Expert practical recommendations for optimising tumour and germline testing of homologous recombination repair gene mutations in metastatic prostate cancer

Abstract

With the approval of PARP inhibitors (PARPi), both as monotherapy and in combination with androgen receptor signalling inhibitors (ARSi), for metastatic castration-resistant prostate cancer (mCRPC) and their rapid development in earlier disease settings, there is an urgent need to integrate homologous recombination repair (HRR) testing into daily clinical practice to identify patients who may benefit from these therapies and to improve patient management. Alterations in HRR genes are found in approximately 15–30% of patients with mCRPC and are associated with poor prognosis. In addition, some alterations are hereditary; therefore, testing also has implications for the identification of hereditary cancer risk.

The successful implementation of HRR gene testing depends not only on access to sequencing technologies but also on the establishment of a comprehensive pre- and post-analytical framework. The aim of this document is to establish a multidisciplinary expert opinion consensus on the optimisation of molecular assessment of BRCA1/2 and other HRR gene alterations in metastatic prostate cancer to support implementation in clinical practice.

Resumen

Con la aprobación de los inhibidores de PARP (iPARP), tanto en monoterapia como en combinación con inhibidores de la señalización del receptor de andrógenos (ARSi), para el cáncer de próstata metastásico resistente a la castración (CPRCm), y su rápido desarrollo en estadios más tempranos de la enfermedad, existe una necesidad urgente de integrar las pruebas de reparación por recombinación homóloga (HRR) en la práctica clínica habitual para identificar a los pacientes que puedan beneficiarse de estas terapias y mejorar su manejo clínico. Las alteraciones en los genes de HRR se encuentran en aproximadamente el 15–30% de los pacientes con CPRCm y se asocian con un peor pronóstico. Además, algunas de estas alteraciones son hereditarias; por lo tanto, las pruebas también tienen implicaciones para la identificación del riesgo de cáncer hereditario.

La implantación eficaz de las pruebas de los genes de HRR depende no solo del acceso a tecnologías de secuenciación, sino también del establecimiento de un marco integral preanalítico y posanalítico. El objetivo de este documento es establecer un consenso de expertos multidisciplinar sobre la optimización de la evaluación molecular de las alteraciones en BRCA1 y BRCA2 y en otros genes de HRR en el cáncer de próstata metastásico, con el fin de facilitar su implantación en la práctica clínica.

Introduction

Prostate cancer (PCa) is the most diagnosed cancer among men worldwide and the second leading cause of cancer-related death in men, with a 5-year survival rate of about 36% for patients with metastatic disease.^1,2 Metastatic castration-resistant prostate cancer (mCRPC), the most advanced and lethal form of this disease, often carries acquired genomic mutations in tumour suppressor genes such as TP53, AR, RB1, and PTEN, as well as in genes involved in DNA damage response, particularly those associated with the homologous recombination repair (HRR) pathway.³

Prognostic and predictive value of HRR alterations

The HRR pathway plays a crucial role in repairing double-strand DNA breaks (DSBs), the most lethal form of DNA damage.⁴ Approximately 15–30% of patients with mCRPC harbour pathogenic alterations in HRR-related genes, which may act as biomarkers of response to PARP inhibitors (PARPi).^3,5–8BRCA2 and ATM (a gene involved in DSB repair regulation) are the most frequently altered, followed by BRCA1, CDK12, PALB2, FANCA, CHEK2, and members of the RAD gene family. Overall, alterations in BRCA1/2 are reported in around 10% of the patients with metastatic PCa (mPCa). Somatic mutations (acquired in tumour cells) are more common than germline mutations (inherited) and occur in nearly two-thirds of cases, depending on the genetic background of the population.^7,8

PCa shows the highest heritability among major male malignancies. Men with a first-degree relative affected by the disease have at least twice the risk of developing PCa.⁹ Germline mutations in BRCA1 and BRCA2 are linked to a three- and seven-fold increased risk of developing PCa, respectively.^10,11 The frequency of germline mutations in HRR genes among patients with mPCa does not differ significantly according to age at diagnosis or family history of PCa; therefore, the absence of a family history does not rule out the presence of a germline mutation.⁶ Patients with BRCA2 mutations often present with higher Gleason scores, are more likely to develop distant metastases, and show increased rates of biochemical recurrence, along with reduced overall survival; however, these genetic variants cannot be excluded in patients without such clinicopathological features.^11–13

Patients with HRR-mutated (HRRm) mPCa have a poor prognosis and experience shorter times to progression with conventional treatments, including androgen receptor signalling inhibitors (ARSi) and taxane-based chemotherapy, but may benefit from treatment with PARPi as monotherapy or in combination with ARSi.^{5,7,8,14–16} PARPi exploit the concept of synthetic lethality, in which simultaneous impairment of the PARP-mediated single-strand break repair pathway and the HRR pathway (already deficient in certain cancer cells) leads to accumulation of unrepaired DNA damage and, ultimately, cell death.^5,17 Emerging evidence from phase 3 trials suggests that treatment with PARPi in combination with ARSi prolongs survival in mCRPC, with a clear benefit in individuals with HRRm, particularly BRCA1/2, and potentially extending to other HRR gene defects such as CDK12 and PALB2 (Appendix 1).^14–16 Notably, several of these agents are currently being evaluated in earlier disease settings, including metastatic hormone-naïve prostate cancer (mHNPC).

