581P - Homologous recombination deficiency (HRD) in 543 ovarian cancer samples: A single center hands-on experience

Background

Biomarkers are investigated to efficiently identify responders to targeted therapy in ovarian cancer (OC). HRD may predict sensitivity to poly-ADP ribose polymerase inhibitors (PARPi) and platin-based regimens. We present our experience on HRD detection by applying a clinically validated commercially available CE-IVD solution.

Methods

A total of 543 formalin-fixed paraffin embedded (FFPE) OC samples were referred for HRD testing over a year. DNA was extracted (QIAamp^Ò DNA Mini kit, Qiagen) followed by library preparation according to the manufacturer (AmoyDx^Ò Focus Panel, Amoy Diagnostics Co., Ltd) and were sequenced on NextSeq550^Ò (Illumina). Bioinformatics were performed by proprietary algorithms (ANDAS v1.0.9-CE-B1.0, Amoy Diagnostics Co., Ltd) to identify BRCA1/2 pathogenic/likely pathogenic (P/PL) mutations (sBRCAmut) and the Genomic Scar Score (GSS) utilizing >24,000 single nucleotide polymorphisms (SNPs).

Results

HRD was called positive when sBRCAmut and/or GSS>50 were identified. A total of 535 samples reached a “verdict” leading to 44.3% HRD detection rate: 5% sBRCAmut, 25.6% GSS+ and 13.7% sBRCAmut/GSS+. Tumor cell content was unknown for 87% of the samples, partially explaining the deviation from “expected” HRD i.e. 50%. Failures were due to unmet quality control (QC) criteria. At least one QC was compromised in >70% of the samples, reflecting low quality FFPE tissue, most likely contributing to false negative results. A total of 8 samples (1.5%) with all QCs unreached were rejected from the analysis. Selected samples that passed all QCs yielded 54.2% HRD.

Conclusions

The method applied could reliably detect HRD positive OC tumors, granting eligibility for targeted therapies such as PARPi and platinum-based chemotherapy. This option may address main concerns related to U.S. Food and Drug Administration (FDA) approved therapy with established companion diagnostics such as cost, test send-out and optimized performance. Our real-life experience underlines the technical challenges related to specimen quality and highlights the necessity for multidisciplinary education and collaboration throughout the patient care pathway.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

The authors.

Funding

Has not received any funding.

Disclosure

All authors have declared no conflicts of interest.