Next-Generation Sequencing (NGS) Testing for Hematologic Malignancies David Viswanatha, M.D. Division of Hematopathology Mayo Clinic, Rochester, Minnesota 2017 MFMER slide-1 DISCLOSURES Relevant Financial Relationship(s) None 2017 MFMER slide 2 Objectives Understand Basic Next Generation Sequencing (NGS) technology and its advantages, shortcomings and costs Overview of current NGS testing applications in hematologic malignancies, with emphasis on myeloid tumors Practical examples: case presentations 2017 MFMER slide 3 1
Part I: Why Next-Gen? 2017 MFMER slide 4 What Is Precision Medicine? The oncologist The pathologist The regulator The payer The patient 2017 MFMER slide 5 Mutation Burden By Cancer The prevalence of somatic mutations across human cancer types LB Alexandrov et al. Nature 500, 415 421 (22 August 2013) doi:10.1038/nature12477 2017 MFMER slide 6 2
Ley T et al. NEJM2013;368:2059 74 Papaemmanuil E, et al. Blood 2013;122:3616 27 2017 MFMER slide 7 Sanger Sequencing and NGS Sanger Next Gen Single target 1 target/sample Sensitivity ~20% Limited range (500-600 bp) CNV, translocations, large deletions not detected Complex sequences difficult to evaluate Multi-target Multiple samples Sensitivity ~5-10% Extended range (1000 s bp) CNV, translocations, large deletions not well detected Complex sequences possibly easier to evaluate 2017 MFMER slide 8 NGS Universe Targeted panels WES WGS RNA Seq Immune Seq CHIP Seq Methylome Seq 2017 MFMER slide 9 3
Key Considerations for NGS Design: Balancing Competing Constraints Comprehensive Analysis Large panel, WES, WGS Higher expense/sample Lower depth & slower TAT Small (targeted) panel Lower expense/sample Higher depth & faster TAT 1 Sample Throughput 100 2017 MFMER slide 10 Overview of NGS Method Adapted from: Jill M. Johnsen et al. Blood 2013;122:3268-3275 2013 by American Society of Hematology 2017 MFMER slide 11 NGS Workflow Wet Bench Process Dry Bench Informatics Library Prep 1 0 Demultiplexing Base calling Pre analytic Template Prep 2 0 Alignment Variant calling Specialized applications Sequencing 3 0 Variant Annotation Variant significance* 2017 MFMER slide 12 4
NGS: Pre-analytic Variables Sample type (e.g. fresh vs. FFPE), anticoagulant, etc. Tumor percentage requirement Need for microdissection (FFPE) Cell sorting DNA (or RNA) extraction and minimum Use of paired normal sample Adjacent normal, buccal, sorted T-cells, etc. 2017 MFMER slide 13 NGS: Library Preparation Amplicon-based Highly multiplexed PCR Less DNA input required Faster library generation Limited target region coverage CNV assessment more difficult Best for hot spot/regions Cheaper? Capture-based Bait (oligonucleotide) probe pull-down of fragmented high quality DNA More DNA input required Longer library preparation Better and more even target region coverage CNV detection easier Note: Sample throughput and accuracy similar 2017 MFMER slide 14 NGS: Platforms Illumina MiSeq V2 Bridge amplification Paired end reads possible Sequencing-by-synthesis Slower (~24 hrs) Less artifact Higher output* More expensive* Ion Torrent PGM Emulsion droplet PCR Single end reads Longer read lengths Semiconductor chip sequencing Homopolymer artifacts Faster (<24 hrs) Less expensive 2017 MFMER slide 15 5
