“Next Generation Sequencing: What Will

it Mean for the Pathologist?”

Christopher Corless, MD, PhD

Professor of Pathology, Oregon Health & Science University

Chief Medical Officer, Knight Diagnostic Laboratories

Disclosures

Company Activity

Ion Torrent/ThermoFisher Travel support; reagents

Ventana/Roche Travel support

Genentech/Roche Honoraria & consulting fees

Novartis Consulting fees

I will mention off-label uses of cancer therapeutics

Topics

• Brief introduction to next-gen sequencing (NGS)

• Critical pre-analytical factors

• Usefulness of quantitative output from NGS

– Detecting mutations – correlation with specimen

– Fusion gene detection

– Copy number assessment

• Data overload and implications for pathologists

42 year old male with ‘glioblastoma’ treated with surgery, temozolomide and radiation

Bone and lymph node mets appeared at 20 months (what is this tumor?)

Admitted to OHSU to manage pain, monitor pending cord compression

Another round of chemo failed

BRAF V600E mutation identified

Patient started on dabrafenib

Excellent clinical response

Individualized Cancer Medicine - An Example -

Garraway, JCO epub April 15, 2013

SHC GRB2

SOS

BRAF

MEK

ERK

P

P

P

P P P

P

P

P

P

PI3K PDK

AKT

P

mTOR

S6K

P

P

P

PTEN

>500 Targeted Therapeutics in Development

Receptor tyrosine kinases

ALK or ROS1

Erlotinib

Lapatinib

Sorafenib

Imatinib

Sunitinib

Afatinib

Dovitinib

Vemurafenib

Dabrafenib

LGX818

Trametinib

Selumetinib

Benitinib

Everolimus

Temsirolimus

Buparlisib

BGT226

BYL719

Crizotinib

Ceritinib

MK2206

SR13668

Trastuzumab

Cetuximab

Panitumumab

METMab

CDK4/6 P1446A-05

LEE001

NRAS

KRAS

HRAS

BEZ235

DNA Sequencing Traditional Sanger Method

Capillary electrophoresis Final Sequence

BRAF V600E

Next-Generation DNA Sequencing • Massively parallel sequencing (many

sequencing reactions performed simultaneously)

ACTGGTCCTGCTGGTTAG

ACTGGTCCTGCTGGTTAG

ACTGGTCCTGCTGGTTAG

ACTGGTCCTGCTGGTTAG

ACTGGTCCTGCTGGTTAG

ACTGGTCCAGCTGGTTAG

ACTGGTCCAGCTGGTTAG

ACTGGTCCAGCTGGTTAG

ACTGGTCCAGCTGGTTAG

G T C

V

842

Frameshift

G A C A T C A T G C ……

G T C A T A T G C …… [Stop]

V A L

PDGFRA D842V and M844fs*16

Gastrointestinal Stromal Tumor (GIST)

Known GIST driver mutation Truncates kinase domain

A S P

Applications of Next-Gen Sequencing

• Broad-based approaches

– Whole genome: all 3.2 billion base pairs

– Whole transcriptome: all mRNAs in a sample

– Whole exome: ~1.6% of genome that encodes proteins

• Targeted approaches

– Panels of genes (10-400)

– Limited to targets regarded as actionable or prognostic

Ion Torrent Proton Illumina NextSeq 500

Ion Torrent PGM Illumina MiSeq (FDA-approved version available)

NGS Sequencers in Clinical Labs

Next-Gen Sequencing Read Out

1. Base-calling: sequencer generates millions of ‘reads’ (50 to 200 base pair stretches of DNA)

2. Alignment: software matches the reads to the known genome

3. Variant calling: software determines where the sequence differs from a standard reference genome

4. Variant annotation: software programs + manual review to determine whether a variant is a:

– Benign polymorphism

– Variant of unknown significance

– Mutation (single nucleotide change, deletion, etc)

NGS Sequencing Methodology

Hybrid-Capture Amplicon

Applications Whole exome; larger panels

Smaller panels

Input DNA 100 ng 10-20 ng

Overnight hybridization Yes No

Gene copy number Yes Yes

Fusion gene detection Yes Yes (RNA)

• 5 x 4 μm unstained sections of 1 cm core biopsy

is sufficient for amplicon-based method • New extraction method may work with just 4 mm2 from a single section

How Much Tissue Is Needed?

