General Information
Conflict of Interest
Request Clinical Trials
Associated Trial(s):- NCT02489318 - A Phase 3 Randomized, Placebo-controlled, Double-blind Study of Apalutamide Plus Androgen Deprivation Therapy (ADT) Versus ADT in Subjects With Metastatic Hormone-sensitive Prostate Cancer (mHSPC)
- NCT01715285 - A Randomized, Double-blind, Comparative Study of Abiraterone Acetate Plus Low-Dose Prednisone Plus Androgen Deprivation Therapy (ADT) Versus ADT Alone in Newly Diagnosed Subjects With High-Risk, Metastatic Hormone-naive Prostate Cancer (mHNPC)
Request Clinical Trials
Data Request Status
Status: Approved Pending DUA SignatureResearch Proposal
Project Title: mHSPC -- Phase 3 -- Doublet versus Triplet: Early PSA-Response Benchmarks, Kinetics, and Prognostic Associations under Standard of Care
Scientific Abstract:
Background: Deep prostate-specific antigen (PSA) responses correlate with survival in metastatic hormone-sensitive prostate cancer (mHSPC). However, benchmarks for early PSA kinetics and whether they merely reflect underlying prognosis or genuinely modify the treatment effect (ARPI vs ADT alone) remain poorly characterized under modern standard-of-care.
Objective: To derive early PSA-response benchmarks for standard-of-care (ARPI + ADT) in mHSPC and characterize their relationship with overall survival (OS) and radiographic progression-free survival (rPFS).
Study Design: Participant-level data meta-analysis of randomized Phase 3 trials. We will use control arms to distinguish baseline prognosis from treatment-effect modification, and use findings to simulate future early-phase trial designs.
Participants: mHSPC patients from the TITAN and LATITUDE trials (requested via YODA), to be pooled with ARCHES, ARANOTE, and ARASENS data (requested concurrently via Vivli).
Primary and Secondary Outcome Measure(s): Primary outcomes are OS and rPFS. Predictors include PSA clearance (<0.2 ng/mL), ultra-low clearance (<0.02 ng/mL), PSA50, PSA90, and PSA kinetics (slope, time to nadir) at 3, 6, and 9 months.
Statistical Analysis: We will estimate pooled benchmarks and use longitudinal mixed-effects models for kinetics. Associations with OS/rPFS will be assessed via landmark-based Cox regression and time-varying covariate methods (accounting for immortal-time bias). Interaction testing will separate prognostic from treatment-modifying effects.
Brief Project Background and Statement of Project Significance:
Background: Prostate-specific antigen (PSA) decline is universally tracked during mHSPC treatment. While profound PSA responses are widely accepted as favorable prognostic indicators, modern standard of care (SOC) now dictates treatment with an androgen receptor pathway inhibitor combined with androgen deprivation therapy (ARPI + ADT). Despite this shift, rigorous, pooled benchmarks for early PSA kinetics under ARPI + ADT remain under-defined. Furthermore, the precise relationship between the depth and speed of PSA response (e.g., <0.2 ng/mL vs. <0.02 ng/mL) and long-term survival outcomes (OS and rPFS) has not been comprehensively modeled across the diverse mHSPC populations represented in pivotal Phase 3 trials.
Significance: This project is highly significant for both clinical practice and trial methodology. Clinically, establishing robust early-response benchmarks will empower physicians to contextualize patient progress by months 3 to 6 of therapy, identifying high-risk patients who fail to meet SOC milestones. Methodologically, defining these quantitative benchmarks is critical to accelerating future drug development. Next-generation therapies for mHSPC urgently require reliable early intermediate endpoints to support single-arm, early-phase trial designs with strict Go/No-Go criteria, rather than waiting years for survival readouts. Critically, because the requested trials are randomized, we will leverage the control (ADT-alone or ADT+docetaxel) arms to unconfound two fundamental forces: how much of the survival benefit seen in "PSA responders" is purely prognostic (i.e., these patients biologically do better regardless of the ARPI), versus how much is a true treatment-effect modifier (where the ARPI drives the differential benefit). Understanding this distinction is vital to validating PSA metrics for clinical trial decision-making. By pooling TITAN and LATITUDE (requested here) with three additional Phase 3 trials requested concurrently via Vivli (ARCHES, ARANOTE, ARASENS), this analysis achieves the statistical power and broad representativeness necessary to compute highly precise benchmarks and conduct nuanced outcome modeling. These findings will materially enhance scientific knowledge by de-risking novel trial designs and advancing the precision management of mHSPC.