Methods

An expert multidisciplinary panel was selected based on their experience in HRR gene testing and PCa management. The panel performed a comprehensive literature review using PubMed and Consensus, and held a focused discussion in November 2024 to reach agreement. All experts agreed on the recommendations to optimise HRR gene mutation testing in clinical practice.

Guideline recommendations for tumour and germline testing

The inclusion of PARPi in clinical practice underlines the need for molecular profiling to identify patients with PCa who may benefit from these treatments. Major guidelines recommend genomic (tumour) and genetic (germline) testing for HRR variants in mPCa (Table 1).^18–22

Table 1.

Guideline recommendations for germline and tumour HRR testing.

Guideline	Germline testing	Tumour testing
NCCN 2025	Patients with a family history, high-risk, regional, or mPCa, Ashkenazi Jewish ancestry, or a personal history of breast cancer	Recommended for all patients with mPCa; may be considered for regional PCa
ESMO 2026	Patients with a family history of cancer [III,B]; recommended for all patients mHNPC [III,A].Germline testing is also recommended for patients with mCRPC if not previously performed [III, B].	Consider tumour testing for HRR genes [III,A] in all patients with CRPC, ideally before first-line treatment for mCRPC.Testing for mismatch repair deficiency (or microsatellite instability) is recommended for all patients with mCRPC after ARSI treatment [III,A]
EAU 2025	Men with mPCa eligible for targeted therapy; those with BRCAm identified on somatic testing; those with multiple family members diagnosed with PCa < 60 years or with a family member who died from PC, and with a FH of high-risk germline mutations or multiple cancers on the same familial side	Recommended for all patients with mPCa, preferably before initiating first-line treatment for mCRPC
ASCO 2025	Recommended for all patients with mPCa	Recommended for patients with mPCa being considered for biomarker-directed systemic therapy

FH, family history; mHNPC, metastatic hormone-naïve prostate cancer; mCRPC, metastatic castration resistant prostate cancer; mPCa, metastatic prostate cancer; PC, prostate cancer.

Currently, as PARPi are only approved for patients with mCRPC, tumour testing should be a priority for this group. Broader testing is likely to become increasingly relevant as targeted therapies expand into earlier disease stages.²³

The expert panel recommends performing HRR testing whenever the results may influence clinical management, at a minimum in all mCRPC patients who may be candidates for targeted therapies, and ideally in all patients with mPCa at the time of mHNPC diagnosis. The analysis should enable the incorporation of the results into clinical decision-making for first-line treatment.

Germline versus tumour testing

Tumour testing provides both prognostic and predictive information and can influence treatment decisions for patients with mPCa by determining eligibility for biomarker-directed therapies, informing prognosis, and identifying candidates for clinical trials.^18,24 Germline genetic testing may provide similar prognostic and predictive information, while also identifying the risk of hereditary cancers that may affect patients and their family members. However, even when germline testing identifies an actionable pathogenic mutation, tumour testing may still be useful to confirm loss of heterozygosity and/or identify other actionable somatic alterations.²¹

Performing germline testing alone could miss up to two thirds of patients with HRRm (those with somatic alterations), depending on the population's genetic background and genes analysed.²⁴ Conversely, tumour testing does not necessarily distinguish whether a variant is present in the germline or is somatic, and may reveal findings suggestive of germline mutations that require confirmatory germline testing. Tumour-only testing may also fail to optimally detect certain clinically actionable germline variants in HRR genes due to technical challenges associated with targeted next-generation sequencing (NGS) analysis. Although the risk is relatively small, a pan-cancer study estimated that up to 7% of patients with germline variants may yield false-negative results on tumour testing.²⁵ Therefore, if tumour testing is negative but clinical suspicion or family history suggests a germline alteration, germline testing should be considered.^24,25 The advantages and disadvantages of germline and tumour testing are summarised in Table 2.

Table 2.

Advantages and disadvantages of germline and tumour testing.

Advantages	Disadvantages
Germline testing
• Identifies hereditary cancer risk, allowing cascade testing for family members • Informs risk of other malignancies • Provides prognostic and therapeutic information	• Does not capture acquired somatic mutations or tumour evolution• A negative result does not exclude somatic HRR mutations in the tumour
Tumour testing
• Provides prognostic and therapeutic information • May indicate the presence of germline mutations • Allows re-evaluation at disease progression	• Tumour heterogeneity can affect test accuracy• Results may be highly affected by pre-analytical conditions of the biospecimen.• Detection of large deletions may be challenging in tissue analysis

HRR, homologous recombination repair.

The expert panel recommends that, at a minimum, all patients with mPCa should undergo tumour testing, with germline testing performed if HRR mutations are detected or if a clinical suspicion or a relevant family history exists. Ideally, both tumour and germline testing should be performed for all patients with mPCa.

