Vera M. Witjes, Joanne A de Hullu, Angela van Remortele, Lilian Vreede, Efraim H. Rosenberg, Saskia A.G.M. Cillessen, Floris H. Groenendijk, Elisabeth M. P. Steeghs, Laura Moonen, Arjen R. Mensenkamp, Arja ter Elst, Wendy W. J. de Leng, Nicoline Hoogerbrugge, Marjolijn J L Ligtenberg
Homologous recombination deficiency (HRD) testing may be used to stratify ovarian carcinoma (OC) patients for PARP inhibitor therapy. In the Netherlands, different NGS-based assays are used to assess genomic instability as a hallmark of HRD. We evaluated the uniformity of HRD testing. Firstly, interlaboratory assessments of 10 tumors were performed in 8 centers. 71 out of the 77 (92%) successful tests were concordant. Results were more consistent in OC with a pathogenic variation (PV) or promoter methylation of a homologous recombination repair (HRR) gene (97%) than in those without (87%). Secondly, concordance between BRCA1/RAD51C promoter methylation and HRD was assessed in 244 samples without a PV in HRR genes. BRCA1/RAD51C promoter methylation was present in 38 out of 100 (38%) samples classified as HRD, and absent in all (n = 144) non-HRD samples (p < 0.001). Lastly, pathology reports from 765 HRD tests were reviewed to evaluate routine diagnostics. Testing was successful in 695 (91%) cases. HRD detection rates were higher in high-grade serous OC compared to other histological subtypes (49% versus 12%, p < 0.001). The five HRD assays varied significantly in HRD detection rates in high-grade serous OC. The results support the applicability of genomic instability analyses to assess HRD, while also highlighting the need to improve harmonization across different assays when HRD is used for therapeutic decision making.