Bayesian trial design to identify a sensitive subpopulation in non-proportional hazard settings

Fuente: Zenodo
Gespeichert in:
Bibliographische Detailangaben
1. Verfasser: Nakakura, Akiyoshi
Format: Recurso digital
Veröffentlicht: Zenodo 2025
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866901893771427840
author Nakakura, Akiyoshi
author_facet Nakakura, Akiyoshi
contents <p>In molecular-targeted drug development, biomarkers are increasingly integrated into clinical trial design to enable personalized treatment evaluation. Identifying subpopulations that derive the greatest benefit from novel agents facilitates efficient development. Traditional Bayesian subgroup identification designs rely on hazard ratios and thus require the proportional hazards assumption. However, this assumption often fails for modern therapies such as immune checkpoint inhibitors and targeted agents, which exhibit delayed or time-varying effects.</p> <p>To address this limitation, we extend Morita et al.’s Bayesian phase II trial framework by incorporating the restricted mean survival time (RMST) as the primary endpoint. RMST provides an interpretable measure of average survival up to a fixed time horizon and does not depend on the PH assumption. The proposed Bayesian RMST-based design allows flexible and reliable identification of biomarker-defined subgroups even under non-proportional hazard conditions. Through extensive simulation studies covering delayed and crossing-hazard scenarios, we demonstrate that the RMST-based approach offers superior robustness, stable operating characteristics, and broad applicability to contemporary oncology trials.</p>
format Recurso digital
id zenodo_https___doi_org_10_5281_zenodo_17381727
institution Zenodo
language
publishDate 2025
publisher Zenodo
record_format zenodo
spellingShingle Bayesian trial design to identify a sensitive subpopulation in non-proportional hazard settings
Nakakura, Akiyoshi
<p>In molecular-targeted drug development, biomarkers are increasingly integrated into clinical trial design to enable personalized treatment evaluation. Identifying subpopulations that derive the greatest benefit from novel agents facilitates efficient development. Traditional Bayesian subgroup identification designs rely on hazard ratios and thus require the proportional hazards assumption. However, this assumption often fails for modern therapies such as immune checkpoint inhibitors and targeted agents, which exhibit delayed or time-varying effects.</p> <p>To address this limitation, we extend Morita et al.’s Bayesian phase II trial framework by incorporating the restricted mean survival time (RMST) as the primary endpoint. RMST provides an interpretable measure of average survival up to a fixed time horizon and does not depend on the PH assumption. The proposed Bayesian RMST-based design allows flexible and reliable identification of biomarker-defined subgroups even under non-proportional hazard conditions. Through extensive simulation studies covering delayed and crossing-hazard scenarios, we demonstrate that the RMST-based approach offers superior robustness, stable operating characteristics, and broad applicability to contemporary oncology trials.</p>
title Bayesian trial design to identify a sensitive subpopulation in non-proportional hazard settings
url https://doi.org/10.5281/zenodo.17381727