Received: 16 September 2022 | Accepted: 20 November 2022 DOI: 10.1002/pros.24469 ORIGINAL ARTICLE A correlative biomarker study and integrative prognostic model in chemotherapy‐naïve metastatic castration‐resistant prostate cancer treated with enzalutamide María P. Fernandez‐Perez PhD1 | Enrique Perez‐Navarro BSc2 | 3 4 Teresa Alonso‐Gordoa MD | Vicenza Conteduca MD | Albert Font MD5 | 6 7 Sergio Vázquez‐Estévez MD | Aránzazu González‐del‐Alba MD | Daniel Wetterskog PhD8 | Emmanuel S. Antonarakis MD9 | Begona Mellado MD, PhD10 | Ovidio Fernandez‐Calvo MD11 | María J. Méndez‐Vidal MD12 | Miguel A. Climent MD13 | Ignacio Duran MD, PhD14 | Enrique Gallardo MD15 | Angel Rodriguez Sanchez MD16 | Carmen Santander MD17 | Maria I. Sáez MD18 | Javier Puente MD, PhD19 | Julian Tudela MD20 | Alberto Martínez BSc21 | Maria J. López‐Andreo PhD22 | José Padilla BSc1 | Rebeca Lozano MD, PhD23,24 | David Hervas BSc25 | Jun Luo PhD26 | Ugo de Giorgi MD4 | Daniel Castellano MD2 | Gerhardt Attard MD, PhD8 | Enrique Grande MD27 | Enrique Gonzalez‐Billalabeitia MD, PhD1,2,28 1 Department of Haematology and Medical Oncology, Hospital Universitario Morales Meseguer, IMIB, Murcia, Spain 2 Department of Medical Oncology, Instituto de Investigación, Hospital Universitario 12 de Octubre, Madrid, Spain 3 Department of Medical Oncology, Hospital Ramón y Cajal, Madrid, Spain 4 Department of Medical Oncology, Istituto Scientifico Romagnolo per lo Studio e la Cura dei Tumori (IRST) "Dino Amadori" IRCCS, Meldola, Italy 5 Department of Medical Oncology, Catalan Institute of Oncology, Badalona Applied Research Group in Oncology (BARGO), Badalona, Spain 6 Department of Medical Oncology, H. Universitario Lucus Augusti, Lugo, Spain 7 Department of Medical Oncology, H.U. Son Espases, Palma de Mallorca, Spain 8 University College London Cancer Institute, Paul O'Gorman Building, London, UK 9 Department of Oncology, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA 10 Department of Medical Oncology, IDIBAPS, Hospital Clinic, Universidad de Barcelona, Barcelona, Spain 11 Department of Medical Oncology, Complejo Hospitalario Universitario Ourense, Orense, Spain 12 Department of Medical Oncology, Hospital Universitario Reina Sofía (HURS), Maimonides Institute for biomedical research of Córdoba (IMIBIC), Córdoba, Spain 13 Servicio de Oncología Médica, Instituto Valenciano de Oncología, Valencia, Spain 14 Instituto de Biomedicina de Sevilla, IBiS/Hospital Universitario Virgen del Rocío/CSIC/Universidad de Sevilla, Seville, Spain 15 Department of Medical Oncology, Servicio de Oncología Médica, Parc Taulí Hospital Universitari, Institut d'Investigació i Innovació Parc Taulí I3PT, Universitat Autònoma de Barcelona, Sabadell, Spain 16 Department of Medical Oncology, Hospital Universitario de León, León, Spain 17 Department of Medical Oncology, Hospital Universitario Miguel Servet, Zaragoza, Spain 18 Medical Oncology Intercenter Unit, Regional and Virgen de la Victoria University Hospitals, IBIMA, Málaga, Spain This is an open access article under the terms of the Creative Commons Attribution‐NonCommercial‐NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non‐commercial and no modifications or adaptations are made. © 2022 The Authors. The Prostate published by Wiley Periodicals LLC. 376 | wileyonlinelibrary.com/journal/pros The Prostate. 2023;83:376–384. FERNANDEZ‐PEREZ ET AL. | 377 19 Department of Medical Oncology, Hospital Clínico San Carlos, Instituto de Investigación Sanitaria del Hospital Clínico San Carlos (IdISSC), CIBERONC, Madrid, Spain 20 Department of Pathology, Hospital Morales Meseguer, Murcia, Spain 21 Biobanco de la región de Murcia, IMIB, Nodo 3, Murcia, Spain 22 Department of Molecular Biology, SAI‐IMIB‐Universidad de Murcia, Murcia, Spain 23 Prostate Cancer Clinical Research Unit, Spanish National Cancer Research Centre, Madrid, Spain 24 Genitourinary Translational Research Group, Instituto de Investigación Biomédica de Málaga, Málaga, Spain 25 Data Science Unit, Instituto de Investigación Sanitaria La Fe, Valencia, Spain 26 Department of Urology, Johns Hopkins University School of Medicine, Baltimore, Maryland, USA 27 MD Anderson Cancer Center Madrid, Madrid, Spain 28 Universidad Católica San Antonio de Murcia‐UCAM, Murcia, Spain Correspondence Enrique Gonzalez‐Billalabeitia, MD, PhD, Abstract Department of Medical Oncology, Hospital Background: There is a considerable need to incorporate biomarkers of resistance to Universitario “12 de Octubre”, Madrid 28041, Spain. new antiandrogen agents in the management of castration‐resistant prostate Email: egbillalabeitia@salud.madrid.org cancer (CRPC). Funding information Methods: We conducted a phase II trial of enzalutamide in first‐line chemo‐naïve Astellas; Instituto de Salud Carlos III, asymptomatic or minimally symptomatic mCRPC and analyzed the prognostic value of Grant/Award Number: PI18/00883; SEOM‐CRIS Cancer Foundation