Bayesian dose-finding method using Bayesian model averaging

2022 Fiscal Year Final Research Report

• Project/Area Number: 20K23318
• Research Category: Grant-in-Aid for Research Activity Start-up
• Allocation Type: Multi-year Fund
• Review Section: 1001:Information science, computer engineering, and related fields
• Research Institution: Tokyo Medical and Dental University
• Principal Investigator: Sato Hiroyuki 東京医科歯科大学, 医学部附属病院, 助教 (70793595)
• Project Period (FY): 2020-09-11 – 2023-03-31
• Keywords: 用量探索デザイン / がん臨床試験 / 分子標的薬 / ベイズ統計学

Outline of Final Research Achievements

Model-based dose-finding methods that assume a statistical model for dose-efficacy relationships of molecular-targeted agents in oncology Phase I clinical trial have been challenged by the selection of appropriate statistical models from limited information. To address this issue of model misspecification, we incorporate Bayesian model averaging into the existing model-based dose-finding method and expand it to generate model calibration methods that account for model uncertainty in the dose-efficacy relationship of molecular-targeted agents. We conducted computer simulation studies assuming various dose-efficacy relationships for molecular-targeted agents to compare the performance of the proposed methods with that of existing dose-finding methods. Based on these simulation study results, the proposed dose-finding method showed performance comparable to or better than the existing model-based dose-finding method and was found to be robust against uncertainty in model selection.

Free Research Field

臨床統計学

Academic Significance and Societal Importance of the Research Achievements

近年，個別化医療を実現するためにがん分子標的薬が活発に開発されているが，最適用量として特定された用量では薬剤の有効性が最大化されない可能性があるという報告がある．本研究では，統計モデルを利用した用量探索法におけるモデル選択の不確実性という課題に着目し，ベイズ統計学を利用した新しい臨床試験デザインを開発し，既存の用量探索法と比較して同等以上の性能を示した．この成果は，本邦のがん分子標的薬の開発に貢献するものであると考える．