−−

−

−−− −−−

−

− −−

−−

−

−−

− −

−

−

−0

25

50

75

100

10−2 10−1 100 101 102

cfDNA VAF (%)

Tiss

ue V

AF (%

)

Clonal Statusclonal: called in cfDNA (n=130, 90%)clonal: not called in cfDNA (n=15, 10%)subclonal: called in cfDNA (n=35, 46%)subclonal: not called in cfDNA (n=41, 54%)undetermined: called in cfDNA (n=23, 79%)undetermined: not called in cfDNA (n=6, 21%)

Metastatic Breast Cancer

−−

−−−

−−−

−

−

−−

−−

−

−−

−

−

− −−−

−−

−

−

−−−

− − −

−

−−

−

−0

25

50

75

100

10−2 10−1 100 101 102

cfDNA VAF (%)

Tiss

ue V

AF (%

)Clonal Status

clonal: called in cfDNA (n=164, 80%)clonal: not called in cfDNA (n=42, 20%)subclonal: called in cfDNA (n=39, 59%)subclonal: not called in cfDNA (n=27, 41%)undetermined: called in cfDNA (n=40, 82%)undetermined: not called in cfDNA (n=9, 18%)

Metastatic Lung Cancer

−

−−

−

−

−

−

−−

−

−

−

−

−

−−−−−

−

−

−

−−

0

25

50

75

100

10−2 10−1 100 101 10cfDNA VAF (%)

Tiss

ue V

AF (%

)

Clonal Statusclonal: called in cfDNA (n=87, 77%)clonal: not called in cfDNA (n=26, 23%)subclonal: called in cfDNA (n=15, 52%)subclonal: not called in cfDNA (n=14, 48%)undetermined: called in cfDNA (n=20, 77%)undetermined: not called in cfDNA (n=6, 23%)

Metastatic Prostate Cancer

2

− hotspot (n = 23) − hotspot (n = 38) − hotspot (n = 24)

Performance of a high-intensity 508-gene circulating-tumor DNA (ctDNA) assay in patients with metastatic breast, lung, and prostate cancerPedram Razavi1, Bob T. Li1, Wassim Abida1, Alex Aravanis9, Byoungsok Jung9, Ronglai Shen2, Chenlu Hou9, Ino De Bruijn4, Sante Gnerre9, Raymond S. Lim4, Earl Hubbell9, Dalicia Reales3, Tara Maddala9, Michael F. Berger1, 4, 6, Gregory J. Riely7, Howard I. Scher1, 8, William F. Novotny9, David B. Solit1, 3, 6, Mark Lee9, Jorge S. Reis-Filho4, Jose Baselga1, 3

1. Department of Medicine, Memorial Sloan Kettering Cancer Center (MSKCC), 1275 York Avenue, New York, NY, USA. 2. Department of Epidemiology and Biostatistics, Memorial Sloan Kettering Cancer Center. 3. Human Oncology and Pathogenesis Program, Memorial Sloan Kettering Cancer Center. 4. Department of Pathology, Memorial Sloan Kettering Cancer Center. 5. Department of Pathology, Molecular Diagnostics Service, Memorial Sloan Kettering Cancer Center. 6. Marie-Josée and Henry R. Kravis Center for Molecular Oncology, Memorial Sloan Kettering Cancer Center. 7. Thoracic Oncology Service, Division of Solid Tumor Oncology, Department of Medicine, Memorial Sloan Kettering Cancer Center. 8. Genitourinary Oncology Service, Department of Medicine, Memorial Sloan Kettering Cancer Center 9. GRAIL, Menlo Park, CA 94402 USA

Background• ctDNA assays can noninvasively assess

tumor burden and biology by identifying tumor-derived somatic alterations

• To date, ctDNA studies have focused primarily on detecting driver mutations to inform treatment strategies in advanced disease or monitoring disease burden in patients with established cancer diagnoses. Platforms used for these purposes target individual variants or limited genomic regions informed by sequencing of tumor tissue. (Wan 2017)

• Analysis of plasma cell-free DNA may enable early cancer detection in previously undiagnosed individuals but will require de novo variant calling (in the absence of tissue) as well as sufficient genomic coverage to address the spectrum of variant profiles that are cancer-defining. (Aravanis 2017)

• We propose that a high-intensity approach (ultra-deep sequencing of plasma cell-free DNA with broad genomic coverage) will add a new dimension to our understanding of intra-patient and population-level heterogeneity.

Objectives• Assess concordance of variants detected in

tissue with MSK-IMPACT™ versus detected in plasma cell-free DNA.

• Assess the cell-free DNA variant detection rate based on observing at least one MSK-IMPACT™ tissue variant in the same patient.

• Assess concordance using tissue as a reference according to:

- Clinical actionability - Clonality

Methods: Patient Population• Metastatic breast, lung, or prostate cancer,

either de novo or with progressive disease on current therapy.

• All patients have provided written informed consent to an MSK institutional protocol (NCT01775072) allowing research of cfDNA and clinical tumor sequencing.

• Blood and tissue were prospectively collected within 6 weeks of each other with no intervening therapy change.

- Two tubes of blood collected in Streck. - Tissue from surgical resection or biopsy.

• Blood and tissue were analyzed independently and blinded to the results of each.

Results: Patient Disposition

Methods: AnalysiscfDNA variant calling pipeline• The variant calling pipeline includes the following steps:

- Read alignment, error correction (consisting of read collapsing by position and UMI, as well as stitching paired reads), de novo assembly, variant calling, and variant filtering.

- Variants are filtered using two approaches (1) heuristics applied based on the surrounding sequence context and type and quality of reads supporting the variant and (2) empirical noise levels observed in a set of healthy samples.

- Blood samples from an independent cohort of 24 non-cancer (self-reported) individuals (vendor sourced) were used to set a baseline noise level in this study.

