Genomic Alterations of Anaplastic Thyroid Carcinoma Detected by Targeted Massive Parallel Sequencing in a BRAF V600E Mutation-Prevalent Area

Abstract

Background:

Anaplastic thyroid carcinoma (ATC), the most aggressive type of thyroid cancer, has no effective therapy. Due to its dismal prognosis, it is vital to understand the genetic alterations of ATC and identify effective molecular targets. Targeted next-generation sequencing was performed to investigate the mutational profile of ATC using a massive parallel sequencing approach.

Methods:

DNA from formalin-fixed, paraffin-embedded archival samples of 11 ATCs and normal matched pairs were used. A total of 48 genetic alterations were identified by targeted exome sequencing. These alterations were validated by mass spectrometric genotyping and direct Sanger sequencing.

Results:

The most commonly mutated gene was BRAF, identified in 10 samples (91%), all showing the V600E point mutation. A KRAS point mutation was observed in the one sample (9%) without the BRAF^V600E mutation. All 11 ATCs harbored BRAF or RAS mutations, reflecting the possibility that differentiated thyroid carcinomas progress to ATCs after the accumulation of mutations. A loss of function mutation of TP53 was observed in eight samples (73%), a PIK3CA mutation was observed in two samples (18%), and a frameshift mutation of PTEN was observed in one sample (9%). Twenty-eight novel mutated genes were found that had not previously been associated with ATC. Of these, loss of function mutations of NF2, KMT2D, and PKHD1 were repeatedly seen in three samples (27%), two samples (18%), and two samples (18%), respectively. Using direct Sanger sequencing, two samples (18%) were also found with a RASAL1 mutation. KMT2D and RASAL1 mutations were significantly associated with shorter ATC patient survival.

Conclusions:

This comprehensive analysis of ATCs using targeted massive parallel sequencing identified several novel mutations in ATCs, such as loss of function mutations of NF2 or KMT2D. Future studies are needed to confirm the role of these novel mutations as independent drivers of ATC development.

Introduction

Anaplastic thyroid carcinoma (ATC), the most aggressive type of thyroid cancer, has a median survival of three to five months (1,2). ATC is highly proliferative, easily invades the trachea, and metastasizes to distant organs. ATC is usually resistant to radioactive iodine or conventional chemotherapy, and there is no effective therapy. As a result, the cancer-specific mortality of ATC at one year is >80%, even with maximal intervention (2 –4). Even though the prevalence of ATC is <1% of all thyroid cancers, it is important to understand the genetic alterations in ATC and find effective molecular targets due to its dismal prognosis.

Genetics studies of ATC have revealed frequent somatic mutations in BRAF and RAS, which are also commonly mutated in papillary thyroid cancer (PTC) and follicular thyroid cancer (FTC), as well as additional common PI3KCA, PTEN, TP53, CTNNB1, ALK, and TERT promoter mutations (2 –7). These findings suggest that ATC might develop from existing well-differentiated thyroid carcinoma with activation of both the mitogen activated protein kinase (MAPK) and phosphoinositide 3-kinase (PI3K)-AKT pathways by the accumulation of various genetic alterations. However, the precise molecular mechanisms involved in the dedifferentiation and progression of cancer are largely unknown. Recent advances in sequencing technologies might enable the comprehensive analysis of genetic alterations and identification of new targets.

In this study, targeted next-generation sequencing was performed to investigate the mutational profile of ATC using a massive parallel sequencing approach (8).

Materials and Methods

Patients and tissue samples

Eleven patients with ATC were enrolled who underwent thyroid surgery between 2009 and 2013 at the Asan Medical Center, Seoul, Korea. DNA from formalin-fixed, paraffin-embedded (FFPE) archival samples was used for targeted exome sequencing and validation. Only FFPE tissue samples from the initial thyroid surgery were used; no samples had been collected after radioactive iodine treatment. All pathological specimens were reviewed by an experienced pathologist (D.E.S.), and ATC diagnosis was confirmed according to the World Health Organization classification (9). The pathologist selected adequate tissue blocks for isolation of DNA from ATC tissues and matched normal tissues. The proportion of ATC cells in the FFPE tissues is indicated in Table 1. The protocol of this study was approved by the Institutional Review Board of the Asan Medical Center.

Table 1.

