NGS-BASED GENE FUSION DETECTION HOUSTON METHODIST HOSPITAL EXPERIENCE

USCAP - Pulmonary Pathology Society Bryce Portier MD, PhD Director Solid Tumor Molecular Diagnostics

DISCLOSURES Financial Disclosures: • No financial association with any of the companies or products in this presentation

ALK TRANSLOCATIONS AND RESPONSE TO TARGETED THERAPY

Kwak et al. N Engl J Med. 2010;363:1693‐1703.. 

MOLECULAR DIVERSITY CAUSES DIAGNOSTIC CHALLENGES

Modified from: Horn and Pao , J Clin Oncol. 2009 Sep 10;27(26):4232‐5. 

MULTIPLE METHODS FOR GENE FUSION DETECTION DNA

RNA

Protein

cDNA

FISH

RT‐PCR

IHC

NGS

Reverse Transcriptase PCR

Immunohistochemistry

Next Generation Sequencing

Fluorescent in situ hybridization

Limitations of Current Methods: • Throughput limitations due to limited multiplexing capability (FISH and IHC) • Subjective results, cannot determine breakpoint or fusion partner (FISH and IHC) • Requires knowledge of fusion partners and breakpoints for assay design (RT‐PCR) • Decreased sensitivity due to poor quality of RNA from FFPE (RT‐PCR) Murakami Y, Mitsudomi T, Yatabe Y., Front Oncol 2012

MULTIPLE METHODS FOR GENE FUSION DETECTION DNA

RNA

Protein

cDNA

FISH

RT‐PCR

IHC

NGS

Fluorescent in situ hybridization

Reverse Transcriptase PCR

Immunohistochemistry

Next Generation Sequencing

Advantages of Current Methods: • Screening method used for clinical trials (FISH and IHC) • Established techniques and workflows in many labs (FISH and IHC) • Applicable to archival specimens (FISH and IHC) • Rapid turn around time and inexpensive (RT‐PCR) Murakami Y, Mitsudomi T, Yatabe Y., Front Oncol 2012

MOLECULAR DIVERSITY CAUSES DIAGNOSTIC CHALLENGES

MLL (KMT2A) Gene Rearrangements 66 Partner genes Catalogue of somatic mutations in cancer v71, 2015  http://cancer.sanger.ac.uk/cosmic/gene/overview?ln=ALK

MOLECULAR DIVERSITY CAUSES DIAGNOSTIC CHALLENGES

ALK Gene Rearrangements 19 Partner genes Catalogue of somatic mutations in cancer v71, 2015  http://cancer.sanger.ac.uk/cosmic/gene/overview?ln=ALK Sasaki et al, Eur J Cancer 2010

MOLECULAR DIVERSITY CAUSES DIAGNOSTIC CHALLENGES

ALK, RET, ROS1 Gene Rearrangements ~36 Partner gene fusions >100 with multiple fusion break points

www.archerdx.com

NGS-BASED FUSION DETECTION WORKFLOW

RNA Extraction

Frampton, et. al Nature Biotechnology Nov 2013 (modified)

Sequencing

Analysis & Interpretation

NGS-BASED FUSION DETECTION WORKFLOW

RNA Extraction

Library Generation 1‐ Target (hybridization) Capture “Foundation Medicine” 2‐ Opposing Primers “Ion Life Technologies” 3‐ Anchored Multiplex PCR “Archer”

Sequencing

Analysis & Interpretation

Limitations Higher input requirements (not ideal for biopsies/limited FFPE samples) Fusion partner and breakpoint must be known Opposing primer pair needed for every fusion partner and breakpoint Fixed start site for every amplicon (problematic with fragmented FFPE)

Frampton, et. al Nature Biotechnology Nov 2013 (modified)

ADVANTAGES OF NGS GENE FUSION DETECTION BY ANCHORED MULTIPLEX PCR Features of  Anchored Multiplex PCR (AMP™)  Low input requirements (nanograms)  ‐ FFPE, fresh frozen tissue or blood  Molecular indexing measures number of unique fusion events  Random start sites improve coverage  Up to 600 amplicons can be multiplexed in a single tube ‐ DNA or RNA targets  Thermal cycling maintain linearity during amplification  ‐ Enabling quantitation  Bar codes available for MiSeq or PGM sequencing

www.archerdx.com

ADVANTAGES OF NGS GENE FUSION DETECTION BY ANCHORED MULTIPLEX PCR

Library Input

Library Uniformity 55ng RNA input

27.5ng RNA input

5ng RNA input

• Library generated with as little as 1 ng of RNA • Library complexity increases with input levels

• Uniform read coverage and excellent library diversity across sample input quantities