- Variant calls are further filtered using matching WBC for each patient. - Bioinformatics filters included removing non-COSMIC dbSNP variants as well as restricting to protein coding regions.

MSK tissue variant calling pipeline has been described extensively in the past (Cheng 2015, Zehir 2017)

FACETS algorithm and clonality analysis:• To classify the variants identified in tissue into clonal and subclonal mutations, the cancer cell fraction (CCF) was

calculated based on mutant variant allele frequency (VAF) adjusting for tumor purity, ploidy, and local allele-specific copy number obtained from the FACETS algorithm (Shen 2016).

• Exact confidence intervals (CI) were calculated around the point estimate of CCF: - Mutations with lower bound of 95% CI ≥75% were classified as “clonal”. - Mutations with CCF ≥ 80% and lower bound of 95% CI below 75% were classified as “likely clonal”. - Mutations with CCF < 80% and lower bound of 95% CI below 75% were classified as “subclonal”.

Statistical Methods• Only regions covered by both panels were used for concordance analysis and synonymous variants were excluded from

the concordance analysis since they are not called by the MSK-IMPACT™ pipeline.• Concordance is calculated as positive percent agreement (PPA) with tissue as a reference (cfDNA/tissue).

Methods: Sample Workflows and Assays Results: Per Patient Variant Detection | Primary Analysis – Independent cfDNA and Tissue Assessment

Actionable Mutations in Tissue and Plasma

Consistency of Mutational Signature | APOBEC Hypermutation in Both Tissue and Plasma

• Patient mutational signatures were identified by deconvolving the observed triplet mutation profile of patients using constrained linear regression onto COSMIC mutational signatures (Alexandrov 2013). Patients with evidence for signatures compatible with APOBEC hypermutation were prioritized for cfDNA signature analysis.

• Seven patients (6 breast, 1 prostate) exhibited evidence for increased signature 2 and/or 13 in both tissue and plasma, which are both associated with APOBEC hypermutation; Two patients (MSK-VB-0023 and MSK-VB-0046) shown here.

Patient Characteristics Cancer Characteristics

Enrollment

Analysis

Assigned pt ID (N = 232)

Breast80

Lung82

Prostate70

EligibleEligibleClinically Eligible and Lab Evaluable (N = 161)

NBreast

53

Lung

53

Prostate

55

EligibleClinically Eligible and Lab Evaluable (N = 161)

Breast53

Lung53

Prostate55

cfDNA Assay Evaluable (N = 151)Breast

48Lung

49Prostate

54

Tissue assay unevaluable (N = 27)

Breast

95

Lung

84

Prostate

101cfDNA assay unevaluable (N = 10)

Excluded subjects (N = 71)

Clinical exclusion*Lab exclusions for insufficient sampleTissue or cfDNA assay unevaluable

Breast

27

9108

Lung

29

20**81

Prostate

15

708

* Clinical exclusion includes new systemic therapy, disease progression not

Tissue Assay Evaluable (N = 134)Breast

44Lung

45Prostate

45

Concordance Cohort (N = 124)

Breast39

Lung41

Prostate44

Both Tissue and cfDNA Evaluable

** Includes 2 pseudoprogressionconfirmed on review, etc

Percent of patients for whom at least 1 of the variants detected in tumor tissue was also detected in plasma: Breast cancer - 97%, Lung cancer - 85%, Prostate cancer - 84%.

*Analytical sensitivity and specificity determined by cell line and cfDNA titrations, respectively

cfDNA and tissue assays share 1.2 Mb of targeted sequence.

Most actionable mutations detected in tissue were also detected in plasma (54/71, 76%; SNVs only: 28/31, 90%). A subset of driver mutations were observed in plasma but not tissue, including some with potential therapeutic implications:• Breast: PIK3CA (E453K, E542K, E545K, E726K, M1043I), ERBB2 I767M• Lung: EGFR T790M• Prostate: AR amp, AR (L702H, T878A), BRCA1 trunc, BRCA2 trunc, MLH1, PIK3CA E545K

Num

ber o

f Var

iant

s

0

5

10

15

20

Metastatic Breast Cancer Metastatic Lung Cancer Metastatic Prostate Cancer

AK

T1−E

17K

ER

BB

2 A

mp

FGFR

1 A

mp

FGFR

2 A

mp

PIK

3CA

−E54

2K

PIK

3CA

−E54

5K

PIK

3CA

−E72

6K

PIK

3CA

−H10

47R

PIK

3CA

−M10

43I

PIK

3CA

−N34

5K

EG

FR E

xon

19

EG

FR−G

719A

EG

FR−L

858R

EG

FR−L

861Q

EG

FR−T

790M

ME

T A

mp

RO

S1

Fusi

on

AR

Am

p

AR

−H87

5Y

AR

−L70

2H

BR

CA

1 tru

n

BR

CA

2 tru

n

PIK

3CA

−E54

2K

MSK−IMPACT onlyMSK−IMPACT & GRAIL

MSK−VB−0023 cfDNA

# M

utat

ions

(cfD

NA

)

0

200

400

600

800

1000

C>A C>G C>T T>A T>C T>G

MSK−VB−0023 Tumor

# M

utat

ions

(Tum

or)

0

2

4

6

8

C>A C>G C>T T>A T>C T>G

0.0

0.2

0.4

0.6

0.8

1.0

0.0

0.2

0.4

0.6

0.8

1.0

MSK−VB−0046 cfDNA

# M

utat

ions

(cfD

NA

)

0

10

20

30

40

50

60

C>A C>G C>T T>A T>C T>G

MSK−VB−0046 Tumor

# M

utat

ions

(Tum

or)