Clinical Characteristics of the Study Patients with ATC

No.	Source (%)^a	Age	Sex	Presentation	Operation	Primary tumor size (cm)	N stage	M stage	Metastases at diagnosis	Survival	Survival duration (months)	Additional thyroid disease
1	FFPE (55)	69	M	Hoarseness	TT	1.7	N1b	1	Lung, bone	Expired	2.8	Follicular adenoma, nodular hyperplasia, lymphocytic thyroiditis
2	FFPE (80)	49	M	Neck mass	TT	3.5	N1b	1	Lung	Expired	1.1	None
3	FFPE (85)	80	F	Dyspnea	TT	3.5	N1a	0	None	Expired	7.8	Nodular hyperplasia, lymphocytic thyroiditis
4	FFPE (85)	80	F	Hoarseness	Debulking	Unknown	Unknown	1	Bone	Expired	2.3	None
5	FFPE (75)	80	F	Dysphagia	TT	4.0	N1b	1	Bone	Expired	4.8	Nodular hyperplasia
6	FFPE (35)	75	F	Neck mass	TT	6.0	N1a	1	Lung, bone	Expired	2.2	PTMC
7	FFPE (75)	75	F	Neck mass	TT	5.0	N1b	0	None	Expired	2.2	PTMC
8	FFPE (60)	80	F	Neck mass	TT	5.0	N1b	0	None	Expired	2.2	None
9	FFPE (40)	68	F	Voice change	TT	7.0	N1b	1	Lung	Expired	4.0	None
10	FFPE (85)	88	F	Neck mass	TT	4.5	0	0	None	Expired	8.7	None
11	FFPE (70)	74	M	Neck mass	Debulking	Unknown	Unknown	0	None	Expired	3.4	None

Age denotes age at diagnosis of ATC. N and M stages were classified by the 6th TNM staging system.

Proportion of ATC tumor in the FFPE tissue.

ATC, anaplastic thyroid carcinoma; FFPE, formalin-fixed, paraffin-embedded tissue; F, female; M, male; TT, total thyroidectomy; y, years; PTMC, papillary thyroid microcarcinoma.

DNA extraction methods

Matched hematoxylin and eosin–stained slides from each FFPE tissue were reviewed by a pathologist under the microscope. Genomic DNA was then extracted from between two and five 6-μm-thick sections per FFPE tissue; the number of sections used depended on the size and cellularity of the tumor. After de-paraffinization with xylene and ethanol, genomic DNA was isolated with a NEXprep FFPE Tissue Kit (#NexK-9000; Geneslabs, Seongnam, Korea) according to the manufacturer's recommendations. Briefly, tissue pellets were completely lysed by incubation with proteinase K in lysis buffer overnight at 56°C, followed by additional incubation for 3 min with magnetic beads and solution A at room temperature. After incubation for 5 min on a magnetic stand, the supernatant was removed and the beads were washed three times with ethanol. The beads were then dried for 5 min, DNA was eluted in 50 μL of DNase- and RNase-free water, and quantification was performed using a Quant-iT^™ PicoGreen dsDNA Assay kit (Invitrogen, Carlsbad, CA).

Targeted next-generation sequencing

For evaluation of the somatic mutations in the 11 ATC tissues, targeted next-generation sequencing was performed using the MiSeq platform (Illumina, Inc., San Diego, CA) with OncoPanel version 2 (OP_v2) to capture the exons of 505 cancer-related genes plus partial introns from 15 genes often rearranged in cancer (Supplementary Table S1; Supplementary Data are available online at www.liebertpub.com/thy). Known ATC- or PTC-related genes not included in this panel are listed in Supplementary Table S2 (10 –12). This panel only included probes that could identify gene fusions in introns 19 and 20 of the ALK gene. The EML4/ALK rearrangement in intron 17 was not evaluated. Genomic DNA (200 ng) was fragmented to 250 bp by sonication (Covaris Inc., Woburn, MA), followed by size selection using Agencourt AMPure XP beads. A DNA library was prepared by ligation of 50 ng of purified DNA with TruSeq adaptor using a SureSelect XT Reagent Kit (Agilent Technologies, Santa Clara, CA). Each library was made with sample specific barcodes 6 bp in size, quantified using PicoGreen, and four libraries were pooled to a total of 600 ng for hybrid capture using an Agilent SureSelectXT custom kit (OP_v2 RNA bait, 2.9 Mb; Agilent Technologies). The concentration of the enriched target was measured by quantitative polymerase chain reaction (PCR; Kapa Biosystems, Inc., Woburn, MA), and samples were loaded onto the MiSeq platform (Illumina) for paired-end sequencing.

