the role of molecular markers in thyroid nodule...

Yuri E. Nikiforov, MD, PhDDivision of Molecular & Genomic PathologyUniversity of Pittsburgh Medical Center

The role of Molecular Markers in Thyroid Nodule Management:

Then and Now

Disclosures

• Quest Diagnostics (consultant)• Service agreement between UPMC and

CBLPath to offer ThyroSeq for commercial use

Progress in Identifying Driver Mutations in Thyroid Cancer

Beginning of the molecular era…

Genetic Basis of Thyroid Cancer

Major pathways involved: MAPK, PI3K Main mutation mechanisms: Point 

mutations and chr. rearrangements

Genomic Landscape of PTC

The Cancer Genome Atlas Research Network. Cell 159:676-690 (2014)

74%

15%

9%

Point mutations

Gene fusions

Copy number variations

Driver mutations define carcinogenic mechanisms BRAFV600E-like and RAS-like scores

The Cancer Genome Atlas Research Network. Cell 159:676-690 (2014)

Thyroid Differentiation Status in PTC

The Cancer Genome Atlas Research Network. Cell 159:676-690 (2014)

Lawrence MS et al. Nature. 2013 Jul 11; 499(7457): 214–218

Frequencies of Somatic Mutation in Different Tumors

Molecular Markers for Cancer Diagnosis

Diagnostic Category Risk                 Usualof cancer  management

I.   Nondiagnostic or Unsatisfactory 1‐4% Repat FNA with US

II.  Benign  0‐3% Clinical follow‐up

III. Atypia of Undetermined Significance   5‐15% Repeat FNAor Follicular Lesion of UndeterminedSignificance (AUS/FLUS)

IV. Follicular Neoplasm or Suspicious  15‐30% Surgical lobectomyfor a Follicular Neoplasm (FN/SFN)

V. Suspicious for malignancy  60‐75% Total or lobectomy

VI.Malignant     97‐99% Total thyroidectomy

Baloch ZW et al. Diagn Cytopathol 2008;36:425Cibas ES & Ali SZ. Am J Clin Pathol 2009;132:658

The Bethesda System for Reporting Thyroid Cytopathology

• Gene mutations

• Gene expression (mRNA)

• Circulating TSHR mRNA

• miRNAs

• Proteomics

• Combination of marker types 

Molecular Markers for Thyroid Cytology

Progress in Composition of Gene Mutation Panels

Single genes

Conventional sequencing

7-gene panel

Multi-gene panel

Next-Generation sequencing

Conventional sequencing

• Meta‐analysis of 47 studies• 9,924 FNA tested for BRAF V600E• Specificity – ~100%• Sensitivity – 52% (95%CI 39%‐64%)

Single gene tests - BRAF V600E

High specificity/PPV, very low sensitivity/NPV – good “rule in” test

J Clin Endocrinol Metab 94: 2092–2098, 2009

Small (7-gene) panels

Gene Fusions (RNA)

RET/PTC1

RET/PTC3

PAX8/PPARG

Gene Mutations (DNA)

BRAF

NRAS HRAS

KRAS

none

Small (7-gene) panels

High specificity/PPV, low sensitivity/NPV – good “rule in” test

Nishino M. Cancer Cytopathol 2016;124:14-27

Afirma Gene Expression Classifier (GEC)

N Engl J Med 2012; 367:705-715

Afirma Gene Expression Classifier

Alexander E et al. N Engl J Med (2012)

• Multi-institutional double-blind prospective study of indeterminate cytology FNA samples

• Sample size – 265 FNAs

Cytologicdiagnosis

n Sensitivity Specificity NPV PPV

AUS/FLUS 129 90% 53% 95% 38%

FN/SFN 81 90% 49% 94% 37%

SUSP 55 94% 52% 85% 76%

High sensitivity/NPV, low specificity/PPV – good “rule out” test

Afirma Gene Expression Classifier

Genomic Revolution:Next Generation (Deep) Sequencing

Next Generation Sequencing

• Sequence up to 106-109 bases• Cost - $0.05-1 per 106 bases• Sensitivity - 3-5% of mutant allele

