HRR and BRCA Alterations in Prostate Cancer: What the Test Tells Us, and What It Doesn't

When a man with metastatic castration-resistant prostate cancer (mCRPC) is found to carry a pathogenic BRCA2 alteration, the clinical question shifts almost immediately — not to whether homologous recombination repair (HRR) testing matters, but to which testing approach actually captured that variant, and which one might have missed it. That distinction, between germline and somatic detection, tissue and plasma, is where much of the real interpretive work in this space lives, and it's why an HRR result can rarely be read off a report in isolation.

What the Test Measures

Homologous recombination repair is one of the cell's high-fidelity systems for mending double-strand DNA breaks. When a gene in this pathway — most importantly BRCA2, but also BRCA1, ATM, and a longer list of others — carries a loss-of-function alteration, the cell loses that accurate repair machinery and leans on error-prone backup pathways. The result is genomic instability, which drives tumor development but also creates a therapeutic vulnerability: cells already crippled in one repair pathway are exquisitely dependent on the remaining ones [1].

Not every gene on an HRR panel carries the same weight. BRCA2 is the most consistently actionable and the best characterized; the evidence for ATM and several others is thinner and, in places, contested. So "HRR-altered" is a useful shorthand, but it flattens a panel of genes whose biology and predictive strength genuinely differ.

How It's Tested

Two complementary strategies are in play, and they answer different questions.

Germline sequencing uses a blood (or saliva) sample and asks whether a pathogenic variant is inherited — present in every cell. This matters not only for the patient's own care but for relatives, since a germline BRCA2 finding carries cascade-testing and cancer-risk implications well beyond the prostate.

Somatic next-generation sequencing interrogates the tumor itself, either from archival FFPE tissue or from circulating tumor DNA (ctDNA) in plasma. Somatic testing captures alterations that arose in the tumor and aren't present in the germline — and it's the only way to detect them.

The two aren't interchangeable. Germline testing will miss a variant that's confined to the tumor; a somatic assay confined to tissue can miss things too, depending on specimen quality. FFPE preanalytics are a recurring headache: older blocks, low tumor cellularity, and fixation-induced DNA damage all degrade sensitivity. Plasma ctDNA sidesteps the block-quality problem but depends on the tumor actually shedding enough DNA — in low-volume or indolent disease, a "negative" plasma result may simply mean insufficient tumor fraction rather than true absence of an alteration. In practice, germline and somatic testing are best treated as partners, and a negative on one doesn't close the question.

Scoring is comparatively straightforward once sequencing is done: the report states whether a pathogenic (or likely pathogenic) HRR-gene alteration is present or absent. Variants of uncertain significance sit in a gray zone and generally aren't treated as actionable.

What Each Result State Means

An HRR/BRCA alteration present result identifies a tumor with a defect in high-fidelity DNA repair. Among these, BRCA2 is the most actionable and carries the strongest predictive signal; it also informs prognosis and has implications for screening in the patient's family. An ATM or other lower-tier alteration still flags an HRR defect, but the strength of the predictive association is weaker and less uniform across the evidence base.

An alteration absent result means no pathogenic HRR variant was detected by the assay used. That phrasing matters — "not detected" is not the same as "not present." A germline-only workup can't exclude a somatic variant, and a low-yield ctDNA sample can't confidently exclude anything. The specimen and the assay define the boundaries of what a negative can claim.

What It Determines for Treatment Eligibility

The central clinical function of HRR testing in mCRPC is predictive: it informs eligibility for the PARP-inhibitor drug class, and in some indications for PARP inhibitors combined with androgen-receptor-pathway inhibitors [1, 2]. PARP inhibitors exploit synthetic lethality — a tumor already deficient in homologous recombination can't tolerate the additional repair blockade, and the accumulating damage becomes lethal to those cells [1].

Consider how a result actually flows through a decision. A pathogenic germline BRCA2 alteration is about as unambiguous as this biomarker gets: it establishes HRR deficiency, is captured by both germline and somatic assays, and typically places the tumor squarely within the population studied in the PARP-inhibitor trials — the PROfound program being the pivotal example [2]. It simultaneously triggers genetic-counseling and cascade-testing considerations for relatives, so a single blood test can carry consequences well beyond the index patient.

Contrast that with an isolated somatic ATM variant found on tumor NGS. Here the interpretive footing is softer. The alteration signals a possible HRR defect and may satisfy a companion-diagnostic definition, but the evidence that ATM-altered tumors respond like BRCA-altered ones is weaker and inconsistent across trials [1]. The report says "HRR alteration present," yet the expected clinical benefit isn't the same, and that gap is exactly what a treating team must weigh. It's a reminder that a positive HRR result is a gate, not a guarantee — and that the specific gene matters as much as the binary.

None of this constitutes a treatment recommendation. The biomarker establishes whether a patient's disease falls into a molecularly defined group for which a drug class has been studied and approved; the actual decision belongs to the treating clinician and patient.

Caveats and What's Evolving

The most important caveat is the one that runs through everything above: the HRR panel is not a monolith. BRCA2 anchors the actionable end, while genes like ATM sit in a region where predictive value is genuinely contested and continues to shift as trial data mature [1]. Reading "HRR-altered" as a single category obscures this, and it's the mistake most likely to lead a report astray.

That heterogeneity compounds the assay-level limitations already noted — the germline/somatic gap, FFPE degradation, and ctDNA yield. A confident interpretation therefore depends on knowing not just what was found but how it was tested and what the assay could not have seen. Variant classification adds another moving part: reclassification of a VUS over time can retroactively change a patient's eligibility status, so today's "absent" is not always permanent.

Where the field is heading is toward better resolution within the panel — distinguishing which non-BRCA HRR genes truly confer sensitivity, refining plasma-based detection thresholds so a negative ctDNA result carries more weight, and pairing sequencing with functional readouts of HR deficiency rather than relying on gene status alone. Until those tools mature, the sharpest clinical value remains concentrated in BRCA2, and the most common error remains treating every gene on the panel as if it spoke with the same voice.

References

1. PARP Inhibitors in Metastatic Castration-Resistant Prostate Cancer. PMC / PubMed Central, 2024. PMCID: PMC10890352.

2. PARP inhibitors in HRR-altered mCRPC (primary literature, including the PROfound program). Current.