0

2

4

6

8

C>A C>G C>T T>A T>C T>G

0.0

0.2

0.4

0.6

0.8

1.0

0.0

0.2

0.4

0.6

0.8

1.0

Metastatic Prostate CancerTissue calls detection rate by cfDNA

Metastatic Lung CancerTissue calls detection rate by cfDNA

Metastatic Breast CancerTissue calls detection rate by cfDNA

S30S18S15S13S10S7S6S2S1

MS

K−V

B−0

013

MS

K−V

B−0

004

MS

K−V

B−0

024

MS

K−V

B−0

053

MS

K−V

B−0

005

MS

K−V

B−0

006

MS

K−V

B−0

008

MS

K−V

B−0

030

MS

K−V

B−0

059

MS

K−V

B−0

066

MS

K−V

B−0

076

MS

K−V

B−0

016

MS

K−V

B−0

032

MS

K−V

B−0

036

MS

K−V

B−0

062

MS

K−V

B−0

001

MS

K−V

B−0

054

MS

K−V

B−0

063

MS

K−V

B−0

069

MS

K−V

B−0

072

MS

K−V

B−0

084

MS

K−V

B−0

077

MS

K−V

B−0

033

MS

K−V

B−0

037

MS

K−V

B−0

031

MS

K−V

B−0

034

MS

K−V

B−0

012

MS

K−V

B−0

041

MS

K−V

B−0

058

MS

K−V

B−0

040

MS

K−V

B−0

044

MS

K−V

B−0

067

MS

K−V

B−0

029

MS

K−V

B−0

045

MS

K−V

B−0

073

MS

K−V

B−0

023

MS

K−V

B−0

057

MS

K−V

B−0

050

MS

K−V

B−0

046

Metastatic Breast Cancer

Num

ber o

f Var

iant

s

0

5

10

15

20

25

30

35

MS

K−V

L−00

73M

SK

−VL−

0002

MS

K−V

L−00

10M

SK

−VL−

0015

MS

K−V

L−00

64M

SK

−VL−

0009

MS

K−V

L−00

83M

SK

−VL−

0005

MS

K−V

L−00

13M

SK

−VL−

0051

MS

K−V

L−00

67M

SK

−VL−

0069

MS

K−V

L−00

71M

SK

−VL−

0007

MS

K−V

L−00

21M

SK

−VL−

0028

MS

K−V

L−00

03M

SK

−VL−

0061

MS

K−V

L−00

76M

SK

−VL−

0077

MS

K−V

L−00

82M

SK

−VL−

0012

MS

K−V

L−00

42M

SK

−VL−

0044

MS

K−V

L−00

56M

SK

−VL−

0027

MS

K−V

L−00

45M

SK

−VL−

0054

MS

K−V

L−00

17M

SK

−VL−

0057

MS

K−V

L−00

22M

SK

−VL−

0038

MS

K−V

L−00

80M

SK

−VL−

0074

MS

K−V

L−00

75M

SK

−VL−

0048

MS

K−V

L−00

55M

SK

−VL−

0008

MS

K−V

L−00

60M

SK

−VL−

0065

MS

K−V

L−00

35

Metastatic Lung Cancer

Num

ber o

f Var

iant

s

0

5

10

15

20

25

30

35

MS

K−V

P−0

030

MS

K−V

P−0

040

MS

K−V

P−0

002

MS

K−V

P−0

005

MS

K−V

P−0

007

MS

K−V

P−0

028

MS

K−V

P−0

029

MS

K−V

P−0

034

MS

K−V

P−0

035

MS

K−V

P−0

003

MS

K−V

P−0

008

MS

K−V

P−0

015

MS

K−V

P−0

016

MS

K−V

P−0

018

MS

K−V

P−0

026

MS

K−V

P−0

037

MS

K−V

P−0

042

MS

K−V

P−0

045

MS

K−V

P−0

051

MS

K−V

P−0

056

MS

K−V

P−0

001

MS

K−V

P−0

031

MS

K−V

P−0

032

MS

K−V

P−0

059

MS

K−V

P−0

004

MS

K−V

P−0

009

MS

K−V

P−0

010

MS

K−V

P−0

014

MS

K−V

P−0

033

MS

K−V

P−0

038

MS

K−V

P−0

046

MS

K−V

P−0

047

MS

K−V

P−0

050

MS

K−V

P−0

061

MS

K−V

P−0

006

MS

K−V

P−0

021

MS

K−V

P−0

054

MS

K−V

P−0

019

MS

K−V

P−0

024

MS

K−V

P−0

025

MS

K−V

P−0

058

MS

K−V

P−0

057

MS

K−V

P−0

041

MS

K−V

P−0

023

Metastatic Prostate Cancer

Num

ber o

f Var

iant

s

0

5

10

15

20

25

30

35

Tissue assay onlyTissue & cfDNA assay

Post-hoc Analysis | Association between Cancer Cell Fraction (CCF) in Tumor and Detection Rate in Plasma

• Clonal variants in tissue were more likely to be detected in plasma than subclonal variants (p<.0001).• The greater the representation of cancer cell fraction (CCF) from FACETS in the tumor, the higher the detection in plasma.