Conventional Sequencing

• Sequence up to 102-103 bases• Cost - $2400 per 106 bases• Sensitivity - 20-30% of mutant allele

Expansion of Molecular Panels Using Next-Gen Sequencing

PAX8/PPARγ

RASRET/PTC1RET/PTC3

BRAF

RET

PTEN

BRAF other

TSHRPIK3CA

AKT1

CTNNB1

TP53

13%

Novel mutations(eg. TERT, EIF1AX) 

Novel gene fusions(eg. ALK, NTRK3)12%

90%65%

ThyroSeq® v.2 NGS Mutation Panel

Gene Fusions(mRNA) RETPPARG

NTRK1

NTRK3

BRAF

ALK

Other

Gene Mutations(DNA)BRAF RET

NRAS  TSHR

HRAS AKT1

KRAS TP53

PIK3CA GNAS

PTEN CTNNB1

TERT EIF1AX

• 14 genes for mutations; 42 fusion types; 16 genes for expression

Gene expression(mRNA)PGK1KRT7TGTTF1NISCalcitonin

PTH

KRT20

Other

– pan-cell marker

Thyroid epithelial cells

– MTC– parathyroid

– metastatic

• 465 consecutive FNA samples with AUS/FLUS (Bethesda III)  cytology at UPMC from May 2014‐March 2015 

• Prospective study• Surgical outcome known for 96 patients

• Cancer prevalence after surgery – 22.5%

Nikiforov et al. Thyroid 2015;25:1217-23

ThyroSeq v2 Performance in AUS/FLUS (Bethesda III) Cytology Nodules

Sensitivity 91% (CI: 78-100%)Specificity 92% (CI: 86-98%)NPV 97% (CI: 93-100%)PPV 77% (CI: 61-93%)Accuracy 92% (CI: 86-97%)

Nikiforov et al. Thyroid 2015;25:1217-23

High specificity/PPV and high sensitivity/NPV – good “rule in” and “rule out” test

Cancer 2014, 120:3627-34

• Patients with FN/SFN (Bethesda IV)  cytology and known surgical outcome seen at UPMC from Oct. 2013‐May 2014

• Retrospective and prospective groups  

• Sample size: 143 consecutive FNA samples

• Cancer prevalence after surgery – 39/143 (27.3%)

143 consecutive FN/SFN nodules with surgery

Overall test performance

Retrospective groupn=91

Mutation NEGATIVE

n=64

Mutation POSITIVE

n=27

CANCERn=2

BENIGNn=62

BENIGNn=4

Sensitivity 92%Specificity 94%PPV 85%NPV 97%Accuracy 93%

CANCERn=23

Prospective groupn=52

Mutation NEGATIVE

n=37

Mutation POSITIVE

n=15

BENIGNn=35

BENIGNn=3

Sensitivity 86%Specificity 92%PPV 80%NPV 95%Accuracy 90%

CANCERn=2

CANCERn=12

Sensitivity 90% (CI: 80-99%)Specificity 93% (CI: 88-98%)NPV 96% (CI: 92-95%)PPV 83% (CI: 72-95%)

Accuracy 92% (CI: 88-97%)

Nikiforov et al. Cancer 2014,120:3627-34

ThyroSeq v2 NPV

Afirma NPV

Afirma PPV

ThyroSeq v2 PPV

Nikiforov et al. Thyroid 2015;25:1217-23

Test Performance in AUS/FLUS (Bethesda III) Cytology Nodules

ThyroSeq v2

Afirma

*Unless AF>30%

BRAFALKNTRK1,3PPARGRET/PTCTERTTP53

NRASHRASKRAS

PTENEIF1AX

TSHRGNAS

Cancer risk >95% 80% 20‐40% Very low*

Cancer risk associated with individual mutations

Cancer Risk in Nodules with RAS Mutations

RASmutation‐positive “benign” nodules 

79%

Evidence for clonal neoplasm and early transformation to cancer

Nikiforov et al. Thyroid 2015;25:1217-23

Cancer Risk in Nodules with RAS Mutations

Medicine 2015;94:1-6

Total FNA samples  198

Total RAS+ 31

RAS+ Cancer 7 (23%)

