Introduction• The potential utility of circulating tumor DNA (ctDNA) to detect minimal residual disease (MRD) following surgery and/or adjuvant therapy

and to monitor metastatic disease has been demonstrated in breast cancer.1-4

• Plasma ctDNA levels have been shown to correlate with changes in tumor burden, thereby providing an earlier measure of treatment response5,6 and to discriminate patients with and without eventual clinical recurrence post-surgery.2,3

• We previously reported early detection of lung cancer recurrence using a personalized tumor-specific approach;7,8 an improved version of this assay for patient-specific ctDNA detection was developed and made available for research use only (Signatera™ RUO).

Objectives The aim of the study was to determine the ‘lead interval’ between detection of ctDNA in plasma and clinical detection of overt metastatic

disease.

Methods

Clinical Protocol• Patients diagnosed with non-metastatic breast cancer with high risk of reccurence were prospectively recruited starting in Dec 2013.

• Eligible patients were followed up with semi-annual blood sampling for ctDNA analysis, along with concomitant clinical examination, and biochemical measurements, including CA 15-3 (Figure 1).

– All patients had completed adjuvant chemotherapy within 3 years of entering the study.

Figure 1. Schematic of Clinical Sample Collection

Conclusions• Plasma ctDNA was detected ahead of clinical or radiological relapse in 16/18 (89%) relapsed

patients. – Of the two relapsed patients who were not detected in the study, one patient had a local recurrence while the other had recently completed chemotherapy, and neither of the relapses were confirmed by biopsy.

• Metastatic relapse was predicted with a lead time of up to 2 years (median=8.9 months; range: 0.5–24.0 months).

• None of the 31 non-relapsing patients were ctDNA-positive at any time point (across 156 plasma samples).

• Our results demonstrate the use of a scalable patient-specific ctDNA approach capable of identifying residual disease in breast cancer patients with a high degree of sensitivity ahead of clinical or radiological detection of metastatic recurrence.

References

SAN ANTONIO BREAST CANCER SYMPOSIUM 2018San Antonio, Texas, USA | December 4–8, 2018

ctDNA Status at 1st Timepoint

0 25 50 75 100

75

Months After Surgery

Rel

aps-

Free

Sur

viva

l

100

50

25

0

| | ||| | | | |

| | | | | | | | | | | | |

ctDNA Status at Any TimepointA B

HR: 11.78 (4.28–32.46)p-value <0.001

| | | | | | | | | | | | | | | | | | | | | |

75

100

50

25

00 25 50 75 100

Months After Surgery

ctDNA NegativectDNA Positive

Rel

apse

Fre

e Su

rviv

al

HR: 35.84 (7.9 –161.32)p-value <0.001

ctDNA NegativectDNA Positive

Personalized, 16-plex assays accurately detect ctDNA ahead of clinical relapse. The figure depicts a summary of each patient’s (n=49) treatment regimen along with results of serial plasma samples (n=208) analyzed. Patients are categroized into three groups based on cancer subtype. Circles represent ctDNA status; solid black circles (ctDNA positive) indicate samples with at least two positve targets.

Personalized ctDNA detection in serial plasma samples predicts recurrence-free survival. A. Relapse-free survival according to the detection of ctDNA in the first post-surgical plasma sample. B. Relapse-free survival according to the detection of ctDNA in any follow-up plasma sample post-surgery. Data are from n=49 patients.

Figure 6. Personalized ctDNA Detection in Serial Plasma Samples Predicts Recurrence-Free Survival

Abstract number #1266

Lead Time: 721 Days

A

0.01

1.00

470 652 1078876 1246

Varia

nt A

llele

Fre

quen

cy (%

)

0

50

100

150

200

250

0.001

0.1

≤ 0.001

CA 15-3

Lead Time: 611 DaysE026: HR+

CA 15-3 Baseline

Days After Surgery Days After Surgery

E

CA 15-3 Baseline

110 306 518 690

Lead Time: 570 Days

E033: TNBC

CA

15-3

Days After Surgery

Var

iant

Alle

le F

requ

ency

(%)

