A new analysis from the phase III PROpel trial highlights the complementary role of circulating tumor DNA testing alongside tumor tissue testing for identifying homologous recombination repair gene mutations in metastatic castration-resistant prostate cancer.
The original report, titled “Detection of Homologous Recombination Repair Gene Mutations by Tumor Tissue and Circulating Tumor DNA Testing in Prostate Cancer in the Phase III PROpel Trial,” was published in JCO Precision Oncology on May 11, 2026. The study evaluated biomarker testing in patients enrolled in the phase III PROpel trial, focusing on homologous recombination repair gene mutations and BRCA1/BRCA2 mutations using both tumor tissue and plasma-derived circulating tumor DNA testing.
Authors: Andrew J. Armstrong, Fred Saad, Mototsugu Oya, Neal Shore, Giuseppe Procopio, Cagatay Arslan, Niven Mehra, Emma Brown, Jae Young Joung, Mikio Sugimoto, Mustafa Özgüroğlu, Craig Gedye, Oliver Sartor, Christian Poehlein, Ping Qiu, Yu-Zhen Liu, Xiaodun Li, Elizabeth A. Harrington, David McGuinness, Jinyu Kang, Alan Barnicle, and Noel Clarke.
Why This Analysis Matters
Molecular testing has become an important part of treatment decision-making in metastatic castration-resistant prostate cancer. Alterations in homologous recombination repair genes, particularly BRCA1 and BRCA2, may help identify patients for whom olaparib-containing treatment strategies are clinically relevant.
Securing usable tumor material for sequencing is notoriously difficult in this disease setting. A large share of men with advanced prostate cancer present with disease confined largely or entirely to bone, and the archival biopsies that are available are often too aged or too sparse to support reliable molecular workup. These constraints have driven growing interest in plasma-based ctDNA assays as a less invasive alternative source of genomic data.
The PROpel trial previously showed that olaparib plus abiraterone improved radiographic progression-free survival compared with placebo plus abiraterone in the first-line mCRPC setting. In this exploratory biomarker analysis, investigators assessed how tumor tissue testing and ctDNA testing performed in identifying HRR mutations and BRCA mutations among patients in PROpel.
You can read about Prostate Cancer Cure Rate on OncoDaily.
Methods
Patients in PROpel were randomly assigned to receive either olaparib 300 mg twice daily plus abiraterone 1,000 mg once daily or placebo plus abiraterone. Both groups also received prednisone or prednisolone 5 mg twice daily.
Patients were enrolled regardless of HRR mutation status. Tumor tissue samples, either archival or newly collected, and baseline blood samples were prospectively collected. Testing was performed for qualifying mutations in 14 HRR-related genes: BRCA1, BRCA2, ATM, BARD1, BRIP1, CDK12, CHEK1, CHEK2, FANCL, PALB2, RAD51B, RAD51C, RAD51D, and RAD54L.
Tumor tissue was analyzed using an assay based on FoundationOne CDx, while blood-derived ctDNA was analyzed using an assay based on FoundationOne Liquid CDx. The investigators reported biomarker results separately for each testing method and also as an aggregate analysis, where patients were classified using combined tumor tissue and ctDNA results.
Results
Among the 796 patients randomly assigned in PROpel, tumor tissue samples were available for 782 patients, and blood samples were available for 794 patients. Tumor tissue testing successfully assessed HRRm/BRCAm status in 535 of 782 patients, corresponding to 68%. In comparison, ctDNA-based testing successfully assessed biomarker status in 734 of 794 patients, corresponding to 92%.
By combining tumor tissue and ctDNA testing, aggregate HRR mutation status was obtained for 778 of 796 patients, or 98% of the randomized population. This aggregate approach identified 226 patients with HRR mutations, including 85 patients with BRCA1 and/or BRCA2 mutations.
In the aggregate population, the most frequently reported single HRR gene mutations were in BRCA2, ATM, and CDK12. Overall, HRR mutations were identified in 29.0% of patients using the aggregate approach, while BRCA mutations were identified in 10.9%.
The investigators also evaluated concordance between tumor tissue and ctDNA testing in 491 patients who had both results available. Using tumor tissue as the reference, the overall percent agreement was 85.1% for HRR mutation status and 93.9% for BRCA mutation status. The negative predictive value was 93.7% for HRR mutations and 97.3% for BRCA mutations.
The analysis suggested a low estimated rate of potential false-negative results among biomarker-negative subgroups when ctDNA-based testing was used without an available tissue result. Among aggregate non-BRCAm patients, approximately 6 of 226 patients classified as non-BRCAm using ctDNA only may have had an unidentified BRCA mutation, representing approximately 1% of the total aggregate non-BRCAm subgroup, or 6 of 693 patients.
You can also read about Top Prostate Cancer Trials to Watch at ASCO 2026 on OncoDaily.
ctDNA Fraction and Discordance
Although concordance between tumor tissue and ctDNA testing was generally high, some discordance was observed. One important factor was the ctDNA tumor fraction. When ctDNA fraction was low or not evaluable, ctDNA testing was less likely to detect certain alterations identified in tumor tissue.
This was particularly relevant for structural alterations such as homozygous deletions and large rearrangements, including some BRCA1/2 alterations. At the variant level, ctDNA testing showed high sensitivity for frameshift, nonsense, and splice alterations, while sensitivity was lower for homozygous deletions and large rearrangements.
The authors also noted that some mutations detected in ctDNA but not in tumor tissue, particularly involving genes such as ATM and CHEK2, may be related to clonal hematopoiesis of indeterminate potential. However, they emphasized that this interpretation remains speculative without paired testing of myeloid cells.
Conclusion
Overall, this biomarker analysis from PROpel indicates that pairing the two testing modalities substantially expands the proportion of mCRPC patients for whom a definitive HRR or BRCA status can be established. While tissue-based sequencing continues to serve as the diagnostic benchmark, blood-based ctDNA profiling offers a meaningful fallback in the common scenarios where archival samples are missing, degraded, or yield no usable result.
By layering both methods, investigators were able to classify nearly all randomized participants — 98% — with only a small fraction at risk of being miscategorized. The authors argue this combined workflow should be more widely adopted in routine mCRPC care, since it sharpens the ability to distinguish patients carrying actionable alterations from those who do not.
The data position liquid biopsy as an adjunct to, rather than a substitute for, conventional tissue analysis. When a tissue sample cannot be obtained or fails to yield interpretable results, ctDNA sequencing offers clinicians a practical second route to the molecular information needed for prognostic assessment and therapy selection.