• Specimen decalcification

• Acid is often used to soften the bone for sectioning and this degrades nucleic acids

• EDTA-based solutions are better

Result

Mimics a C>T mutation

• DNA deamination is common in older paraffin blocks

Pre-Analytic Factors Influencing NGS

• Tumor fraction: % of specimen comprised of tumor

Impact of Tumor Purity

Mutant Allele Ratio in Tumor Cells

Percent Tumor in Specimen

Mutant Allele Ratio in Sequenced DNA

50% 90% 45%

50% 70% 35%

50% 50% 25%

50% 30% 15%

50% 10% 5%

• Each cellular nucleus in a specimen contributes DNA • % Tumor fraction is based on tumor nuclei, not area • The lower limit of sensitivity for most NGS panels is ~2-3%

Selecting Tumor-Rich Material

for DNA Extraction

Coring

a block Scraping

slides

Metastatic

Melanoma in Lung

Tumor purity: ~90%

Metastatic

Colorectal

Adenocarcinoma

in Mesentery

Tumor purity: ~10-20%

Transbronchial Biopsy

Adenocarcinoma of the Lung

Please,

get another

sample!

• NRAS p.Q61R (27% mutant allele)

• TP53 p.R282W (26% mutant allele)

• Tumor fraction: 52-54%

Metastatic Colorectal Adenocarcinoma

Estimated Tumor Fraction

• Microscope: 60%

Quantitating

Tumor Nuclei

Using

CellProfiler

Semi-automated

using

machine learning

algorithm

• Tumor nuclei

• Normal nuclei

KRAS

EGFR

ALK BRAF

PIK3CA ERBB2

MAPK1

NTRK1

NTRK3

MET ROS1

HRAS

RET

DDR2

PTEN

ARAF

FGFR3

FGFR2

FGFR1

Unknown

Molecular Subtypes of NSCLC Before 2000

KRAS

EGFR

ALK BRAF

PIK3CA ERBB2

MAPK1

NTRK1

NTRK3

MET ROS1

HRAS

RET

DDR2

PTEN

ARAF

FGFR3

FGFR2

FGFR1

Unknown

Molecular Subtypes of NSCLC in 2003

KRAS

EGFR

ALK BRAF

PIK3CA ERBB2

MAPK1

NTRK1

NTRK3

MET ROS1

HRAS

RET

DDR2

PTEN

ARAF

FGFR3

FGFR2

FGFR1

Unknown

Molecular Subtypes of NSCLC in 2015

FDA approved drugs

FDA approved drugs in other cancers

In clinical trials

NSCLC Case Example

• 47 y/o woman with treatment refractory

bronchioalveolar carcinoma

• Genotyping: BRAF V600E mutation

• Phase I study combining BRAF + MEK inhibitors: 6 month response

• Disease controlled for 7 months; she died about 1 year later

Baseline 2 months

http://www.oregonlive.com/health/index.ssf/

2012/12/new_oregon_health_science_univ_1.html

Interim results of phase II study BRF113928 of dabrafenib in BRAF V600E

mutation–positive non-small cell lung cancer (NSCLC) patients.

J Clin Oncol 31, 2013 (suppl; abstr 8009)

KRAS

EGFR

ALK BRAF

PIK3CA ERBB2

MAPK1

NTRK1

NTRK3

MET ROS1

HRAS

RET

DDR2

PTEN

ARAF

FGFR3

FGFR2

FGFR1

Unknown

GENE FUSIONS

Molecular Subtypes of Lung Cancer 2015

ROS1 fusions in lung adenocarcinoma

• Discovery published in Jan. 2012

• Trial published Nov. 2014

• FDA-approved tx in Apr. 2015

• 2nd Line drug published in Oct. 2014

Kinase

Fusion FISH IHC RNAseq

ALK Yes Yes Yes

AXL Yes ?? Yes

BRAF Yes No Yes

FGFR1 Yes No Yes

FGFR2 Yes No Yes

FGFR3 Yes No Yes

MET Yes No Yes

NTRK1 Yes ?? Yes

NTRK3 Yes ?? Yes

PGDFRA Yes No Yes

RET Yes ?? Yes

ROS1 Yes Yes Yes

Detecting Actionable Gene Fusions In NSCLC

Kohno et al. Trans Lung Cancer Res. Vol 4, No 2 (April 2015)

Target Genes (169 potential fusions / 94 partners)

ALK BRAF FGFR1 FGFR2 FGFR3 MET

NTRK1 NTRK2 NTRK3 PDGFRA PDGFRB RAF1

RET ROS1 TMPRSS2 EGFR EGFRvIII ERBB4

NRG1 AKT3

Gene Expression

ALK ROS1 RET EGFR NTRK1 NTRK2

NTRK3 BRAF Housekeeping genes (5)

Also useful for: • Cholangioca • Bladder ca • HNSQCC

Beadling et al., submitted

Detecting Copy Number Alterations by NGS

FGFR1

ERRB2/HER2

Normal

(2 Copies)