0

50

100

150

200

250

0.01

0.1

1

10

≤ 0.001

C D

Days After Surgery

E040: HR-/HER2+

CA 15-3 Baseline

15711426 1615

Var

iant

Alle

le F

requ

ency

(%)

CA

15-3

0

50

100

150

200

250

≤ 0.001

Lead Time: 313 Days

0.01

1

10

1258

0.1

F

CA

15-3

B

CA 15-3 Baseline

330 673498 890 1093 1268 1400

E017: HR+

0

50

100

150

200

250

≤ 0.001

0.01

0.1

Var

iant

Alle

le F

requ

ency

(%)

476 660 856 961

Varia

nt A

llele

Fre

quen

cy (%

)

E029: TNBC

CA

15-3

0

50

100

150

200

250

0.01

1.00

≤ 0.001

0.1

Lead Time: 258 Days

CA 15-3 Baseline

Days After Surgery

Med

ian

VA

F (%

) of P

ositi

ve T

arge

ts

ctDNANegative ctDNA PositiveMolecular Relapse Clinical Relapse

CA 15-3 Mean VAF

0.1

1.0

10.0

Num

ber o

f Pos

itive

Tar

gets

First Time Point

LastTime Point

2

4

6

8

10

12

14

First Time Point

LastTime Point

Figure 5. Personalized Profiling Detects Rising ctDNA Ahead of Clinical Relapse

Figure 3. Patient Summaries

1 2 3 4

Timepoints

Obtain whole blood samples at longitudinal timepoints (eg, every 3 months)

Cell-free DNA extraction and patient-speci�c 16-plex PCR

Analyze ultra-deep NGS data for each of the 16 patient-speci�c mutations to detect presence of ctDNA

Analyze sequencing from tumor biopsy and matched normal blood at initial timepoint

Identify patient-speci�c, somatic variants and design custom 16-plex panel for each patient

1 2

3 4 5

11

13

16

1

4

7

LEAD TIME

MOLECULAR RELAPSE CLINICAL RELAPSE

VA

RIA

NT

ALL

ELE

FR

EQ

UE

NC

Y (

%)

DAYS AFTER SURGERY

0.01

1.00

≤ 0.001

0.1

First 50 women re-consented and included for Signatera study

Tumour biopsy tissue collected1 excluded due to insufficient

tumour DNA for WESWES to identify somatic

variants (n=49)

Non relapsed patients (n=31) total of 156 plasma samples

analysed

Relapsed patients (n=18) total of 52 plasma samples analyzed

Serial plasma samples collected (n=218)

Serial plasma samples analysed (n=215)

16 individual variants selected per patient

197 women on follow up after surgery and adjuvant therapy

188 women followed up

9 excluded for not meeting inclusion criteria

1 failed sequencing QC6 failed concordance QC

3 plasma samples excluded due to insufficient DNA for WES

Figure 2. Schematic of Molecular Protocol - ctDNA Analysis

The VAF distribution of custom assay in each patient’s primary tumour tissue

Breast Cancer Subtype

Total Patients (n)

Relapsed Patients(n)

Relapses Detected n (%)

PPV(%)

NPV(%)

Median Lead Time (Days)

HR+/HER2- 34 11 9 (82) 100 92 301

HER2+ 8 2 2 (100) 100 100 164

TNBC 7 5 5 (100) 100 100 258

Total 49 18 16 (89) 100 94 285

Table 1. Summary Table by Breast Cancer Subtype

Var

iant

Alle

le F

requ

ency

(%

)

0.0

0.2

0.4

0.6

0.8

1.0

E025

E042

E035

E012

E004

E018

E013

E038

E006

E043

E010

E041

E014

E005

E034

E039

E027

E021

E023

E017

E019

E032

E026

E020

E008

E047

E031

E036

E030

E022

E011

E044

E003

E016

E009

E002

E001

E037

E045

E007

E028

E040

E048

E015

E024

E049

E033

E046

E029

HR+/HER2- HER2+ TNBC

Patient Samples

Molecular Protocol - ctDNA Analysis (SignateraTM RUO)• The workflow of the molecular protocol is outlined in Figure 2.