TMPRSS2‐ERG and other biomarkers, including circulating tumor cells (CTCs), androgen receptor splice variant (AR‐V7) in CTCs and plasma Androgen Receptor copy number gain (AR‐gain). These biomarkers were correlated with treatment response and survival outcomes and developed a clinical–molecular prognostic model using penalized cox‐ proportional hazard model. This model was validated in an independent cohort. Results: Ninety‐eight patients were included. TMPRSS2‐ERG fusion gene was detected in 32 patients with no differences observed in efficacy outcomes. CTC detection was associated with worse outcome and AR‐V7 in CTCs was associated with increased rate of progression as best response. Plasma AR gain was strongly associated with an adverse outcome, with worse median prostate specific antigen (PSA)‐PFS (4.2 vs. 14.7 m; p < 0.0001), rad‐PFS (4.5 vs. 27.6 m; p < 0.0001), and OS (12.7 vs. 38.1 m; p < 0.0001). The clinical prognostic model developed in PREVAIL was validated (C‐Index 0.70) and the addition of plasma AR (C‐Index 0.79; p < 0.001) increased its prognostic ability. We generated a parsimonious model including alkaline phosphatase (ALP); PSA and AR gain (C‐index 0.78) that was validated in an independent cohort. Conclusions: TMPRSS2‐ERG detection did not correlate with differential activity of enzalutamide in first‐line mCRPC. However, we observed that CTCs and plasma AR gain were the most relevant biomarkers. KEYWORDS AR gain, AR‐V7, CTCs, enzalutamide, prostate cancer, TMPRSS2‐ERG 1 | INTRODUCTION of treatment in advanced patients.2 Virtually all patients progress to a castration‐resistant state where androgen signaling is potentially Prostate cancer (PCa) is the second cause of death from cancer in driving tumor progression.3 Enzalutamide is a potent AR‐targeted 1 males. Its growth is dependent upon androgen receptor (AR) agent that competitively binds the AR ligand‐binding domain signaling, and androgen deprivation therapy (ADT) is the mainstay and inhibits AR signaling.4 Treatment with enzalutamide has 378 | FERNANDEZ‐PEREZ ET AL. improved overall survival and quality of life in PCa.5 However benefit (ECOG) 0–1, and were asymptomatic or mildly symptomatic (Brief from this treatment is variable and several clinical and molecular Pain Inventory Short Form question 3 of less than 4). events have been proposed to explain this heterogeneity.6–8 The IRST cohort consisted of patients with histologically Fusion genes involving E26 transformation‐specific (ETS) onco- confirmed prostate adenocarcinoma without neuroendocrine differ- genes are the most common driver events affecting 30%–70% of entiation, progressive disease despite “castration levels” of serum PCa.9–11 The most frequent fusion gene involves transmembrane testosterone (<50 ng/dl), ongoing LHRH analog treatment or prior protease serine 2 (TMPRSS2) on 21q22.3 and v‐ets erythroblastosis surgical castration, and no prior treatment with enzalutamide or virus E26 oncogene homolog (ERG) on 21q22.2, either by intrachro- abiraterone. mosomal deletion or translocation, and as a result of the oncogene Treatment in both cohorts consisted of enzalutamide at a dose of ERG becomes AR‐regulated. TMPRSS2‐ERG is involved in tumor 160 mg once daily. Treatment continued until the occurrence of initiation,12 invasion,13 and progression,14 and has been associated unacceptable side effects or confirmed radiographic progression. with increased efficacy in abiraterone‐treated patients.15,16 Other Written informed consent was obtained from all patients. biomarkers, including plasma AR gain,17–19 circulating tumor cells (CTCs),20,21 and AR‐V7 in CTCs22,23 have demonstrated prognostic and/or predictive value in pretreated metastatic castration‐resistant 2.3 | Procedures prostate cancer (mCPRC). However, the prognostic value of these biomarkers in first‐line mCPRC needs to be further explored. In the PREMIERE trial, tumor tissue and blood samples were shipped Clinical prognostic models in mCRPC, including a prognostic before study entry and a central pathologist reviewed all histological model in patients treated with enzalutamide in first‐line mCRPC24 are FFPE samples. Blood samples were collected before study entry, at intended to deal with this heterogeneity. This latter model needs to 12 weeks and at progression, and included three 5 ml EDTA tubes: be externally validated and can potentially be improved including two were locally processed to obtain plasma that was stored at molecular information. −80°C and centrally shipped at study completion and one was In this phase 2 multicenter biomarker study with enzalutamide in shipped overnight at 4°C to a central laboratory and processed in less first‐line chemo‐naïve mCRPC, we aimed to evaluate the clinical than 24 h for CTC analyses. In the IRST cohort, plasma samples were significance of TMPRSS2‐ERG fusion gene, and to explore other collected before treatment