Post on 24-May-2020
Advance lung liquid biopsy research with Next Generation Sequencing
Nicola Normanno
ISTITUTO NAZIONALE PER LO STUDIO E LA CURA DEI TUMORI
FONDAZIONE G. Pascale – NAPOLI SC Biologia Cellulare e Bioterapie
CENTRO RICERCHE ONCOLOGICHE MERCOGLIANO (AV)
Laboratorio di Farmacogenomica
ESMO 2017- THERMO FISHER SCIENTIFIC SYMPOSIUM
Blood-based biomarkers in cancer
Crowley Nat Rev Cancer 2013
With the term liquid biopsy we refer to the possibility to perform tumor molecular profiling by using tumor-derived biomarkers that can be isolated from the peripheral blood (or any other body fluid) of cancer patients
BIOMARKERS DNA RNA miRNA Proteins
SOURCES Cell-free CTC Exosomes Platelets
Challenges in cfDNA analysis
The absolute levels are low: few nanongrams per ml of plasma
The circulating cell-free DNA (cfDNA) contains both tumor-derived DNA (ctDNA) and normal DNA originating form dividing cells (blood cells, GI tract, skin)
The ctDNA is only a fraction (<0.1% to 50%) of the cfDNA
Levels are usually correlated with tumor burden and are higher in advanced cancer
Highly fragmented, typically 50-200bp range (165bp peak)
Short half-life
Liquid biopsy for mutational profiling of NSCLC
Challenges with research sample availability • In Europe, a surgical specimen or a core-biopsy is available in
<50% of NSCLC cases • In approximately 25% of the cases there is no material
available for molecular profiling or the material is inadequate
• Almost 40% of the cases have available cytology specimens or small biopsies for molecular characterization
• In cases with cytology or small biopsies the samples might not be representative of the disease
Normanno WCLC 2015
The ASSESS trial: sensitivity of plasma-based assays for EGFR sensitizing mutations
Performance of four different plasma assays (38 plasma samples from the AURA trial)
Thress Lung Cancer 2015
Secondary Mutation in Gefitinib/Erlotinib/Afatinib-Resistant NSCLC
• A substitution of methionine for threonine at position 790 (T790M) in the kinase domain in exon 20 increases the ATP affinity of the EGFR
• ~50% of patients with acquired resistance to EGFR TKIs develop the T790M mutation
• Among 155 lung tumour specimens only 1 had pre-existing T790M without prior EGFR-TKI exposure
1. Pao PLos Med 2005; Kobayashi New Engl. J. Med. 2005; Yun PNAS 2008 2. Socinski MA, et al. The oncologist. 2016.
The efficacy of targeted therapy depends on
TUMOR HETEROGENEITY
Burrell Nature 2013
Genetic and phenotypic variation observed between tumors of different tissue and cell types, as well as between individuals with the same tumor type
Subclonal diversity observed within a tumor (tumors are formed of different clones with different genetic and molecular features)
cfDNA analysis recapitulate heterogeneity of disease
• Tumor biopsies obtained from 3 different regions of SCC and cfDNA isolated from plasma
• cfDNA profiling reveals heterogeneous mutations that could be missed by standard single-pass tissue biopsies
Jamal-Hanjani Ann Oncol 2016
BEAMing, Droplet digital PCR (ddPCR)
Assays – sensitivity vs. breadth
Courtesy of K. Thress BREADTH
(No. bases assayed per sample)
RELATIVE SENSITIVITY
(Ability to see increasingly rare mutant alleles)
10% (1/10)
1% (1/100)
0.1% (1/1000)
0.01% (1/10,000)
0.001% (1/100,000)
>109 10 1 102 103 104 105 106 107 108
QPCR
Kilobases Megabases Gigabases (genome)
100% NGS – Unselected whole genome
NGS – Exome
NGS – Targeted panels
Allele specific- and emulsion-PCR methods are highly sensitive but detect only point mutations and short indels on very few loci NGS allows query of many more loci (up to whole genome), detects point mutations, indels, CNV and fusions but at low sensitivity
?
