Use of Bayesian Approaches in Oncology Clinical Trials: A Cross-Sectional Analysis.

Borja G. Lopez-Rey ^1,2 Gerard Carot-Sans ^3,4 Dan Ouchi ² Ferran Torres ^2,4* Caridad Pontes ^4,5,6

• ¹ Agencia Española de Medicamentos y Productos Sanitarios, Madrid, Madrid, Spain
• ² Universitat Autònoma de Barcelona, Barcelona, Spain, Barcelona, Spain
• ³ Catalan Health Service, Barcelona, Catalonia, Spain
• ⁴ Digitalization for the sustainability of the healthcare system (DS3), Barcelona, Spain
• ⁵ Departamento de Farmacología, Terapéutica y Toxicología, Facultad de Medicina, Universidad Autónoma de Barcelona, Barcelona, Catalonia, Spain
• ⁶ Servei de Farmacologia Clínica, Hospital de la Santa Creu i de Sant Pau, Barcelona, Spain, Barcelona, Spain

Introduction: Bayesian approaches can improve trial efficiency and accelerate decision-making, but their adoption is limited due to reliance on frequentist methods. Oncology trials, often addressing severe conditions with few therapeutic options, are well-suited for Bayesian methodologies, particularly in early phases. Objectives: This study describes the use of Bayesian methods in oncology clinical trials over the past 20 years. Methods: A cross-sectional observational study was conducted to identify oncology trials using Bayesian approaches registered in clinicaltrials.gov between 2004 and 2024. Data were extracted from clinicaltrials.gov, PubMed, and manual cross-references, and analyzed using R with manual verification. Results: The Bayesian trials were retrieved and their main characteristics were extracted using R and verified manually. Between 2004 and 2024, 384,298 trials were registered in clinicaltrials.gov; we identified 84,850 oncology clinical trials (22%); of whose 640 (0.75%) used Bayesian approaches. The adoption of Bayesian methods increased significantly after 2011, but while half of all Bayesian studies started in the last 5 years, this paralleled the overall increase in oncological research rather than increased proportion of Bayesian trials. Most Bayesian trials were phase 1 and phase 2 studies, and two-thirds of Bayesian trials with efficacy objectives had single-arm designs, often utilizing binary endpoints, such as overall response, as the primary measure. Conclusions: Bayesian methods remain underutilized in oncology trials, primarily applied to treatment efficacy analyses in single-arm designs with binary endpoints. Their broader adoption offers significant potential, especially in small populations and severe conditions with unmet needs.