malig-
nancy prevalence with test properties and performance of the gene
expression classifier in indeterminate thyroid nodules [published
online ahead of print December 15, 2016]. JAMA Otolaryngol Head
Neck Surg. doi: 10.1001/jamaoto.2016.3526.
26. Yip L, Wharry LI, Armstong M. A clinical algorithm for fine-needle
aspiration molecular testing effectively guides the appropriate extent
of initial thyroidectomy. Ann Surg. 2014;260:163-168.
27. Kloos RT, Monroe RJ, Traweek ST, Lanman RB, Kennedy GC. A
genomic alternative to identify medullary thyroid cancer preopera-
tively in thyroid nodules with indeterminate cytology. Thyroid.
2016;26:785-793.
28. Santhanam P, Khthir R, Gress T, et al. Gene expression classifier for
the diagnosis of indeterminate thyroid nodules: a meta-analysis
[serial online]. Med Oncol. 2016;33:14.
29. Samulski TD, LiVolsi VA, Wong LQ, Baloch Z. Usage trends and
performance characteristics of a “gene expression classifier” in the
management of thyroid nodules: an institutional experience. Diagn
Cytopathol. 2016;44:867-873.
30. Villabona CV, Mohan V, Arce KM, et al. Utility of ultrasound ver-
sus gene expression classifier in thyroid nodules with atypia of unde-
termined significance. Endocr Pract. 2016;22:1199-1203.
31. Chaudhary S, Hou Y, Shen R, Hooda S, Li Z. Impact of the Afirma
gene expression classifier result on the surgical management of thy-
roid nodules with category III/IV cytology and its correlation with
surgical outcome. Acta Cytol. 2016;60:205-210.
32. Wu JX, Young S, Hung ML, et al. Clinical factors influencing the
performance of gene expression classifier testing in indeterminate
thyroid nodules. Thyroid. 2016;26:916-922.
33. Labourier E. Utility and cost-effectiveness of molecular testing in
thyroid nodules with indeterminate cytology. Clin Endocrinol (Oxf).
2016;85:624-631.
Cancer March 1, 2018
Molecular Testing for Thyroid Nodules/Roth et al
34. Abeykoon JP, Mueller L, Dong F, Chintakuntlawar AV, Paludo J,
Mortada R. The effect of implementing gene expression classifier on
outcomes of thyroid nodules with indeterminate cytology. Horm
Cancer. 2016;7:272-278.
35. Londero SC, Jespersen ML, Krogdahl A, et al. gene-expression classi-
fier in papillary thyroid carcinoma: validation and application of a
classifier for prognostication. Study from the Danish Thyroid Cancer
Group–DATHYRCA (part of the DAHANCA organization). Anti-
cancer Res. 2016;36:749-756.
36. Sipos JA, Blevins TC, Shea HC, et al. Long-term nonoperative rate
of thyroid nodules with benign results on the Afirma gene expression
classifier. Endocr Pract. 2016;22:666-672.
37. Noureldine SI, Olson MT, Agrawal N, Prescott JD, Zeiger MA,
Tufano RP. Effect of gene expression classifier molecular testing on
the surgical decision-making process for patients with thyroid nod-
ules. JAMA Otolaryngol Head Neck Surg. 2015;141:1082-1088.
38. Celik B, Whetsell CR, Nassar A. Afirma GEC and thyroid lesions:
an institutional experience. Diagn Cytopathol. 2015;43:966-970.
39. Wu JX, Lam R, Levin M, Rao J, Sullivan PS, Yeh MW. Effect of malig-
nancy rates on cost-effectiveness of routine gene expression classifier test-
ing for indeterminate thyroid nodules. Surgery. 2016;159:118-126.
40. Yip L, Nikiforova MN, Yoo JY, et al. Tumor genotype determines
phenotype and disease-related outcomes in thyroid cancer: a study of
1510 patients. Ann Surg. 2015;262:519-525; discussion 525-525.
41. Witt RL. Outcome of thyroid gene expression classifier testing in
clinical practice. Laryngoscope. 2016;126:524-527.
42. Wong KS, Angell TE, Strickland KC, et al. Noninvasive follicular
variant of papillary thyroid carcinoma and the Afirma gene-
expression classifier. Thyroid. 2016;26:911-915.
43. Lastra RR, Pramick MR, Crammer CJ, LiVolsi VA, Baloch ZW.
Implications of a suspicious Afirma test result in thyroid fine-needle
aspiration cytology: an institutional experience. Cancer Cytopathol.
2014;122:737-744.
44. Brauner E, Holmes BJ, Krane JF, et al. Performance of the Afirma gene
expression classifier in Hurthle cell thyroid nodules differs from other
indeterminate thyroid nodules. Thyroid. 2015;25:789-796.