Philip C. Mack, PhD, UC Davis Comprehensive Cancer Center and California Northstate University
ctDNA utility in under-genotyped non-squam NSCLC
Tissue Genotyping Status
383 of 1288 (30%)
Biomarker Positive
879 (68%) Quantity Insufficient (QNS) or
Undergenotyped (UG)
ctDNA NGS Increased Biomarker Yield by 65%
ctDNA analysis
identified 252 additional actionable biomarkers
(19% of 1288) (29% of 879)
not previously detected in tissue QNS/UG cases
The cost of the Guardant Health panel is 5600$
Philip C. Mack, PhD, UC Davis Comprehensive Cancer Center and California Northstate University
5
Updated from Lanman et al. 2015 PLoS One
Mutant Allele Fraction (MAF) is typically very low P
erce
ntag
e of
var
iant
s
MAF (cell-free DNA %)
All variants detected
heterozygous SNPs
homozygous SNPs
somatic SNVs
MAF (cell-free DNA %)
Per
cent
age
of v
aria
nts
Reported somatic variants
Min: 0.03% 25th: 0.19% Median: 0.44% 75th: 2.02% Max: 97.62% Mean: 3.27%
Differentiation of Somatic vs. Germline Variants
Half of Variants reported occur below 0.44% MAF
NGS for Liquid Biopsies: Sensitivity vs Specificity
Newman Nat Biotechnol 2016
Duplex NGS (complementary strand
variant matching)
Molecular barcoding
“Data polishing”
For research use only. Not for use in diagnostic procedures
Safe-Sequencing System “Safe-SeqS”
Kinde PNAS 2011 Bettegowda Sci Transl Med 2014
• Assignment of a unique identifier (UID) to each DNA template molecule
• Amplification of each uniquely tagged template molecule to create UID families
• Redundant sequencing of the amplification products
• PCR fragments with the same UID are considered mutant (“supermutants”) only if ≥95% of them contain the identical mutation
Cancer Personalized Profiling by Deep Sequencing (CAPP-Seq)
• Panel targeting recurrently mutated regions in NSCLC
• Diverse mutation types: point mutations (SNVs), indels, copy number variants (CNVs), and rearrangements
• Optimized library preparation for cfDNA and low inputs (ng)
• Custom bioinformatics for variant detection <0.1% (≈0.02%)
• Ultra-deep sequencing: >10,000X
Newman Nat Med 2014
Barcoding and Polishing Eliminate Distinct Populations of Background Errors of CAPP-Seq
Newman Nat Biotechnol 2016
Integrated digital error suppression (iDES)-enhanced CAPP-Seq detected tumor-derived DNA down to 0.0025% (2.5 in 105 molecules)
24 For Research Use Only. Not for use in diagnostic procedures.
Oncomine cfDNA assays
Thermo Fisher All Rights Reserved .
25 For Research Use Only. Not for use in diagnostic procedures.
Oncomine Lung cfDNA Assay
Content developed in collaboration with the OncoNetwork consortia Assays designed to detect
• Primary tumor drivers • Resistance mutations
169 hotspot mutations like EGFR: T790M, L858R, Exon19 del, C797S KRAS: G12X, G13X, Q61X ALK: 1151Tins, L1152R, C1156Y BRAF: V600E Detect rare variants present down to 0.1% with
Mean Sensitivity – 90% Mean Specificity – 98%
ALK BRAF EGFR ERBB2 KRAS MAP2K1
MET NRAS PIK3CA ROS1 TP53
Gene List
Thermo Fisher All Rights Reserved .