NGS: Analysis Bioinformatics processing Index demultiplexing, base call and quality, sequencer performance data Secondary pipelines: initial alignment to reference, SNV and indel detection, variant calling Additional analyses: breakpoint analysis, structural variant (SV) detection, specialized applications (RNAseq, immunology, microbial, etc.) Data and computationally intensive Interpretation Final variant annotation and curation Pathogenicity determination (5 level) Potential germline association(s) Detailed report creation 2017 MFMER slide 16 Variant Classification is Multidisciplinary! NGS Analysis Team QC review Visual BAM file review Annotation & preliminary variant classification Initial report generation Consultant HP Final review and editing of mutation effect and pathogenic class Final report Genetic Counselors In depth variant classification and comment edits Evaluate possible germline findings 2017 MFMER slide 17 NGS Sequencing Cost $700 1000/sample! $1k per genome 2017 MFMER slide 18 6
NGS Panel Turn-around Time 24 48 hrs 24 36 hrs thermofisher.com 24 hrs genomics.agilent.com 2 to 4 hrs/case illumina.com 48 96 hrs/case wikipedia.com Typical TAT: 7 10 days 2017 MFMER slide 19 INFORMATION OVERLOAD!!! 2017 MFMER slide 20 Although not detracting from the excitement and opportunity that the current situation affords us, we would introduce a note of caution. Like modern day archeologists confronted with ancient Egyptian hieroglyphs, our current ability to confidently translate the presence of these mutations into clinical gains for individual patients significantly lags behind our ability to detect them. Grimwade D, et al. Blood 2016;127(1):29 41 2017 MFMER slide 21 7
Pros and Cons of NGS Panel Testing Advantages Platform convergence Multiplexing capacity Multiple applications Comprehensive variant detection Information density (comprehensive analysis) Integrated reporting Disadvantages Cost and TAT* Technical complexity Interpretive complexity Customization Limited platform availability Information density (uncertainty) and scale Reimbursement 2017 MFMER slide 22 Part II: Applications of NGS to Hematologic Malignancies 2017 MFMER slide 23 2017 MFMER slide 24 8
Targeted Therapies and Cancer (I) Disease Target Therapies Notes Colon Ca EGFR cetuximab Contraindicated if K/NRAS, RAF mutation PD1 pembrolizumab MSI status NSCLC EGFR cetuximab T790M mutation resistance RET cabozantinib Rare, young age, non smoker ROS1, ALK crizotinib Rare, female, younger Other solid T s PD1 pembrolizumab H&N squamous Ca; GU tumors, breast 2017 MFMER slide 25 Targeted Therapies and Cancer (II) Disease Target Therapies Notes CHL PD1 pembrolizumab ABVD Rx effective NHL BCR signaling (BTK, PIK3CD) CD20, CD22, CD25 Ibrutinib, idelasilib Rituximab, epratuzumab, alemtuzumab, blinatumomab ABC DLBCL, LPL/WM, CLL Humanized antibodies, BiTE agents, etc. BCL2 Venetoclax CD19 CAR T ALL, expanding MM 26S proteasome Bortezomib +2 nd generation drugs CRBN pathway Lenalidomide (IMiD) +2 nd generation drugs 2017 MFMER slide 26 Targeted Therapies and Cancer (III) Myeloid neoplasms: BCR-ABL (CML, Ph+ B-LBL) PML-RARA (APL) Highly efficacious targeted drugs for most other myeloid cancers are not significantly prevalent Why? 2017 MFMER slide 27 9
