You are here

Article Text

Article menu

Download PDFPDF

Ovarian cancer
2022-RA-913-ESGO Clinical performance evaluation of a novel deep learning solution for homologous recombination deficiency detection
Free
  1. Adrien Buisson1,
  2. Pierre Saintigny2,
  3. Alexandre Harle3,
  4. Davide Vacirca4,
  5. Massimo Barberis4,
  6. Pauline Gilson3,
  7. Cristin Roma5,
  8. Francesca Bergantino5,
  9. Christoph Grimm6,
  10. Leonhard Müllauer7,
  11. Philipp Harter8,
  12. Sandro Pignata9,
  13. Antonio Gonzalez-Martin10,
  14. Christian Schauer11,
  15. Keiichi Fujiwara12,
  16. Ignace Vergote13,
  17. Nicoletta Colombo4,
  18. Eric Pujade-Lauraine14,
  19. Isabelle Treilleux15,
  20. Isabelle Ray-Coquard16,
  21. group Consortium. on behalf of the PAOLA Study Group and ENGOT Inter
  1. 1Département de Biopathologie, Centre Léon Bérard-Cheney, Lyon, France
  2. 2Department of Medical Oncology, Centre Léon Bérard and University Claude Bernard Lyon I, Lyon, France
  3. 3Service de Biopathologie, Université de Lorraine, Institut de Cancérologie de Lorraine, Nancy, France
  4. 4European Institute of Oncology, Milan, Italy
  5. 5Istituto Nazionale Tumori-IRCCS-Fondazione G. Pascale, I-80131, Naples, Italy
  6. 6Division of General Gynecology and Gynecologic Oncology, Gynecology Cancer Unit, Medical University of Vienna, Vienna, Austria
  7. 7Department of Pathology, Medical University of Vienna, Währinger Gürtel 18–20, A-1090 Vienna, Vienna, Austria
  8. 8Kliniken Essen Mitte, Essen, Germany
  9. 9Department of Urology and Gynecology, Istituto Nazionale Tumori IRCCS Fondazione G. Pascale, and Multicenter Italian Trials in Ovarian Cancer and Gynecologic Malignancies (MITO), Naples, Italy
  10. 10Clínica Universidad de Navarra, Madrid and GEICO, Madrid, Spain
  11. 11Hospital Barmherzige Brüder Graz, Graz, Austria
  12. 12Saitama Medical University International Medical Center, Saitama, Japan
  13. 13Leuven Cancer Institute, University Hospital Leuven, Leuven, Belgium
  14. 14Association de Recherche Cancers Gynécologiques (ARCAGY), Paris, France, Paris, France
  15. 15Centre de Recherche en Cancérologie de Lyon, Centre Léon Bérard, Université de Lyon, Lyon, France
  16. 16Centre Léon BERARD, and University Claude Bernard Lyon I, Lyon and Groupe d’Investigateurs Nationaux pour l’Etude des Cancers Ovariens (GINECO), Lyon, France

Abstract

Introduction/Background Comprehensive identification of ovarian cancer (OvCa) patients that can benefit from poly (ADP-ribose) polymerase inhibitors (PARPi) treatment is currently based on genome-wide enumeration of known Homologous Recombination Deficiency (HRD) biomarkers and requires deep sequence profiles (>30x). The cost and challenges of implementing available solutions currently hinders clinical adoption of HRD testing. We present the interim results for the ongoing multicenter evaluation study of SOPHiA DDM Dx HRD (SOPHiA GENETICS, SA), a novel deep learning-based solution that leverages the impact of HRD on the coverage profiles from low-pass whole-genome sequencing (lpWGS, 1x) data.

Methodology In this multicenter study, we processed and analyzed according to manufacturer’s instructions, DNA from 319 formalin-fixed paraffin-embedded (FFPE) OvCa samples. We assessed the concordance between SOPHiA DDM Dx HRD solution and Myriad myChoice® CDx results. For the subset of our cohort (206 samples) included in the PAOLA-1 clinical trial, we carried out survival analysis to investigate differences in progression-free survival (PFS) in the olaparib and placebo arms of the study between patients, with HRD positive or non-positive tumors.

Results We found a high overall percentage agreement, 90.48% (95% confidence interval [CI], 86.58–93.33), between SOPHiA DDM Dx HRD and Myriad myChoice® CDx status. The median PFS time for patients with HRD positive tumors was 20.8 months higher in the olaparib arm (hazard ratio [HR], 0.44; 95% CI, 0.26–0.76, P=0.003). No significant difference in PFS was observed between treatment arm in patients without HRD positive tumors (HR, 0.92; 95% CI, 0.59–1.43; P=0.69). The effect of the interaction between olaparib and HRD status on PFS, in the interim study, was similar for the two stratification methods (P=0.20).

Conclusion The interim results of SOPHiA DDM Dx HRD Solution evaluation study support the value of lpWGS data for patient stratification, making it suitable for HRD testing in the clinical setting.

https://doi.org/10.1136/ijgc-2022-ESGO.591

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.