HRR gene panel: which genes should be included?

Most clinical guidelines do not explicitly define which HRR genes should be included in standard clinical testing but recommend using gene panels that can both assess familial cancer risk and provide data relevant to treatment decisions and clinical trial eligibility.

The National Comprehensive Cancer Network Clinical Practice Guidelines in Oncology (NCCN Guidelines) recommend multigene tumour testing for alterations in HRR genes, including but not limited to BRCA1, BRCA2, ATM, PALB2, FANCA, RAD51D, CHEK2, and CDK12, in patients with mCRPC. In addition, tumour testing for MSI-H or mismatch repair deficiency (dMMR) is recommended because of its therapeutic implications. Germline testing should include MLH1, MSH2, MSH6, and PMS2 (associated with Lynch syndrome), as well as the HRR genes BRCA1, BRCA2, ATM, PALB2, and CHEK2, given their association with hereditary cancer risk.¹⁸

The European Society for Medical Oncology Precision Oncology Working Group (POWG) recommends assessing the mutational status of at least BRCA1/2 with the option of including additional genes (such as ATM and PALB2).²⁶ The ESMO guidelines further recommend testing for BRCA1, BRCA2, CDK12 and PALB2, as well as for dMMR.²⁰

Clinical trials evaluating PARPi in prostate cancer have consistently included testing for BRCA1, BRCA2, CDK12, PALB2, ATM, and CHEK2. Additional genes evaluated in these trials were BARD1, BRIP1, CHEK1, FANCL, RAD51, RAD54, FANCA, HDAC2, ATR, MLH1, MRE11A, NBN, PPP2R2A, RAD51B, RAD51C, RAD51D, and RAD54L.^{14–16,27,28} Although additional genes may be included in NGS panels for research purposes, the use of larger gene panels for germline testing may increase detection of variants of uncertain significance (VUS).²⁴

The homologous recombination deficiency (HRD) signature used in ovarian cancer is not recommended for predicting response to PARPi in mPCa, as it has not been validated because of fundamental differences in HRD biology, clonality, and functional relevance at the time of treatment.²⁹

The expert panel recommends testing at least those genes with a currently approved therapeutic indication (e.g. in Spain, access to PARPi is linked to germline and/or somatic mutations in BRCA1/2), as well as germline mutations that may influence clinical follow-up or genetic testing of the patient and their relatives due to hereditary implications. Ideally, a broader panel of HRR genes should be included, particularly those with demonstrated predictive value for PARPi response, such as CDK12 and PALB2, or ATM, FANCA, RAD51D, and CHEK2. The panel also recommends including additional genes with prognostic value, such as RB1, TP53, PTEN, and AR, as well as genes associated with dMMR.

Specimen selection

Tumour testing may be performed using tumour tissue samples (either fresh or formalin-fixed, paraffin-embedded (FFPE) biopsies or surgical specimens) or plasma circulating tumour DNA (ctDNA) isolated from peripheral blood samples. An evidence-based algorithm should guide the selection of the optimal specimen type, and diagnostic laboratories must clearly define and communicate both specimen requirements and assay limitations to referring clinicians.²⁴

Tumour tissue analysis remains the gold standard for molecular testing; however, its routine implementation presents challenges related to specimen availability and quality. Failure rates of archival primary prostate tumour biopsies may reach 30–40%, due to factors such as low tumour cellularity, inadequate DNA yield, or poor DNA quality.^15,30 DNA extraction, fragmentation and library preparation are key processes that might also affect results. Analyses of paired samples from primary prostate tumours and metastases at the time of castration resistance suggest that HRR gene mutations occur early in the evolution of advanced PCa.^31,32 Therefore, evaluation of the dominant tumour in archival diagnostic specimens is appropriate for molecular diagnostics even after progression to mCRPC, provided that these specimens retain good-quality DNA.²³

Among the various tissue specimen sites, lymph node biopsies demonstrate the highest success rates, because of high tumour cellularity and the availability of high-quality, non-decalcified tissue suitable for comprehensive NGS, whereas bone-derived specimens show the lowest success rates, primarily due to the negative impact of decalcification on DNA quality.^30,31

When tissue biopsy is unsafe or unfeasible, ctDNA testing offers a non-invasive alternative for tumour testing. The analysis of HRR gene alterations in matched tumour tissue and ctDNA samples from patients with mPCa screened for the PROfound study showed 81% positive agreement and 92% negative agreement for BRCA and ATM. However, at the variant subtype level, concordance was notably lower for homozygous deletions (27%),³³ largely because of the lower tumour content in plasma compared with tumour samples, as tumour content is a key factor for accurate estimation of gene copy number. Liquid biopsy testing presents several challenges, including false-positive results caused by the detection of clonal haematopoiesis of indeterminate potential (CHIP) mutations in circulating cell-free DNA. These may originate from expanded haematopoietic cells rather than from the tumour and may therefore be mistakenly interpreted as targetable tumour variants. False-negative results may arise from both technical factors (e.g., sample processing or limited sensitivity) and biological factors (e.g., low-shedding tumours resulting in low tumour fraction in circulating cell-free DNA).²⁴ Plasma for ctDNA analysis should be preferably collected at the time of clear disease progression to maximise diagnostic yield, as actively progressing cancers release more ctDNA into the circulation.^18,24

Gene alterations detected in both ctDNA and FFPE samples may sometimes reflect germline alterations from normal cells.²¹ However, neither specimen type has been validated for the reporting or interpretation of germline variants. The advantages and disadvantages of tissue and liquid biopsy for tumour testing in mPCa are summarised in Table 3.