Signal Transduction: JAK2 KIT FLT3 CSF3R K/NRAS MPL CBL PTPN11 RNA Splicing: SF3B1 SRSF2 U2AF1 ZRSR2 Transcription Factor: GATA1/2 RUNX1 CEBPA ETV6 NOTCH1 BCOR WT1 PHF6 Chromosome stability: TERT Cohesin complex (STAG1, SMC3, RAD21) Tumor Suppressor: TP53 Epigenetic: IDH1/2 TET2 EZH2 ASXL1 DNMT3A Other: SETBP1 CALR NPM1 2017 MFMER slide 28 Blood. 2015 Feb 26;125(9):1367 76 Blood. 2016;128(5):686 698 2017 MFMER slide 29 Patterns of AML Mutations SRSF2 SF3B1 U2AF1 ZRSR2 PTPN11 EZH2 CBL ASXL1 STAG2 BCOR K/NRAS FLT3 TP53 IDH1/2 TET2 DNMT3A GATA2 RUNX1 NPM1 CEBPA CBF r KMT2A r Ref: Blood. 2015;125(9):1367 1376 2017 MFMER slide 30 10
N Engl J Med 2016;374:2209 21 2017 MFMER slide 31 References 1. Blood. 2016;127(20):2391 2405 2. Blood. 2017;129(4):424 447 3. J Natl Cancer Net 2017;15:926 957 4. Arch Pathol Lab Med. doi: 10.5858/arpa.2016 0504 CP 2017 MFMER slide 32 Gene Mutation Profiling: Myeloid Tumors AML: Tier 1 Tier 2 FLT3 ITD, IDH2, TP53, CEBPA dm Spliceosome/epigenetic, TP53, KIT, RUNX1,?DNMT3A Tier 3 Germ line associated: GATA2, DDX41, ANKRD26, CEBPA... MDS: SF3B1, TP53, other (mutation burden) MDS/MPN: ASXL1, TET2, SRSF2, JAK2, SETBP1, ETNK1, RAS pathway MPN (PMF): JAK2, CALR, MPL, ASXL1, TP53, CSF3R 2017 MFMER slide 33 11
Part III: Case Presentations 2017 MFMER slide 34 Case 1 62 year old male: macrocytic anemia and decreased platelets CBC: Hgb 10 g/dl, MCV 108.4 fl, WBC 10.3 X10 9 /L, platelets 141 X10 9 /L WBC: Ne 50%, Ly 38%, Mo 12%, 0% blasts Bone marrow evaluated at Mayo Clinic: Hypercellular (50-60%): myelomonocytic hyperplasia with 15% blasts and promonocytes Cytogenetics: 46,XY[20] Diagnosis: Chronic myelomonocytic leukemia-2 (CMML-2) 2017 MFMER slide 35 Case 1 2017 MFMER slide 36 12
Case 1 2017 MFMER slide 37 Case 1 2017 MFMER slide 38 Case 1 2017 MFMER slide 39 13
Case 1 NM_002520.6 NPM1 NM_022552.4 DNMT3A 2017 MFMER slide 40 Case 1 NGS findings: DNMT3A: c.2645g>a; p.arg882his (40%) NPM1: c.860_863dup; p.trp288cysfs*12 (26%) Conservative Rx (decitabine) with follow-up at another institution 6 months later: 22% PB blasts NGS results can inform or refine difficult morphologic diagnosis and/or suggest alternative biologic consequences 2017 MFMER slide 41 CMML: Molecular Pathogenesis 50 45 40 35 30 25 20 15 10 5 0 Adapted from Itzykson R and Solary E Leukemia 2013;27:1441 1450 RAS CBL JAK2 TET2 SRSF2 ASXL1 RUNX1 2017 MFMER slide 42 14
Case 2 33 year old female with diffuse back and hip pain July 2015 CBC: Hgb 12 g/dl, WBC 7.6 X10 9 /L, platelets 24 X10 9 /L WBC: Ne 13%, Ly 18%, blasts 66% Bone marrow biopsy: 92% myeloblasts Flow cytometry: CD45, CD13, CD15, CD33, CD117, HLA-DR, CD7, CD36, CD64, CD38 Cytogenetics 46,XX[20]; AML FISH negative 2017 MFMER slide 43 Case 2 NGS findings: FLT3: c.2503g>t; p.asp835tyr (18%) NPM1: c.860_863dup; p.trp288cysfs*12 (39%) PTPN11: c.218c>t; p.thr73ile (26%) WT1: c.938c>a; p.ser313* (44%) Rx: HiDAC induction chemotherapy with CR1; MUD allogeneic SCT in 10/2015 Remains in complete remission >12 months NGS results may indicate gene mutation patterns with variable, conflicting or uncertain