initiation and analyzed at a central relevant biomarkers, including plasma AR, CTCs, and AR‐V7 in CTCs, laboratory. and its contribution to the available clinical prognostic models. TMPRSS2‐ERG fusion gene was studied in FFPE tumor tissue from the PREMIERE trial. In brief, an 8‐μm FFPE slide was macrodissected and DNA/RNA was extracted and quantitative PCR (qPCR) was 2 | MATERIALS AND METHODS performed in duplicate with the TaqMan primer probe for TMPRS2‐ ERG fusion transcript (Hs03063375 ft, Applied Biosystems) following 2.1 | Study design and conduct manufacturer's instructions. ERG immunohistochemistry was per- formed using ERG rabbit monoclonal antibody (EPR3864, Epitomics; The PREMIERE trial is a translational multicenter single‐arm open‐ dilution 1:100) and correlated with qPCR results. Fluorescence in situ label phase 2 clinical trial (NCT02288936) of enzalutamide in first‐ hybridization (FISH) for TMPRSS2‐ERG fusion gene was indepen- line mCRPC, originally designed to analyze the prognostic value of dently performed as previously described25 at two different institu- the gene fusion TMPRSS2‐ERG and other correlative laboratory tions (HMM and CNIO). studies, including plasma DNA and CTC analysis. The study was approved by a central independent review board (IRB). – Plasma AR copy number. Plasma was collected within 30 days The validation cohort consisted of a single institution cohort before treatment initiation and plasma aliquots were stored at from patients participating in a protocol approved by the Istituto −80°C and centrally analyzed upon study completion. ddPCR Scientifico Romagnolo per lo Studio e la Cura dei Tumori (IRST), assays were carried out as previously described.18 In brief, for Meldola, Italy (REC 2192/2013) with plasma samples collected each individual sample, AR CN was estimated using each of the prospectively with the primary aim of biomarker evaluation. reference genes NSUN3, ElF2C1, and AP3B1 and using ZXDB at Xp11.21 as a control gene to determine X chromosome CN. An AR gain cutoff of ≥1.92 was considered as AR gain, as previously 2.2 | Participants published.19 – CTC and AR‐V7 analysis. The CTC analyses were conducted using The PREMIERE trial consisted of patients with histologically the commercially available AdnaTest platform (Qiagen) following confirmed adenocarcinoma of the prostate with documented the manufacturer's instructions with a minor modification metastases and tumor progression and a serum testosterone level previously described.22 Custom primers, as previously described, of 50 ng per deciliter or less with continued androgen‐deprivation were used to detect AR‐V7 mRNA. Sanger sequencing confirmed therapy. Eligible patients had Eastern Cooperative Oncology Group the accuracy of the PCR product. AR‐V7 cross‐validation was FERNANDEZ‐PEREZ ET AL. | 379 performed between PREMIERE central laboratory (CL) and Johns TABLE 1 Patients' characteristics Hopkins University (JH). Patient characteristics All (n = 98) Age median (range) 77 (57–95) | Outcomes Metastases, no. (%) 2.4 Bone 80 (82%) The primary endpoint for the PREMIERE clinical trial was prostate Visceral 16 (16%) specific antigen (PSA) progression‐free survival (PSA‐PFS) and Liver 4 (4%) secondary endpoints included PSA response, radiographic progression‐ Lymph nodes 47 (48%) free survival (rad‐PFS), and overall survival (OS). PSA and blood tests were assessed monthly and radiographic disease was evaluated with Bone metastasis, no. (%) the use of computed tomography (CT) and bone scan at the time of <4 50 (51%) screening and every 12 weeks thereafter. Response was evaluated ≥4 30 (31%) according to the Prostate Cancer Clinical Trials Working Group 2 ECOG, no. (%) (PCWG2) criteria26 and soft tissue disease was assessed on the basis of Response Evaluation Criteria in Solid Tumors (RECIST), version 0 53 (54%) 27 1.1. CTC conversion was defined considering the analysis at basal 1 45 (46%) and after 12 weeks of treatment. Pain, N (%) No pain 45 (46%) 2.5 | Statistical analyses Mild (≤3 BPI score) 52 (54%) PSA, ng/dl, median (range) 24.95 (0.59–4319) Statistical analyses were performed using the R software, version Albumin (µg/dl) 4.16 (3.29–5.00) 3.3.0. The statistical plan for biomarkers included descriptive and Hemoglobin median (range) 13.20 (7.50–17.30) prognostic analyses of TMPRSS2‐ERG, CTC, AR‐V7 in CTCs, and AR gain using the primary and secondary outcomes of the trial, as ALP ratio median (range) 0.71 (0.24–17.46) previously described. Post hoc exploratory analyses included High (>UNL) 28 (29%) correlation with primary progression. Qualitative variables were LDH ratio median (range) 0.84 (0.29–3.36) compared using the Fisher exact test. Time variables were evaluated High (>UNL) 31 (32%) using Kaplan‐Meier analysis and cox‐proportional hazards models. All tests were two‐sided, and an alpha‐error of less than 0.05 was NLR median (range) 2.14 (0.52–12.33) required to be considered statistically significant. Cohen's Kappa test CTC positive, no. (%) 35 (36%) was used to study concordance between the central and the external Plasma DNA,ng/ml, median (range) 18.64 (0.05–1585) laboratory. A Cox‐proportional hazards model via penalized maxi- AR gain, no. (%) 11 (11%) mum likelihood, using the package glmnet 2.0‐18,28 was used to generate a parsimonious clinical–molecular model. Abbreviations: ALP, alkaline phosphatase; AR, androgen receptor; BPI, Brief Pain Inventory; CTC, circulating tumor cell; ECOG, Eastern Cooperative Oncology Group; LDH, lactate dehydrogenase; NLR, neutrophil to lymphocyte ratio. 