Sequenced reads were aligned to the human reference genome (NCBI build 37) with Burrows-Wheeler Aligner (0.5.9) (13) with default options, and de-multiplexing was performed with MarkDuplicates of the Picard package to remove PCR duplicates. De-duplicated reads were realigned at known indel positions with GATK IndelRealigner (14), then base quality was recalibrated using GATK TableRecalibration. Somatic variant calling for single nucleotide variants and short indels was performed with matched normal tissue using MuTect (1.1.7) (15) and SomaticIndelocator in GATK, respectively. Germline variants from candidates of somatic variants were filtered out with common dbsnp (build 141; found in ≥1% of samples) and a panel of normal samples. Final somatic variants were annotated using Variant Effect Predictor (v79) and were then converted to maf file format using vcf2maf. Structural variations were identified using BreaKmer and then were filtered out with a panel of normal samples (16).

Validation of mutations using MassARRAY

To validate genetic alterations detected by targeted exome sequencing, mass spectrometric genotyping was performed using the Sequenom MassARRAY technology platform (Sequenom, San Diego, CA). Four different multiplex amplification pools for 45 mutation candidates were designed by the Assay Designer unit of the MassARRAY TYPER software (Supplementary Table S3). While the detection limit of mutant alleles varies between assays and samples, it is usually considered to be 5–10% (17). After multiplex PCR, treatment with shrimp alkaline phosphatase from the iPLEX-Pro kit (10142-2; Sequenom) was done to inactivate the residual deoxynucleotides, which were then subjected to single-base extension. After spotting of the desalted product onto a 384-format SpectroCHIP II, the spectrum profiles generated by matrix-assisted laser desorption/ionization time of flight mass spectrometry were acquired and interpreted using the TYPER 4.0 software (Sequenom), and then further reviewed manually by two independent researchers to undo any uncertain calls resulting from clustering artifacts.

Sanger sequencing validation

Seven somatic mutation candidates for the KMT2D, NF2, and PKHD1 genes were additionally validated by direct Sanger sequencing. Mutations in RASAL1, which was recently defined as a candidate driver gene of ATC, were also additionally evaluated by direct Sanger sequencing using the 11 ATC tissues (12). DNA was PCR amplified using a 2720 Thermal Cycler (Applied Biosystems, Foster City, CA). The 20 μL reaction volume contained a 250 μM mixture of dATP, dGTP, dTTP, and dCTP, 0.25 IU of AmpliTaq-Gold DNA polymerase, and 0.5 μM of each primer pair shown in Supplementary Table S4. The amplification protocol consisted of initial denaturation at 95°C for 15 min, followed by 45 cycles of denaturation at 95°C for 30 sec, annealing at 60°C for 30 sec, extension at 72°C for 30 sec, and a final extension at 72°C for 5 min. Amplified PCR products were sequenced using the forward or reverse PCR primers and Big Dye Terminator v3.0 cycle sequencing reagents (Applied Biosystems). PCR amplification consisted of 25 cycles of denaturation at 96°C for 10 sec, annealing at 50°C for 5 sec, and extension at 60°C for 4 min. Each DNA sequence was read on an ABIPRISM 3100 automated sequencer (Applied Biosystems).

Statistical analysis

Survival curves were constructed using the Kaplan–Meier method, and the log-rank test was used to evaluate differences in survival according to genetic alterations. All p-values were two sided, with p < 0.05 considered statistically significant.

Results

Clinical and pathological characteristics of the ATC patients

The clinical and pathological characteristics of the ATC patients in this study are presented in Table 1. The mean age of the 11 patients at thyroid surgery was 74.4 years (median 75 years), and eight patients (73%) were women. Nine patients underwent total thyroidectomy, and two patients underwent debulking surgery. The mean maximal diameter of the ATC was 4.5 cm (median 4.5 cm), and distant metastasis of the ATC was confirmed in six patients (55%) at initial diagnosis. Papillary microcarcinomas in the contralateral thyroid lobe were incidentally found in two patients. All patients died during follow-up, and the median survival time was 2.8 months.

Targeted next-generation sequencing

A total of 461 somatic variants were identified, and 76 of them were non-synonymous mutations: 43 coding missense mutations, 18 frameshift mutations, 10 non-sense mutations, 4 mutations at splicing sites, and 1 deletion. The overall mean target coverage, on target rate, and percentage of bases covered >20 times for the 11 ATC tumors were 147 reads, 36%, and 96%, respectively.

Among the 76 non-synonymous mutations, 28 were manually filtered out because of the low depth of each variant allele (the thresholds for MuTect and Somatic Indelocator were 5 and 10 counts, respectively), and the remaining 48 variations were considered true positive. These 48 variations were validated using the Sequenom MassARRAY system, and all mutations except one were confirmed. The mutant allele frequency of the unconfirmed genetic variation in targeted exome sequencing was 8%, close to the borderline of the detection limit of the MassARRAY system. The results of the targeted exome sequencing validated by the MassARRAY system are detailed in Figure 1 and Table 2.