Num

ber o

f Var

iant

s

0

5

10

15

20

Metastatic Breast Cancer Metastatic Lung Cancer Metastatic Prostate Cancer

AK

T1−E

17K

ER

BB

2 A

mp

FGFR

1 A

mp

FGFR

2 A

mp

PIK

3CA

−E54

2K

PIK

3CA

−E54

5K

PIK

3CA

−E72

6K

PIK

3CA

−H10

47R

PIK

3CA

−M10

43I

PIK

3CA

−N34

5K

EG

FR E

xon

19

EG

FR−G

719A

EG

FR−L

858R

EG

FR−L

861Q

EG

FR−T

790M

ME

T A

mp

RO

S1

Fusi

on

AR

Am

p

AR

−H87

5Y

AR

−L70

2H

BR

CA

1 tru

n

BR

CA

2 tru

n

PIK

3CA

−E54

2K

MSK−IMPACT onlyMSK−IMPACT & GRAIL

MSK−VB−0023 cfDNA

# M

utat

ions

(cfD

NA

)

0

200

400

600

800

1000

C>A C>G C>T T>A T>C T>G

MSK−VB−0023 Tumor

# M

utat

ions

(Tum

or)

0

2

4

6

8

C>A C>G C>T T>A T>C T>G

0.0

0.2

0.4

0.6

0.8

1.0

0.0

0.2

0.4

0.6

0.8

1.0

MSK−VB−0046 cfDNA#

Mut

atio

ns (c

fDN

A)

0

10

20

30

40

50

60

C>A C>G C>T T>A T>C T>G

MSK−VB−0046 Tumor

# M

utat

ions

(Tum

or)