45. McIver B, Castro MR, Morris JC, et al. An independent study of a gene
expression classifier (Afirma) in the evaluation of cytologically indetermi-
nate thyroid nodules. J Clin Endocrinol Metab. 2014;99:4069-4077.
46. Krane JF. Lessons from early clinical experience with the Afirma
gene expression classifier. Cancer Cytopathol. 2014;122:715-719.
47. Harrell RM, Bimston DN. Surgical utility of Afirma: effects of high
cancer prevalence and oncocytic cell types in patients with indetermi-
nate thyroid cytology. Endocr Pract. 2014;20:364-369.
48. Kloos RT, Reynolds JD, Walsh PS, et al. Does addition of BRAF
V600E mutation testing modify sensitivity of the Afirma gene
expression classifier in cytologically indeterminate thyroid nodules?
J Clin Endocrinol Metab. 2013;98:E761-E768.
49. Alexander EK, Schorr M, Klopper J, et al. Multicenter clinical expe-
rience with the Afirma gene expression classifier. J Clin Endocrinol
Metab. 2014;99:119-125.
50. Duick DS, Klopper JP, Diggans JC, et al. The impact of benign gene
expression classifier test results on the endocrinologist-patient decision
to operate on patients with thyroid nodules with indeterminate fine-
needle aspiration cytopathology. Thyroid. 2012;22:996-1001.
51. Edge SB, Byrd DR, Compton CC, Fritz AG, Greene FL, Trotti A,
eds. AJCC Cancer Staging Manual. 7th ed. New York: Springer;
2010.
52. Angell TE, Frates MC, Medici M, et al. Afirma benign thyroid nod-
ules show similar growth to cytologically benign nodules during fol-
low-up. J Clin Endocrinol Metab. 2015;100:E1477-E14583.
53. Zhu QL, Faquin WC, Samir AE. Relationship between sonographic
characteristics and Afirma gene expression classifier results in thyroid
nodules with indeterminate fine-needle aspiration cytopathology.
AJR Am J Roentgenol. 2015;205:861-865.
54. Sohn SY, Park WY, Shin HT, et al. Highly concordant key genetic
alterations in primary tumors and matched distant metastases in dif-
ferentiated thyroid. Thyroid. 2016;26:672-682.
55. Xing M, Westra WH, Tufano RP, et al. BRAF mutation predicts a
poorer clinical prognosis for papillary thyroid cancer. J Clin Endocri-
nol Metab. 2005;90:6373-6379.
897
Review Article
56. Elisei R, Viola D, Torregrossa L, et al. The BRAF(V600E) mutation
is an independent, poor prognostic factor for the outcome of
patients with low-risk intrathyroid papillary thyroid carcinoma:
single-institution results from a large cohort study. J Clin Endocrinol
Metab. 2012;97:4390-4398.
57. Kim TY, Kim WB, Rhee YS, et al. The BRAF mutation is useful
for prediction of clinical recurrence in low-risk patients with conven-
tional papillary thyroid carcinoma. Clin Endocrinol (Oxf). 2006;65:
364-368.
58. Li C, Lee KC, Schneider EB, Zeiger MA. BRAF V600E mutation and
its association with clinicopathological features of papillary thyroid can-
cer: a meta-analysis. J Clin Endocrinol Metab. 2012;97:4559-4570.
59. Fernandez IJ, Piccin O, Sciascia S, et al. Clinical significance of
BRAF mutation in thyroid papillary cancer. Otolaryngol Head Neck
Surg. 2013;148:919-925.
898
60. Tufano RP, Teixeira GV, Bishop J, Carson KA, Xing M. BRAF
mutation in papillary thyroid cancer and its value in tailoring initial
treatment: a systematic review and meta-analysis. Medicine (Balti-
more). 2012;91:274-286.
61. Melo M, da Rocha AG, Vinagre J, et al. TERT promoter mutations
are a major indicator of poor outcome in differentiated thyroid car-
cinomas. J Clin Endocrinol Metab. 2014;99:E754-E765.
62. Xing M, Liu R, Liu X, et al. BRAF V600E and TERT promoter
mutations cooperatively identify the most aggressive papillary thyroid
cancer with highest recurrence. J Clin Oncol. 2014;32:2718-2726.
63. Song YS, Lim JA, Choi H, et al. Prognostic effects of TERT pro-
moter mutations are enhanced by coexistence with BRAF or RAS
mutations and strengthen the risk prediction by the ATA or TNM
staging system in differentiated thyroid cancer patients. Cancer.
2016;122:1370-1379.
Cancer March 1, 2018