Circulating tumor DNA (ctDNA) and circulating tumor cells (CTCs) are being investigated as potential prognostic biomarkers in metastatic castration-resistant prostate cancer (mCRPC). Beyond identifying genomic alterations that may inform treatment selection, liquid biopsy approaches offer opportunities to assess disease burden, monitor molecular changes during therapy, and better understand patient prognosis.
Although tumor tissue sequencing remains the standard approach for molecular profiling, obtaining adequate tissue can be challenging, particularly when archival samples are unavailable or no longer reflect the evolving genomic landscape of metastatic disease.
A new exploratory analysis of the phase III TALAPRO-2 trial examined the concordance between tissue and ctDNA testing, alongside the prognostic significance of baseline ctDNA burden and early changes in ctDNA and CTCs during treatment with talazoparib plus enzalutamide or enzalutamide alone.
The findings were published on October 28, 2026, in The Lancet Oncology, in an article titled: “Talazoparib plus enzalutamide versus placebo plus enzalutamide in metastatic androgen pathway modulator-resistant (castration-resistant) prostate cancer (TALAPRO-2): a biomarker analysis of circulating tumour DNA and cells for prognosis and molecular profiling.”
The study was authored by Steven M Yip, A Douglas Laird, Neeraj Agarwal, Stefanie Zschäbitz, Karim Fizazi, Nobuaki Matsubara, Jae Young Joung, Lawrence I Karsh, Joan Carles, André P Fay, Josep M Piulats, Peter C C Fong, Glenn Liu, Lee A Albacker, Lincoln W Pasquina, Jijumon Chelliserry, Cynthia G Healy, Xun Lin, Matko Kalac, and Arun A Azad.
Study Design and Biomarker Assessment
TALAPRO-2 was a randomized, double-blind, placebo-controlled phase III trial evaluating first-line talazoparib plus enzalutamide versus placebo plus enzalutamide in patients with mCRPC. The current analysis focused on the intention-to-treat cohort unselected for homologous recombination repair (HRR) gene alterations.
Between January 7, 2019, and September 17, 2020, 805 patients were randomized to receive talazoparib 0.5 mg plus enzalutamide 160 mg once daily (n=402) or placebo plus enzalutamide (n=403). Eligible patients had asymptomatic or mildly symptomatic mCRPC, an Eastern Cooperative Oncology Group performance status of 0–1, and no previous life-prolonging systemic treatment for castration-resistant disease.
Previous TALAPRO-2 analyses had demonstrated improved radiographic progression-free survival (rPFS; HR 0.63, 95% CI 0.51–0.78) and overall survival (HR 0.80, 95% CI 0.66–0.96) with the addition of talazoparib in the unselected cohort.
HRR alterations were assessed across 12 genes, including BRCA1, BRCA2, ATM, CDK12, and PALB2, using assays based on FoundationOne CDx for tumor tissue and FoundationOne Liquid CDx for plasma samples. Serial ctDNA tumor fraction was assessed at baseline and week 9 using FoundationOne Monitor, a research-use-only assay, while CTC enumeration using CELLSEARCH was performed at baseline and subsequent treatment timepoints. The analysis was exploratory, with a data cutoff of August 16, 2022.
Concordance Between ctDNA and Tumor Tissue Testing
One objective was to determine whether ctDNA testing could complement tissue-based molecular profiling in identifying HRR gene alterations. Retrospective ctDNA test results from screening samples were reported for 739 of 805 patients (92%). Among the 501 patients who could be classified as HRR-deficient or HRR-non-deficient by both prospective tissue testing and retrospective ctDNA testing, concordance was 84% (421 of 501).
In the smaller group with evaluable prospective tissue and ctDNA results, agreement was 81% (51 of 63). Concordance was higher in patients with a plasma tumor fraction of at least 1%. Agreement reached 89% (216 of 242) when tumor fraction was at least 1%, compared with 79% (205 of 259) when tumor fraction was below 1%. These findings support the complementary role of ctDNA testing, particularly when tissue-based molecular profiling is unsuccessful or insufficient.
The prevalence of individual gene alterations also differed between ctDNA and tumor tissue testing. BRCA2 alterations were detected in 6.7% of evaluable ctDNA samples versus 9.1% of tumor tissue samples, while CDK12 alterations were identified in 5.8% versus 8.5%, respectively.
In contrast, ATM alterations were more frequent in ctDNA than tissue (10.5% vs 6.0%), as were CHEK2 alterations (4.7% vs 1.9%). These differences may partly reflect clonal hematopoiesis, which can complicate the interpretation of plasma-based genomic findings. A particularly notable difference involved androgen receptor (AR) alterations, which were detected in 33.8% of ctDNA samples compared with 4.3% of tumor tissue samples.
The investigators suggested that this difference may reflect genomic changes associated with disease progression and previous treatment exposure, particularly because much of the available tumor tissue was archival.
Baseline ctDNA Burden and Radiographic Progression-Free Survival
Among the 687 patients evaluable for baseline ctDNA burden, detectable ctDNA was associated with substantially shorter rPFS in both treatment groups.