3 | RESULTS Ninety‐eight chemotherapy‐naïve mCRPC patients initiated enzalu- Enzalutamide was well tolerated with no unexpected toxicities, as tamide as first‐line treatment between February and November 2015 shown in Supporting Information: Table S4. at 16 Spanish institutions in the PREMIERE trial. Patients' character- istics are described in Table 1. Survival outcomes for all patients included in the study are shown in Supporting Information: Figure S1. 3.1 | TMPRSS2‐ERG status With a median follow‐up of 37 months, median PSA‐PFS was 14.1 months (95% confidence interval [CI]: 10.2–20.2), median rad‐PFS TMPRSS2‐ERG fusion gene was detected in 32 patients (33%). All was 25.2 months (95% CI: 21.7–32.1) and median OS was 37.5 positive samples expressed high ERG in the nucleus by IHC (r = 0.93; months (95% CI: 33.7 vs. NR). Treatment responses are described in p < 0.0001). Baseline patients' characteristics were similar between both Supporting Information: Table S2. PSA‐50 (decrease ≥50%) was groups and no differences were observed in any efficacy outcome observed in 82%, and PSA‐90 (decrease ≥90%) in 53%, with based on the detection of TMPRSS2‐ERG. Further details are described radiographic response observed in 49% of the patients (N = 21). in Supporting Information: Tables S1–S3 and Figures S2 and S3. 380 | FERNANDEZ‐PEREZ ET AL. 3.2 | Plasma AR gain (HR 4.68, 95% CI 2.64–8.28; p < 0.0001). CTC conversion analysis using landmark survival analysis is shown in Figure 2D. Conversion of Baseline plasma AR gain was detected in 11 patients (11%). Plasma CTC detection from positive to negative was associated with AR‐based survival analyses are shown in Figure 1A–C. AR gain was improved evolution in all survival outcomes, including PSA‐PFS associated with worse survival outcomes than AR normal: median (p < 0.001), rad‐PFS (p < 0.001), and OS (p < 0.001). No association PSA‐PFS 4.2 versus 14.7 months (hazard ratio [HR] 4.03, 95% was observed between TMPRSS2‐ERG expression and CTC detection confidence interval [CI] 1.87–8.72; p < 0.001), median rad‐PFS was (p = 0.825), as described in Supporting Information: Table S5. 4.5 versus 27.6 months (HR 9.83, 95% CI 4.50–21.44; p < 0.0001) and median OS of 12.7 versus 38.1 months (HR 6.65, 95% CI 3.18–13.91; p < 0.0001), respectively. AR gain was also associated 3.4 | AR‐V7 expression in CTCs with worse PSA response, as shown in Figure 1D. AR gain was detected in 19.4% (6/31) versus 7.6% (5/66) in AR‐V7 was detected in 16% of CTC‐positive evaluable patients TMPRSS2‐ERG positive and negative, respectively (p = 0.166), as (5/32). Progression as best response at 12 weeks was observed in shown in Supporting Information: Table S5. 60% (3/5) in AR‐V7 positive patients, compared with 30% (8/27) in AR‐V7 negative and 8% (5/63) in CTC negative patients (p = 0.0014). However, two externally validated AR‐V7 patients had tumor 3.3 | CTCs by Adna‐Test® responses lasting for 14 and 25 months, as shown in Supporting Information: Figure S4A. AR‐V7‐based survival analyses are shown in Baseline CTCs were present in 35 patients (36%). Survival outcomes Figure 2B–D. No differences were observed for survival outcomes. by baseline CTC status are described in Figure 2A–C. The presence of Intriguingly, at baseline all AR‐V7 positive patients were TMPRSS2‐ CTC at baseline was associated with worse evolution in all survival ERG negative (5/21 vs. 0/11, p = 0.134), as shown in Supporting outcomes. Median PSA‐PFS for positive patients was 7.4 versus 20.2 Information: Table S5. A numerical increase, not statistically signifi- months (HR 3.37, 95% CI 2.01–5.66; p < 0.0001), median rad‐PFS cant, was observed for AR gain in AR‐V7 positive when compared was 11.5 versus 33.1 months (HR 5.21, 95% CI 2.82–9.64; with AR‐V7 negative and CTC negative patients: 40% (2/5), 16% p < 0.0001) and median OS was 25.4 months versus not reached (4/24), and 8% (5/66), respectively. F I G U R E 1 Survival outcomes and response by AR gain in plasma DNA. Kaplan–Meier and Cox‐regression models were used for survival analyses: (A) PSA progression‐free survival; (B) radiographic progression‐free survival; (C) overall survival; and (D) waterfall plot representing maximum PSA decrease by plasma AR status. AR, androgen