Real-World Analysis of HRD Assay Variability in High-Grade Serous Ovarian Cancer: Impacts of BRCA1/2 Mutation Subtypes on HRD Assessment

Abstract

The HRD (Homologous Recombination Deficiency) test is considered a genomic alteration useful for guiding therapeutic decisions in patients with ovarian cancer. Some commercial and in house alternative "academic" tests are available. Recent findings indicate that not all BRCA1/2 mutations determine the magnitude of HRD and that some patients carrying BRCA1/2 mutations may exhibit indeterminate or even negative HRD scores. Furthermore, certain therapies (e.g., olaparib and bevacizumab) offer particularly pronounced benefits for high-grade serous ovarian cancer (HGSOC) patients harboring mutations in the DNA-binding domain (DBD) of BRCA1/2. The aim of the present study is to investigate the relationship between the HRD scores and BRCA1/2 status of 51 HGSOC patients (50 BRCA1/2 mutated and 1 wild type). The HRD status was assessed by means of shallow whole-genome sequencing and BRCA1/2 status by the NGS pipeline. We did not find a correlation between the HRD status and type of BRCA1/2 alterations. A strong correlation between the HRD score and age was found. Our paper underlines the need to introduce other biological factors within the algorithms of the HRD evaluation in order to better tailor the HRD status, harmonize the metrics of the HRD assessment, and personalize therapies.

Keywords: BRCA; HRD; ovarian cancer.

Figure 1

Lollipop plot of BRCA1 and BRCA2 mutational domains. (A) BRCA1: Among the 24 samples with a Class 5 or Class 4 (pathogenic or likely pathogenic) mutation, 22 could be positioned within specific mutational domains. Two samples (IDs 10 and 20) carried large deletions that prevented precise domain mapping. (B) BRCA2: Among the nineteen samples with a Class 5 or Class 4 mutation, sixteen were successfully mapped to defined protein domains, while three samples (IDs 13, 14, and 52) harbored large deletions that could not be localized within the domain structure.

Figure 2

Distribution of the HRD status among ClinVar mutational classes (left panel) and across BRCA1/2 functional domains (right panel). Percentages within the bars represent the proportion of each mutational class/BRCA domain within the corresponding HRD status group. The right panel considers all 51 HRD tests, including also sample 38 (benign) with an HRD-N status.

Figure 3

CA15.3 distribution among patients with 3 different HR scores. 0 = HR-N (red); 1 = HR-M (green); 2 = HR-D (blue). Standard values are below 23.4 U/mL.

Figure 4

Log odds weight of the HR status across diagnostic age groups of HGSOC patients. (A) Log odd ratio distribution of the HR status stratified by the diagnostic age within the current cohort. (B) The analysis includes patients from our real-world ovarian cancer cohort combined with publicly available data from patients with BRCA1/2 mutations from The Cancer Genome Atlas (TCGA). Patients were stratified into five diagnostic age categories (<40, 40–51, 52–65, 65–70, and 70+ years). Log odds values are shown separately for the HR-negative (HR-N, left) and HR-deficient (HR-D, right) subgroups, highlighting contrasting age-associated patterns of HRD prevalence. The right-side chart uses a different scale (−1.5 to 1.5) compared to the left chart (−0.6 to 0.6), indicating different magnitudes of effects being measured.