Introduction/Background Extensive genomic instability and heterogeneity are characteristics of ovarian high-grade serous cancer (HGSC). Matching best treatment options for patients is difficult due to diverse resistance mechanisms and paucity of predictive biomarkers. Somatic HR-deficiency (HRD) testing is an approved clinical test to stratify patients for PARP-inhibitor treatment. Here, we investigate the effect of spatial and temporal heterogeneity on HRD status.

Methodology Patients (n=59) at Hammersmith Hospital (HH) underwent primary cytoreductive surgery for advanced HGSC, with tumour biopsies collected (range 4–15) plus any paired relapse samples (n=11). Tumour DNA was extracted (n=5 tumours per case, plus relapse), genotyping performed for all cases with targeted sequencing (20 HR-related genes) performed for a subset. Multi-site CN data was also accessed from two cohorts (GSE38787, GSE40546). Tumours that failed QC or had an aberrant cell fraction <0.3 were excluded. HR scores were estimated in each cohort using a genomic instability score with a cut-off of ≥42 denoting HRD tumours. Mutations were called with Mutect2 and Haplotypecaller, following GATK (4.4.0.0) best practices.

Results We detected variation in HR scores in each cohort with approximately one-fifth of patients presenting with a mixed HR score, displaying both HRD and HR-Proficient tumour scores: HH (20%, 11/54 patients); GSE38787 (17%, 4/24 patients); and GSE40546 (28%, 4/14 patients). In the targeted sequencing cases, data was correlated with HR scores to delineate the underlying mutational status of each HR group (HRD, HRP and HR-mixed). Multiple repair-related mutations were flagged as possible causes for the mixed-HR scores (notably BARD1, FANCM), while also demonstrating the intra-patient heterogeneity, even in non-mixed patients.

Conclusion Variability in HRD/HRP scores indicates that HR status for a subset of ovarian HGSC patients is not uniform across their disseminated tumour burden, thus suggesting that sampling a single tumour site may not accurately represent a patient’s tumour biology, leading to incorrect treatment stratification.

Disclosures N/A