RAS+ Benign 24 (77%)

BioMed Res Int 2015;2015:697068

Total FNA samples  132

Total RAS+ 27

RAS+ Cancer 26 (96%)

RAS+ Benign 1 (4%)

JCEM 2013;98:E829

all RAS 40%

J Clin Endocrinol Metab 100: 2743–2750, 2015

• Combination of 7-gene mutation panel (miRInform) and miRNA expression classifier based on 10 miRNA genes (ThyraMIR)

• Validated in 109 FNA samples with AUS/FLUS and FN/SFN cytology• Cancer prevalence 32%• Multi-center study (12 endocrinology centers)

J Clin Endocrinol Metab 100: 2743–2750, 2015

Performance of combined 7-gene mutation panel (miRInform) and expression of 10 miRNA genes test

Cancer Cytopathol 2016;124:14-27

Nishino M. Cancer Cytopathol 2016;124:14-27

*Depending on the type of mutation  Nikiforov et al. Cancer 2014, 120:3627-34Nikiforov et al. Thyroid 2015, 25:1217-23Data on file; UPMC MGP lab

FNA

FLUS (BC III)

FN/SFN(BC IV)

SUSP(BC V)

Cancer Risk 10‐15% 50‐70%

ThyroSeq v2 ‐ +

Therapeutic Surgery

Modified Cancer Risk 3% 75‐99%*

Cytology

20‐30%

Observation

‐ +

Therapeutic Surgery

4% 75‐99%*

Observation

‐ +

Therapeutic Surgery

20% 90‐99%*

LobectomyRecommended Management

Cancer risk and proposed patient management based on combination of cytology and ThyroSeq test

Molecular Markers for Cancer Prognostication

Yip et al. Ann Surg 262:519‐25 (2015)

Mean follow‐up 33 ± 21.2 months

BRAF V600E status and tumor recurrence and mortality

Gene Panels for Tumor Prognostication

Xing M et al. JAMA (2013)Xing M et al. JCEM (2005) Kim TY et al. Clin Endocrinol(2006)

N=219 N=203

All Variants of PTC included Conventional PTC Only All Patients with PTC

N=1849

BRAF V600E mutation

N Sensitivity Specificity NPV PPVMortalityTufano et al. Medicine (2012)

2167 83% 51% 99% 5%

RecurrenceXing M al. JAMA (2013)

1849 66% 54% 87% 25%

Thyroid Cancer Prognostication Aggressive Tumors Have Multiple Mutations

Ricarte-Filho JC et al. Cancer Res (2009)

Papillary Carcinomas

Nikiforova M et al. JCEM (2013)

n=57BRAFNRASHRASKRAS

PIK3CATSHRTP53

0                                       20                                      40  

• 469 patients with FCDTC• Mean follow‐up 7.8±5.8 years

Melo M et al. JCEM (2014)

Aggressive Tumors Have TERT Mutations

Follow‐up (years)

FCDTC Papillary Carcinoma Follicular Carcinoma

Cum Survival

TERT wt TERT wtTERT wt

TERT mutTERT mut TERT mut

Univariateanalysis

Multivariate analysis

Univariateanalysis

Multivariate analysis

Presence (%) OR (95%CI) P value OR  (95%CI) P value Presence (%) OR  (95% CI) P value OR  (95%CI) P value

TERT wt 34 (12.5) 1 <0.001 1 0.002 48  (24.6) 1 0.001 1 0.007

TERT mut 10 (43.5) 5.36 (2.18‐13.18)

4.60 (1.73‐12.21)

10 (62.5) 5.10 (1.76‐14.78)

4.68 (1.54‐14.27)

Thyroid Cancer Prognostication

Xing M. et al. JCO (2014)

PTC (all types) PTC (conventional type)

BRAF+TERTBRAF+TERT

III. Aggressive Tumors Have TERT MutationsThyroid Cancer Prognostication

Molecular Signature

• BRAF + TERT• Multiple driver mutations  (eg. BRAF and PIK3CA)