• For the 49 women with breast cancer who were monitored in this study, exomic alterations were determined through paired-end sequencing of FFPE tumor-tissue specimens and matched normal DNA.

• Patient-specific somatic variants were identified by analyses of primary tumor and matched normal whole exome sequencing samples.

• Serial plasma samples were analyzed with the corresponding custom 16-plex assay panels using the SignateraTM RUO workflow in a blinded manner; a total of 208 samples were analyzed for ctDNA detection.

• Samples were considered ctDNA positive if at least two patient-specific targets were called and met the qualifying confidence score threshold.

Figure 4. Tumor Variant Allele Frequency Distribution in Signatera Patient-Specific Assays

A-E. Plasma levels of ctDNA across serial plasma time points for five breast cancer patients (one per panel). Mean VAFs are denoted by dark blue circle and solid lines represent the average VAF profile over time. The lead time is calculated as the time interval between clinical relapse (red triangle) and molecular relapse (blue triangle). CA 15-3 levels are graphed over time (teal circle) and the baseline levels are marked in light blue. F. Summary of percent variant allele frequency (VAF) and number of targets detected at molecular and clinical relapse for all ctDNA positive samples. Data are from 13 relapsed patients, excluding 3 patients with only one plasma time point.

1. Beaver JA et al. Detection of cancer DNA in plasma of patients with early-stage breast cancer. Clin Cancer Res. 2014;20(10):2643-2650.2. Garcia-Murillas I et al. Mutation tracking in circulating tumor DNA predicts relapse in early breast cancer. Sci Transl Med. 2015;7(302):302ra133.3. Olsson E et al. Serial monitoring of circulating tumor DNA in patients with primary breast cancer for detection of occult metastatic disease. EMBO Mol Med. 2015;7(8):1034-1047.4. Shaw JA et al. Genomic analysis of circulating cell-free DNA infers breast cancer dormancy. Genome Res. 2012;22(2):220-231.5. Dawson SJ et al. Analysis of circulating tumor DNA to monitor metastatic breast cancer. N Engl J Med. 2013;368(13):1199-1209.6. Ma F et al. ctDNA dynamics: a novel indicator to track resistance in metastatic breast cancer treated with anti-HER2 therapy. Oncotarget. 2016;7(40):66020-66031.7. Abbosh C et al. Phylogenetic ctDNA analysis depicts early-stage lung cancer evolution. Nature. 2017;545(7655):446-451.8. Jamal-Hanjani M et al. Detection of ubiquitous and heterogeneous mutations in cell-free DNA from patients with early-stage non-small-cell lung cancer. Ann of Oncol. 2016;27(5):862-867.

Results

PPV, Positive Predicitive Value [True Positive/(True Positive+False Positive)]; NPV, Negative Predictive Value [True Negative/(False Negative+True Negative)]

Early Detection of Residual Breast Cancer through a Robust, Scalable and Personalized Analysis of Circulating Tumour DNA (ctDNA) Antedates Overt Metastatic RecurrenceRaoul C. Coombes1, Anne Armstrong2, Samreen Ahmed3, Karen Page4, Robert K. Hastings2, Raheleh Salari5, Himanshu Sethi5, Anna-Rita Boydell1, Svetlana V. Shchegrova5, Daniel Fernandez-Garcia4, Kelly LT Gleason1, Kate Goddard1, Dave S. Guttery4, Zoe J. Assaf5, Mustafa Balcioglu5, David A. Moore6, Lindsay Primrose4, Samantha L. Navarro5, Alexey Aleshin5, Farah Rehman1, Bradley J. Toghill 4, Maggie C. Louie 5, Cheng-Ho Jimmy Lin5, Bernhard G. Zimmermann5, Jacqueline A. Shaw4

1Imperial College London, United Kingdom; 2The Christie Foundation NHS Trust, United Kingdom; 3Leicester Royal Infirmary, United Kingdom; 4The University of Leicester, United Kingdom; 5Natera, San Carlos, CA, United States; 6University College London, United Kingdom.

This presentation is the intellectual property of the author/presenter. Contact Jacqueline Shaw at js39@leicester.ac.uk for permission to reprint and distribute.

RESEARCH USE ONLY