prognostic information 2017 MFMER slide 44 Study Outcome Summary Virappane 2008: 470 CN-AML (94% de novo) WT1+ 10% Paschka 2008: 196 young (<60) with de novo CN-AML WT1+ 10.7%; Associated with FLT3-ITD and higher WBC Median F/U: 122.4 mo RFS: 5 yr: 22% vs 44% (HR 2.16; p 0.005) OS:5 years: 26% vs 47% HR 1.91 p=0.007 Median F/U : 50.4 mo RFS: 3 yr DFS: 13% vs 50% (p<0.001) OS: 10% vs 56% (p<0.001) Worse OS, RFS independent of NPM1 and FLT3 ITD+; Remained significant for OS, RFS at ASCT Poorer DFS (HR 2.7 (p=0.009) controlling for ERG, FLT3 ITD, CEBPA Poorer OS HR 3.2, p=<0.001) controlling for CEBPA, FLT3 ITD/NPM1, WBC Gaidzik 2009: 617 De novo, s-aml, t-aml WT1+ 12.6%; Associated with younger age, FLT3-ITD+ and CEBPA, higher blasts, inverse NPM1+ Med F/U 44 mo RFS: NO difference WT1+/- but WT+/FLT3-ITD + subgroup P=0.006 OS: NO difference, but WT1+/FLT-ITD + subgroup P<0.001 WT1 mutations had NO significant impact. NPM1, FLT3 ITD, CEBPA, MLL were significant Hou 2010: 470 non-m3 AML WT1+ 6.8% (33% M6) Krauth 2015: 3157 AML (1178 CN-AML) WT1+ 6%; FLT3-ITD+ 8.5% Bi-CEBPA 13.6% Median F/U: 53 mo RFS: 6 mo vs 14 mo (p<0.001) OS: 14 mo vs. 29.5 mo (p=0.021) RFS: 10.8 mo vs 17.9 mo (p=0.008) OS: NO significant impact Poorer OS and RFS in all and young CN-AML pts. (RR 3.752, p=0.024; RR 3.8, p-0.003) NPM1+/FLT3 ITD- favorable Poorer EFS controlling for FLT3 ITD, age HR 1.64 (p=0.002) NPM1+/FLT3 ITD- HR 0.68 (P<0.001) 2017 MFMER slide 45 15
Case 3 75 year old with recent Dx lung adenocarcinoma History of thrombocytopenia, apparently for several years, previous normal bone marrow Hgb 12.4 g/dl, MCV 99.2 fl, WBC 3.5 X10 9 /L, platelets 45 X10 9 /L WBC: Ne 50%, Ly 743%, Mo 4%, Eo 3% Bone marrow 12/15: hypercellular (50%) with 5% ring sideroblasts, slightly decreased megakaryocytes and no increase in blasts Diagnostic morphologic features of MDS not established Cytogenetics: 46,XX,del(20)(q11.2q13.2)[1/46,XY[19] 2017 MFMER slide 46 2017 MFMER slide 47 2017 MFMER slide 48 16
Case 3 NGS findings: NM_001025203.1 U2AF1: c.4708a>c; p.gln157pro (16%) 2017 MFMER slide 49 2017 MFMER slide 50 Jaiswal S. N Engl J Med 2014;371:2488 98 2017 MFMER slide 51 17
CHIP-ICUS-CCUS-Cancer Spectrum CHIP ICUS CCUS Individuals without history or concurrent hematologic cancer Somatic mutations in one or sometimes >1 gene detected at >2% variant allele fraction (VAF) TET2, DNMT3A, ASXL1, SF3B1, TP53 Increasing incidence with age Identified from WES/WGS of large populations Variable risk of progression to myeloid cancer (0.5 1% per year; 10X OR risk) Unexplained cytopenia(s) in absence of cytogenetic and NGS detectable clonal population Uncertain risk of neoplastic progression, likely low CHIP + ICUS Risk of progression to myeloid cancer is higher and becoming more quantifiable 2017 MFMER slide 52 CHIP: Associations CHIP is highly prevalent (>90%) after age 50 with high sensitivity techniques {Nat Commun. 2016;7:12484} As defined, CHIP increases risk for developing a hematologic neoplasm Risk increases if multiple mutations, higher VAF% Risk of all-cause mortality and cardiovascular disease Therapy-related myeloid neoplasms: higher risk for development of T-MN with pre-existing CHIP {Lancet Oncol 2017; 18: 100 11; Lancet Oncol 2017; 18: 112 21} Presence of CHIP in donor HSCT allograft can result in impaired graft function in recipient but not donor-derived leukemia {Blood 2017;130:91-94} Not all CHIP is equal {Blood. 2017;130(6):753-762} TET2 and DNMT3A very common, yet not clearly associated with adverse effects on hematopoietic function, or risk of myeloid neoplasia Clonal size expands with age in some individuals:? Compensation for failing wild type HSCs 2017 MFMER slide 53 CCUS: Risk of Neoplastic Progression Earlier studies: highly suggestive of increased risk for ICUS with clonal hematopoietic mutations to progress to overt myeloid cancers {Blood. 2015;126(21):2362-2365; Blood. 2015;126(21):2355-2361} Recent large studies confirm and delineate the risk more specifically {Blood. 2017;129(25):3371-3378; Am. J. Hematol. 91:1234 1238, 2016} Longitudinal follow-up VAF>10%; >2 mutations and spliceosome mutations with classic CHIP alterations have very high PPV for progression to myeloid neoplasm HR for progression 13.9 Certain mutation patterns may be more strongly predictive and can provide presumptive evidence for early myeloid neoplasia in the absence of definitive morphologic criteria or cytogenetic abnormalities Conversely the absence of mutations (with sufficiently comprehensive screening) has very high NPV for neoplasia, even in the setting of morphologic atypia 2017 MFMER slide 54 18
Suggested Indications for NGS Testing Useful New AML Relapse AML New Dx MDS MDS/MPN (e.g. CMML) MPN-PMF or triple negative Patient with suspicion of GL syndrome Unexplained cytopenias ( ICUS ) Not Indicated/Uncertain Palliative situation Mastocytosis (KIT D816V, KIT seq) MPN-PV, MPN-ET (MPN-R) MPN-CNL (CSF3R) CML (BCR-ABL) CEL/HES and other neoplastic eosinophilias (FISH) 2017 MFMER slide 55 NGS and Myeloma GENETIC CATEGORY RISK STATUS Hyperdiploid/CNA HD (odd chromosome # s) Standard (35 40%) CNA at +1q, del1p, del17p High (5 10% each) Translocations t(11;14) CCND1, t(6;14) CCND3 Standard (15%) t(14;16) and t(14;20) MAF;MAFB High (<5%) t(4;14) FGFR3; NSD2 (MMSET, WHSC1) Intermediate (10 15%) 8q24 MYC High (20%) Notes: Based on Mayo msmart 2.0 High risk GEP signature is also high risk feature 2017 MFMER slide 56 NGS and Myeloma Modify or additionally stratify current FISH-defined risk classification Focus on additional prognostic alterations and therapeutic/resistance genotypes Ideal state is a single test to capture most relevant alterations in SMM, MM (e.g. SNV, SV, CNV) Technically difficult Possible utility for refractory/relapsed disease In early stages of application and clinical utility determination (indications) 2017 MFMER slide 57 19
NGS and Myeloma MAPK N/KRAS BRAF Other FGFR3 CCND1 FAM46C DIS3 IMiD CRBN CUL4B IKZF1 IRF4 NFKB TRAF3 CYLD DNA DR TP53 ATM ATR EGR1 2017 MFMER slide 58 Summary NGS is a comprehensive and convergent technology platform for hematologic cancer genetics NGS is currently relatively expensive and information dense Useful for diagnosis, prognosis and limited theranostics Evolving understanding of co-mutation patterns and clonal hematopoiesis 2017 MFMER slide 59 Questions & Discussion 2017 MFMER slide 60 20