26 For Research Use Only. Not for use in diagnostic procedures.
Core Technology and Variant Analysis Method
cfDNA molecule with tumor variant Wild type cfDNA molecule
Variant
Gene Specific Primer Molecular tags
Assay
Analysis
Amplification, tagging and sequencing
Variant calling
27 For Research Use Only. Not for use in diagnostic procedures.
Oncomine Lung cfDNA Assay - R&D reproducibility results
Test site Users Sample Sensitivity Specificity
sensitivity each usr average ( n=4)
average ( n=16) specificity usr average
( n=4) all average
( n=16)
Site A usr1
0.5%MM
100
100
100
98
98
98
100 99 100 97 100 100
Site B
usr2
100
100
98
98 100 97 100 98 100 97
usr3
100
100
99
99 100 98 100 99 100 98
usr4
100
100
99
99 100 100 100 99 100 97
Site A usr1
0.1% MM
95
90
90
99
99
98
85 98 88 99 93 99
Site B
usr2
90
92
99
98 88 99 95 99 95 98
usr3
93
93
99
98 90 97 95 99 95 99
usr4
85
85
99
98 98 98 88 99 70 97
28 For Research Use Only. Not for use in diagnostic procedures.
Sample Repeat
Sensitivity Specificity
single point in run average ( n=4)
all runs average ( n=12) single point in run average
( n=4) all runs average
( n=12)
0.5%MM
run1
100.0%
100.0%
100.0%
99.2%
99.6%
99.6%
100.0% 100.0% 100.0% 99.2% 100.0% 100.0%
run2
100.0%
100.0%
100.0%
99.4% 100.0% 99.2% 100.0% 99.2% 100.0% 99.2%
run3
100.0%
100.0%
100.0%
99.6% 100.0% 100.0% 100.0% 99.2% 100.0% 99.2%
0.1%MM
run1
97.5%
97.5%
98.3%
100.0%
99.8%
99.8%
97.5% 100.0% 97.5% 99.2% 97.5% 100.0%
run2
100.0%
99.4%
99.2%
99.8% 100.0% 100.0% 97.5% 100.0%
100.0% 100.0%
run3
100.0%
98.1%
100.0%
99.8% 95.0% 99.2% 97.5% 100.0%
100.0% 100.0%
Oncomine Lung cfDNA Assay – R&D repeatability results
29 For Research Use Only. Not for use in diagnostic procedures.
OncoNetwork Consortia Prof. Harriet Feilotter
Dr Paul Park Queen's University, Ontario
Canada
Dr. Jose Costa IPATIMUP, Medical Faculty of
Porto. Portugal
Marjolijn J.L. Ligtenberg Radboud University Nijmegen Medical Centre, Netherlands
Dr. Nicola Normanno Centro Ricerche Oncologiche
Mercogliano, Italy
Prof. Orla Sheils St James's Hospital Dublin,
Ireland
Prof. Ian Cree UHCW
United Kingdom
Prof. Pierre Laurent Puig Université Paris Descartes,
HEGP Paris, France Prof. Aldo Scarpa ARC-NET University of
Verona Italy
Dr. Henriette Kurth VIOLLIER AG Basel ,
Switzerland
Prof. Kazuto Nishio Faculty of Medicine, Kinki University Osaka, Japan
Cecily P. Vaughn ARUP- Institute for Clinical and
Experimental Pathology Utah, USA
Dr. Michael Hummel Institute of Pathology Charité
Berlin, Germany
30 For Research Use Only. Not for use in diagnostic procedures.
• Each lab performed two runs using Oncomine Lung cfDNA Assay on Horizon Multiplex cfDNA Reference Standard Set.
• Each run included 5%, 1% and 0.1% and WT samples.