receptor. [Color figure can be viewed at wileyonlinelibrary.com] FERNANDEZ‐PEREZ ET AL. | 381 F I G U R E 2 Survival outcomes by CTC and CTC conversion. Kaplan–Meier and Cox‐regression models were used for survival analyses: (A, D) PSA progression‐free survival in CTCs and by CTC conversion; (B, E) radiographic progression‐free survival in CTCs and by CTC conversion; and (C, F) overall survival in CTCs and by CTC conversion. CTC, circulating tumor cell. [Color figure can be viewed at wileyonlinelibrary.com] 3.5 | Development of an integrated of patients treated at IRST with enzalutamide obtaining a C‐Index of clinical–molecular model 0.71. We then designed a nomogram capable to predict the survival probability at 12, 24, and 36 months, as shown in Figure 3. We evaluated in our series the clinical prognostic model previously developed based on the PREVAIL trial externally validating it in our series, with a Concordance Index (C‐Index) of 0.70. Interestingly, 4 | D IS CU SS IO N most patients in our study were included in the low (61%) and intermediate (37%) risk groups, as shown in Supporting Information: We here show the final results of a biomarker phase 2 multicentre Figure S5. clinical trial of enzalutamide in first‐line chemo‐naïve mCRPC. The A multivariate analysis of all clinical variables included in the primary aim of the study was to evaluate the association between clinical model, CTC, and AR is shown in Table 2. We then integrated TMPRSS2‐ERG and the efficacy of enzalutamide and we did not find AR gain with the clinical variables included in the clinical model into a any difference. However, we observed that CTCs by AdnaTest® and new comprehensive clinical and molecular model that obtained a AR gain are strong and independent prognostic variables. In addition, bootstrap‐validated C‐Index of 0.79. The addition of AR to the clinical we externally validated a clinical prognostic model developed in the model increased the prognostic ability of the clinical model PREVAIL trial and observed that both CTCs and AR gain are able to (p < 0.001). Using a penalized regression model, we obtained a improve the prognostic ability of the model. We then developed a clinical–molecular parsimonious model that included three variables: parsimonious model including clinical variables and AR gain, that ALP ratio, PSA, and AR gain Table 3, and a bootstrap‐validated validated an independent data set of mCRPC patients treated with C‐Index of 0.78. This model was validated in an independent cohort enzalutamide in the same clinical scenario. 382 | FERNANDEZ‐PEREZ ET AL. TABLE 2 Multivariable analysis for the clinical and molecular Previous results regarding the predictive value of TMPRSS2‐ERG to variables new antiandrogens were controversial with abiraterone, with an early Premiere phase I/II trial of abiraterone that reported increased PSA responses in Prognostic HR (95% CI) p* TMPRSS2‐ERG associated tumors.15,29 Our results with enzalutamide‐ treated patients, together with previous studies with abiraterone, Albumin 0.57 (0.21–1.57) 0.276 support that TMPRSS2‐ERG fusion gene has limited value as ALP 0.44 (0.23–0.87) 0.018 predictive biomarker in mCRPC treated with anti‐androgen therapies. Number of bone metastases 0.94 (0.82–1.08) 0.355 We also evaluated plasma AR‐gain and other promising Hemoglobin 1.00 (0.79–1.27) 0.985 biomarkers including CTCs measured by AdnaTest® and AR‐V7 expression in CTCs. We previously published the association of LDH 0.66 (0.35–1.26) 0.207 plasma AR‐gain with an adverse outcome to enzalutamide with a NLR 0.77 (0.36–1.64) 0.499 median follow‐up of 11 months. Here we present the final updated Pain score 2.13 (1.12–4.05) 0.021 survival, with a median follow‐up of 37 months that confirms its Pattern of spread 0.63 (0.14–2.83) 0.543 strong independent prognostic value in first‐line mCRPC. AR‐gain prognostic value was independent of other clinical and molecular Loge PSA 1.39 (1.10–1.74) 0.005 prognostic variables. We observed a numerical increase in AR gain in Time from diagnosis to randomization 1.00 (0.99–1.01) 0.863 AR‐V7 positive patients compared with AR‐V7 negative and CTC AR gain 8.25 (3.17–21.45) <0.001 negative (40% vs. 17% vs. 8%). This result, although limited by the CTC 2.74 (1.34–5.61) 0.006 low numbers, is in agreement with previous publications in PCa metastatic tissue, that observe an association between AR gain and Abbreviations: ALP, alkaline phosphatase; AR, androgen receptor; BPI, AR‐V7 expression in PCa metastases.30 Brief Pain Inventory; CI, confidence interval; CTC, circulating tumor cell; ECOG, Eastern Cooperative Oncology Group; HR, hazard ratio; LDH, We show that CTCs can be detected using AdnaTest® in almost lactate dehydrogenase; NLR, neutrophil to lymphocyte ratio. one third (35%) of low‐intermediate risk mCRPC patients, and that it *p Value was calculated using Cox regression. is a strong and independent prognostic