General Information
Conflict of Interest
Request Clinical Trials
Associated Trial(s): What type of data are you looking for?: Individual Participant-Level Data, which includes Full CSR and all supporting documentationRequest Clinical Trials
Data Request Status
Status: OngoingResearch Proposal
Project Title: Machine Learning-Driven Identification of Clinical Subgroups with Heterogeneous Responses to Apalutamide in mCSPC
Scientific Abstract:
Background: The TITAN trial demonstrated that apalutamide plus ADT significantly improves radiographic progression-free survival (rPFS) and overall survival (OS) in mCSPC. However, traditional subgroup analyses may miss complex interactions between clinical variables.
Objective: To identify patient subgroups with heterogeneous treatment effects (HTE) using machine learning and validate their clinical relevance.
Study Design: Post hoc analysis of the TITAN trial data.
Participants: 1,052 patients randomized to apalutamide + ADT or placebo + ADT.
Primary Outcome: Restricted mean survival time (RMST) difference in rPFS between treatment arms within identified subgroups.
Secondary Outcomes: OS, PSA progression, and safety profiles.
Statistical Analysis: Causal forest models for HTE detection, Cox regression for validation, and nomogram development for clinical translation.
Brief Project Background and Statement of Project Significance: The TITAN trial established the efficacy of apalutamide in improving overall survival (OS) and radiographic progression--free survival (rPFS) in mHSPC (Chi et al., 2019). However, conventional subgroup analyses based on single variables (e.g., disease volume, age) may fail to capture complex interactions among clinical features. Machine learning methods, such as causal forests, enable hypothesis-free exploration of heterogeneous treatment effects (HTE) through multidimensional stratification (Athey & Imbens, 2016). This study proposes to leverage the TITAN clinical dataset to identify patient subgroups with enhanced or diminished responses to apalutamide, guiding personalized therapy.
Specific Aims of the Project: The specific aims and objectives of this research project are to develop a causal forest model to quantify HTE of apalutamide using baseline clinical variables. The robustness of the identified subgroups will be validated through sensitivity analyses. Finally, a clinically actionable decision rule, such as a nomogram, will be generated to stratify patients into high/low benefit subgroups.
Study Design: Individual trial analysis
What is the purpose of the analysis being proposed? Please select all that apply.: New research question to examine treatment effectiveness on secondary endpoints and/or within subgroup populations
Software Used: RStudio
Data Source and Inclusion/Exclusion Criteria to be used to define the patient sample for your study:
Inclusion: All randomized patients with complete baseline data.
Exclusion: None .
Primary and Secondary Outcome Measure(s) and how they will be categorized/defined for your study:
Primary: rPFS (time to radiographic progression/death).
Secondary: OS, PSA progression, Second progression-free survival.
Main Predictor/Independent Variable and how it will be categorized/defined for your study:
Baseline variables: Age, disease volume, Gleason score, prior docetaxel and other treatment, baseline PSA, tumor zone, etc.
Other parameters, such as prostate - specific antigen doubling time (PSADT), PSA velocity, and the like. Clinical parameters should be as detailed and comprehensive as possible.
Other Variables of Interest that will be used in your analysis and how they will be categorized/defined for your study: See above.
Statistical Analysis Plan:
1. Data Preparation: Extract TITAN trial data (baseline demographics, disease characteristics, treatment history, outcomes).
2. Machine Learning Modeling:
- Implement causal forests to estimate HTE using RMST for rPFS.
- Compute variable importance to prioritize predictors .
3. Validation:
- Compare subgroup survival using Cox models and Kaplan-Meier curves.
- Assess robustness via bootstrap resampling.
4. Clinical Translation: Develop a nomogram integrating key predictors for bedside use.
Narrative Summary: This study aims to identify subgroups of patients with metastatic castration-sensitive prostate cancer (mCSPC) who benefit most from apalutamide combined with androgen deprivation therapy (ADT). Using data from the TITAN trial, we will apply advanced machine learning methods to analyze how combinations of clinical factors (e.g., disease volume, age, Gleason score) influence treatment outcomes. By uncovering hidden patterns in the data, this research will help doctors personalize treatment decisions, ensuring patients receive therapies most likely to improve their survival and quality of life.
Project Timeline:
Months 1--3: Data preparation and model development.
Months 4--9: Subgroup validation and sensitivity testing.
Months 10--12: Nomogram development and manuscript drafting.
Dissemination Plan:
Target Journals: Journal of Clinical Oncology, European Urology.
Audience: Oncologists, urologists, and precision medicine researchers.
Bibliography:
1. Chi KN, Agarwal N, Bjartell A, et al. Apalutamide for Metastatic, Castration-Sensitive Prostate Cancer. *N Engl J Med*. 2019;381(1):13-24.
2. Athey S, Imbens G. Recursive Partitioning for Heterogeneous Causal Effects. *PNAS*. 2016;113(27):7353-7360.