FIG. 1.

Somatic mutations in anaplastic thyroid carcinomas (ATCs) detected by targeted massive parallel sequencing. Tumor samples are arranged from left to right. Somatic gene mutations found in 11 ATCs are annotated for each sample according to the color panel below the image. The somatic mutation frequencies for each mutated gene are plotted in the right panel. The mutation number, transition/transversion ratio, and type of base-pair substitution are displayed in the top, middle, and bottom panels, respectively.

Table 2.

Summary of Somatic Mutations Frequently Detected in the 11 ATCs in this Study

Gene	Mutation (amino acid)	Nucleotide change	Mutation type	No. of cases
ATC-related genes
BRAF	V600E	1799T>A	Missense	10
KRAS	G12V	35G>T	Missense	1
TP53	K132R	395A>G	Missense	1
	T125K	374C>A	Missense	1
	R248Q	743G>A	Missense	1
	Q331 ^*	991C>T	Non-sense	1
	Q317 ^*	949C>T	Non-sense	1
	X331	933 + 1G>T	Splicing site	1
	N131d	393_395delCAA	Deletion	1
	F328Sfs^17*	983del T	Frame shift	1
PIK3CA	G364R	1090G>A	Missense	1^a
	G364E	1091G>A	Missense	1^a
	E365K	1093G>A	Missense	1^a
	E545K	1633G>A	Missense	1
PTEN	V45Yfs^9*	133delG	Frame shift	1
Recurrently mutated novel genes
NF2	S228 ^*	863C>G	Non-sense	1
	Q470 ^*	1408C>T	Non-sense	1
	X200	600-2A>G	Splicing site	1
KMT2D	V5245G	15734T>G	Missense	1
	R3714K	11141G>A	Missense	1
PKHD1	Q1665 ^*	4993C>T	Non-sense	1
	E4004K	12010G>A	Missense	1

These three different mutations in PIK3CA were detected in one sample.

ATC-related gene mutations

The top 20 recurrently mutated genes were defined according to the COSMIC database of ATC as the ATC-related genes (18,19). All 11 ATCs had ATC-related gene mutations, such as BRAF, KRAS, TP53, PIK3CA, and PTEN mutations (Fig. 1 and Table 2). The most commonly mutated gene was BRAF, seen in 10 samples (91%), all showing the V600E point mutation. The KRAS point mutation was observed in the one sample (9%) without the BRAF^V600E mutation. A loss of function mutation of TP53 was observed in eight samples (73%): three missense mutations, two non-sense mutations, one frameshift mutation, one deletion, and one mutation at a splicing site. PIK3CA missense mutations were observed in two samples (18%). A frame shift deletion mutation of PTEN was observed in the one sample (9%) Two samples did not harbor additional ATC-related gene mutations other than the BRAF^V600E mutation.

Novel mutated genes in ATC

There were 32 somatic mutations in 28 novel genes in the 11 ATCs (Fig. 1 and Table 2). Of the 28 novel somatic mutated genes, NF2 was the most frequently mutated gene with a frequency of 27% (3/11 cases). Two of the NF2 mutations were non-sense mutations that were previously reported in the cancer genome atlas studies of lung cancers (20,21), and another new mutation was in a splice site. Novel KMT2D missense mutations were seen in two samples (18%). Novel PKHD1 mutations were also present in two samples (18%): one missense mutation and one non-sense mutation. The other 25 gene mutations were observed in one sample each.

All novel somatic mutations in the NF2, KMT2D, and PKHD1 genes were confirmed by direct Sanger sequencing (Supplementary Fig. S1). Additionally, RASAL1 mutations were evaluated by direct Sanger sequencing, and two missense mutations were detected among the 11 ATCs (18%; patients 1 and 11). These alterations were all located in the RAS GTPase-activating domain, as previously reported (12) (Supplementary Fig. S1).

Survival differences according to genetic alterations

Survival curves were constructed, and the differences were compared according to the mutations in the BRAF, TP53, NF2, KMT2D, PKHD1, and RASAL1 genes. The presence of a KMT2D or RASAL1 mutation was significantly associated with shorter ATC patient survival (p = 0.004 and p = 0.031, respectively). The presence of a TP53 or NF2 mutation also tended to affect the survival curves, but the results were not statistically significant (Fig. 2).

FIG. 2.

