NGS-BASED GENE FUSION DETECTIONHOUSTON METHODIST HOSPITAL EXPERIENCE
USCAP - Pulmonary Pathology SocietyBryce Portier MD, PhDDirector 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
FISHFluorescent in situ hybridization
RT‐PCRReverse Transcriptase PCR
IHCImmunohistochemistry
DNA ProteinRNA
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)
NGSNext Generation Sequencing
cDNA
Murakami Y, Mitsudomi T, Yatabe Y., Front Oncol 2012
MULTIPLE METHODS FOR GENE FUSION DETECTION
FISHFluorescent in situ hybridization
RT‐PCRReverse Transcriptase PCR
IHCImmunohistochemistry
DNA ProteinRNA
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)
NGSNext Generation Sequencing
cDNA
Murakami Y, Mitsudomi T, Yatabe Y., Front Oncol 2012
MOLECULAR DIVERSITY CAUSES DIAGNOSTIC CHALLENGES
MLL (KMT2A) Gene Rearrangements66 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 Rearrangements19 Partner genes
Catalogue of somatic mutations in cancer v71, 2015 http://cancer.sanger.ac.uk/cosmic/gene/overview?ln=ALKSasaki 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
Frampton, et. al Nature Biotechnology Nov 2013 (modified)
RNAExtraction Sequencing
Analysis &Interpretation
Library Generation1‐ Target (hybridization) Capture
“Foundation Medicine”2‐ Opposing Primers
“Ion Life Technologies”3‐ Anchored Multiplex PCR
“Archer”
LimitationsHigher input requirements (not ideal for biopsies/limited FFPE samples)
Fusion partner and breakpoint must be knownOpposing primer pair needed for every fusion partner and breakpointFixed start site for every amplicon (problematic with fragmented FFPE)
Frampton, et. al Nature Biotechnology Nov 2013 (modified)
RNAExtraction Sequencing
Analysis &Interpretation
NGS-BASED FUSION DETECTION WORKFLOW
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
Bar codes available for MiSeq or PGM sequencing
Molecular indexing measures number of unique fusion events
Up to 600 amplicons can be multiplexed in a single tube‐ DNA or RNA targets
Thermal cycling maintain linearity during amplification ‐ Enabling quantitation
Random start sites improve coverage
www.archerdx.com
• 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
55ng RNA input
27.5ng RNA input
5ng RNA input
Library Input Library Uniformity
ADVANTAGES OF NGS GENE FUSION DETECTION BY ANCHORED MULTIPLEX PCR
CLINICAL ASSAY VALIDATION PLANAccuracyCorrelation to FISH results
Analytical Sensitivity (LOD)A dilution series utilizing a translocation positive Geneblock mixed with Ambion normal lung.
Analytical Specificity/Interfering SubstancesAnalytical specificity‐ correlation with FISH and known Geneblock constructsInterfering 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.
PrecisionIntra‐ and inter‐run reproducibilityInter‐technologist reproducibility
Reference values and reportable rangesReference range: Negative for translocation Reportable range: Positive or negative (qualitative) for translocation
CAP Molecular Pathology ChecklistsCLSI guidelinesArch Pathol Lab Med, Test verification and validation for molecular diagnostic assays, 2012.
CLINICAL ASSAY VALIDATION SAMPLES
Anonymized Samples25 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
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:R12NCOA4‐RET N6:R12
• ROS1‐ CD74‐ROS1 C6:R32CD74‐ROS1 C6:R34
Homo sapiens mRNA for fusion protein EML4‐ALK variant 1, complete cds (Variant 1 = EML4 ex 13 and ALK ex20) GenBank: AB274722.1
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 1: ID Variant of Interest Break Point
Step 2: Download cDNA sequence (Genebank)
Step 3: Locate Break Point
Step 4: Order Oligonucleotide “Engineered Control”
Controls:
PREBUILT BIOINFORMATIC PIPELINEArcher Analysis Pipeline Version 3.0
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
Analysis &Interpretation
AUTOMATED BIOINFORMATICNGS ANALYSIS
1 2
3
AUTOMATED BIOINFORMATICNGS ANALYSIS
4
5
VALIDATION DATA SUMMARY
Anonymized Samples25 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
0
20
40
60
80
100
2013‐2014 2011‐2012 2005‐2010
% of C
ases
Passing QC
Cases by Surg‐Path Accession Year
0
20
40
60
80
100
2013‐2014 2011‐2012 2005‐2010
% of C
ases
Passing QC
Cases by Surg‐Path Accession YearNGS RT‐PCR
SUMMARYNGS 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 YOUNGS-BASED GENE FUSION DETECTIONHOUSTON METHODIST HOSPITAL EXPERIENCE
AcknowledgmentsPhilip Cagle MDRandall Olsen MD, PhDNeal Lindeman MDLynette Sholl MDJohn Iafrate MD, PhDKirtee Raparia MBBSJan Nowak MD, PhD