Multicenter study- repeatability and reproducibility
• Horizon controls at different percentages and WT in a duplicate run Repeatability
• Same experiment in 11 laboratories Reproducibility
31 For Research Use Only. Not for use in diagnostic procedures.
Mutations Screened In the Multicenter Study
Additional WT level variants in cell line detected by the cfDNA assays include:
EGFR p.G719S; PIK3CA p.H1047R; MAP2K1 p.Q56P; BRAF p.V600E These variants should be expected at approximately 20-40% AF
Multiplex I cfDNA Reference Standard Set (5%, 1%, 0,1%, 0%)
32 For Research Use Only. Not for use in diagnostic procedures.
Lab Sensitivity Specificity NPV PPV
ARCNET 89,58% 99,53% 99,61% 87,76%
CROM 95,83% 100,00% 99,84% 100,00%
IPATIMUP 95,83% 100,00% 99,84% 100,00%
Radboud 89,58% 99,84% 99,61% 95,56%
Queens 97,92% 99,92% 99,92% 97,92%
St James 95,83% 100,00% 99,84% 100,00%
UHCW 95,83% 99,76% 99,84% 93,88%
Charite 95,83% 99,76% 99,84% 93,88%
Viollier* 97,50% 99,82% 99,91% 95,12%
HEGP 93,75% 99,69% 99,76% 91,84%
Kindai* 95,83% 99,69% 99,84% 92,00%
Total 94,81% 99,82% 99,80% 95,93%
Sensitivity and Specificity at overall results
33 For Research Use Only. Not for use in diagnostic procedures.
Lab Sensitivity Specificity NPV PPV
ARCNET 68,75% 100,00% 98,43% 100,00%
CROM 87,50% 100,00% 99,37% 100,00%
IPATIMUP 87,50% 100,00% 99,37% 100,00%
Radboud 68,75% 99,68% 98,43% 91,67%
Queens 93,75% 100,00% 99,68% 100,00%
St James 87,50% 100,00% 99,37% 100,00%
UHCW 87,50% 100,00% 99,37% 100,00%
Charité 87,50% 100,00% 99,37% 100,00%
Viollier 87,50% 100,00% 99,37% 100,00%
HEGP 81,25% 99,36% 99,05% 86,67%
Kindai 87,50% 99,68% 99,37% 93,33%
Total 83,93% 99,88% 99,19% 99,13%
Sensitivity and Specificity at 0.1% detection limit
34 For Research Use Only. Not for use in diagnostic procedures.
Sensitivity and specificity – OncoNetwork
Allele Frequency Sensitivity Specificity
0.1%-5% 94.81% 99.82%
0.1% 83.93% 99.88%
Data collected from 11 laboratories: UHCW, CROM, ARCNET, IPATIMUP, Radboud University, Queen’s University,