Table 3.

Advantages and disadvantages of tissue biopsy and liquid biopsy for tumour testing in metastatic prostate cancer.

Advantages	Disadvantages
Tissue biopsy specimen
• Allows integration of genomic testing with histological evaluation • Archival tissue may be available for testing • Well-established method in clinical practice	• Specialised expertise required for tissue acquisition• Biopsy procedures are invasive and associated with procedural risks; some disease sites are challenging (i.e. bone, retroperitoneum)• May not entirely reflect tumour heterogeneity• Diagnostic biopsy may not capture tumour evolution• Serial biopsies are impractical for longitudinal monitoring of treatment response
Liquid biopsy specimen
• Minimally invasive procedure for sample collection • Faster turnaround time than tissue biopsy • Suitable for serial sampling/real-time monitoring of treatment response and resistance • Potential to capture spatial and temporal tumour heterogeneity	• May yield false negative results if there is not enough ctDNA• Potential for false positive results (CHIP)• May miss homozygous somatic deletions• Not routinely used in early disease stages due to insufficient disease burden

ctDNA, circulating tumour DNA; CHIP, clonal hepatopoiesis of indeterminate potential.

Adapted from Selvarajah et al.²⁴

The expert panel recommends selecting the most recent tumour tissue sample with the highest tumour content that meets the minimum requirements for the test being used (generally, point or frameshift mutations > 20%; copy number variations (CNV) ≥ 30%) and that has the best quality among the available specimens. Good practice dictates that specimen selection should be a multidisciplinary decision involving both the pathologist and the treating physician.

• •

  The primary tumour tissue sample is suitable provided that sufficient quality and cellularity are present. When testing the primary tumour using core needle biopsy material, the panel recommends selecting the sample with the highest Gleason score.

• •

  If a high-quality primary tumour sample is available, performing a biopsy of metastatic sites is unlikely to improve the assessment of HRR genes. Re-biopsy should be considered only when no optimal sample is available or when there is a discordance between the clinical behaviour and the patient's molecular characteristics. When possible, lymph node tissue should be preferred over bone metastases. The study of late-stage metastatic biopsies can uncover cancer temporal evolution and heterogeneity features, but their impact in clinical assessment of HRR mutations is minimal.

• •

  ctDNA represents an alternative approach, particularly when obtaining a new biopsy is unsafe or unfeasible. However, clinicians should remain aware of the potential for both false-positive and false-negative results.

Pre-analytical phase

Sample collection

Optimising tissue acquisition at the time of diagnosis is vital for the success of molecular testing in PCa. The expert panel recommends following the current recommendations from the European Association of Urology (EAU) 2025 guidelines,¹⁹ favouring transperineal targeted biopsy of visible lesions on magnetic resonance imaging (MRI), including perilesional samples. In the absence of MRI, a minimum of 12 cores should be collected at initial diagnosis. These procedures should aim to maximise tumour tissue capture while minimising contamination by necrotic or stromal areas.

Biopsy cores should be processed separately according to the sample site indicated on the specimen container.³⁴

Technical recommendations for sample processing and block selection

Proactive collaboration with pathology services is crucial. HRR testing should ideally be performed reflexively at diagnosis or, at the very least, through a personalised approach.

The expert panel recommends that suitable specimens for HRR testing should contain enough tumour cellularity (i.e. 20–30% neoplastic content) to ensure variants can be reliably detected and distinguished from deamination or oxidation artefacts and background noise (Table 4).³⁵ Low tumour content limits the detection of somatic mutations with low allele frequency and compromises accurate assessment of CNVs. Selecting tumour-rich areas increases the likelihood of obtaining viable DNA and reduces the risk of testing failure.

Table 4.

Expert panel recommendations for the processing and storage of FFPE samples for NGS analysis.

Factor	Recommendation
Tumour cellularity	>20–30% neoplastic content, depending on the test performed^a
Fixation method	10% neutral buffered formalin
Fixation time	As short as possible (i.e., 3–6 h for core needle biopsies; maximum 24 h)
No. of cores per block	≤2 cores/block
Decalcification	For bone metastasis biopsies, use EDTA rather than acid decalcification to preserve DNA integrity
Block identification and labelling	Include in the diagnostic report the most appropriate block(s) for molecular testing, indicating the percentage of tumour cellularity. Clearly label blocks for easy identification
Storage of FFPE blocks	Controlled environment (i.e., low humidity, 18–25 °C) to minimise nucleic acid oxidation and degradation

CNV, copy number variations; EDTA, ethylenediaminetetraacetic acid.

a

Minimum criteria for HRR genes should be reviewed according to test specifications (point mutations and frameshift mutations > 20%; CNV ≥ 30%).

To preserve the molecular integrity of samples, fixation protocols must be carefully controlled. The recommended fixative is 10% neutral buffered formalin. Core needle biopsies should be fixed for a minimum of 3–6 h and no longer than 24 h. Transurethral resection and prostatectomy specimens should be fixed for 24–48 h. Over-fixation can lead to excessive DNA crosslinking, fragmentation, and chemical modifications such as deamination, which may interfere with PCR amplification and sequencing quality. Experts recommend embedding no more than two cores per FFPE block to facilitate precise sectioning and efficient macro-dissection, and to avoid sample exhaustion.

Processing bone metastasis biopsies adds complexity due to the need for decalcification. Traditional acid-based decalcification can severely damage nucleic acids, rendering samples unusable for genomic analysis. Therefore, EDTA-based decalcification is strongly recommended, despite its longer processing time, as it significantly improves DNA yield and quality.³⁶

Macro-dissection is advised to enrich tumour areas, particularly when cores are heterogeneous. Pooling multiple cores from more than one biopsy may increase DNA yield; however, it carries the risk of obscuring inter-lesion heterogeneity.

The most appropriate blocks for molecular testing, based on tumour content, should be clearly identified in pathology diagnostic reports and clearly labelled to facilitate retrieval at the time of genomic testing. Attention should be also paid to tissue preservation, absence of necrosis, and minimal contamination with benign tissue. Standardising the block selection process within pathology laboratories at the time of diagnosis can streamline the workflow and reduce diagnostic delays.

Storage

DNA stability over time is influenced by environmental conditions and sample handling. Proper storage conditions include a stable temperature (18–25 °C) and low humidity, which can greatly extend the usability of FFPE samples by minimising oxidative damage and hydrolytic degradation (Table 4).

In the PROfound study, although success rates declined with increasing sample age, nearly 50% of archived samples stored for more than 10 years still produced reliable sequencing data,³⁰ highlighting the feasibility of long-term use when storage and processing are optimised.

Optimisation of tissue use

Strategic planning of tissue use is vital to ensure adequate material for molecular testing. Immunohistochemistry panels should be optimised to minimise exhaustion of the tissue block, and only the minimum amount of tissue necessary should be used for diagnosis.

Diagnostic workflows should include checkpoints to identify cases at risk of failing molecular analysis due to pre-analytical shortcomings. Institutions should implement standard operating procedures (SOPs) and local guidelines that incorporate genomic testing into recommended procedures for mPCa management, considering the need to preserve tissue under appropriate conditions and designating a responsible individual to ensure traceability and accessibility of samples over time. When developing efficient local sample workflows, it is essential to consider the time and cost implications of test failures. Such failures, or the need for re-biopsy, can delay access to targeted therapies, a critical concern for patients with mPCa.²⁴

Analytics

NGS allows for the simultaneous sequencing of millions of DNA fragments and is currently used for HRR testing.²⁶ NGS workflows generally comprise four main steps: sample preparation, library preparation, sequencing and data analysis.

Sample preparation in the molecular laboratory

Tissue sample preparation involves extraction and fragmentation of nucleic acids to produce fragments with a uniform size distribution. Primary quality control of DNA samples, including quantification of double-stranded DNA yield and fragment size assessment, is essential to minimise post-library preparation test failures.²³ Archival FFPE samples are most likely to fail during DNA extraction; therefore, optimisation of the extraction procedure is important to reduce sample transfers and prevent material loss across multiple steps.³⁰ During this process, it is critical to avoid cross-contamination between samples by changing scalpel blades between tissue dissections, wiping work surfaces frequently, and ensuring that samples are handled individually.³⁷ Due to the deaminating effects of formalin, an enzymatic repair step is recommended prior to DNA purification for FFPE specimens to improve sequencing accuracy.³⁸

The expert panel recommends the use of validated DNA extraction and mechanical shearing protocols, for both tissue or ctDNA, to ensure sufficient DNA quantity and quality for the chosen methodology. For FFPE samples, using a DNA repair protocol is advised. Quality control checks should be performed on both extracted DNA and generated libraries to prevent the unnecessary use of resources. Scalpel blades must be changed between each tissue dissection. Five to ten sections, with a thickness of 5–10 μm, are recommended, depending on tissue size and cellularity.