Survival curves of patients with anaplastic thyroid carcinomas according to genetic alterations. Presence of KMT2D and RASAL1 mutations was significantly associated with shorter survival of anaplastic thyroid carcinoma patients. (A) TP53, (B) NF2, (C) KMT2D, and (D) RASAL1.

Gene rearrangements in ATC

The presence of rearrangements were tested in 15 genes, including RET, ALK, BRAF, and AKT. However, there were no gene rearrangements in the 11 ATCs.

Comparison of the types of nucleotide substitutions between somatic and germline mutants

In the current study of 11 ATC cases, the transition (Ti)/transversion (Tv) ratio of germline single nucleotide polymorphisms (SNPs) was 2.41 (19,923/8264). However, in the case of somatic single nucleotide variants, the Ti/Tv ratio was dramatically decreased to 1.10 (76/69; Fig. 1).

Discussion

Targeted next-generation sequencing was performed using a massive parallel screening method. With this up-to-date technique, five ATC-related gene mutations and 28 novel gene mutations were detected in 11 ATCs. These data confirm the heterogeneous genetic background of ATCs.

The most frequently observed mutation in the ATCs was the BRAF^V600E mutation, with a prevalence of 91%. This prevalence is markedly higher than that of previous reports, which indicated that BRAF mutations occur in about 10–38% of ATCs (Table 3) (22 –24). The reason for this difference is unknown. However, the prevalence of the BRAF^V600E mutation of PTC in Korea has been reported to be higher than in other geographic regions, possibly due to differences in genetic backgrounds or iodine intake (25,26). All BRAF mutations were mutually exclusive with the RAS mutation, and only one KRAS codon 12 mutation was observed in this study. All 11 ATCs harbored a BRAF or RAS mutation, suggesting that PTC or FTC might progress to ATC with the accumulation of mutations. The additional loss of function mutation of TP53 is one of the mechanisms promoting the dedifferentiation of well-differentiated thyroid cancer (5,27), and it was found in 73% of the current ATC cases. Additional PI3KCA and PTEN mutations were observed in 18% and 9% of the ATCs, respectively. However, there were no additional mutations of known ATC-related genes other than BRAF or RAS in two ATCs, and the dedifferentiation mechanisms in these cases are uncertain.

Table 3.

Summary of Previous Report of BRAF Mutations in ATC

Author (publication year, reference number)	No. of ATCs evaluated	Prevalence of BRAF mutation (%)
Ricarte-Filho et al. (2009, 22)	18	38%
Xing et al. (2004, 23)	10	20%
Nikiforova et al. (2003, 24)	31	10%
COSMIC database (2015)	837	25%

A loss of function mutation of the NF2 gene was the most frequent novel mutation in the 11 ATC patients (3/11; 27%). The NF2 mutation is associated with neurofibromatosis, acoustic neuroma, and meningioma, as well as breast, colorectal, hepatic, and prostate cancer (28). NF2 is a tumor suppressor gene that encodes the protein merlin, which acts as a scaffold protein that modulates receptor-mediated signaling pathways such as receptor tyrosine kinase, small GTPase, PI3K/AKT/mTOR (mammalian target of rapamycin), and hippo pathways (28). One previous study reported a frequent allelic loss of NF2 in aggressive medullary thyroid carcinomas (29), and another recent study that performed whole exome sequencing of ATCs also reported that mutations in NF2 genes were novel driver mutations in ATCs (18). Furthermore, one recent study reported that NF2 inactivation increased MAPK output by promoting RAS transcription (30).

Mutations in KMT2D were also repeatedly observed in the ATCs in the present study. KMT2D, formerly named MLL2, is a histone methyltransferase that regulates gene transcription by modulating chromatin accessibility. KMT2D is frequently mutated in medulloblastoma, renal, bladder, gastric, hepatic, colorectal, and lung cancer, and previous studies have indicated that KMT2D is a tumor suppressor (31 –34). To the authors' knowledge, the current study is the first to report KMT2D mutations in thyroid cancer. Furthermore, patients with ATCs harboring KMT2D mutations had significantly shorter survival than those with ATCs without KMT2D mutations, suggesting a possible prognostic role of KMT2D mutations in ATC patients. Further studies are needed to uncover the detailed molecular mechanisms how loss of function mutations in NF2 or KMT2D affect the pathogenesis of ATC.

Inactivation of RASAL1, which encodes a RAS GTPase-activating protein, may promote ATC development by upregulating the PI3K pathway (12). The present study also found two samples with a RASAL1 mutation among the 11 ATCs by direct Sanger sequencing. These findings confirm the importance of the RASAL1 mutation in the pathogenesis of ATC. RASAL1 mutations were also associated with shorter patient survival in the present analysis. Further studies are needed to determine the prognostic role of RASAL1 mutations in ATC patients.