biomarker. We also observe, using landmark survival analysis, that CTC conversion at 12 weeks of treatment is associated with treatment outcome, including a TABLE 3 Clinical–molecular parsimonious model favorable outcome when it turns negative and an adverse outcome Coefficient p* when it becomes positive. However, prospective randomized clinical ALP ratio (log) 0.73 <0.001 trials are needed to fulfill the Prentice criteria31 and fully qualify as a meaningful biomarker for regulatory matters. PSA ng/dl (log) 0.36 <0.001 AR‐V7 in CTCs has previously demonstrated clinical value in AR gain (yes vs. no) 2.15 <0.001 mCRPC patients treated with a new anti‐androgen therapy.22,23 In this *p Value was calculated using Cox regression. first‐line study, we observed a low detection rate for AR‐V7 (5% of all F I G U R E 3 Nomogram for the clinical– molecular parsimonious model. Nomogram developed using the integrated parsimonious three variables prognostic model. ALP ratio and PSA were included in logarithmic scale. This nomogram predicts survival probability at 12, 24, and 36 months. ALP, alkaline phosphatase. FERNANDEZ‐PEREZ ET AL. | 383 patients; N = 5/98) and low PSA50 response rate (40% vs. 82%). These participated in advisory boards for Astellas, Janssen, Sanofi, and Bayer. results are consistent with a recently published phase 3 trial32 in this EG has received travel grants from Astellas, Janssen, Sanofi, Bayer, same scenario, with a detection rate of 7%–10% (N = 38/953) and a Roche, BMS, and Eisai, honoraria for speaking engagements for Astellas, response rate of 42% in 19 patients treated with enzalutamide. Janssen, Sanofi, Bayer, Pfizer, Roche, BMS, Novartis, Rovi, Daiichi Investigators from the pivotal trial of enzalutamide in first‐line Sankyo, Leo Pharma, Menarini, Eisai, MSD, Boehringer Ingelheim, mCRPC published a clinical prognostic model able to stratify patients Merck, and EUSA Pharma and participated in advisory boards for based on pretreatment clinical variables. This model was limited by the Astellas, Janssen, Sanofi, Bayer, Pfizer, Roche, Novartis, Eisai, EUSA lack of external validation and the absence of molecular variables. We Pharma, BMS, AstraZeneca, Merck, Rovi, Daiichi Sankyo, and Techdow. here validated this clinical prognostic model, obtaining a C‐index of 0.70 JP has received research funding from Astellas and Roche and received and observed that both CTCs and AR‐gain were independent variables honoraria for speaker engagements, advisory roles, or continuous that were able to improve the prognostic ability of the model. We then medical education from Astellas, Astra Zeneca, Janssen, MSD, Bayer, built a parsimonious clinical–molecular model, composed of three Pfizer, Eisai, Ipsen, Sanofi, Roche, BMS, and Merck. BM has served variables: AR gain, ALP, and PSA that was independently validated. LDH advisory role from Roche, Sanofi, Janssen, Astellas, Pfizer, Novartis, has been previously selected in multiple prognostic models in mCRPC, Bristol‐Myers Squibb, and Ipsen, research funding from Roche, Bayer, including models that include CTCs.33,34 Intriguingly, LDH was not and Janssen, and accommodation expenses from Pfizer and Janssen. significant in our penalized regression model. The association of LDH MA Climent has received travel grants from Astellas, Jansen, Sanofi, with high‐risk features, that were not represented in our model, might Pfizer, Roche, and Ipsen. And honoraria for speaker engagements, explain the lack of statistical significance in our study. advisory boards, and continuous medical education from Janssen, Astellas, Sanofi, Bayer, Roche, Ipsen, BMS, MSD, Pfizer, and Astra Zeneca MJMV has received travel grants from Astellas and Janssen and 5 | C ONC LUS I ON S has participated in advisory boards for Janssen, Astellas, Sanofi, and Bayer. María Isabel Sáez Medina has received travel grants from Sanofi In this phase 2 biomarker trial, we report additional evidence with long and Roche and participated in advisory boards from Sanofi, Ipsen, and follow‐up on the prognostic impact of AR gain together with other Astellas. DC has received research funding from Astellas, educational biomarkers, including CTCs and AR‐V7 chemo‐naïve mCRPC patients. grants from Pfizer and Janssen, travel grants from Pfizer and BMS, and We demonstrate its ability to improve clinical prognostic models and has participated in advisory boards for Pfizer, Astellas, Janssen, Sanofi, propose a new parsimonious model including plasma AR gain. Further Bayer, and Roche. JL is an inventor of AR‐V7‐related technologies that studies, including comprehensive biomarker panels and well‐annotated have been licensed to A&G Pharmaceuticals and Qiagen. GA reported clinical data sets, are required to redefine prognostic models in PCa. receiving grants, personal fees, nonfinancial support, and speaker fees from Janssen, Astellas, and Sanofi; personal fees, nonfinancial