Online Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 441 - Bayesian (and Other) Clinical Trials Designs
Type: Contributed
Date/Time: Thursday, August 12, 2021 : 4:00 PM to 5:50 PM
Sponsor: Biopharmaceutical Section
Abstract #317928
Title: Bayesian Logistic Regression with Covariates in Oncology Dose Escalation
Author(s): Arnab Kumar Maity* and MARZIEH GOLMAKANI and LADA ALEKSANDROVNA MARKOVTSOVA
Companies: Pfizer and PFIZER and PFIZER
Keywords: BLRM; EWOC; Oncology; Phase I; MTD; DLT
Abstract:

The traditional dose finding design guided by the Bayesian logistic regression model (BLRM) with escalation with overdose control (EWOC) models the probability of dose limiting toxicity (DLT) as a function of doses received by the trial participants. Then the probability of DLT is categorized into under dosing probability, target toxicity probability, and overdosing probability to provide statistical recommendation for dose escalation or dose de-escalation. When taking the final decision toward declaring the maximum tolerated dose (MTD) or the recommended phase II dose (RP2D) the trial investigators also consider the biomarkers measurements obtained during the study. However, these biomarkers are only studied in subjective manner and generally not included in the dose finding model. In this work, we consider the problem of including the biomarkers in the Bayesian logistic regression as covariates. We discuss the potential advantages and challenges and demonstrate a few avenues to overcome them using real data examples.


Authors who are presenting talks have a * after their name.

Back to the full JSM 2021 program