One limitation of the current study is that it involved a small series of 11 ATCs. Furthermore, not all mutated genes could be detected because this panel included just 505 known cancer-related genes. However, the number of subjects was comparable to previous studies of ATC because of its extremely low prevalence (6,35), and the present study is noteworthy because comprehensive analysis of ATCs was performed using targeted next-generation sequencing analysis with the MassARRAY process validation. Genetic changes detected as germline mutations were also excluded by comparing the results of the mutational profile in the ATCs with those of matched normal tissues. Significant differences in the Ti/Tv ratio between germline mutations (2.41) and somatic mutations (1.10) in the current study series confirmed the adequacy of the sequencing analysis. With regard to single nucleotide substitutions, all substitutions can be classified by six patterns of base substitutions: two types of Ti (C:G>T:A, A:T>G:C) and four types of Tv (C:G>A:T, C:G>G:C, A:T>C:G, A:T>T:A). During evolution, Ti are much more frequent than Tv, and the number of Ti over the number of Tv, referred to as the Ti/Tv ratio, in known SNPs is expected to be between two and four (36,37).

In conclusion, this comprehensive analysis of ATC using targeted massive parallel sequencing identified several novel mutations in ATCs, such as loss of function mutations in NF2 and KMT2D. Future studies are needed to confirm the role of these mutations as independent drivers in the development of ATC.

Footnotes

Acknowledgments

This study was supported by the National Research Foundation (NRF) of Korea Research Grant (NRF-2015R1C1A1A02036597).

Author Disclosure Statement

The authors have no competing interests to disclose.

References

Smallridge

. 2012. Approach to the patient with anaplastic thyroid carcinoma. J Clin Endocrinol Metab, 97:2566–2572.

Smith

, Nucera

. 2015. Personalized therapy in patients with anaplastic thyroid cancer: targeting genetic and epigenetic alterations. J Clin Endocrinol Metab, 100:35–42.

Guerra

, Di Crescenzo

, Garzi

, Cinelli

, Carlomagno

, Tonacchera

, Zeppa

, Vitale

. 2013. Genetic mutations in the treatment of anaplastic thyroid cancer: a systematic review. BMC Surg, 13:S44.

Ragazzi

, Ciarrocchi

, Sancisi

, Gandolfi

, Bisagni

, Piana

. 2014. Update on anaplastic thyroid carcinoma: morphological, molecular, and genetic features of the most aggressive thyroid cancer. Int J Endocrinol, 2014:790834.

Xing

. 2013. Molecular pathogenesis and mechanisms of thyroid cancer. Nat Rev Cancer, 13:184–199.

Murugan

, Xing

. 2011. Anaplastic thyroid cancers harbor novel oncogenic mutations of the ALK gene. Cancer Res, 71:4403–4411.

Liu

, Bishop

, Shan

, Pai

, Liu

, Murugan

, Sun

, El-Naggar

, Xing

. 2013. Highly prevalent TERT promoter mutations in aggressive thyroid cancers. Endocr Relat Cancer, 20:603–610.

Wagle

, Berger

, Davis

, Blumenstiel

, Defelice

, Pochanard

, Ducar

, Van Hummelen

, Macconaill

, Hahn

, Meyerson

, Gabriel

, Garraway

. 2012. High-throughput detection of actionable genomic alterations in clinical tumor samples by targeted, massively parallel sequencing. Cancer Discov, 2:82–93.

DeLellis

, Lloyd

, Heitz

, Eng

(Eds) 2004. Pathology and Genetics of Tumours of Endocrine Organs. IARC Press, Lyon.

10.

Murugan

, Bojdani

, Xing

. 2010. Identification and functional characterization of isocitrate dehydrogenase 1 (IDH1) mutations in thyroid cancer. Biochem Biophys Res Commun, 393:555–559.

11.

Hemerly

, Bastos

, Cerutti

. 2010. Identification of several novel non-p.R132 IDH1 variants in thyroid carcinomas. Eur J Endocrinol, 163:747–755.

12.

Liu

, Yang

, Bojdani

, Murugan

, Xing

. 2013. Identification of RASAL1 as a major tumor suppressor gene in thyroid cancer. J Natl Cancer Inst, 105:1617–1627.

13.

, Durbin

. 2009. Fast and accurate short read alignment with Burrows–Wheeler transform. Bioinformatics, 25:1754–1760.

14.