Breast Carcinoma With FGFR1 and HER2 Amplification

Validation

Samples

•Known

FISH status

•Microarray

results

Grasso et al. J Mol Diag 17:53-63; 2015

FFPE Ductal Carcinoma of the Breast

EGFR

EGFR Amplification in Breast Carcinoma

68 year old F

Tumor metastatic

to the liver

HER2-negative

[6% of breast

carcinomas have

EGFR amplif]

Bhargava et al. Mod Pathol. 2005 Aug;18(8):1027-33

Example of Variant List For a 37-Gene Panel Head & Neck Squamous Cell Carcinoma

Chrom Position_Start Position_End Ref Variant Type Consequence Zygosity Var_Freq Gene p_AA_change

chr4 55968089 55968089 T G SNP nonsynonymous Het 61.29 KDR p.K747N

chr9 21971096 21971096 C A SNP nonsynonymous Het 83.33 CDKN2A p.E88*

chr17 7577548* 7577548 CGCCCA C DEL stop gain Het 30.52 TP53 p.G245G

chr17 7577557* 7577557 A C SNP stop gain Het 31.28 TP53 p.C242G

chr17 7577558* 7577558 G C SNP stop gain Het 32.33 TP53 p.S241S

chr7 55229255 55229255 G A SNP nonsynonymous Het 23.13 EGFR p.R521K

chr17 7579472 7579472 G C SNP nonsynonymous Het 31.25 TP53 p.P72R

chr17 37884037 37884037 C G SNP nonsynonymous Het 60 ERBB2 p.P1170A

chr17 37879588 37879588 A G SNP nonsynonymous Het 65.48 ERBB2 p.I655V

chr2 29416366 29416366 G C SNP nonsynonymous Het 92.68 ALK p.D1529E

chr2 29416572 29416572 T C SNP nonsynonymous Hom 99.34 ALK p.I1461V

chr9 139418260 139418260 A G SNP synonymous Het 7.37 NOTCH1 p.N104N

chr2 29455267 29455267 A G SNP synonymous Het 8.61 ALK p.G845G

chr7 55238874 55238874 T A SNP synonymous Het 23.68 EGFR p.T629T

chr17 29508775 29508775 G A SNP synonymous Het 24.79 NF1 p.L234L

chr4 55602765 55602765 G C SNP synonymous Het 34.95 KIT p.L862L

chr17 29553485 29553485 G A SNP synonymous Het 52.07 NF1 p.P678P

chr12 25362777 25362777 A G SNP synonymous Het 57.24 KRAS p.D173D

chr7 116436097 116436097 G A SNP synonymous Het 80 MET p.P1382P

chr7 116435768 116435768 C T SNP synonymous Het 82.81 MET p.D1304D

chr7 116436022 116436022 G A SNP synonymous Het 85.36 MET p.A1357A

chr2 29416615 29416615 G A SNP synonymous Het 85.88 ALK p.T1446T

chr2 29445458 29445458 G T SNP synonymous Het 86.96 ALK p.G1125G

chr2 30143499 30143499 G C SNP synonymous Hom 97.3 ALK p.L9L

chr7 55214348 55214348 C T SNP synonymous Hom 99.26 EGFR p.N158N

chr7 55266417 55266417 T C SNP synonymous Hom 99.5 EGFR p.T903T

chr12 25368462 25368462 C T SNP synonymous Hom 99.61 KRAS p.R161R

chr2 29940529 29940529 A T SNP synonymous Hom 99.62 ALK p.P234P

chr2 29543663 29543663 T C SNP synonymous Hom 99.65 ALK p.Q500Q

chr7 55249063 55249063 G A SNP synonymous Hom 99.87 EGFR p.Q787Q

chr4 1807894 1807894 G A SNP synonymous Hom 100 FGFR3 p.T651T

Gene

Amino acid

Change

KDR K747N

CDKN2A E88*(stop)

TP53 G245G

TP53 C242G

KDR encodes VEGFR2

Head & Neck SQCC

Before Treatment Treatment Day 15

TKI With VEGFR2 Activity

“We need to provide knowledge, not just data” Dr. Kojo Elenitoba-Johnson, Nov. 2013

• 40 year old male with thymic carcinoma and no good treatment options

• Tumor sequenced by a commercial laboratory (large gene panel)

–No actionable mutations reported

–Variants of unknown significance at very bottom of report included “KIT Y646D”

