Integrating tumor burden with survival outcome for treatment effect evaluation in oncology trials

2025-06-09

AI Summary

Early-phase cancer trials face challenges of sparse data and difficulty quantifying treatment effects reliably. Method: We propose a Bayesian utility-based estimand that jointly models longitudinal tumor burden dynamics and time-to-event outcomes (e.g., OS/PFS), integrating biological response (e.g., RECIST criteria or continuous tumor shrinkage) with clinical benefit via an interpretable utility function. Contribution/Results: Built upon a joint longitudinal-survival modeling framework and Bayesian inference, the estimand enables early, robust treatment effect assessment. Simulation studies demonstrate strict control of Type I error even in small samples, substantially improved statistical power over conventional endpoints (e.g., ORR, PFS), and seamless scalability from exploratory Phase I/II to confirmatory Phase III analyses. This approach bridges the gap between early biomarker-driven evaluation and late-stage clinical outcome validation.

Technology Category

• Reasoning under Uncertainty: Decision/Utility Theory
• Machine Learning: Calibration & Uncertainty Quantification
• Intelligent Robots: State Estimation

Application Category

• User Modeling, Personalization and Recommendation: Metrics for user behavior and evaluating success
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics

Abstract

In early-phase cancer clinical trials, the limited availability of data presents significant challenges in developing a framework to efficiently quantify treatment effectiveness. To address this, we propose a novel utility-based Bayesian approach for assessing treatment effects in these trials, where data scarcity is a major concern. Our approach synthesizes tumor burden, a key biomarker for evaluating patient response to oncology treatments, and survival outcome, a widely used endpoint for assessing clinical benefits, by jointly modeling longitudinal and survival data. The proposed method, along with its novel estimand, aims to efficiently capture signals of treatment efficacy in early-phase studies and holds potential for development as an endpoint in Phase 3 confirmatory studies. We conduct simulations to investigate the frequentist characteristics of the proposed estimand in a simple setting, which demonstrate relatively controlled Type I error rates when testing the treatment effect on outcomes.

Problem

• Evaluating treatment effects in early-phase cancer trials with limited data
• Integrating tumor burden and survival outcomes for efficacy assessment
• Developing a Bayesian approach for efficient treatment signal detection

Innovation

• Utility-based Bayesian approach for oncology trials
• Joint modeling of tumor burden and survival
• Novel estimand for early-phase efficacy signals