0

2

4

6

8

C>A C>G C>T T>A T>C T>G

0.0

0.2

0.4

0.6

0.8

1.0

0.0

0.2

0.4

0.6

0.8

1.0

Metastatic Prostate CancerTissue calls detection rate by cfDNA

Metastatic Lung CancerTissue calls detection rate by cfDNA

Metastatic Breast CancerTissue calls detection rate by cfDNA

S30S18S15S13S10S7S6S2S1

MS

K−V

B−0

013

MS

K−V

B−0

004

MS

K−V

B−0

024

MS

K−V

B−0

053

MS

K−V

B−0

005

MS

K−V

B−0

006

MS

K−V

B−0

008

MS

K−V

B−0

030

MS

K−V

B−0

059

MS

K−V

B−0

066

MS

K−V

B−0

076

MS

K−V

B−0

016

MS

K−V

B−0

032

MS

K−V

B−0

036

MS

K−V

B−0

062

MS

K−V

B−0

001

MS

K−V

B−0

054

MS

K−V

B−0

063

MS

K−V

B−0

069

MS

K−V

B−0

072

MS

K−V

B−0

084

MS

K−V

B−0

077

MS

K−V

B−0

033

MS

K−V

B−0

037

MS

K−V

B−0

031

MS

K−V

B−0

034

MS

K−V

B−0

012

MS

K−V

B−0

041

MS

K−V

B−0

058

MS

K−V

B−0

040

MS

K−V

B−0

044

MS

K−V

B−0

067

MS

K−V

B−0

029

MS

K−V

B−0

045

MS

K−V

B−0

073

MS

K−V

B−0

023

MS

K−V

B−0

057

MS

K−V

B−0

050

MS

K−V

B−0

046

Metastatic Breast Cancer

Num

ber o

f Var

iant

s

0

5

10

15

20

25

30

35

MS

K−V

L−00

73M

SK

−VL−

0002

MS

K−V

L−00

10M

SK

−VL−

0015

MS

K−V

L−00

64M

SK

−VL−

0009

MS

K−V

L−00

83M

SK

−VL−

0005

MS

K−V

L−00

13M

SK

−VL−

0051

MS

K−V

L−00

67M

SK

−VL−

0069

MS

K−V

L−00

71M

SK

−VL−

0007

MS

K−V

L−00

21M

SK

−VL−

0028

MS

K−V

L−00

03M

SK

−VL−

0061

MS

K−V

L−00

76M

SK

−VL−

0077

MS

K−V

L−00

82M

SK

−VL−

0012

MS

K−V

L−00

42M

SK

−VL−

0044

MS

K−V

L−00

56M

SK

−VL−

0027

MS

K−V

L−00

45M

SK

−VL−

0054

MS

K−V

L−00

17M

SK

−VL−

0057

MS

K−V

L−00

22M

SK

−VL−

0038

MS

K−V

L−00

80M

SK

−VL−

0074

MS

K−V

L−00

75M

SK

−VL−

0048

MS

K−V

L−00

55M

SK

−VL−

0008

MS

K−V

L−00

60M

SK

−VL−

0065

MS

K−V

L−00

35

Metastatic Lung Cancer

Num

ber o

f Var

iant

s

0

5

10

15

20

25

30

35

MS

K−V

P−0

030

MS

K−V

P−0

040

MS

K−V

P−0

002

MS

K−V

P−0

005

MS

K−V

P−0

007

MS

K−V

P−0

028

MS

K−V

P−0

029

MS

K−V

P−0

034

MS

K−V

P−0

035

MS

K−V

P−0

003

MS

K−V

P−0

008

MS

K−V

P−0

015

MS

K−V

P−0

016

MS

K−V

P−0

018

MS

K−V

P−0

026

MS

K−V

P−0

037

MS

K−V

P−0

042

MS

K−V

P−0

045

MS

K−V

P−0

051

MS

K−V

P−0

056

MS

K−V

P−0

001

MS

K−V

P−0

031

MS

K−V

P−0

032

MS

K−V

P−0

059

MS

K−V

P−0

004

MS

K−V

P−0

009

MS

K−V

P−0

010

MS

K−V

P−0

014

MS

K−V

P−0

033

MS

K−V

P−0

038

MS

K−V

P−0

046

MS

K−V

P−0

047

MS

K−V

P−0

050

MS

K−V

P−0

061

MS

K−V

P−0

006

MS

K−V

P−0

021

MS

K−V

P−0

054

MS

K−V

P−0

019

MS

K−V

P−0

024

MS

K−V

P−0

025

MS

K−V

P−0

058

MS

K−V

P−0

057

MS

K−V

P−0

041

MS

K−V

P−0

023

Metastatic Prostate Cancer

Num

ber o

f Var

iant

s

0

5

10

15

20

25

30

35

Tissue assay onlyTissue & cfDNA assay

Post-hoc Analysis | ctDNA VAF versus Tissue VAF by Clonal Status

Num

ber o

f Var

iant

s

0

5

10

15

20

Metastatic Breast Cancer Metastatic Lung Cancer Metastatic Prostate Cancer

AK

T1−E

17K

ER

BB

2 A

mp

FGFR

1 A

mp

FGFR

2 A

mp

PIK

3CA

−E54

2K

PIK

3CA

−E54

5K

PIK

3CA

−E72

6K

PIK

3CA

−H10

47R

PIK

3CA

−M10

43I

PIK

3CA

−N34

5K

EG

FR E

xon

19

EG

FR−G

719A

EG

FR−L

858R

EG

FR−L

861Q

EG

FR−T

790M

ME

T A

mp

RO

S1

Fusi

on

AR

Am

p

AR

−H87

5Y

AR

−L70

2H

BR

CA

1 tru

n

BR

CA

2 tru

n

PIK

3CA

−E54

2K

MSK−IMPACT onlyMSK−IMPACT & GRAIL

MSK−VB−0023 cfDNA

# M

utat

ions

(cfD

NA

)

0

200

400

600

800

1000

C>A C>G C>T T>A T>C T>G

MSK−VB−0023 Tumor

# M

utat

ions

(Tum

or)

0

2

4

6

8

C>A C>G C>T T>A T>C T>G

0.0

0.2

0.4

0.6

0.8

1.0

0.0

0.2

0.4

0.6

0.8

1.0

MSK−VB−0046 cfDNA

# M

utat

ions

(cfD

NA

)

0

10

20

30

40

50

60

C>A C>G C>T T>A T>C T>G

MSK−VB−0046 Tumor

# M

utat

ions

(Tum

or)

