Você está aqui: Página Inicial / Seminários / 14/08/2020 - Comparison of inferential seamless phase II/III designs for acute stroke trials with biomarkers as surrogate endpoints - Palestrante: Márcio Augusto Diniz (Cedars-Sinai)

14/08/2020 - Comparison of inferential seamless phase II/III designs for acute stroke trials with biomarkers as surrogate endpoints - Palestrante: Márcio Augusto Diniz (Cedars-Sinai)

Quando 14/08/2020
das 14h00 até 16h00
Adicionar evento ao calendário vCal
iCal

Seminário Conjunto UFSCar/USP

Data e Horário: 14/08/2020 às 14h

Vídeo da Apresentação: https://drive.google.com/file/d/1qihKiE-lmIGwHkZssl5_wXcMtfrosGlZ/view?usp=sharing

Link com Apresentação: Apresentação Marcio A. Diniz

Título: Comparison of inferential seamless phase II/III designs for acute stroke trials with biomarkers as surrogate endpoints

Palestrante: Márcio Augusto Diniz (Cedars-Sinai)

Resumo:

Inferential seamless phase II/III designs have been proposed as an alternative path to
traditional drug development in order to decrease failure in large phase III trials. Seamless phase II/III
designs allow investigators to accommodate randomized phase II designs that require larger samples
sizes instead of single-arm designs that depend on historical data with often different case-mix
populations. Response adaptive randomization (RAR) is often implemented aiming to allocate more
subjects to the apparently superior arm, stop the lead-in phase early for efficacy, drop a poorly
performing arm, and thus, may resulting in more patients being treated successfully. However, RAR
algorithms require a short-term response relative to the accrual rate, otherwise, allocation probabilities
will not be updated sufficiently quick such that patients could be benefited. Several authors have
proposed the use of biomarkers as a surrogate for a long-term endpoint when calculating allocation
probabilities without considering their diagnostic performance. In this work, we study the loss in power,
type I error and trial duration for an acute stroke trial with a long-term ordinal endpoint as function of
sensitivity and specificity for biomarkers that are used as surrogate when calculating allocation
probabilities and/or stopping rules.

Autores: Márcio Augusto Diniz, Patrick Lyden and Mourad Tighiouart - Cedars-Sinai Medical Center

 

registrado em: