[PPT] - prediction and assessment of clonal evolution Davide Rossi, M.D., PowerPoint Presentation

SLIDE 1

Role of circulating tumor DNA in response prediction and assessment of clonal evolution

Hematology IOSI - Oncology Institute of Southern Switzerland IOR - Institute of Oncology Research Bellinzona - Switzerland

Davide Rossi, M.D., Ph.D.

SLIDE 2

Conflict of interest

Research Support: Gilead, Abbvie, Janssen, Cellestia Employee No Consultant No Major Stockholder No Speakers Bureau No Honoraria Gilead, Abbvie, Janssen, Roche, AstraZeneca Scientific Advisory Board Gilead, Abbvie, Janssen, AstraZeneca, MSD

SLIDE 3

Agenda

General notions and practicalities
ctDNA in DLBCL
ctDNA in cHL

SLIDE 4

Agenda

General notions and practicalities
ctDNA in DLBCL
ctDNA in cHL

SLIDE 5

Allele frequency of ctDNA mutations ctDNA mutations Mutations VAF>1% Mutations VAF<1% 100% 10% 1% 0%

ctDNA is of low abundance: Optimization of sensitivity and specificity of NGS is mandatory

Spina V, et al. Blood 2018 Allele frequency in cfDNA

Variant frequency (%) Variant position (NM_000546.5) Sensitivity threshold

Background noise of NGS True mutations

0% 1% 10% 100%

Allele frequency in gDNA

SLIDE 6

500 200

Pre-analytics is critical

cfDNA of poor quality: gDNA contamination cfDNA sample of good quality: peak sized between 100 and 200 bp cfDNA cfDNA gDNA

SLIDE 7

Hohaus S et al. Ann Oncol. 2009;20(8):1408-1413

cfDNA circulates in small amounts

SLIDE 8

Challenges in the identification

f small abundant ctDNA variants by NGS
Input DNA (at least 32 ng)
Library preparation chemistry (capture based,

molecular barcoding)

Coverage (>2000X >80% target region)
Bioinformatic pipeline for variant calling

(catalogue of systematic errors)

SLIDE 9

The origin of cell free DNA in healthy subjects and cancer patients

Snyder, Cell 2016

In healthy individuals cfDNA derives from apoptosis of normal hematopoietic cells
In tumor patients cfDNA is released by tumor apoptotic cells
ctDNA is distinguished from other cfDNA by the presence of somatic mutations

representative of tumor biology absent in normal cells

SLIDE 10

Agenda

General notions and practicalities
ctDNA in DLBCL
ctDNA in cHL

SLIDE 11

Diffuse large B-cell lymphoma vs classical Hodgkin lymphoma

Tumor cells are rare in the mass Exome sequencing data from only 10 cases

Reichel J, et al. Blood 2015

Tumor cells are enriched in the mass Exome sequencing data from >1000 cases

DLBCL cHL

Pasqualucci L, et al. Semin Hematol 2015

SLIDE 12

0% 10% 20% 30%

N=30

GC Non-GC

Rossi D, et al. Blood 2017

0% 20% 40% 60% 80% 100% Sensitivity

N=87 N=21 N=18 82.8%

Mutation frequency

Mutation identified both in gDNA and in cfDNA Mutation identified in cfDNA only Mutation identified in gDNA only

CREBBP CCND3 EZH2 KMT2D MYC PIM1 STAT6 TBL1XR1 TNFAIP3 TP53

missense mutation truncating mutation

ctDNA mirrors the genetics of DLBCL cells

SLIDE 13

Circulating tumor DNA resolves the spatial heterogeneity of lymphomas

Scherer F, et al. Sci Transl Med 2016

SLIDE 14

Scherer F, et al. Sci Transl Med 2016

Longitudinal cfDNA genotyping allows Non invasive detection of ibrutinib resistance mutations

SLIDE 15

LymphoSIGHT™ platform

CTGGCCCCAGTAGTCATACCAACTAGCG TTGGCCCCAGAAATCAAGACCATCTAAA ACGGCCCCAGAGATCGAAGTACCAGTGT TTGGCCCCAGACGTCCATATTGTAGTAG CTGGCCCCAGAAGTCAGACCGGCTAACA

1) Collect 10cc peripheral blood 2) Extract DNA 3) Amplify VDJ with multiplex PCR 4) Prepare for sequencing with common PCR 5) Sequence ~1M 100bp reads Genomic DNA PCR amplicons Sequencing library Sequence data Serum

SLIDE 16

5-year TTP of 94.6% vs. 11.8%

MRD at the end of treatment predicts progression

Roschewski M et al. Lancet Oncol, 2015

SLIDE 17

Recovery rate of the tumor IG rearrangement from DLBCL tissue biopsies

Limitations of Ig-HTS as tumor fingerprint in the Liquid biopsy

Kurtz DM et al. Blood, 2015.

SLIDE 18

Mutational profile as tumor fingerprint

PB granulocytes Plasma

4 10 17 26 31

cHL 72

CD36 STAT6 IRF8 PIK3CA ATM ATM FOXO1 MAP3K14 CD79B

Log fold change in ctDNA

5
3
1

1 ND

FOXO1 PIK3CA

Ultra deep sequencing Resolution of the tumor mutation profile

SLIDE 19

The prognostic value of molecular response is independent of interim imaging

Kurtz, J Clin Oncol 2018

SLIDE 20

Agenda

General notions and practicalities
ctDNA in DLBCL
ctDNA in cHL

SLIDE 21

Understanding of the genetics of DLBCL vs cHL

Tumor cells are rare in the mass Exome sequencing data from only 10 cases

Reichel J, et al. Blood 2015

Tumor cells are enriched in the mass Exome sequencing data from >1000 cases

DLBCL cHL

SLIDE 22

ctDNA mirrors the genetics of HRS cells

0% 20% 40% 60% 80%

STAT6 ITPKB TNFAIP3 B2M GNA13 HIST1H1E CIITA IRF8 ARID1A BTG1 IRF4 PCBP1 PIM1 STAT3 ATM BCL6 BTK CCND3 CD58 CXCR4 ID3 KMT2D MYC NFKBIE NOTCH1 PRDM1 SPEN TET2 TNFRSF14 TP53 TRAF3 XPO1

% of mutated cases 5 10 15 20

missense nonsense frameshift splicing start loss 3’-UTR

N. of mutations

80% (12/15) 27% (4/15) 20% (3/15) 13% (2/15) 53% (8/15) 5 10 15 20

N. of mutations

Patient

0% 20% 40% 60% 80% 100%

87.5%

Biopsy confirmed mutations Mutation identified both in gDNA and in ctDNA Mutation identified in ctDNA 7% (1/15)

OTU ZF ZF ZF ZF ZF ZF ZF 790 1

TNFAIP3

1 CATALYTIC domain

ITPKB

*chemorefractory sample

946

a b c d

1 847 STAT_int STAT alpha STAT_bind SH2 STAT6_C

STAT6

missense nonsense frameshift splicing

e

84 13 12 Mutation identified in gDNA 0% 50% 100%

100% 100% 100% 100% 100% 100% 100% 100% 100% 91% 75% 66.6% 66.6% 50% 40%

Biopsy confirmed mutations identified in ctDNA

SLIDE 23

Mutational landscape of newly diagnosed cHL

N=80

Spina V, et al. Blood 2018

STAT6 TNFAIP3 ITPKB GNA13 B2M ATM SPEN KMT2D XPO1 TP53 ARID1A HIST1H1E BTG1

SLIDE 24

NF-κB PI3K-AKT Cytokine signaling Epigenetic genes

46.2% (37/80) 46.2% (37/80) 37.5% (30/80) 35% (28/80) 27.5% (22/80)

immune surveillance genes NOTCH pathway

20% (16/80)

Mutated pathways in newly diagnosed cHL

Spina V, et al. Blood 2018

SLIDE 25

1 2 3 5

1. No uptake 2. FDG < MBP 3. FDG >MBP ≤ liver 4. FDG > liver 5. FDG >> liver

4

Baseline PET Interim PET

False positive rate = 19% False negative rate = 3% Interim PET/CT accuracy in cHL

Terasawa, et al J Clin Oncol 2009 27:1906-1914

SLIDE 26

Changes in tumor cfDNA complement iPET

3 4 4 3 3 3 3 1 2 2 2 4 4 5 4 1 2 1 1 1 1 1 3 1 PD PD PD PD PD PD CR CR CR CR CR CR CR CR CR CR CR CR CR CR CR CR CR CR Outcome

5
4
3
2
1

1 cHL 66 UPN6 UPN25 cHL 85 cHL 67 cHL 52 cHL 19 UPN4 cHL 23 cHL 8 UPN3 cHL 63 cHL 78 cHL 64 cHL 65 cHL 13 UPN11 cHL 74 cHL 75 cHL 76 cHL 79 cHL 80 cHL 83 cHL 86

Deauville Score

iPET positive – Progressive disease iPET negative – Progressive disease iPET positive – Cured iPET negative – Cured Log fold change in tumor ctDNA ND

5
4
3
2
1

1 50 100 150 200 Days from start of therapy Log fold change in tumor ctDNA ND

p<0.001 > -2 log fold reduction < -2 log fold reduction

18 15 2 2 6

a c b d e

4.5 -4.0 -3.5
3.0 -2.5 -2.0
1.5

2.0 2.5 3.0 3.5 4.0 4.5 Log Standardized log-rank statistic

p=0.001

Spina V, et al. Blood 2018

SLIDE 27

Conclusions

Molecular markers informing treatment: BCR signaling

mutations, EZH2 mutations

Clonal evolution: mutation resistance monitoring
Molecular markers informing response to therapy:

minimal residual disease monitoring

SLIDE 28

PET0 ctDNA0 Tx Cycle 1-2 PET2 ctDNA2 Tx Cycle 3-6 Intensification Biological agent PETEOT ctDNAEOT Follow-up Salvage +/+ +/-

/+
/-

+/-

/+

+/+

/-

Clinical trial design incorporating ctDNA assessment

SLIDE 29

Lymphoma Unit Bernhard Gerber Alden Moccia Anastasios Stathis Georg Stüssi Emanuele Zucca Nuclear Medicine Luca Ceriani Michele Ghielmini Clara Deambrogi Lorenzo De Paoli Fary Diop Luca Nassi Gianluca Gaidano Experimental Hematology Alessio Bruscaggin Claudia Cirillo Adalgisa Condoluci Francesca Guidetti Gabriela Forestieri Valeria Spina Lodovico Terzi di Bergamo Lymphoma & Genomics Francesco Bertoni Franco Cavalli Martina Di Trani Silvia Locatelli Carmelo Carlo-Stella Hematology Annarosa Cuccaro Stefan Hohaus Pathology Maurizio Martini Luigi Larocca

Unrestricted research grant from Gilead and Abbvie

SLIDE 30