In the talazoparib plus enzalutamide group, median rPFS was 21.0 months in patients with detectable baseline ctDNA, compared with not reached in those with undetectable ctDNA (HR 2.46, 95% CI 1.68–3.60; p<0.0001). A similar association was observed in the placebo plus enzalutamide group, where median rPFS was 11.0 months versus 33.2 months, respectively (HR 4.03, 95% CI 2.85–5.72; p<0.0001).
Detectable baseline ctDNA was therefore associated with less favorable outcomes in both treatment groups. The results also highlight the distinction between ctDNA as a prognostic biomarker and its potential use for predicting benefit from a particular treatment.
Early ctDNA Conversion and Treatment Outcomes
The investigators also evaluated whether changes in ctDNA status after treatment initiation were associated with subsequent outcomes. At week 9, 283 patients receiving talazoparib plus enzalutamide and 296 receiving placebo plus enzalutamide were evaluable for ctDNA conversion. In both groups, patients whose ctDNA converted from detectable to undetectable had longer rPFS than those who remained ctDNA-positive.
In the talazoparib plus enzalutamide group, median rPFS was 25.9 months among patients who converted to ctDNA-negative, compared with 10.8 months among those who remained positive (HR 1.66, 95% CI 1.02–2.68; p=0.037, for persistent positivity versus conversion).
In the placebo plus enzalutamide group, median rPFS was 16.5 months among patients who converted from ctDNA-positive to ctDNA-negative versus 5.5 months among those who remained ctDNA-positive (HR 3.77, 95% CI 2.39–5.94; p<0.0001, for persistent positivity versus conversion). Patients who remained ctDNA-negative from baseline through week 9 had more favorable rPFS than those who converted from positive to negative.
Exploratory analyses also assessed whether ctDNA conversion provided additional information beyond prostate-specific antigen (PSA) responses. In the talazoparib group, ctDNA status improved prognostic stratification among patients who had not achieved a PSA50 or PSA90 response, whereas its additional utility was limited among those who had achieved these responses.
Circulating Tumor Cell Changes and Prognosis
Alongside ctDNA, post-hoc analyses evaluated whether early reductions in CTC counts at weeks 9 and 17 were associated with radiographic progression-free survival. CTC conversion was defined as a reduction from at least five CTCs per 7.5 mL of blood to four or fewer. CTC0 referred to a reduction from at least one detectable CTC to zero. At week 9, CTC conversion was associated with longer rPFS in both treatment groups. Among patients receiving talazoparib plus enzalutamide, median rPFS was 22.1 months in those achieving CTC conversion versus 5.5 months in those who did not (HR 0.13, 95% CI 0.06–0.32; p<0.0001).
In the placebo plus enzalutamide group, median rPFS was 16.6 months versus 3.6 months, respectively (HR 0.16, 95% CI 0.07–0.40; p<0.0001). Similarly, achieving CTC0 at week 9 was associated with more favorable rPFS. In the talazoparib group, median rPFS was not reached among patients achieving CTC0, compared with 8.4 months among those who did not (HR 0.33, 95% CI 0.19–0.57; p<0.0001).
In the placebo group, median rPFS was 19.4 months versus 10.0 months, respectively (HR 0.41, 95% CI 0.24–0.69; p=0.0006). The prognostic associations of CTC conversion and CTC0 were also observed at week 17. Early CTC reductions were therefore associated with longer rPFS in both treatment groups.
Clinical Implications and Limitations
The concordance between tissue and ctDNA testing supports the use of plasma-based genomic profiling as a complementary approach, particularly when adequate tumor tissue cannot be obtained or genomic testing is inconclusive. At the same time, baseline ctDNA burden, early ctDNA conversion, and reductions in CTCs were consistently associated with rPFS outcomes.
However, the biomarkers evaluated were not demonstrated to predict differential benefit from talazoparib plus enzalutamide. Their prognostic associations should therefore not be interpreted as evidence that they can independently guide PARP inhibitor selection or treatment modification.
Several limitations are important. This exploratory analysis was not powered to detect statistically significant differences in these biomarker comparisons, and all reported p-values were exploratory. Biomarker data were unavailable or uninformative for some patients, and ctDNA tumor fraction was assessed at only one on-treatment timepoint. The investigators also did not conduct a multivariable analysis. Independent validation will be necessary before these measures can be incorporated into routine clinical decision-making.
Conclusion
The TALAPRO-2 biomarker analysis highlights the potential prognostic value of baseline ctDNA burden and early ctDNA and CTC changes in mCRPC. Tissue and ctDNA testing showed substantial agreement for HRR gene alteration status, supporting their complementary roles in molecular profiling. Detectable ctDNA at baseline was associated with shorter rPFS, while conversion to undetectable ctDNA and early reductions in CTCs were associated with more favorable outcomes in both treatment groups.
Although these findings do not establish ctDNA or CTC changes as predictive biomarkers for talazoparib benefit, they support continued investigation of blood-based molecular monitoring as part of prognostic assessment in mCRPC.
