Statistics and Its Interface

Volume 11 (2018)

Number 1

Adaptive oncology phase I trial design of drug combinations with drug-drug interaction modeling

Pages: 109 – 127

DOI: https://dx.doi.org/10.4310/SII.2018.v11.n1.a10

Authors

Yang Yang (Department of Mathematics and Statistics, University of Maryland Baltimore County, Baltimore, Md., U.S.A.)

Hong-Bin Fang (Department of Biostatistics, Bioinformatics and Biomathematics, Georgetown University, Washington, D.C., U.S.A.)

Anindya Roy (Department of Mathematics and Statistics, University of Maryland Baltimore County, Baltimore, Md., U.S.A.)

Ming Tan (Department of Biostatistics, Bioinformatics and Biomathematics, Georgetown University, Washington, D.C., U.S.A.)

Abstract

The goal of a Phase I trial is to find the maximum tolerated dose (MTD). In a single-agent dose finding Phase I trial, the key underlying assumption is that toxicity probability increases monotonically with the dose level. However, in multi-agent trials, this assumption may not hold because the drug-drug interaction potentially can either decrease or increase the joint toxicity as compared to either one used alone, which may lead to an unforeseen toxicity probability surface. Thus there exists multiple MTDs. We first develop a novel adaptive dose-finding approach which can be applied to these kinds of multi-drug combination trials. With this approach, drug-drug interaction and toxicity probability are modeled jointly through a Bliss independence model. The main goal of our dose finding scheme is to search for maximum tolerated region (MTR), as opposed to maximum tolerated dose (MTD), in single agent phase I trials. The method allows exploration of more combinations in the phase I stage, which is of particular relevance in oncology since phase I trials on the combinations may be the only opportunity before launching a costly phase III trial, comparing selected combination(s) with a standard of care. Dose escalation/de-escalation decision rules are determined by the posterior estimates of both joint toxicity probability and the corresponding drug-drug interaction, which can be continuously updated by sequentially assigning new patients into the trial while more data is being observed. We evaluate the operating characteristics of the proposed method through extensive simulation studies under various scenarios. The proposed method demonstrates satisfactory performance. In addition, the MTR offers several combinations that investigators may choose to advance to future trials based on external information from e.g., preclinical antitumor activities and other trials.

Keywords

adaptive Bayesian design, bliss independence, drug combination, dose finding, interaction index function, maximum tolerated dose (MTD), maximum tolerated region, objective function

The research of Fang and Tan is partly supported by US National Cancer Institute award R01CA164717.

Received 25 August 2016

Published 23 August 2017