McKenna

, Hanna

, Banks

, Sivachenko

, Cibulskis

, Kernytsky

, Garimella

, Altshuler

, Gabriel

, Daly

, DePristo

. 2010. The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res, 20:1297–1303.

15.

Cibulskis

, Lawrence

, Carter

, Sivachenko

, Jaffe

, Sougnez

, Gabriel

, Meyerson

, Lander

, Getz

. 2013. Sensitive detection of somatic point mutations in impure and heterogeneous cancer samples. Nat Biotechnol, 31:213–219.

16.

Abo

, Ducar

, Garcia

, Thorner

, Rojas-Rudilla

, Lin

, Sholl

, Hahn

, Meyerson

, Lindeman

, Van Hummelen

, MacConaill

. 2015. BreaKmer: detection of structural variation in targeted massively parallel sequencing data using kmers. Nucleic Acids Res, 43:e19.

17.

Arcila

, Lau

, Nafa

, Ladanyi

. 2011. Detection of KRAS and BRAF mutations in colorectal carcinoma roles for high-sensitivity locked nucleic acid-PCR sequencing and broad-spectrum mass spectrometry genotyping. J Mol Diagn, 13:64–73.

18.

Kunstman

, Juhlin

, Goh

, Brown

, Stenman

, Healy

, Rubinstein

, Choi

, Kiss

, Nelson-Williams

, Mane

, Rimm

, Prasad

, Hoog

, Zedenius

, Larsson

, Korah

, Lifton

, Carling

. 2015. Characterization of the mutational landscape of anaplastic thyroid cancer via whole-exome sequencing. Hum Mol Genet, 24:2318–2329.

19.

Forbes

, Beare

, Gunasekaran

, Leung

, Bindal

, Boutselakis

, Ding

, Bamford

, Cole

, Ward

, Kok

, Jia

, De

, Teague

, Stratton

, McDermott

, Campbell

. 2015. COSMIC: exploring the world's knowledge of somatic mutations in human cancer. Nucleic Acids Res, 43:D805–811.

20.

Network TCGAR. 2012. Comprehensive genomic characterization of squamous cell lung cancers. Nature, 489:519–525.

21.

Cancer Genome Atlas Research Network. 2014. Comprehensive molecular profiling of lung adenocarcinoma. Nature, 511:543–550.

22.

Ricarte-Filho

, Ryder

, Chitale

, Rivera

, Heguy

, Ladanyi

, Janakiraman

, Solit

, Knauf

, Tuttle

, Ghossein

, Fagin

. 2009. Mutational profile of advanced primary and metastatic radioactive iodine-refractory thyroid cancers reveals distinct pathogenetic roles for BRAF, PIK3CA, and AKT1 . Cancer Res, 69:4885–4893.

23.

Xing

, Vasko

, Tallini

, Larin

, Wu

, Udelsman

, Ringel

, Ladenson

, Sidransky

. 2004. BRAF T1796A transversion mutation in various thyroid neoplasms. J Clin Endocrinol Metab, 89:1365–1368.

24.

Nikiforova

, Kimura

, Gandhi

, Biddinger

, Knauf

, Basolo

, Zhu

, Giannini

, Salvatore

, Fusco

, Santoro

, Fagin

, Nikiforov

. 2003. BRAF mutations in thyroid tumors are restricted to papillary carcinomas and anaplastic or poorly differentiated carcinomas arising from papillary carcinomas. J Clin Endocrinol Metab, 88:5399–5404.

25.

Hong

, Lim

, Kim

, Choi

, Yoo

, Min

, Won

, Lee

, Jung

, Park do

, Park

. 2014. The frequency and clinical implications of the BRAF(V600E) mutation in papillary thyroid cancer patients in Korea over the past two decades. Endocrinol Metab (Seoul), 29:505–513.

26.

Kim

, Kim

, Song

, Rhee

, Gong

, Cho

, Kim

, Hong

, Shong

. 2005. The BRAF mutation is not associated with poor prognostic factors in Korean patients with conventional papillary thyroid microcarcinoma. Clin Endocrinol (Oxf), 63:588–593.

27.

Smallridge

, Marlow

, Copland

. 2009. Anaplastic thyroid cancer: molecular pathogenesis and emerging therapies. Endocr Relat Cancer, 16:17–44.

28.

Petrilli

, Fernandez-Valle

. 2016. Role of Merlin/NF2 inactivation in tumor biology. Oncogene, 35:537–548.

29.

Sheikh

, Tometsko

, Niehouse

, Aldeeb

, Swalsky

, Finkelstein

, Barnes

, Hunt

. 2004. Molecular genotyping of medullary thyroid carcinoma can predict tumor recurrence. Am J Surg Pathol, 28:101–106.