0

2

4

6

8

C>A C>G C>T T>A T>C T>G

0.0

0.2

0.4

0.6

0.8

1.0

0.0

0.2

0.4

0.6

0.8

1.0

Metastatic Prostate CancerTissue calls detection rate by cfDNA

Metastatic Lung CancerTissue calls detection rate by cfDNA

Metastatic Breast CancerTissue calls detection rate by cfDNA

S30S18S15S13S10S7S6S2S1

MS

K−V

B−0

013

MS

K−V

B−0

004

MS

K−V

B−0

024

MS

K−V

B−0

053

MS

K−V

B−0

005

MS

K−V

B−0

006

MS

K−V

B−0

008

MS

K−V

B−0

030

MS

K−V

B−0

059

MS

K−V

B−0

066

MS

K−V

B−0

076

MS

K−V

B−0

016

MS

K−V

B−0

032

MS

K−V

B−0

036

MS

K−V

B−0

062

MS

K−V

B−0

001

MS

K−V

B−0

054

MS

K−V

B−0

063

MS

K−V

B−0

069

MS

K−V

B−0

072

MS

K−V

B−0

084

MS

K−V

B−0

077

MS

K−V

B−0

033

MS

K−V

B−0

037

MS

K−V

B−0

031

MS

K−V

B−0

034

MS

K−V

B−0

012

MS

K−V

B−0

041

MS

K−V

B−0

058

MS

K−V

B−0

040

MS

K−V

B−0

044

MS

K−V

B−0

067

MS

K−V

B−0

029

MS

K−V

B−0

045

MS

K−V

B−0

073

MS

K−V

B−0

023

MS

K−V

B−0

057

MS

K−V

B−0

050

MS

K−V

B−0

046

Metastatic Breast Cancer

Num

ber o

f Var

iant

s

0

5

10

15

20

25

30

35

MS

K−V

L−00

73M

SK

−VL−

0002

MS

K−V

L−00

10M

SK

−VL−

0015

MS

K−V

L−00

64M

SK

−VL−

0009

MS

K−V

L−00

83M

SK

−VL−

0005

MS

K−V

L−00

13M

SK

−VL−

0051

MS

K−V

L−00

67M

SK

−VL−

0069

MS

K−V

L−00

71M

SK

−VL−

0007

MS

K−V

L−00

21M

SK

−VL−

0028

MS

K−V

L−00

03M

SK

−VL−

0061

MS

K−V

L−00

76M

SK

−VL−

0077

MS

K−V

L−00

82M

SK

−VL−

0012

MS

K−V

L−00

42M

SK

−VL−

0044

MS

K−V

L−00

56M

SK

−VL−

0027

MS

K−V

L−00

45M

SK

−VL−

0054

MS

K−V

L−00

17M

SK

−VL−

0057

MS

K−V

L−00

22M

SK

−VL−

0038

MS

K−V

L−00

80M

SK

−VL−

0074

MS

K−V

L−00

75M

SK

−VL−

0048

MS

K−V

L−00

55M

SK

−VL−

0008

MS

K−V

L−00

60M

SK

−VL−

0065

MS

K−V

L−00

35

Metastatic Lung Cancer

Num

ber o

f Var

iant

s

0

5

10

15

20

25

30

35

MS

K−V

P−0

030

MS

K−V

P−0

040

MS

K−V

P−0

002

MS

K−V

P−0

005

MS

K−V

P−0

007

MS

K−V

P−0

028

MS

K−V

P−0

029

MS

K−V

P−0

034

MS

K−V

P−0

035

MS

K−V

P−0

003

MS

K−V

P−0

008

MS

K−V

P−0

015

MS

K−V

P−0

016

MS

K−V

P−0

018

MS

K−V

P−0

026

MS

K−V

P−0

037

MS

K−V

P−0

042

MS

K−V

P−0

045

MS

K−V

P−0

051

MS

K−V

P−0

056

MS

K−V

P−0

001

MS

K−V

P−0

031

MS

K−V

P−0

032

MS

K−V

P−0

059

MS

K−V

P−0

004

MS

K−V

P−0

009

MS

K−V

P−0

010

MS

K−V

P−0

014

MS

K−V

P−0

033

MS

K−V

P−0

038

MS

K−V

P−0

046

MS

K−V

P−0

047

MS

K−V

P−0

050

MS

K−V

P−0

061

MS

K−V

P−0

006

MS

K−V

P−0

021

MS

K−V

P−0

054

MS

K−V

P−0

019

MS

K−V

P−0

024

MS

K−V

P−0

025

MS

K−V

P−0

058

MS

K−V

P−0

057

MS

K−V

P−0

041

MS

K−V

P−0

023

Metastatic Prostate Cancer

Num

ber o

f Var

iant

s

0

5

10

15

20

25

30

35

Tissue assay onlyTissue & cfDNA assay

Patient Characteristics Breast (n=39) Lung (n=41) Prostate (n=44)

Age at enrollment

Mean (SD) 56.5 (11.55) 65.2 (11.18) 67.3 (10.03)

Median 60 67 67

Range 30, 79 33, 83 46, 87

Gender, N (%)

Female 39 (100.0%) 28 (68.3%) N/A

Histology N (%)

Metastatic Breast Cancer (n=39)

Breast Invasive Ductal Carcinoma 32 (82.1%)

Breast Invasive Lobular Carcinoma 2 (5.1%)

Breast Mixed Ductal and Lobular Carcinoma 5 (12.8%)

Metastatic Lung Cancer (n=41)

Lung Adenocarcinoma 38 (92.7%)

Lung Non-adenocarcinoma 3 (7.3%)

Metastatic Prostate Cancer (n=44)

Prostate Adenocarcinoma 39 (88.6%)

Prostate Neuroendocrine 5 (11.4%)

Number of patients

Evaluable With at least one clonal mutation detected in tissue

At least one clonal mutation detected in cfDNA

Metastatic Breast Cancer 39 36 35 (97%)

Metastatic Lung Cancer 41 37 31 (84%)

Metastatic Prostate Cancer 44 40 33 (83%)

Post-hoc Analysis on Clonal Mutations

Pooled Variant Detection | Primary Analysis – Independent cfDNA and Tissue Assessment

PPA = positive percent agreement with tissue as a reference (cfDNA/tissue)In tissue, pooled across patients, 864 variants were detected across the 3 tumor types, with 627 (73%) also detected in plasma: single nucleotide variants/indels - 75%, fusions - 67%, and copy number alterations - 58%. * For breast and lung cancer, defined based on OncoKB, precision knowledge base maintained at MSKCC. In breast cancer, variants known to contribute to progression, level 1, 2, or 3 evidence. In lung cancer, variants shown to predict sensitivity or resistance to targeted therapy, level 1 or 2 evidence. In prostate cancer, variants shown to contribute to progression (Chang 2015, Robinson 2015).

All Variants (SNV, indels, CNA, Fusions) SNV/ indels ONLY Clinically Actionable Mutations*

/tissue PPA (95% CI) cfDNA/tissue PPA (95% CI) cfDNA/tissue PPA (95% CI)

Metastatic Breast Cancer Patient 221 / 300 74% (68, 79) 188 / 250 75% (69, 80) 24 / 27 89% (71, 98)

Metastatic Lung Cancer Patient 256 / 355 72% (67, 77) 243 / 321 76% (71, 80) 11 / 18 61% (36, 83)

Metastatic Prostate Cancer Patient 150 / 209 72% (65, 78) 122 / 168 73% (65, 79) 19 / 26 73% (52, 88)

All Patients 627 / 864 73% (69, 76) 553 / 739 75% (72, 78) 54 / 71 76% (64, 85)

cfDNA Assay Tissue Assay

Inputs PlasmaWBC

FFPE tissue biopsyWBC

Genes 508 410

Breadth (Mb) 2.13 1.36

Raw sequencing coverage >60,000X (~3-4000 error corrected depth) 500 - 1000X

Enrichment Hybridization capture Hybridization capture

Analytical Metrics

SNV/indel detection (30 ng input DNA)Sensitivity*: >95% @ 0.2% MAF >70% @ 0.1% MAF Specificity*: 99.992%

LOD of 2% for hotspot mutations and 5% for non-hotspot mutations

Num

ber o

f Var

iant

s

0

5

10

15

20

Metastatic Breast Cancer Metastatic Lung Cancer Metastatic Prostate Cancer

AK

T1−E

17K

ER

BB

2 A

mp

FGFR

1 A

mp

FGFR

2 A

mp

PIK

3CA

−E54

2K

PIK

3CA

−E54

5K

PIK

3CA

−E72

6K

PIK

3CA

−H10

47R

PIK

3CA

−M10

43I

PIK

3CA

−N34

5K

EG

FR E

xon

19

EG

FR−G

719A

EG

FR−L

858R

EG

FR−L

861Q

EG

FR−T

790M

ME

T A

mp

RO

S1

Fusi

on

AR

Am

p

AR

−H87

5Y

AR

−L70

2H

BR

CA

1 tru

n

BR

CA

2 tru

n

PIK

3CA

−E54

2K

MSK−IMPACT onlyMSK−IMPACT & GRAIL

MSK−VB−0023 cfDNA

# M

utat

ions

(cfD

NA

)

0

200

400

600

800

1000

C>A C>G C>T T>A T>C T>G

MSK−VB−0023 Tumor

# M

utat

ions

(Tum

or)

0

2

4

6

8

C>A C>G C>T T>A T>C T>G

0.0

0.2

0.4

0.6

0.8

1.0

0.0

0.2

0.4

0.6

0.8

1.0

MSK−VB−0046 cfDNA

# M

utat

ions

(cfD

NA

)