Evaluation of NGS technologies

Two principal strategies are employed for the NGS analysis of target genes: hybrid capture-based and amplification-based approaches. Hybrid capture uses biotinylated, sequence-specific probes to enrich genomic regions of interest. In contrast, amplicon-based approaches employ multiplex PCR to enrich target regions. Library preparation varies according to the selected method. Prior to sequencing, generated libraries are assessed for size distribution and quantified for quality control.³⁷

Amplification-based NGS is a versatile, scalable technique that allows construction of libraries of various sizes. It requires less hands-on time and lower DNA yields, making it more suitable for small-panel testing than hybrid capture-based methods.^23,39 However, it is sensitive to limitations in PCR primer design. Mismatches in primer regions due to single nucleotide polymorphisms (SNPs) or indels can lead to allele dropout, reduced coverage, and inaccurate variant calling. Moreover, this method is less effective for GC-rich or repetitive regions and often exhibits reduced sequencing quality at the amplicon ends.³⁷

In hybrid capture-based NGS library preparation, fragment size affects assay performance: shorter fragments improve capture specificity by reducing off-target sequences. Hybrid capture probes are longer than PCR primers, allowing for hybridisation despite the presence of mismatches. This method captures both target regions and surrounding DNA, enabling broader genomic coverage than amplicon-based assays. However, it may also enrich off-target regions, potentially reducing depth of coverage in areas of interest if the library is not properly balanced.³⁷

The minimum depth of coverage required during sequencing depends on the methodology employed. For reliable detection of variants at a 5% variant allele frequency (VAF), approximately 500× coverage is recommended for amplicon-based NGS assays that do not incorporate de-duplication strategies (i.e., unique molecular identifiers), whereas ∼200× coverage is generally sufficient for hybrid capture-based approaches.⁴⁰ Nonetheless, each laboratory should establish its own minimum coverage thresholds according to the analytical sensitivity required and the specific characteristics of the sequencing platform and assay design.³⁷ Statistical tools can aid in calculating the minimum required depth of coverage (available at http://app.olgen.cz/clc/). The large size of the BRCA1 and BRCA2 genes requires comprehensive analysis of all exons and splice sites, making variant detection challenging. Studies have shown that hybrid capture-based NGS approaches outperform amplicon-based methods in detecting BRCA1/2 and other HRR mutations, offering better coverage uniformity, fewer allele dropouts, reduced false positives, and higher single nucleotide variants (SNV), indels and CNV detection rate, particularly in FFPE-derived DNA. Furthermore, these approaches are more suitable for detecting low- frequency variants.^39,41,42 A comparison of hybrid capture-based versus amplicon-based NGS approaches is presented in Table 5.

Table 5.

Comparison of hybrid capture-based vs amplicon-based NGS methods.

Advantages	Disadvantages
Hybrid capture-based
• Broad genomic coverage • Tolerates mismatches • Improved detection of CNVs and low-frequency variants • High reproducibility	• Requires higher DNA input• Longer preparation time• Potential for off-target enrichment
Amplicon-based
• Fast and simple workflow • Requires lower DNA input • Suitable for small panels	• Sensitive to primer mismatches• Allele dropout risk• Less reliable in FFPE or GC-rich regions• Limited detection rate for CNVs

CNV, copy number variations; FFPE, formalin-fixed, paraffin-embedded.

The expert panel recommends validated hybrid capture-based NGS over amplicon-based approaches to maximise detection of HRR mutations. If only amplicon-based NGS approaches are available, it is essential to ensure coverage of the full coding regions and splice sites of the genes under analysis.

Post-analytical phase: variant interpretation

After sequencing, the data analysis pipeline can be divided into four primary operations: base calling, read alignment, variant identification, and variant annotation.³⁷ Clinical validation of bioinformatic analysis pipelines is essential before reporting results, and adherence to Genome Analysis Toolkit (GATK) best practice guidelines for initial data processing is recommended to ensure consistency and reliability.⁴³

Standardised approaches for interpreting HRR testing results are essential to support clinical decision-making. Peer-reviewed literature, clinical guidelines, and cancer mutation databases are key resources for assessing the clinical significance of variants. Interpretation requires both education and experience, and ideally, a molecular tumour board (MTB) should be involved in providing genomic-informed clinical recommendations, particularly for cases exhibiting complex genomic alterations, as discussed by the POWG.²⁴

The ENIGMA guidelines for interpreting BRCA1/2 results are of particular relevance.⁴⁴ Although there are no specific guidelines for general HRR variant interpretation, the standards and guidelines of the American College of Medical Genetics and Genomics (ACMG guidelines) are well-accepted in practice,⁴⁵ and the Standards and Guidelines for the Interpretation and Reporting of Sequence Variants in Cancer (AMP/ASCO/CAP guideline)⁴⁶ are commonly used to support variant classification.

Variants are considered clinically relevant biomarkers only when they predict sensitivity to therapy, serve as inclusion criteria for clinical trials, affect prognosis, aid diagnosis, or have implications for the assessment of hereditary cancer risk.^46,47 It is important to distinguish between pathogenicity (the impact of a variant on the biological function of a protein) and clinical actionability (the potential of a variant to guide patient management decisions). While some HRR genes, particularly BRCA2, have both prognostic and predictive value, other genes identified through NGS panel testing have only prognostic or hereditary implications.^21,24 Thus, not all pathogenic variants are clinically actionable.

