# Functional footprints of homologous recombination deficiency in prostate cancer revealed by ctDNA fragmentation and transcription factor accessibility

## TL;DR

This study introduces a non-invasive blood test using ctDNA to detect homologous recombination deficiency in prostate cancer, offering a more comprehensive and accurate method than current approaches.

## Contribution

The novel contribution is a multimodal ctDNA-based strategy integrating genomic instability and fragmentomics to better identify HRD in prostate cancer.

## Key findings

- BRCA2 was the most frequently altered HRR gene, often co-occurring with PTEN loss.
- High genomic instability scores correlated with BRCA2/RB1 loss, more copy number alterations, and worse survival.
- HRD tumors showed specific mutational signatures and increased dinucleosome-length fragments with reduced zinc finger transcription factor accessibility.

## Abstract

Homologous recombination deficiency (HRD) is a predictive biomarker for response to PARP inhibitors and platinum-based therapies in prostate cancer (PCa). However, current diagnostic approaches, often limited to BRCA1/2 mutation testing or genomic scars, fail to capture the full spectrum of HRD. Tissue-based testing is further hampered by tumour heterogeneity and biopsy limitations in patients with metastatic bone disease. This study aimed to develop a noninvasive, multimodal ctDNA-based strategy for comprehensive HRD profiling in advanced PCa.

We analysed plasma-derived ctDNA from 106 patients with metastatic PCa. The approach integrated targeted sequencing of homologous recombination repair (HRR) genes, low-pass whole genome sequencing for genomic instability scores (GIS), whole-exome sequencing for mutational signature analysis, and cfDNA fragmentomics, including chromatin accessibility profiling.

BRCA2 was the most frequently altered HRR gene, frequently co-occurring with PTEN loss. High GIS was associated with BRCA2/RB1 loss, increased somatic copy number alterations, and poor overall survival. HRD tumours were enriched for mutational signatures SBS3 and ID6, displayed increased dinucleosome-length fragments, and showed reduced accessibility at zinc finger transcription factor binding sites. A fragmentomics-based classifier identified HRD-positive cases with high accuracy.

Our findings support the use of multimodal ctDNA profiling as a non-invasive approach to identify HRD in prostate cancer. The integration of mutation, genomic instability, and fragmentomic features provides a broader functional view of HRD and may enhance patient stratification for targeted therapies.