References:
Wallis CJD, et al. Eur Urol Oncol. 2026;9:251-258.
Roy S, et al. Nat Commun. 2026;17:667.
Chi KN, et al. N Engl J Med. 2019.
Fizazi K, et al. N Engl J Med. 2017.
Specific Aims of the Project:
Aim 1: Benchmark derivation. Derive SOC benchmarks for early PSA response--clearance (<0.2 ng/mL), ultra-low clearance (<0.02 ng/mL), PSA50, and PSA90--at 3-, 6-, and 9-month landmarks within the ARPI+ADT population.
Aim 2: PSA kinetics modeling. Fit longitudinal models to derive benchmark distributions for PSA slope (decline rate) and time to nadir.
Aim 3: Association with outcomes. Evaluate the association of early PSA benchmarks and late clearance (achieving clearance between 3-6 or 6-9 months) with OS and rPFS, using time-varying methods to control for immortal-time bias.
Aim 4: Prognostic vs. treatment-effect--modifying decomposition. Using randomized control arms, distinguish the purely prognostic component of PSA response from the treatment-effect--modifying component (differential ARPI benefit by response status).
Aim 5: Docetaxel interaction. Determine if concurrent chemotherapy fundamentally alters the PSA--outcome relationship by comparing the ARASENS cohort against non-concurrent chemotherapy cohorts.
Aim 6: Trial design simulation. Quantify OS and rPFS hazard ratios implied by varying early PSA clearance rates to inform decision thresholds for future early-phase trials.
Hypothesis: Deeper/earlier PSA responses and late PSA clearance correlate with improved OS and rPFS. Furthermore, a substantial portion of the responder survival benefit reflects underlying baseline prognosis, which Aim 4 will quantify.
Study Design: Meta-analysis (analysis of multiple trials together)
What is the purpose of the analysis being proposed? Please select all that apply.: Participant-level data meta-analysis Meta-analysis using data from the YODA Project and other data sources Develop or refine statistical methods Research on clinical trial methods Research on clinical prediction or risk prediction
Software Used: Python, R, RStudio, STATA
Data Source and Inclusion/Exclusion Criteria to be used to define the patient sample for your study:
For this YODA request, we will utilize two Phase 3 randomized controlled trials in mHSPC: TITAN (NCT02489318) and LATITUDE (NCT01715285).
Inclusion Criteria: All randomized patients with mHSPC from TITAN and LATITUDE. (For LATITUDE, this specifically represents a high-risk mHSPC population).
Exclusion Criteria: None. All intention-to-treat patients with available baseline covariates and longitudinal outcome/PSA data will be included.
Note on Data Pooling Platform Context:
The complete statistical analysis plan requires pooling patient-level data from five Phase 3 trials to ensure adequate statistical power for benchmark derivation and kinetics modeling. This YODA request provides access to the J&J/Janssen datasets (TITAN, LATITUDE). Concurrently, data from three additional trials (ARCHES, ARANOTE, and ARASENS) are being requested through the Vivli data-sharing platform (Vivli ID: 00016865). The participant-level data meta-analysis will be conducted by securely transferring and pooling the datasets in an approved secure research environment (e.g., the Vivli SRE), subject to cross-platform approvals and data use agreements.
Primary and Secondary Outcome Measure(s) and how they will be categorized/defined for your study:
Primary Outcome Measures:
Overall Survival (OS): Defined as the time from randomization to death from any cause, as assessed by the individual trial protocols.
Radiographic Progression-Free Survival (rPFS): Defined as the time from randomization to radiographic disease progression or death from any cause, whichever occurs first, according to individual trial definitions (typically utilizing PCWG2/3 and RECIST 1.1 criteria). Because death without radiographic progression serves as a competing risk for rPFS, the analysis of rPFS will be modeled primarily using cause-specific hazards. Fine-Gray cumulative incidence models will be utilized as a sensitivity analysis.
Secondary Outcome Measures:
None. The objective of this study is strictly focused on linking PSA kinetics to the two gold-standard survival endpoints in mHSPC (OS and rPFS).