0

10

20

30

40

50

60

C>A C>G C>T T>A T>C T>G

MSK−VB−0046 Tumor

# M

utat

ions

(Tum

or)

0

2

4

6

8

C>A C>G C>T T>A T>C T>G

0.0

0.2

0.4

0.6

0.8

1.0

0.0

0.2

0.4

0.6

0.8

1.0

Metastatic Prostate CancerTissue calls detection rate by cfDNA

Metastatic Lung CancerTissue calls detection rate by cfDNA

Metastatic Breast CancerTissue calls detection rate by cfDNA

S30S18S15S13S10S7S6S2S1

MS

K−V

B−0

013

MS

K−V

B−0

004

MS

K−V

B−0

024

MS

K−V

B−0

053

MS

K−V

B−0

005

MS

K−V

B−0

006

MS

K−V

B−0

008

MS

K−V

B−0

030

MS

K−V

B−0

059

MS

K−V

B−0

066

MS

K−V

B−0

076

MS

K−V

B−0

016

MS

K−V

B−0

032

MS

K−V

B−0

036

MS

K−V

B−0

062

MS

K−V

B−0

001

MS

K−V

B−0

054

MS

K−V

B−0

063

MS

K−V

B−0

069

MS

K−V

B−0

072

MS

K−V

B−0

084

MS

K−V

B−0

077

MS

K−V

B−0

033

MS

K−V

B−0

037

MS

K−V

B−0

031

MS

K−V

B−0

034

MS

K−V

B−0

012

MS

K−V

B−0

041

MS

K−V

B−0

058

MS

K−V

B−0

040

MS

K−V

B−0

044

MS

K−V

B−0

067

MS

K−V

B−0

029

MS

K−V

B−0

045

MS

K−V

B−0

073

MS

K−V

B−0

023

MS

K−V

B−0

057

MS

K−V

B−0

050

MS

K−V

B−0

046

Metastatic Breast Cancer

Num

ber o

f Var

iant

s

0

5

10

15

20

25

30

35

MS

K−V

L−00

73M

SK

−VL−

0002

MS

K−V

L−00

10M

SK

−VL−

0015

MS

K−V

L−00

64M

SK

−VL−

0009

MS

K−V

L−00

83M

SK

−VL−

0005

MS

K−V

L−00

13M

SK

−VL−

0051

MS

K−V

L−00

67M

SK

−VL−

0069

MS

K−V

L−00

71M

SK

−VL−

0007

MS

K−V

L−00

21M

SK

−VL−

0028

MS

K−V

L−00

03M

SK

−VL−

0061

MS

K−V

L−00

76M

SK

−VL−

0077

MS

K−V

L−00

82M

SK

−VL−

0012

MS

K−V

L−00

42M

SK

−VL−

0044

MS

K−V

L−00

56M

SK

−VL−

0027

MS

K−V

L−00

45M

SK

−VL−

0054

MS

K−V

L−00

17M

SK

−VL−

0057

MS

K−V

L−00

22M

SK

−VL−

0038

MS

K−V

L−00

80M

SK

−VL−

0074

MS

K−V

L−00

75M

SK

−VL−

0048

MS

K−V

L−00

55M

SK

−VL−

0008

MS

K−V

L−00

60M

SK

−VL−

0065

MS

K−V

L−00

35

Metastatic Lung Cancer

Num

ber o

f Var

iant

s

0

5

10

15

20

25

30

35

MS

K−V

P−0

030

MS

K−V

P−0

040

MS

K−V

P−0

002

MS

K−V

P−0

005

MS

K−V

P−0

007

MS

K−V

P−0

028

MS

K−V

P−0

029

MS

K−V

P−0

034

MS

K−V

P−0

035

MS

K−V

P−0

003

MS

K−V

P−0

008

MS

K−V

P−0

015

MS

K−V

P−0

016

MS

K−V

P−0

018

MS

K−V

P−0

026

MS

K−V

P−0

037

MS

K−V

P−0

042

MS

K−V

P−0

045

MS

K−V

P−0

051

MS

K−V

P−0

056

MS

K−V

P−0

001

MS

K−V

P−0

031

MS

K−V

P−0

032

MS

K−V

P−0

059

MS

K−V

P−0

004

MS

K−V

P−0

009

MS

K−V

P−0

010

MS

K−V

P−0

014

MS

K−V

P−0

033

MS

K−V

P−0

038

MS

K−V

P−0

046

MS

K−V

P−0

047

MS

K−V

P−0

050

MS

K−V

P−0

061

MS

K−V

P−0

006

MS

K−V

P−0

021

MS

K−V

P−0

054

MS

K−V

P−0

019

MS

K−V

P−0

024

MS

K−V

P−0

025

MS

K−V

P−0

058

MS

K−V

P−0

057

MS

K−V

P−0

041

MS

K−V

P−0

023

Metastatic Prostate Cancer

Num

ber o

f Var

iant

s

0

5

10

15

20

25

30

35

Tissue assay onlyTissue & cfDNA assay

Triplet mutational counts and signatures deconvolved from cfDNA for both samples VB-0046 (left), VB-0023 (right)

At least 1 MSK-IMPACT™-detected variant in tumor N PatientscfDNA/tissue

DetectionRate (%) 95% CI

All Patients 110/124 89 (82, 94)

Metastatic Breast Cancer Patients 38/39 97 (87, 100)

Metastatic Lung Cancer Patients 35/41 85 (71, 94)

Metastatic Prostate Cancer Patients 37/44 84 (70, 93)

ConcordanceAnalysis

Normal Cells

FFPE Tissue

PathologyReview

MSK-IMPACT 410-gene panel

TM

500 - 1,000x rawtarget coverage

Variant callingAnnotationFiltering

Blood(Streck BCT)

