RNA-Based Homologous Recombination Deficiency Signature Detects Homologous Recombination Deficiency-RNA+ Patients With and Without Homologous Recombination Repair Gene Pathogenic Alterations in Men With Prostate Cancer.

Homologous recombination deficiency (HRD) is a well-described phenotype of some prostate cancers; however, current biomarkers for HRD are imperfect and rely on detection of single gene alterations in the homologous recombination repair (HRR) pathway, which may not capture the complexity of HRD biology. RNA signature-based methods of HRD identification present a potentially dynamic assessment of the HRD phenotype; however, its relationship with HRR gene alterations is not well characterized in prostate cancer.

A HRD assay on the basis of an RNA signature associated with biallelic BRCA1/2 loss was applied to a retrospective cohort study of 985 men with prostate cancer analyzed on the Tempus xT platform. HRD status was defined by a binary threshold on a continuous scale.

In this cohort, of the 126 (13%) patients found to be HRD+ by RNA signature (HRD-RNA+), 100 (79%) had no coexisting HRR gene alteration. Among samples with biallelic BRCA1/2 loss, 78% (7/9) were classified as HRD-RNA+, while 8% (2/25) of samples with BRCA1/2 monoallelic loss were HRD-RNA+. Biallelic and monoallelic ATM loss exhibited HRD-RNA+ at a lower prevalence: 6.7% (1/15) and 7.1% (1/14), respectively, compared with HRD-RNA+ prevalence among samples without any HRR gene loss (13%; 100/782). HRD-RNA+ was associated with a significantly higher prevalence of TP53 and AR gene alterations relative to HRD-RNA- after correction for multiple comparisons, 59% versus 39% (q = 0.003) and 23% versus 12% (q = 0.024), respectively.

Use of an RNA-based HRD signature significantly expands the fraction of patients with prostate cancer who may derive benefit from poly (ADP-ribose) polymerase inhibitors (PARPis) compared with using HRR gene mutations alone. Further studies are needed to evaluate functional HRD significance and inform future usage as a predictive biomarker for PARPi selection.

JCO precision oncology. 2023 Sep [Epub]

Landon C Brown, Jason Zhu, Elizabeth Mauer, Stephanie N Thiede, Lisa Macera, Michelle M Stein, Timothy Taxter, Derek Raghavan, Earle F Burgess

Levine Cancer Institute, Atrium Health, Charlotte, NC., Tempus Labs, Inc, Chicago, IL.