Abstract 579P
Background
Poly (ADP-ribose) polymerase inhibitors (PARPi) induce synthetic lethality in homologous recombination deficient (HRD) tumors and have revolutionized ovarian cancer treatment. Patient stratification methods currently used are based on genome-wide enumeration of known HRD biomarkers and require deep sequence profiles (> 30x), in some cases from tumor-normal pairs. The cost and challenges introduced by implementing the current solutions hinder the clinical adoption of HRD testing. Here we present a multicenter performance evaluation study of an alternative deep learning-based decentralized solution for HRD detection, the SOPHiA DDM Dx HRD Solution (SOPHiA GENETICS SA), which leverages the impact of HRD on the coverage profiles obtained from low-pass whole-genome sequencing (lpWGS) data (X1 fold coverage).
Methods
We generated and sequenced, in 5 independent clinical laboratories, 153 low pass WGS libraries (X1, ∼10 million reads, Illumina) for 150 Formalin-Fixed Paraffin-Embedded ovarian cancer samples. HRD classification was performed according to the manufacturer’s recommendations. The analytical performance was evaluated as overall (OPA), positive (PPA), and negative (NPA) percentage of agreement with Myriad myChoice CDx status (Myriad Genetic Laboratories, Inc.; US FDA-approved).
Results
The Quality Control criteria for SOPHiA DDM Dx HRD Solution analysis were met for 97,3% of the libraries (n=149). We found a high overall percentage agreement, 93.7% (97.1% NPA and 90.4% PPA) between SOPHiA DDM Dx HRD Solution and Myriad myChoice CDx classification (Table). Additionally, SOPHiA DDM Dx HRD Solution results for samples analyzed in duplicate were 100% reproducible. Table: 579P
Confusion matrix of comparison between Myriad myChoice CDx and SOPHiA DDM Dx HRD Solution results (149 unique samples)
SOPHiA DDM Dx HRD Solution | ||||
Positive | Negative | Fail | ||
Myriad myChoice CDx | Positive | 66 | 9 | - |
Negative | 2 | 67 | 3 | |
Fail | - | 1 | 1 |
Conclusions
We conclude that SOPHiA DDM Dx HRD Solution supports accurate HRD detection from low pass whole genome sequencing data, which is amongst the most cost-effective and easy to implement genome profiling methods.
Clinical trial identification
Editorial acknowledgement
Legal entity responsible for the study
SOPHiA GENETICS SA.
Funding
SOPHiA GENETICS SA.
Disclosure
All authors have declared no conflicts of interest.