support, A C KN O W L E D G M E N T S and speaker fees from AstraZeneca; and personal fees from Novartis The authors would like to acknowledge all the staff at SOGUG for and Bayer outside the submitted work; in addition, GA reported having their support to run the PREMIERE trial, Astellas for supporting this a patent (GB1915469.9; blood signatures for prostate cancer detection) research, ISCIII from the Spanish Ministry of Health, and Cris Cancer pending and is on the Institute of Cancer Research list of rewards to Foundation for their support. This trial was promoted by SOGUG and inventors for abiraterone acetate. EG Research funding from Astellas, received a grant from Astellas. EGB received a travel grant (BA16/ Astra Zeneca, IPSEN, Pfizer, and Roche and has participated in advisory 00038) and funding support from the “Instituto de Salud Carlos III” boards for Adacap, AMGEN, Angelini, Astellas, Astra Zeneca, Bayer, (PI18/00883) and support from SEOM‐CRIS Cancer Foundation. Blueprint, Bristol Myers Squibb, Caris Life Sciences, Celgene, Clovis‐ Oncology, Eisai, Eusa Pharma, Genetracer, Guardant Health, HRA‐ CO NFL I CTS OF I NTEREST Pharma, IPSEN, ITM‐Radiopharma, Janssen, Lexicon, Lilly, Merck KGaA, TG has received travel grants from Sanofi and participated in advisory MSD, Nanostring Technologies, Natera, Novartis, ONCODNA (Biose- boards for Sanofi, Janssen, Astellas, and Bayer. AF has received travel quence), Palex, Pharmamar, Pierre Fabre, Pfizer, Roche, Sanofi‐ grants from Astellas, Astra‐Zeneca, Sanofi and participated in advisory Genzyme, Servier, Taiho, Thermo Fisher Scientific. EGB has received boards for Janssen, Sanofi, Eisai, and Bayer. SVE has received travel travel grants from Astellas, Janssen, and Sanofi and has participated in grants from Astellas, Janssen, and Bayer, and has participated in advisory boards for Astellas, Janssen, Sanofi, Astra‐Zeneca, and Bayer. advisory boards for Janssen, Astellas, Sanofi, and Bayer. AGdA has received research funding from Astellas, travel grants from Astellas, DATA AVAILABILITY STATEMENT Jansen, Sanofi, BMS, Roche, Pfizer, and Ipsen, and honoraria for speaker The authors confirm that the data supporting the findings of this engagements, advisory boards, and continuous medical education from study are available within the article and its Supporting Information. Janssen, Astellas, Sanofi, Bayer, Roche, Ipsen, BMS, MSD, Pfizer, Eusa Pharma, Eisai, and Astra Zeneca. ESA is the codeveloper of a patented ORC I D AR‐V7 biomarker technology that has been licensed to Qiagen. ID has Enrique Perez‐Navarro http://orcid.org/0000-0003-1091-8926 received travel grants and research funding from Astellas and has Teresa Alonso‐Gordoa http://orcid.org/0000-0002-8966-7236 384 | FERNANDEZ‐PEREZ ET AL. Vicenza Conteduca http://orcid.org/0000-0002-6921-714X 21. Scher HI, Heller G, Molina A, et al. Circulating tumor cell biomarker panel Enrique Gonzalez‐Billalabeitia http://orcid.org/0000-0003- as an individual‐level surrogate for survival in metastatic castration‐ resistant prostate cancer. J Clin Oncol. 2015;33(12):1348‐1355. 3143-3143 22. Antonarakis ES, Lu C, Wang H, et al. AR‐V7 and resistance to enzalutamide and abiraterone in prostate cancer. N Engl J Med. REFERENCES 2014;371(11):1028‐1038. 1. Siegel RL, Miller KD, Fuchs HE, Jemal A. Cancer statistics, 2021. CA 23. Scher HI, Graf RP, Schreiber NA, et al. Assessment of the validity of Cancer J Clin. 2021;71(1):7‐33. doi:10.3322/caac.21654 nuclear‐localized androgen receptor splice variant 7 in circulating 2. Attard G, Parker C, Eeles RA, et al. Prostate cancer. Lancet. tumor cells as a predictive biomarker for castration‐resistant 2016;387(10013):70‐82. prostate cancer. JAMA Oncol. 2018;4(9):1179‐1186. 3. Beer TM, Armstrong AJ, Rathkopf DE, et al. Enzalutamide in 24. Armstrong AJ, Lin P, Higano CS, et al. Development and validation of metastatic prostate cancer before chemotherapy. N Engl J Med. a prognostic model for overall survival in chemotherapy‐naïve men 2014;371(5):424‐433. with metastatic castration‐resistant prostate cancer. Ann Oncol. 4. Tran C, Ouk S, Clegg NJ, et al. Development of a second‐generation 2018;29(11):2200‐2207. doi:10.1093/ANNONC/MDY406 antiandrogen for treatment of advanced prostate cancer. Science. 25. Mehra R, Tomlins SA, Shen R, et al. Comprehensive assessment of 2009;324(5928):787‐790. TMPRSS2 and ETS family gene aberrations in clinically localized 5. Beer TM, Armstrong AJ, Rathkopf D, et al. Enzalutamide in men with prostate cancer. Mod Pathol. 2007;20(5):538‐544. doi:10.1038/ chemotherapy‐naïve metastatic castration‐resistant prostate cancer: MODPATHOL.3800769 extended analysis of the phase 3 PREVAIL study. Eur Urol. 26. Scher HI, Halabi S, Tannock I, et al. Design and end points of clinical 2017;71(2):151‐154. trials for patients with progressive prostate cancer and castrate 6. Armenia J, Wankowicz SAM, Liu D, et al. The long tail of oncogenic levels of testosterone: recommendations of the Prostate Cancer drivers in prostate cancer. Nature Genet. 