St James Hospital, Violler, HEGP, Kinday University, Charite’ Hospital
Horizon - Multiplex I cfDNA Reference Standard Set
35 For Research Use Only. Not for use in diagnostic procedures.
Reproducibility at 5% limit detection
EGFR E746_A750del ELREA
PIK3CA E545K
KRAS G12D
n=22 n=22
n=21
36 For Research Use Only. Not for use in diagnostic procedures.
Reproducibility at 1% allele frequenci
EGFR E746_A750del ELREA
PIK3CA E545K
KRAS G12D
n=22
n=22
n=21
37 For Research Use Only. Not for use in diagnostic procedures.
EGFR E746_A750del ELREA
PIK3CA E545K
KRAS G12D
n=20
n=20
n=15
Reproducibility at 0.1% limit detection
38 For Research Use Only. Not for use in diagnostic procedures.
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
Alle
le F
requ
ency
5% ARCNET
CROM
IPATIMUP
St Radboud
Queens
St James
UHCW
Charite
Viollier
HEGP
Kindai
EGFR – T790M at 5%, 1% and 0,1% allele frequency
0.00
0.20
0.40
0.60
0.80
1.00
1.20
1.40
1.60
Alle
le fr
eque
ncy
1% ARCNET
CROM
IPATIMUP
St Radboud
Queens
St James
UHCW
Charite
Viollier
HEGP
Kindai
0.00
0.05
0.10
0.15
0.20
0.25
0.30
Alle
le F
requ
ency
0,1% ARCNET
CROM
IPATIMUP
St Radboud
Queens
St James
UHCW
Charite
Viollier
HEGP
Kindai
39 For Research Use Only. Not for use in diagnostic procedures.
Oncomine™ cfDNA Assays – Content
35 amplicon panel for Lung
Covering key hotspot mutations in 11 genes
169 Hotspot SNVs & indels
Oncomine™ Lung cfDNA Assay
ALK BRAF EGFR ERBB2 KRAS MAP2K1
MET NRAS PIK3CA ROS1 TP53
48 amplicon panel for colon (gastrointestinal) cancer type(s)
Covering key hotspot mutations in 14 genes
244 Hotspot SNVs & indels
Oncomine™ Colon cfDNA Assay
AKT1 BRAF CTNNB1 EGFR ERBB2 FBXW7 GNAS
KRAS MAP2K1 NRAS PIK3CA SMAD4 TP53 APC
26 amplicon panel for Breast
Covering key hotspot mutations in 10 genes
157 Hotspot SNVs & indels
Oncomine™ Breast cfDNA Assay
AKT1 EGFR ERBB2 ERBB3 ESR1
FBXW7 KRAS PIK3CA SF3B1 TP53
Thermo Fisher All Rights Reserved
40
Coming Soon – Enhanced Content to Lung & Breast cfDNA Assays
58 amplicon + 49 Fusion Assays + 3 MET Exon Skipping Assays for new Lung Assay
Covering: • Key hotspot mutations in 11 genes • Fusions – ALK, RET, ROS1 • CNV – MET • MET exon 14 skipping
SNV LOD down to 0.1% with 20 ng input Same Sensitivity & Specificity for SNVs Single library from both DNA & RNA
76 amplicon panel for new Breast Assay
Covering: • Key hotspot mutations in 10 genes • CNVs – CCND1, ERBB2, FGFR1 • More complete coverage of TP53 • de novo variant calling across TP53
SNV LOD down to 0.1% with 20 ng input Same Sensitivity & Specificity for SNVs Single library to detect SNVs and CNVs
• Designed to detect primary tumor drivers and resistance mutations • From a single tube of blood
New Oncomine™ Lung cfTNA Assay
ALK BRAF EGFR ERBB2 KRAS MAP2K1
MET NRAS PIK3CA ROS1 TP53
New Oncomine™ Breast cfDNA Assay
AKT1 EGFR ERBB2 ERBB3 ESR1
FBXW7 KRAS PIK3CA SF3B1 TP53
Thermo Fisher All Rights Reserved For Research Use Only. Not for use in diagnostic procedures.
41
Lung enhanced panel – Fusion and MET skipping (+) controls
Assay ID Type 1% rep 1 1% rep 2 0.5% rep 1 0.5% rep 2 MET-MET.M13M15 RNAExonVariant 740 787 382 382
MET.E11E12.WT RNAExonVariant 628 640 297 323 MET.E6E7.WT RNAExonVariant 893 986 375 477
TBP.ENCTRL.E3E4 ProcControl 286 334 208 272 HMBS.ENCTRL.E8E9 ProcControl 275 383 161 183
Assay ID Type 1% rep 2 1% rep 1 0.5% rep 2 0.5% rep 1 0.25% rep 2 0.25% rep 1 CCDC6-RET.C1R12.COSF1271 Fusion 52 40 37 16 16 20
EML4-ALK.E6A20 Fusion 9 19 5 0 7 3 MET-MET.M13M15 RNAExonVariant 0 1 0 0 0 0
MET.E11E12.WT RNAExonVariant 704 669 370 368 147 212 MET.E6E7.WT RNAExonVariant 1110 1113 633 567 226 306
SLC34A2-ROS1.S4R32.COSF1196 Fusion 209 194 83 80 31 51
TBP.ENCTRL.E3E4 ProcControl 273 208 197 198 112 133 HMBS.ENCTRL.E8E9 ProcControl 231 284 161 191 133 157
Tri-Fusion positive samples
MET exon skipping positive samples
*Tri-Fusion or MET exon skipping positive RNA was spiked into cell free total nucleic acid at the indicated frequencies. Values in table represent molecular
coverage for each amplicon indicated
42
Expanded Variant Type Detection with cfDNA Lung Panel
MET CNV + Fusion/MET Exon 14 Skipping at 0.5% titration
CNV Analysis
Fusion/MET exon Skipping Analysis
Note: Expected MET CNV ratios (w/significant p-values) confirmed. Expected fusion/exon skipping/control targets confirmed
For Research Use Only. Not for use in diagnostic procedures.