• TP53• TERT

• ALK fusions• NTRK1 fusions• NTRK3 fusions• BRAF V600E

• RET/PTC • RAS• PTEN• BRAF K601E• PAX8/PPARG

Risk of Recurrence

Preoperative cancer risk stratification based on molecular profiling

High Risk

Low Risk

Intermediate Risk

Case #170 yo F with 1.2 cm nodule

FNA Cytology: • Atypia of Undetermined significance 

Case #474 yo F with 3.7 cm nodule

Case #262 yo F with 2.1 cm nodule

Case #341 yo M with 1.6 cm nodule

Case #1: 70 year-old woman with incidentally noted 1.2 cm thyroid nodule

Cytology: Satisfactory/FLUS• Cancer risk 5‐15%

Left lobe normal, no risk factors

Molecular Testing:

Observation:Neck exam and ultrasound stable at 5 years follow‐up

ThyroSeq v2 (2014)No mutations identified• Cancer risk ‐ 3%

7‐gene panel (2010)No mutations identified• Cancer risk ‐ 6%

Case #2: 62 year-old woman with 2.1 cm thyroid nodule

Total thyroidectomy:PTC with extrathyroidal extension

Cytology:Follicular lesion of undetermined significance (FLUS) (Bethesda III) due to low cellularity  • Cancer risk 5‐15%

ThyroSeq:BRAF V600E mutation is identified(9% AF/ 18% cells with mutation)• Cancer risk ~99%

Case #3: 41 yo male with left inferior pole thyroid nodule

Cytology:Groups of microfollicles with atypia (Bethesda III‐IV)• Cancer risk for FLUS 5‐15%, FN – 20‐30%

Parathyroid exploration, left thyroidectomy:Parathyroid adenoma, left inferior, intrathyroidal 

US: 1.6 x 1.2 x 1.2 cm predominantly solid, isoechoic nodule in left inferior pole, not convincingly border forming

ThyroSeq:‐Mutations NOT identified‐ High expression (92%) of the parathyroid hormone (PTH) gene DETECTED

• Strong evidence for parathyroid tissue

US: Solitary 3.7 x 2.6 x 2.8 cm, solid, isoechoic, hypervascular, circumscribed nodule with calcifications

Cytology: Hurthle cell nodule (Bethesda III/IV)Cancer risk for FLUS 5‐15%, FN – 20‐30%

Case #4: 74 yo female with solitary right lobe nodule, recently increased in size

ThyroSeq:Positive for NRAS (Q61R); TP53 (R175H); PIK3CA (E545K) mutations

Total thyroidectomy:Oncocytic follicular carcinoma with capsular and multifocal vascular invasion  (3 foci)

Case #170 yo F with 1.2 cm nodule

FNA Cytology: • Atypia of Undetermined significance

Case #474 yo F with 3.7 cm nodule

Case #262 yo F with 2.1 cm nodule

Case #341 yo M with 1.6 cm nodule

Observation Therapeutic surgery

Dx of parathyroid disease

Therapeutic surgery Risk prediction

Precision MedicinePrecision Medicine refers to the tailoring of medical treatment to the individualcharacteristics of each patient. It does not literally mean the creation of drugs ormedical devices that are unique to a patient, but rather the ability to classifyindividuals into subpopulations that differ in their susceptibility to a particulardisease, in the biology and/or prognosis of those diseases, or in their response to aspecific treatment. (National Research Council)

AcknowledgementsNikiforov’s Research LaboratoryLindsey KellyRebecca Leeman-Neill Chan Kwon JungViktoria EvdokimovaManoj GandhiRaffaele Ciampi

University of Pittsburgh PathologyMarina NikiforovaPaul OhoriRaja Seethala

Endocrine SurgerySally CartyLinwah YipMichael Stang (now at Duke)Kelly McCoy

Head and Neck SurgeryRobert FerrisUmamaheswar Duvvuri

EndocrinologyShane LeBeauStephen Hodak (now at NYU)Christopher Coyne

RadiologyMitchell Tublin

PharmacologyDanny Altschuler

Grants/Research support:• NIH, ACS, University of Pittsburgh Cancer Institute• Philanthropic support from grateful patients