2018;50(5):645‐651. Clinical Trials Working Group. J Clin Oncol. 2008;26(7):1148‐1159. 7. Quigley DA, Dang HX, Zhao SG, et al. Genomic hallmarks and structural 27. Eisenhauer EA, Therasse P, Bogaerts J, et al. New response variation in metastatic prostate cancer. Cell. 2018;174(3):758‐769. evaluation criteria in solid tumours: revised RECIST guideline 8. Abida W, Cyrta J, Heller G, et al. Genomic correlates of clinical (version 1.1). Eur J Cancer. 2009;45(2):228‐247. outcome in advanced prostate cancer. Proc Natl Acad Sci USA. 28. Friedman J, Hastie T, Tibshirani R. Regularization paths for 2019;116(23):11428‐11436. generalized linear models via coordinate descent. J Stat Softw. 9. Tomlins SA, Rhodes DR, Perner S, et al. Recurrent fusion of 2010;33(1):1‐22. doi:10.18637/jss.v033.i01 TMPRSS2 and ETS transcription factor genes in prostate cancer. 29. Danila DC, Anand A, Sung CC, et al. TMPRSS2‐ERG status in circulating Science. 2005;310(5748):644‐648. doi:10.1126/science.1117679 tumor cells as a predictive biomarker of sensitivity in castration‐ 10. Cancer Genome Atlas Research Network. The molecular taxonomy resistant prostate cancer patients treated with abiraterone acetate. Eur of primary prostate cancer. Cell. 2015;163(4):1011‐1025. Urol. 2011;60(5):897‐904. doi:10.1016/J.EURURO.2011.07.011 11. Robinson D, Van Allen EM, Wu YM, et al. Integrative clinical genomics 30. Sharp A, Coleman I, Yuan W, et al. Androgen receptor splice variant‐ of advanced prostate cancer. Cell. 2015;161(5):1215‐1228. 7 expression emerges with castration resistance in prostate cancer. 12. Carver BS, Tran J, Chen Z, et al. ETS rearrangements and prostate J Clin Invest. 2019;129(1):192‐208. cancer initiation. Nature. 2009;457(7231):E1. doi:10.1038/nature07738 31. Prentice RL. Surrogate endpoints in clinical trials: definition and 13. Wang J, Cai Y, Yu W, Ren C, Spencer DM, Ittmann M. Pleiotropic operational criteria. Stat Med. 1989;8(4):431‐440. biological activities of alternatively spliced TMPRSS2/ERG fusion 32. Taplin ME, Antonarakis ES, Ferrante KJ, et al. Androgen receptor gene transcripts. Cancer Res. 2008;68(20):8516‐8524. modulation optimized for response‐splice variant: a phase 3, 14. Yu J, Yu J, Mani RS, et al. An integrated network of androgen receptor, randomized trial of galeterone versus enzalutamide in androgen polycomb, and TMPRSS2‐ERG gene fusions in prostate cancer progres- receptor splice variant‐7‐expressing metastatic castration‐resistant sion. Cancer Cell. 2010;17(5):443‐454. doi:10.1016/j.ccr.2010.03.018 prostate cancer. Eur Urol. 2019;76:843‐851. 15. Attard G, Swennenhuis JF, Olmos D, et al. Characterization of ERG, 33. Scher HI, Jia X, de Bono JS, et al. Circulating tumour cells as prognostic AR and PTEN gene status in circulating tumor cells from patients markers in progressive, castration‐resistant prostate cancer: a reanalysis with castration‐resistant prostate cancer. Cancer Res. 2009;69(7): of IMMC38 trial data. Lancet Oncol. 2009;10(3):233‐239. 2912‐2918. doi:10.1158/0008-5472.CAN-08-3667 34. Scher HI, Heller G, Molina A, et al. Circulating tumor cell biomarker 16. Attard G, de Bono JS, Logothetis CJ, et al. Improvements in panel as an individual‐level surrogate for survival in metastatic radiographic progression‐free survival stratified by ERG gene status castration‐resistant prostate cancer. J Clin Oncol. 2015;33(12): in metastatic castration‐resistant prostate cancer patients treated 1348‐1355. doi:10.1200/JCO.2014.55.3487 with abiraterone acetate. Clin Cancer Res. 2015;21(7):1621‐1627. 17. Carreira S, Romanel A, Goodall J, et al. Tumor clone dynamics in SUPP ORTING INFO RM ATION lethal prostate cancer. Sci Transl Med. 2014;6(254):254ra125. 18. Conteduca V, Wetterskog D, Sharabiani MTA, et al. Androgen Additional supporting information can be found online in the receptor gene status in plasma DNA associates with worse outcome Supporting Information section at the end of this article. on enzalutamide or abiraterone for castration‐resistant prostate cancer: a multi‐institution correlative biomarker study. Ann Oncol. 2017;28(7):1508‐1516. 19. Jayaram A, Wingate A, Wetterskog D. Plasmsa androgen receptor How to cite this article: Fernandez‐Perez MP, Perez‐Navarro E, copy number status at emergence of metastatic castration‐resistant Alonso‐Gordoa T, et al. A correlative biomarker study and prostate cancer: a pooled multicohort analysis. JCO Precis Oncol. integrative prognostic model in chemotherapy‐naïve metastatic 2019;3:1‐13. castration‐resistant prostate cancer treated with enzalutamide. 20. Danila DC, Heller G, Gignac GA, et al. Circulating tumor cell number and prognosis in progressive castration‐resistant prostate cancer. The Prostate. 2023;83:376‐384. doi:10.1002/pros.24469 Clin Cancer Res. 2007;13(23):7053‐7058.