No changes to the primary outcome definitions from the original trial protocols will be made in the final analysis.
Main Predictor/Independent Variable and how it will be categorized/defined for your study:
The main independent variables are longitudinal and landmark PSA-response metrics, assessed primarily within the ARPI + ADT populations at the 3-, 6-, and 9-month post-randomization landmarks:
1. PSA Clearance: Categorized as achieving a PSA level of <0.2 ng/mL versus >=0.2 ng/mL at the specific landmark timepoint.
2. Ultra-low PSA Clearance: Categorized as achieving a PSA level of <0.02 ng/mL versus >=0.02 ng/mL. (Values falling below the specific trial assay's lower limit of quantification [LLOQ] will be handled via left-censoring modeling techniques).
3. PSA50 and PSA90: Categorized as achieving a >=50% or >=90% decline in PSA from baseline, respectively, versus not achieving these thresholds.
4. Modeled PSA Kinetics: Continuous variables representing the patient-level PSA decline trajectory, specifically the modeled "PSA slope" (rate of PSA decline) and "time to PSA nadir" (time from randomization to the lowest observed PSA value).
5. Late Clearance Status: A time-varying categorical variable defining patients who were non-responders at an early landmark (e.g., 3 months) but achieved clearance by the next landmark (e.g., 6 months).
Other Variables of Interest that will be used in your analysis and how they will be categorized/defined for your study:
The following variables will be used to characterize the study sample, perform multivariable risk adjustment, and test for key statistical interactions:
1. Randomized Treatment Assignment: Categorized as Experimental Arm (ARPI-containing regimens) versus Control Arm (ADT alone or ADT + docetaxel/placebo). This is fundamentally required for Aim 4 to decompose the prognostic versus treatment-effect--modifying nature of PSA response.
2. Baseline Disease Volume: Categorized as high-volume versus low-volume disease (per CHAARTED criteria).
3. Disease Presentation: Categorized as synchronous (de novo metastatic) versus metachronous (recurrent metastatic).
4. Docetaxel Use: Categorized as prior or concurrent docetaxel use versus no docetaxel use, to support the docetaxel-interaction analysis (Aim 5).
5. Patient Demographics & Clinical Baselines: Age (continuous), Eastern Cooperative Oncology Group (ECOG) performance status (categorical), baseline PSA level (continuous log-transformed), and baseline hemoglobin (continuous).
Statistical Analysis Plan:
The plan below describes the intended analytic approach at a high level. Primary benchmark analyses are conducted within the pooled ARPI + ADT population, since benchmarks are treatment-arm-specific by definition; association and decomposition analyses use both randomized arms as specified.
(1) Benchmark derivation
Estimate the proportion of ARPI + ADT patients achieving each PSA metric (clearance <0.2 ng/mL, ultra-low <0.02 ng/mL, PSA50, PSA90) at the 3-, 6-, and 9-month landmarks, with appropriate confidence intervals, reported both pooled and per-trial with an assessment of between-trial heterogeneity.
(2) PSA kinetics modeling
Characterize individual PSA decline using longitudinal mixed-effects modeling of PSA over time, deriving patient-level estimates of PSA slope and time to nadir -- the time from treatment start to a patient's lowest observed PSA value (the nadir) -- and summarizing their distributions under SOC. The approach will account for the assay's lower limit of quantification so that ultra-low values do not bias the estimates.
(3) Association with outcomes
Assess the association of each PSA metric with OS and rPFS using landmark-based time-to-event analyses with appropriate adjustment for baseline prognostic factors. Kinetics-derived measures will be evaluated as both continuous and categorized predictors. Analyses will account for the competing risk of death where relevant to rPFS.
(4) Late-clearance analysis
We will compare outcomes among early responders, late clearers (3→6 mo and 6→9 mo), and sustained non-responders. Because late-clearance status is only defined for patients who survive to clear, these analyses will use methods that appropriately handle guarantee-time (immortal-time) bias.
(5) Prognostic vs. treatment-effect--modifying decomposition (uses randomization)
Using both randomized arms, assess whether PSA response modifies the treatment effect (a response-by-treatment relationship) or is associated with outcome irrespective of treatment assignment (prognostic). Within the ARPI arm, the responder-vs-non-responder association will additionally be examined before and after adjustment for baseline prognostic factors, to gauge how much of the association reflects baseline prognosis versus a treatment-attributable effect.