Alterations in tumour suppressor HRR genes most frequently include SNVs, indels, and CNVs (mainly somatic in BRCA1/2), whereas rearrangements are rare. VAF is a key metric for interpretation, and low-frequency variants are more challenging to detect. To avoid false-positive or artefactual results, it is essential to report only variants with VAFs above the validated detection limit of the method used, typically around 5% for FFPE samples.²³

HRR testing identifies VUS in nearly 30% of cases; however, clinical decisions should not be based on their presence.⁴⁸

Although recent studies have associated clonal heterozygous deletions in HRR genes with poorer prognosis,^7,8 there is insufficient evidence to support considering monoallelic somatic deletions as clinically actionable.^17,49 Importantly, most PARPi trials in prostate cancer have defined tumours with pathogenic mutations (regardless of evidence for second allele loss) or homozygous deletions as “HRR-altered”, but have excluded cases with single-copy deletions, with the exception of the MAGNITUDE trial, which included all.^5,14–16

In an exploratory analysis of the PROfound trial in patients with alterations in BRCA1/2, olaparib was beneficial in all patients with mutations, regardless of whether loss of the second allele could be demonstrated.⁴⁹ Several studies have shown that most tumours with BRCA2 mutations harbour a second inactivation event, although this is often not detected by the targeted sequencing panels commonly used in clinical practice. Interestingly, retrospective analyses of several PARPi trials in prostate cancer have identified the greatest benefit from PARP inhibition in cases with BRCA2 homozygous deletions, which accounts for 2–4% of all patients with mPCa.^5,49

The expert panel recommends validating bioinformatic analysis pipelines and adhering to the GATK best practices for accurate NGS data analysis. Gene analysis should assess SNVs, indels and CNVs. Clinical interpretation of genomic testing results requires expert input and reliable resources, and a MTB is recommended for complex cases. Training and experience of professionals are key, and multidisciplinary expert meetings are recommended for collaborative variant classification.

The panel advises against reporting somatic monoallelic deletions without evidence of mutation in the other allele as actionable events. While biallelic loss (i.e. a mutation in one allele with loss of the second allele) is required to confer PARPi sensitivity, targeted panels have limited sensitivity to detect loss-of-heterozygosity at the individual gene level, particularly in the context of low tumour content. In such cases, referral for germline testing may be useful for confirmatory analysis.

Reporting results

Laboratories reporting NGS test results should ensure that findings are communicated clearly and are readily interpretable by clinicians. The ESMO recommends that genomic reports include the following information: (i) patient and sample characteristics, including estimates of tumour or ctDNA content to provide context for interpretation, (ii) NGS assay and data analysis characteristics and limitations, (iii) sample-specific assay performance and quality control metrics, (iv) genomic alterations and their functional annotation, (v) assessment of clinical actionability, including potential therapeutic implications as well as diagnostic and prognostic relevance, (vi) a summary of the main clinically relevant findings; and (viii) appendices with detailed information and references.⁵⁰

Reporting of genomic variants should follow standard nomenclature, using genomic coordinates or the Human Genome Variation Society (HGVS) guidelines⁴⁶ whenever possible. Test limitations, including the inability to detect larger chromosomal rearrangements, should be clearly stated.^23,50

Only alterations with expected impact on gene function (pathogenic or likely pathogenic) should be included in the main report and evaluated for clinical actionability. If VUS are reported, they should be presented in a separate section to avoid confusion.^23,50 Reports should clearly outline the clinical actionability of variants, highlighting those with therapeutic relevance, as well as variants that may indicate hereditary cancer risk. They should also indicate the need for follow-up or confirmatory tests, including the potential for germline events.^24,50

Effective communication between requesting physicians and professionals reporting genomic data is key to minimise uncertainty and optimise the impact of genomic testing.⁵⁰

The expert panel recommends clear and structured genomic reports, with explicit mention of NGS test limitations, clear distinction between pathogenicity and clinical actionability of the genomic finding, and provision of clear information to support interpretation of the results.

Integrating HRR testing into the patient journey

A harmonised, patient-centred algorithm for tumour and germline HRR testing should be developed, considering cost-effectiveness, logistical considerations, funding, and technology availability. Timely testing is essential to ensure equitable access to PARPi therapy, particularly in resource-limited settings.

All patients with mPCa should undergo tumour testing to identify alterations in HRR genes. Molecular testing is typically initiated at the time of diagnosis of metastatic disease, in line with access to biomarker-targeted therapies. However, for patients with a strong family history of cancer, germline testing for cancer predisposition genes may be considered even in the setting of localised or regional disease.^18–20Fig. 1 presents a germline and tumour testing algorithm for patients with mPCa.

Fig. 1.

Tumour and germline testing algorithm for metastatic prostate cancer. ARSi, androgen receptor pathway inhibitor; ctDNA, circulating tumour DNA; LP, likely pathogenic; P, pathogenic; PARPi, PARP inhibitor QC, quality control. *Select the most recent sample with the highest percentage of tumour content, highest Gleason score, and best overall quality. Tissue blocks most suitable for molecular diagnosis should be clearly identified and properly stored.

Patient consent and post-test considerations

Before obtaining consent for molecular analyses, clinicians must provide patients with clear and comprehensive information, including the purpose of testing, its limitations, and its implications for treatment, eligibility for clinical trials, and familial risk.²¹

Genetic counselling should be offered when pathogenic variants of putative germline origin are identified in the tumour.²⁴ Patients should update their clinicians on their family cancer history, as this may affect risk assessment over time.¹⁸

Barriers and solutions

The routine and equitable implementation of HRR testing in mPCa is hindered by a lack of standardised guidelines, limited access to testing pathways, technical challenges, a shortage of molecular diagnostic specialists, strained genetic counselling services, and insufficient funding.