CLINICAL ASSAY VALIDATION PLAN Accuracy Correlation to FISH results Analytical Sensitivity (LOD) A dilution series utilizing a translocation positive Geneblock mixed with Ambion normal lung.  Analytical Specificity/Interfering Substances Analytical specificity‐ correlation with FISH and known Geneblock constructs Interfering substances‐ to minimize the possibility of interference, validation and testing will be restricted to RNA isolated  from formalin fixed paraffin embedded tissue from samples with >20% tumor burden quantified by H&E examination. Precision Intra‐ and inter‐run reproducibility Inter‐technologist reproducibility Reference values and reportable ranges Reference range:  Negative for translocation  Reportable range:  Positive or negative (qualitative) for translocation

CAP Molecular Pathology Checklists CLSI guidelines Arch Pathol Lab Med, Test verification and validation for molecular diagnostic assays, 2012. 

CLINICAL ASSAY VALIDATION SAMPLES Anonymized Samples 25 samples negative for translocation • 1 Cell line • 7 Geneblocks • 17 FISH negative patient samples 25 samples positive for translocation • 1 Cell line  • 7 Geneblocks • 17 FISH positive patient samples  Total number of samples: 50

GENERATION OF CONTROLS FOR CLINICAL ASSAY VALIDATION Controls:

Step 1: ID Variant of Interest Break Point 

Cell lines:

• HCC‐78 (SLC34A2‐ROS1 S4:R32 & S4:R34) • H2228 (EML4‐ALK E6:A20) • Normal Lung RNA (Ambion)  Engineered Controls: • ALK‐ EML4‐ALK E13:A20 EML4‐ALK E20:A20 • RET‐ CCDC6‐RET C1:R12 KIF5B‐RET K15:R12 NCOA4‐RET N6:R12 • ROS1‐ CD74‐ROS1 C6:R32 CD74‐ROS1 C6:R34

Step 2: Download cDNA sequence (Genebank) Homo sapiens mRNA for fusion protein EML4‐ ALK variant 1, complete cds (Variant 1 = EML4 ex 13  and ALK ex20)  GenBank: AB274722.1

Step 3: Locate Break Point 1621 gggaaatatg aaaagccaaa atttgtgcag tgtttagcat tcttggggaa tggagatgtt 1681 cttactggag actcaggtgg agtcatgctt atatggagca aaactactgt agagcccaca 1741 cctgggaaag gacctaaagt gtaccgccgg aagcaccagg agctgcaagc catgcagatg 1801 gagctgcaga gccctgagta caagctgagc aagctccgca cctcgaccat catgaccgac 1861 tacaacccca actactgctt tgctggcaag acctcctcca tcagtgacct gaaggaggtg

Step 4: Order Oligonucleotide “Engineered Control”

PREBUILT BIOINFORMATIC PIPELINE Archer Analysis Pipeline Version 3.0

Analysis & Interpretation

Virtual machine based prebuilt pipeline • Download with all dependences included • Locked down version for clinical use • Future updates can be loaded as independent virtual machine images for testing • Password protected data and login • All data and analysis done locally

AUTOMATED BIOINFORMATIC NGS ANALYSIS 1

2

3

AUTOMATED BIOINFORMATIC NGS ANALYSIS 4

5

VALIDATION DATA SUMMARY Anonymized Samples 25 samples negative for translocation • 1 Cell line No Fusion Detected 1 of 1 • 7 Geneblocks No Fusion Detected 7 of 7 • 17 FISH negative patient samples No Fusion Detected 17 of 17 25 samples positive for translocation • 1 Cell line Detected 1 of 1 • 7 Geneblocks Detected 7 of 7 • 17 FISH positive patient samples Low Detection Rate

NGS

RT‐PCR

Cases by Surg‐Path Accession Year

100

% of Cases  Passing QC

% of Cases  Passing QC

Cases by Surg‐Path Accession Year 80 60 40 20 0

100 80 60 40 20 0

2013‐2014

2011‐2012

2005‐2010

2013‐2014

2011‐2012

2005‐2010

SUMMARY NGS BASED FUSION DETECTION

Methodology selection for ALK Fusion detection will depend on multiple factors • Sample volume • Recent vs. archive cases • In‐house expertise • Cost/reimbursement

Selection of sequencing platform

RNA degradation in FFPE • Best results: Extract RNA from fresh cut block

THANK YOU NGS-BASED GENE FUSION DETECTION HOUSTON METHODIST HOSPITAL EXPERIENCE Acknowledgments Philip Cagle MD Randall Olsen MD, PhD Neal Lindeman MD Lynette Sholl MD John Iafrate MD, PhD Kirtee Raparia MBBS Jan Nowak MD, PhD