RESULTS � Pathogenic germline BRCA1/2 mutations were identified in 56 individuals (42%) by the reference laboratory.

� Tumor BRCA1/2 sequencing was successful for all available tissue samples.

� 2 gBRCA positive samples had insufficient tumor available for tBRCA testing.

� All 54 available tissue samples from germline mutation carriers were also classified as carrying pathogenic BRCA1/2 mutations by tBRCA analysis (Table 1).

� tBRCA sequencing also identified an additional 12 (9%) pathogenic BRCA1/2 mutations which were presumed to be somatic mutations (Table 1).

� This results in an increase of ~22% in the number of patients identified with a BRCA1/2 mutation, despite the fact that the cohort is enriched for germline mutations.

� A total of 5 LRs were identified during tBRCA testing (examples in Figure 1).

� 2 germline, 3 somatic.

� All 62 of the remaining non-mutant/VUS samples identified by gBRCA analysis were also classified as non-mutant following tBRCA analysis (Table 1).

� 2 gBRCA negative samples had insufficient tumor available for tBRCA testing.

� Patients identified as having a pathogenic mutation in BRCA1/2 treated with olaparib had fewer progression-related events than the control group (Table 2).

� Patients with no pathogenic BRCA1/2 mutations exhibited similar disease progression in both the treatment and placebo groups (Table 2).

� Patients with somatic mutations had an older median age-at diagnosis (58.6 years) relative to patients with germline mutations (53.0 years) (Table 3).

� Similar efficacy was observed for patients with either germline or somatic mutations, regardless of age-at-diagnosis.

Next generation sequencing of BRCA1 and BRCA2 genes in ovarian tumors captures germline mutations and expands the potential treatment group for the PARP inhibitor olaparib

Kirsten M. Timms,1 Chris Neff,1 Brian Morris,1 J. Carl Barrett,2 Darren R. Hodgson,3 Maria Orr,3 Zhongwu Lai,2 Anitra Fielding,3 Brian Dougherty,2 Stuart Spencer,3 Jonathan Ledermann,4 Alexander Gutin,1 Jerry S. Lanchbury1

1. Myriad Genetics, Inc., Salt Lake City, UT, USA 2. AstraZeneca, Waltham, MA, USA 3. AstraZeneca, Macclesfield, UK 4. University College London, London, UK

BACKGROUND � Clinical studies have demonstrated that the PARP inhibitor olaparib is active in patients with platinum-sensitive relapsed high grade serous ovarian cancer (HGSOC) who carry BRCA1/2 mutations.

� Consistent with the current understanding of the mode of action of PARP inhibitors, clinical activity is observed regardless of whether the primary mutation is inherited (germline) or somatically acquired.

� Previously we have shown that BRCA1/2 testing at the tumor level expands the fraction of HGSOC patients eligible for PARP inhibitor clinical trials or treatment with an approved agent.1

� In this study, we examined the ability of a tumor-based NGS test to detect germline BRCA1/2 mutations in HGSOC patients.

METHODSPatients

� 132 HGSOC patients who were eligible for treatment as part of Study 19 (D0810C00019 and NCT00753545) were assessed here. This represents the subset of patients for which germline and tumor mutation status was available.

� Patients in this cohort were aged 18 years or older and had recurrent ovarian cancer. Full details of this cohort have been previously described.1

� Patients either received 400mg olaparib (capsule formulation) twice daily or a placebo.

� This cohort was enriched for patients with germline mutations.

Tumor BRCA1/2 (tBRCA) Testing � Tumor BRCA1/2 (tBRCA) mutation analysis was carried out on DNA extracted from formalin fixed paraffin embedded (FFPE) primary ovarian tumor after library generation and hybridization capture of BRCA1/2 exons using a custom Agilent SureSelect assay.

� Captured DNA was sequenced to a median read depth of 425X using an Illumina HiSeq sequencer. Inclusion criteria required that 99% of bases have a read depth of at least 100X per base, with a minimum inclusion criteria of 50X per base.