In Europe, the framework for access to genomic tests and biomarkers remains heterogeneous. In Spain, a recently published national catalogue of genetic and genomic assays aims to harmonise access to BRCA1/2 testing; however, regional and institutional disparities persist owing to variability in resources, training, and workflow integration.

Overcoming these barriers will require standardised protocols, professional education, multidisciplinary collaboration, and increased investment in technical and personnel resources to ensure equitable access to molecular testing in mPCa.

Conclusions

The expert panel recommends HRR testing in all patients with mPCa when results may influence clinical decision-making, including first-line treatment. Ideally, both tumour and germline testing should be performed at the diagnosis of metastatic disease, using the most recent, high-quality specimen available. Sample and test selection should be guided by multidisciplinary collaboration. Re-biopsy should be limited to necessary cases, with lymph node tissue preferred over bone. ctDNA represents a powerful minimally invasive alternative, particularly when tumour tissue is inadequate or unavailable and re-biopsy is not feasible. Validated DNA extraction protocols must be employed to ensure sample integrity and avoid cross-contamination. Hybrid capture-based NGS approaches are preferred over amplicon-based methods. Genomic data should be interpreted by experienced professionals within a multidisciplinary MTB, and only clinically relevant findings should be reported in a clear manner.

Finally, it is essential that laboratories performing HRR testing adhere to established quality standards, such as ISO 15189 or equivalent accreditation frameworks, and meet appropriate quality control requirements to ensure the reliability and accuracy of the entire diagnostic process.

The development and implementation of a harmonised algorithm for tumour and germline testing remain critical. At the health system level, the establishment of dedicated reference centres for HRR testing should be prioritised to minimise variability in testing quality and increase availability across regions, supporting more consistent and timely clinical decision-making.

Author contributions

Pilar González Peramato, conceptualisation, review and editing of successive drafts.

Eugenia García Fernádez, conceptualisation, review and editing of successive drafts.

Ana Jambrina Prieto, writing of the original draft, review and editing of successive drafts.

Javier Freire Salinas, conceptualisation, review and editing of successive drafts.

Javier Hernández Losa, conceptualisation, review and editing of successive drafts.

Estefanía Linares Espinós, conceptualisation, writing of the original draft, review and editing of successive drafts.

Federico Rojo, conceptualisation, review and editing of successive drafts.

Joaquin Mateo conceptualisation, review and editing of successive drafts.

David Olmos conceptualisation, writing, review and editing of successive drafts.

Funding

Editorial support for this manuscript was funded by Pfizer.

Ethical statements

This article represents an expert consensus and review of published literature and does not include any original study performed by the authors.

The NCCN makes no warranties of any kind regarding their content, use or application, and disclaims any responsibility for its application or use in any way.

Conflicts of interest

All authors were paid consultants to Pfizer in connection with the development of this manuscript.

Pilar Gonzalez-Peramato has served as an advisor top Pfizer, and received speaker fees from Bristo-Myers-Squibb and MSD.

Javier Freire Salinas has received advisory fees from ThermoFisher, Pfizer, Sophia Genetics.

Javier Hernández Losa has received speaker fees, lecture fees, and advisory fees from Pfizer, Roche, Amgen, Biocartis, Lilly, Diaceutics, AstraZeneca, Thermofisher, Owkin, Janssen, and grants from Lilly.

Federico Rojo has received speaker fees, lecture fees, and advisory fees from Roche, GSK, Novartis, MSD, BMS, Merck, Janssen, Lilly, Menarini, AstraZeneca, Astellas, Agilent, and Pfizer; and grants from Roche, Novartis, Menarini, AstraZeneca and Pfizer.

Joaquin Mateo has served as an advisor to AstraZeneca, Amunix/Sanofi, Daichii-Sankyo, Janssen, MSD; Pfizer and Roche. He is member of the scientific advisory board of Nuage Therapeutics and has participated as investigator in several industry-sponsored clinical trials. He is also the principal investigator of research grants funded by AstraZeneca, Amgen and Pfizer to VHIO (institution).

David Olmos has served as an advisor to AstraZeneca, Bayer, Daichii-Sankyo, Janssen/Johnson&Johnson, MSD, PharmaMar, Pfizer and Roche. He has received speaker fees from AstraZeneca, Bayer, Janssen/Johnson&Johnson, MSD, Novartis, PharmaMar, and Pfizer. Roche. He has received travel support to attend meetings from Astrazeneca, Bayer, Johnson&Johnson, and Pfizer. He is also the principal investigator of research grants funded by AstraZeneca, Bayer, Johnson&Johnson and Pfizer to Imas12 (institution).

Acknowledgements

We would like to thank Dr. Bernadette Pfang, on behalf of Springer Health+, for editorial support in the preparation of this manuscript, and Belén Sanz and Vicente Madero from Pfizer for medical writing support.