30.

Garcia-Rendueles

, Ricarte-Filho

, Untch

, Landa

, Knauf

, Voza

, Smith

, Ganly

, Taylor

, Persaud

, Oler

, Fang

, Jhanwar

, Viale

, Heguy

, Huberman

, Giancotti

, Ghossein

, Fagin

. 2015. NF2 loss promotes oncogenic RAS-induced thyroid cancers via YAP-dependent transactivation of RAS proteins and sensitizes them to MEK inhibition. Cancer Discov, 5:1178–1193.

31.

Guo

, Chen

, Huang

, Chang

, Wang

, Pirozzi

, Qin

, Bao

, Greer

, McLendon

, Yan

, Keir

, Bigner

, He

. 2013. KMT2D maintains neoplastic cell proliferation and global histone H3 lysine 4 monomethylation. Oncotarget, 4:2144–2153.

32.

Zang

, Cutcutache

, Poon

, Zhang

, McPherson

, Tao

, Rajasegaran

, Heng

, Deng

, Gan

, Lim

, Ong

, Huang

, Chin

, Tan

, Ng

, Yu

, Wu

, Lee

, Wu

, Poh

, Wan

, Rha

, So

, Salto-Tellez

, Yeoh

, Wong

, Zhu

, Futreal

, Pang

, Ruan

, Hillmer

, Bertrand

, Nagarajan

, Rozen

, Teh

, Tan

. 2012. Exome sequencing of gastric adenocarcinoma identifies recurrent somatic mutations in cell adhesion and chromatin remodeling genes. Nat Genet, 44:570–574.

33.

Liu

, Morrison

, Wang

, Xiong

, Vedell

, Cui

, Hua

, Ding

, Lu

, James

, Ebben

, Xu

, Adjei

, Head

, Andrae

, Tschannen

, Jacob

, Pan

, Zhang

, Van den Bergh

, Xiao

, Lo

, Patel

, Richmond

, Watt

, Albert

, Selzer

, Anderson

, Wang

, Starnes

, Yang

, You

. 2012. Identification of somatic mutations in non-small cell lung carcinomas using whole-exome sequencing. Carcinogenesis, 33:1270–1276.

34.

Yin

, Yang

, Lin

, Deng

, Zhang

, Yi

, Shi

, Tao

, Cai

, Wu

, Zhao

, Hurst

, Zhang

, Hu

, Kong

. 2014. Exome sequencing identifies frequent mutation of MLL2 in non-small cell lung carcinoma from Chinese patients. Sci Rep, 4:6036.

35.

Begum

, Rosenbaum

, Henrique

, Cohen

, Sidransky

, Westra

. 2004. BRAF mutations in anaplastic thyroid carcinoma: implications for tumor origin, diagnosis and treatment. Mod Pathol, 17:1359–1363.

36.

DePristo

, Banks

, Poplin

, Garimella

, Maguire

, Hartl

, Philippakis

, del Angel

, Rivas

, Hanna

, McKenna

, Fennell

, Kernytsky

, Sivachenko

, Cibulskis

, Gabriel

, Altshuler

, Daly

. 2011. A framework for variation discovery and genotyping using next-generation DNA sequencing data. Nat Genet, 43:491–498.

37.

Alexandrov

, Nik-Zainal

, Wedge

, Aparicio

, Behjati

, Biankin

, Bignell

, Bolli

, Borg

, Borresen-Dale

, Boyault

, Burkhardt

, Butler

, Caldas

, Davies

, Desmedt

, Eils

, Eyfjord

, Foekens

, Greaves

, Hosoda

, Hutter

, Ilicic

, Imbeaud

, Imielinski

, Jager

, Jones

, Knappskog

, Kool

, Lakhani

, Lopez-Otin

, Martin

, Munshi

, Nakamura

, Northcott

, Pajic

, Papaemmanuil

, Paradiso

, Pearson

, Puente

, Raine

, Ramakrishna

, Richardson

, Richter

, Rosenstiel

, Schlesner

, Schumacher

, Span

, Teague

, Totoki

, Tutt

, Valdes-Mas

, van Buuren

, van ’t Veer

, Vincent-Salomon

, Waddell

, Yates

, Zucman-Rossi

, Futreal

, McDermott

, Lichter

, Meyerson

, Grimmond

, Siebert

, Campo

, Shibata

, Pfister

, Campbell

, Stratton

. 2013. Signatures of mutational processes in human cancer. Nature, 500:415–421.