43
Results from 4 Retrospective Samples
Sample Tissue at diagnosis Plasma at progression after TKI Detection Method
9308 Not available p.Glu746_Ala750del Therascreen/ ddPCR
7947b p.Glu746_Ala750del p.Glu746_Ala750del Therascreen/ ddPCR
7679b p.L858R L858R and T790M Therascreen/ ddPCR
8151 p.L858R L858R and T790M Therascreen/ ddPCR
Oncomine Lung Cell-Free Total Nucleic Acid Assay results*
9308
7679b
7947b
8151
For Research Use Only. Not for use in diagnostic procedures.
Liquid biopsy can represent temporal and spatial heterogeneity in cancer progression
Burrell & Swanton Mol Oncol 2014
Laboratory of Pharmacogenomics Anna Maria Rachiglio Matilde Lambiase Francesca Fenizia Raffaella Pasquale Claudia Esposito Cristin Roma Laura Forgione Rino E. Abate Alessandra Sacco Alessia Iannacone Francesca Bergantino
Cell Biology and Biotherapy Unit Antonella De Luca Amelia D’Alessio Monica R. Maiello Marianna Gallo Daniela Frezzetti Nicoletta Chicchinelli Michele Grassi
CENTRO RICERCHE ONCOLOGICHE MERCOGLIANO (AV)
Laboratorio di Farmacogenomica
ISTITUTO NAZIONALE PER LO STUDIO E LA CURA DEI TUMORI
FONDAZIONE G. Pascale – NAPOLI
Surgical Pathology Unit Gerardo Botti Fabiana Tatangelo
External Collaborators Carmine Pinto, IRCCS Reggio Emilia Domenico Galetta, IRCCS, Istituto Tumori "Giovanni Paolo II ", Bari Bruno Daniele, Ospedale G. Rummo, Benevento Francesco Ferraù, Ospedale San Vincenzo, Taormina Vienna Ludovini e Lucio Crinò, Ospedale S. Maria della Misericordia, Perugia
Dept. Thoracic Oncology Gaetano Rocco Alessandro Morabito
Clinical Trial Unit Francesco Perrone Maria C. Piccirillo
Acknowledgments
OncoNetwork Consortium
Jose Costa - IPATIMUP, Medical Faculty of Porto
Orla Sheils - St James's Hospital, Dublin
Pierre Laurent Puig - Université Paris Descartes, Paris
Michael Hummel - Institute of Pathology, Charité, Berlin
Aldo Scarpa - ARC-NET University of Verona
Kazuto Nishio - Faculty of Medicine, Kinday University, Osaka
Henriette Kurth - VIOLLIER AG, Basel
Marjolijn Ligtenberg - Radboud University, Njimegen
Ian Cree - UHCW, Warwick
Harriet Feilotter - Queen's University, Ontario
Chris Allen
Kelli Bramlett
Dima Brinza
Tom Bittick
Andrea Luchetti
Rosella Petraroli
Thermo Fisher Scientific and its affiliates are not endorsing, recommending, or promoting any use or application of Thermo Fisher Scientific products presented by third parties during this seminar. Information and materials presented or provided by third parties are provided as-is and without warranty of any kind, including regarding intellectual property rights and reported results. Parties presenting images, text and material represent they have the rights to do so.