|
Activity Number:
|
166
- Non-Clinical Statistics, Personalized Medicine, and Other Topics
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 10, 2021 : 10:00 AM to 11:50 AM
|
|
Sponsor:
|
Biopharmaceutical Section
|
|
Abstract #318212
|
|
|
Title:
|
Utility-Based Approach in Individualized Optimal Dose Selection Using Machine Learning Methods
|
|
Author(s):
|
Pin Li* and Jeremy M.G. Taylor and Philip Boonstra and Theodore S. Lawrence and Matthew J. Schipper
|
|
Companies:
|
University of Michigan and University of Michigan and University of Michigan and University of Michigan and University of Michigan
|
|
Keywords:
|
Gaussian Process;
Random forest;
Utility matrix
|
|
Abstract:
|
The goal in personalized medicine is to individualize treatment using patient characteristics to improve health outcomes. Selection of optimal dose must balance the effect of dose on both efficacy(E) and toxicity(T) outcomes. We propose to use flexible machine learning methods such as random forest (RF) and Gaussian process (GP) to build models for binary E and T depending on dose and biomarkers. A copula is used to model the joint distribution of E and T and the estimates are constrained to have non-decreasing dose-efficacy and dose-toxicity relationships. Numerical utilities are elicited from clinicians for each potential bivariate outcome. For each patient, the optimal dose is chosen to maximize the posterior mean of the utility function. The proposed methods are evaluated in a simulation study to compare expected utility outcomes under various estimated optimal dose rules. GP tended to have better performance than RF. Enforcing monotonicity during modeling provided small benefits. Whether and how, correlation between E and T, was modeled, had little effect on performance. The proposed methods are illustrated with a study of liver cancer patients treated with radiation therapy.
|
Authors who are presenting talks have a * after their name.