Plasma/Buffy CoatSeparation

Optimized single-tube library prepUnique Molecular Identifiers for error suppression

Error suppressionVariant callingAnnotationFiltering

508-gene cancer panel (2.1 Mb)

>60,000x raw target coverage for both cfDNA and gDNA

High-intensity Sequencing

DNA Extraction

Library Prep

Target Capture

Sequencing

Analysis

Tissue Assay cfDNA Assay

Tumor D

NA

Norm

al cell DN

A

Plasma cfD

NA

WB

C gD

NA

# of lines of therapy, N (%)

0 20 (51.3%) 25 (61.0%) 12 (27.3%)

1 2 (5.1%) 9 (22.0%) 14 (31.8%)

2 2 (5.1%) 3 (7.3%) 9 (20.5%)

>=3 15 (38.5%) 4 (9.8%) 9 (20.5%)

Tissue Sampled for MSK-IMPACT™, N (%)

Metastatic 35 (89.7%) 28 (68.3%) 44 (100.0%)

Primary 4 (10.3%) 13 (31.7%) 0

Receptor Status, N (%)

HR+/HER2+ 3 (7.7%) N/A N/A

HR+/HER2- 26 (66.7%) N/A N/A

HR-/HER2+ 2 (5.1%) N/A N/A

Triple Negative 8 (20.5%) N/A N/A

Summary • Tumor tissue variants identified by MSK-IMPACT™,

a validated tumor tissue profiling platform, enabled a demonstration of high overall detection rates (>70%) of the same variants in cfDNA.

• In the majority of patients, at least one mutation detected in tissue was also detected in plasma cfDNA of that same patient (97%, 85%, and 84% in breast, lung, and prostate cancer patients).

• Post-hoc analysis focused on the subset of patients with clonal variants in tissue:

- Based on de-novo variant calls, at least one clonal mutation was detected in cfDNA: 97%, 84%, and 83% in breast, lung, and prostate cancer patients.

• The majority of clinically actionable mutations detected in tissue were also detected in plasma (54/71, 76%; SNVs only: 28/31, 90%).

• The breadth of detected variants in plasma cfDNA enables greater insight into tumor biology, including observation of hypermutation signatures.

Conclusions• This novel, high-intensity cell-free DNA sequencing

assay incorporates unprecedented breadth (10X number of genes) compared to previous assays at these sequencing depths, and demonstrated high levels of concordance for both clonal and non-clonal variants between plasma and tissue.

• By interpreting concordance as strong evidence for tumor DNA detection, an extremely high level of tumor DNA detection in plasma was demonstrated.

• Clinically actionable non-biopsy somatic alterations were detected, which may represent tumor heterogeneity not detectable in a single tissue biopsy. Ongoing work is being conducted to distinguish technical noise from the assay and biological signal for the variants detected in plasma but not in tissue.

• The breadth of the panel enabled the first exploration of mutational signature analysis in plasma, revealing samples with APOBEC signatures.

• This study is part of a larger program to evaluate high-intensity sequencing approaches (e.g. whole genome) to characterize potential cancer-defining signals in cell-free nucleic acids, with an ultimate goal of enabling detection of cancer at early curable stages.

ReferencesctDNAWan, J.C., Massie, C., Garcia-Corbacho, J., Mouliere, F., Brenton, J.D., Caldas, C., … Rosenfeld, N. (2017). Liquid biopsies come of age: towards implementation of circulating tumour DNA. Nat Rev Cancer, 17, 223-38. https://doi.org/10.1038/nrc.2017.7

Aravanis, A.M., Lee, M., Klausner, R.D. (2017). Next-generation sequencing of circulating tumor DNA for early cancer detection. Cell, 168, 571-74. https://doi.org/10.1016/j.cell.2017.01.030

Clinically Actionable VariantsChang, M.T., Asthana, S., Gao, S.P., Lee, B.H., Chapman, J.S., ... Taylor, B.S. (2015). Identifying recurrent mutations in cancer reveals widespread lineage diversity and mutational specificity. Nature Biotechnology, 34, 155-63. https://doi.org/10.1038/nbt.3391

Robinson, D., Van Allen, E.M., Wu, Y.M., Schultz, N., Lonigro, R.J., ... Chinnaiyan, A.M. (2015). Integrative Clinical Genomics of Advanced Prostate Cancer. Cell 161:1215-1228. https://doi.org/10.1016/j.cell.2015.05.001

Mutation SignatureAlexandrov, L.B., Nik-Zainal, S., Wedge, D.C., Aparicio, S. a J. R., Behjati, S., Biankin, A.V, … Stratton, M.R. (2013). Signatures of mutational processes in human cancer. Nature, 500, 415–21. https://doi.org/10.1038/nature12477

MSK-IMPACTCheng, D.T., Mitchell, T.N., Zehir, A., Shah, R.H., Benayed, R., Syed, A., … Berger, M.F. (2015). Memorial Sloan Kettering-integrated mutation profiling of actionable cancer targets (MSK-IMPACT): A hybridization capture-based next-generation sequencing clinical assay for solid tumor molecular oncology. Journal of Molecular Diagnostics, 17(3), 251–264. https://doi.org/10.1016/j.jmoldx.2014.12.006

Zehir, A., Benayed, R., Shah, R.H., Syed, A., Middha, S., … Berger, M.F. (2017). Mutational landscape of metastatic cancer revealed from prospective clinical sequencing of 10,000 patients. Nature Med, https://doi.org/0.1038/nm.4333

FACETSShen, R., & Seshan, V. E. (2016). FACETS: Allele-specific copy number and clonal heterogeneity analysis tool for high-throughput DNA sequencing. Nucleic Acids Research, 44(16), 1–9. https://doi.org/10.1093/nar/gkw520

Tissue Assay cfDNA Assay