KIT-Mutant Thymic Carcinoma

Patient Exon Mutation Treatment Response Reference

54 yr M KIT 11 Deletion V560D

Imatinib Mixed 6 mo

N Engl J Med. 2004 Jun 17;350(25):2625-6

46 yr M KIT 17 D820E Sorafenib PR 15+ mo

J Thorac Oncol. 2009 Jun;4(6):773-5

47 yr F KIT 11 Deletion 557-559

Sorafenib PR 6+ mo

Lung Cancer. 2011 Jan;71(1):109-12

48 yr M KIT 11 Y553N Imatinib PR 9+ mo

J Clin Oncol. 2011 Nov 20;29(33):e803-5

55 yr F KIT 11 Deletion V560D

Imatinib SD 12+ mo

J Thorac Oncol. 2014 Feb;9(2):e12-6

58 yr M KIT 13 K642E Sorafenib PR 6 mo

Onco Targets Ther. 2014, 7:697-702

• Y646D is a non-conservative substitution located near K642E

• Recommendation to oncologist: try a KIT inhibitor

• KIT Mutations occur in ~10% of thymic carcinomas

Sorafenib x 12 days

Patient remains on treatment and is doing well at 15 months

Illumina NextSeq500

Clinical Genomics Database

Clinical Reports

Public Databases COSMIC

dbSNP, Clinvar PCT (MD Anderson) My Cancer Genome

CIViC

Exacloud Computing Cluster

LIS

Ion Torrent PGM

Turning Data Into Knowledge

Whole Exome Sequencing (Sequencing All Protein-Coding Regions Across 20,000 Genes)

OHSU/Intel Collaboration

Clinical Sequence Data

Research Genomic Data

Exacloud Computing

Center Variant

Annotation

Sequence alignment Variant calling

Data Reporting

Clinical Genomics Database

7,000 Processors for data crunching 220 Terabytes of storage for clinical datasets

1200 Terabytes of storage for research samples

‘Liquid’ Biopsies DNA in Plasma and Urine

Commercial Labs

• Small fragments of DNA are present in normal plasma and urine

• Increased levels occur in patients with advanced cancer

• New, high sensitivity tests based on next-gen sequencing can be used to detect mutations in such samples

• May be useful in monitoring treatment responses

Summary

• Advances in DNA sequencing technology are supporting the new era of ‘precision’ cancer care

• NGS supports: o Determination of exact allele frequency o Detection of gene fusion events o Assessment of gene copy number

• The pathologist plays a critical role in selecting material appropriate for testing

• NGS panels are becoming very routine and can be integrated into practice as another tool alongside IHC, FISH, etc.

Acknowledgements

Carol Beadling, PhD

Richard Press, MD, PhD

Tanaya Neff, MA

Fei Yang, MD

Rebecca O’Gara

Marina Pukay

Andrea Warrick

Cara Poage

Amy Schilling

Catie Grasso, PhD

With Assistance From:

• Knight Cancer Institute

• Ion Torrent/ThermoFisher

• Intel

• GIST Cancer Research Fund

• LifeRaft Group

AKT1 1%

BRAF 44%

CTNNB1 2%

GNAQ 1% KIT

1% MAP2K1

1%

NRAS 16%

PIK3CA 1%

TP53 1%

No Mutation

34%

Cutaneous Melanoma Cases (n=140)

Updated from Beadling et al. J Molec Diag, 2011 Sep;13(5):504-13

BRAF V600E(K)

• Vemurafenib

• Dabrafenib

• Trametinib

Best Response to Vemurafenib in

Metastatic BRAFV600E Melanoma

Flaherty , et al. N Engl J Med. 2010 Aug 26;363(9):809-19

KRAS p.A146T

Metastatic BRAFV600E Papillary Thyroid Carcinoma

Treated With Vemurafenib

AKT1 1%

BRAF 44%

CTNNB1 2%

GNAQ 1% KIT

1% MAP2K1

1%

NRAS 16%

PIK3CA 1%

TP53 1%

No Mutation

34%

Cutaneous Melanoma Cases (n=140)

Updated from Beadling et al. J Molec Diag, 2011 Sep;13(5):504-13

Would imatinib

work in

melanoma?

KIT-Mutant Mucosal Melanoma

Hodi et al. J Clin Onc 2008 Apr 20;26(12):2046-51

Response to Imatinib

Targeting KIT-Mutant Melanoma

Phase II Clinical Trials

Drug

# Pts

Clinical Benefit Rate

Carvajal et al. JAMA 2011;305(22):2327-2334

Imatinib 25 72%

Guo et al. JCO 2011;29(21):2904-2909

Imatinib 43 53%

Minor et al. Clin Cancer Res 2012; 18(5):1457-63

Sunitinib 4 75%

Hodi et al. JCO (epub Jul 8, 2013)

Imatinib 13 77%

Left Lower Lobe Left Upper Lobe

72 Year Old Female Smoker

KRAS G12C KRAS G12V