� Mutation results were reported down to a mutation frequency of 5% to accommodate samples with low tumor content.

� Genomic DNA derived from tumor tissue was analyzed by NGS dosage analysis to determine copy-number abnormalities indicative of multi-exon duplications or deletions, or large rearrangements (LRs).

Reference Germline BRCA1/2 (gBRCA) Testing � Germline BRCA1/2 testing was carried out on DNA extracted from blood samples using Sanger sequencing, with LR analysis performed by multiplex quantitative PCR analysis.

tBRCA versus gBRCA Comparison � Blinded analyses were performed by the tBRCA laboratory and reference laboratory.

� Samples that received a laboratory classification of Deleterious or Suspected Deleterious were considered pathogenic (mutation positive).

� Classifications were performed independently for tBRCA and gBRCA test results and the comparison was performed after completion of testing by both laboratories.

Table 1. BRCA1/2 Mutations Identified by gBRCA and tBRCA

BRCA1/2 Mutation Status

gBRCA Reference Laboratory

tBRCAStudy Laboratory

Positive (germline) 56 54*Positive (somatic) 0 12Negative 76 62*Total 132 128**Four samples had insufficient tumor available for tBRCA analysis (2gBRCA positive, 2 gBRCA negative)

Figure 1. Representative example of large rearrangements detected by tBRCA testing in BRCA1 (top) and BRCA2 (bottom). Each square represents the average read coverage for each exon. Exons at normal dosages appear at about 2 on the y-axis.

Table 2. Progression-Related Events

Subgroup Treatment NProgression-

Related Events*N (%)

Germline Mutation (gBRCA positive)

olaparib 31 8 (25.8%)

placebo 23 17 (73.9%)

Somatic Mutation(tBRCA positive only)

olaparib 6 2 (33.3%)

placebo 6 4 (66.7%)

No Pathogenic Mutation(tBRCA, gBRCA VUS/WT)

olaparib 25 16 (64.0%)

placebo 37 26 (70.3%)Note: VUS - Variant of Unknown Significance; WT - Wild Type*Disease progression or death (duration of follow-up ranged from 5 to 22 months)

CONCLUSIONS � A tumor-based next generation sequencing test is able to accurately detect germline pathogenic mutations from FFPE ovarian tumor material with 100% sensitivity.

� The tumor test detects mutations of all germline classes, including large rearrangements.

� Tumor BRCA1/2 testing identified additional somatic mutations beyond what would have been identified by germline testing alone.

� Patients with germline and tumor BRCA1/2 mutations showed similar trends in treatment response, suggesting that tumor testing identifies patients appropriate for treatment with PARP inhibitors.

BRCA1 Deletion

BRCA2 Deletion

Table 3. Age-At-Diagnosis

Subgroup Treatment N MedianAge-at-Diagnosis SD Min, Max

Germline Mutation (gBRCA positive)

olaparib 31 54.0 10.47 37, 81

placebo 23 53.0 9.37 32, 64

TOTAL 54 53.0 10.06 32, 81

Somatic Mutation(tBRCA positive only)

olaparib 6 54.0 9.63 46, 70

placebo 6 58.5 6.15 49, 65

TOTAL 12 58.6 7.79 46, 70

No Pathogenic Mutation(tBRCA, gBRCA VUS/WT)

olaparib 25 59.0 11.05 41, 78

placebo 37 60.0 8.81 43, 75

TOTAL 62 59.5 9.74 41, 78Note: VUS - Variant of Unknown Significance; WT - Wild Type

REFERENCES1. Ledermann J, Harter P, Gourley C. et al. Lancet Oncol. 2014;15:852-861.

ACKNOWLEDGEMENTSMedical writing assistance was provided by Krystal Brown, an employee of Myriad Genetics, Inc.

Presented at ESMO - 9/26/2015