The Landscape of RAD51D in Chinese Ovarian Cancer Patients: Prevalence, Correlation with HRD Score, and Correlation with Efficacy

T. Sun, Mengpei Zhang, Xinyun Xu, Yiming Liang, Jinghong Chen, Qingli Li, Jing Zeng, Zhe Li, Yu Dong, Rutie YinmedRxiv 2024 · DOI

TL;DR

This study identifies RAD51D as a critical, non-BRCA biomarker in Chinese ovarian cancer patients, appearing in 2.9% of cases and often occurring as a bi-allelic loss-of-function (BILOF). These mutations are mutually exclusive with BRCA1/2 and serve as a strong predictor for improved progression-free survival (PFS) following first-line chemotherapy.

What Problem Does This Paper Address?

Ovarian cancer remains a leading cause of cancer-related death in women, with high-grade serous ovarian cancer (HGSOC) being particularly aggressive. For years, clinical focus has centered on BRCA1 and BRCA2 mutations, which impair the cell's ability to perform Homologous Recombination Repair (HRR)—a high-fidelity mechanism for fixing double-stranded DNA breaks.

However, many patients lack BRCA mutations but still exhibit Homologous Recombination Deficiency (HRD), making them sensitive to platinum-based chemotherapies and PARP inhibitors. This study investigates the "landscape" of RAD51D, a lesser-known HRR gene, specifically within the Chinese population. The goal is to determine if RAD51D can serve as a reliable independent biomarker to better stratify patients and predict treatment outcomes.

Methods

The researchers conducted a post-hoc analysis of the FACT cohort (First-line Adjuvant ChemoTherapy), which is part of the larger EOC-HRD study.

  • Cohort: 380 Chinese patients with epithelial ovarian, fallopian tube, or primary peritoneal cancer.
  • Genomic Profiling: They utilized a panel of 28 HRR-related genes (the HRR28 list).
  • Definitions: A key focus was placed on BILOF (bi-allelic loss-of-function), where both copies of a gene are inactivated, versus MOLOF (mono-allelic loss-of-function).
  • Analysis: They correlated mutation status with HRD scores (a genomic signature of repair failure) and Progression-Free Survival (PFS) using Kaplan-Meier estimates and Cox proportional hazards models.

Datasets

The study relied on the FACT cohort, consisting of medical records and genomic data from patients treated at Peking Union Medical College Hospital and West China Second University Hospital between 2009 and 2020. No new public benchmarks were introduced, but the study provides a specialized look at a specific ethnic population often underrepresented in global genomic databases.

Key Insights and Ideas

  • Mutual Exclusivity: One of the most striking findings was that RAD51D mutations and BRCA1/2 mutations did not overlap in any patients. This suggests that RAD51D is an independent driver of HRD, and testing for it catches "missing" patients that a BRCA-only screen would overlook.
  • The Importance of "Double Loss" (BILOF): While any RAD51D mutation was beneficial, patients with bi-allelic loss (BILOF) showed significantly higher HRD scores and even better survival outcomes. This reinforces the "two-hit hypothesis" in tumor suppressor genes: losing both copies is what truly drives the clinical phenotype.
  • The "HRR3" Concept: The authors propose a simplified "HRR3" gene list—BRCA1, BRCA2, and RAD51D. They found that this three-gene panel was a slightly stronger predictor of chemotherapy efficacy than BRCA1/2 alone and increased the "actionable" mutation rate from 35% to 37.9%.
  • A "Clear Winner" Among Non-BRCA Genes: While other genes like RAD51C and PALB2 are often grouped with RAD51D, this study found that RAD51D had a much stronger correlation with HRD scores and survival in this specific cohort.

Key Results

  • Prevalence: 2.9% (11/380) of patients had RAD51D mutations. Of these, 72.7% were BILOF.
  • Survival Boost: For the general population, RAD51D mutations increased median PFS from 13 months to 25.2 months. For those with the BILOF variant, median PFS reached 36.5 months (HR=0.571).
  • BRCA-Wildtype Benefit: In patients who were BRCA1/2 "wild-type" (no mutation), having a RAD51D mutation still significantly predicted better response to chemotherapy (P=0.031).
  • HRD Correlation: RAD51D mutations elevated the median HRD score from 56 to 65.

Limitations and Future Work

The primary limitation is the small sample size of the RAD51D mutant group (only 11 patients), which is a common hurdle when studying rarer genetic variants. Because this was a retrospective study, there is a risk of selection bias or "randomness" in the findings.

The authors call for larger, prospective clinical trials to validate these findings. Furthermore, while this study focused on first-line chemotherapy, future research should explore how RAD51D BILOF status specifically impacts the efficacy of PARP inhibitors in the maintenance setting for Chinese patients.

Summary updated

Abstract

ABSTRACT Human genome is continuously threatened by endogenous and exogeneous sources of DNA damage, DNA damage repair (DDR) mechanisms are vital for genome instability and cancer prevention. Homologous recombination repair (HRR) is one of the most important DDR mechanisms because of its high-fidelity. BRCA1/2 have long been the best characterized HRR genes, but other non-BRCA1/2 HRR genes should not be overlooked. In this study, we illustrate that RAD51D is an important non-BRCA1/2 HRR gene in Chinese ovarian cancer patients, from perspectives including prevalence (2.9% patients mutated in RAD51D, 72.7% of RAD51D mutations were bi-allelic loss-of-function [BILOF]), correlation with HRD score (P=0.295 for RAD51D mutation, P=0.087 for RAD51D BILOF), and correlation with efficacy (progression-free survival: HR=0.685 P=0.069 for RAD51D mutation, HR=0.571 P=0.054 for RAD51D BILOF).
BibTeX
@misc{sun2024landscape,
  author = {T. Sun and Mengpei Zhang and Xinyun Xu and Yiming Liang and Jinghong Chen and Qingli Li and Jing Zeng and Zhe Li and Yu Dong and Rutie Yin},
  title = {{The Landscape of RAD51D in Chinese Ovarian Cancer Patients: Prevalence, Correlation with HRD Score, and Correlation with Efficacy}},
  year = {2024},
  doi = {10.1101/2024.03.19.24304522},
  howpublished = {medRxiv},
}
Figure