(6) Docetaxel-interaction analysis
To determine how concurrent chemotherapy affects the PSA--outcome relationship, we will test for an interaction between docetaxel context (the ARASENS concurrent-docetaxel cohort vs. the non-concurrent cohorts) and each PSA-response metric. This assesses whether the prognostic value of PSA response differs in the presence of concurrent docetaxel.
(7) Simulation for trial design
Use simulation to characterize the OS and rPFS hazard ratios implied by different PSA clearance-rate scenarios, drawing component survival estimates from the analyses above. Results will be reported as a range that reflects key sources of uncertainty (including immortal-time handling, the PSA-mediated versus direct component of the treatment effect, and the prognostic-selection contribution from Aim 5), to inform expected effect sizes and decision thresholds for a future single-arm early-phase trial.
Narrative Summary: This study defines benchmarks for how prostate-specific antigen (PSA) responds to standard mHSPC treatment (ARPI + ADT) and links these responses to long-term outcomes. We will measure PSA response depth (e.g., clearance, 50/90% reductions), model decline speed, and assess if PSA clearance predicts improved survival and delayed progression. Using control arms, we will determine if PSA response reflects underlying prognosis or modifies treatment effect. Finally, simulations will translate PSA clearance rates into survival benefits to guide future early-phase trial designs. Analyzing five pooled Phase 3 datasets provides crucial statistical power. We require anonymized data on survival, progression, longitudinal PSA, and baseline characteristics for both treatment and control arms.
Project Timeline:
Anticipated Project Start Date (Data Access Granted): December 7, 2026
Anticipated Analysis Completion Date: December 7, 2027 (Aligning with the conclusion of the initial 12-month Data Use Agreement period).
Anticipated Date Manuscript Drafted: March 2028 (Within 3--4 months following analysis completion).
Anticipated Date Manuscript First Submitted for Publication: June 2028
Anticipated Date Results Reported Back to the YODA Project: June 2028 (Concurrent with manuscript submission, a summary of findings will be provided to YODA, followed by the final citation upon publication).
Dissemination Plan:
Anticipated Products:
We anticipate producing two primary outputs from this research:
Abstract/Presentation: Initial findings will be submitted as an abstract for presentation (oral or poster) at a major international oncology or urology congress (e.g., the American Society of Clinical Oncology [ASCO] Annual Meeting, the European Society for Medical Oncology [ESMO] Congress, or the American Urological Association [AUA] Annual Meeting).
Peer-Reviewed Manuscript: A comprehensive manuscript detailing the derived benchmarks, kinetics modeling, and prognostic/treatment-effect decomposition.
Target Audience:
The target audience includes medical oncologists, urologists, and researchers managing patients with mHSPC, as well as biostatisticians, clinical trialists, and regulatory scientists focused on novel trial designs and intermediate endpoints in prostate cancer.
Potentially Suitable Journals:
We will target high-impact, peer-reviewed oncology and urology journals. Suitable targets include the Journal of Clinical Oncology (JCO), European Urology, Annals of Oncology, or The Lancet Oncology.
Bibliography:
- Wallis CJD, et al. Association of deep and durable PSA responses with outcomes in mHSPC: insights from ARASENS and ARANOTE trials. Eur Urol Oncol. 2026;9:251--258.
- Roy S, et al. Early favorable prostate-specific antigen response prediction in metastatic hormone sensitive prostate cancer. Nat Commun. 2026;17:667.
- Armstrong AJ, et al. ARCHES: enzalutamide + ADT in mHSPC. J Clin Oncol. 2019.
- Chi KN, et al. TITAN: apalutamide + ADT in mHSPC. N Engl J Med. 2019.
- Fizazi K, et al. LATITUDE: abiraterone + ADT in high-risk mHSPC. N Engl J Med. 2017.
- Saad F, et al. ARANOTE: darolutamide + ADT in mHSPC. J Clin Oncol. 2024.
- Smith MR, et al. ARASENS: darolutamide and docetaxel added to ADT in mHSPC. N Engl J Med. 2022;386:1132--1142.