Toggle Main Menu Toggle Search

Open Access padlockePrints

Adaptive designs for clinical trials assessing biomarker-guided treatment strategies.

Lookup NU author(s): Professor James WasonORCiD, Janice Dunn



This work is licensed under a Creative Commons Attribution 4.0 International License (CC BY 4.0).


The Biomarker Strategy Design has been proposed for trials assessing the value of a biomarker in guiding treatment in oncology. In such trials, patients are randomised to either receive the standard chemotherapy treatment or a biomarker-directed treatment arm, in which biomarker status is used to guide treatment. Motivated by a current trial, we consider an adaptive design in which two biomarkers are assessed. The trial is conducted in two stages. In the first stage, patients in the biomarker-guided arm are assessed using a standard and an alternative cheaper biomarker, with the standard biomarker guiding treatment. An analysis comparing biomarker results is then used to choose the biomarker to use for the remainder of the trial. The new biomarker is used if the results for the two biomarkers are sufficiently similar. We show that in practical situations the first-stage results can be used to adapt the trial without type I error rate inflation. We also show that there can be considerable cost gains with only a small loss in power in the case where the alternative biomarker is highly concordant with the standard one. Adaptive designs have an important role in reducing the cost and increasing the clinical utility of trials evaluating biomarker-guided treatment strategies.

Publication metadata

Author(s): Wason J, Marshall A, Dunn J, Stein RC, Stallard N

Publication type: Article

Publication status: Published

Journal: British Journal of Cancer

Year: 2014

Volume: 110

Issue: 8

Pages: 1950-1957

Print publication date: 15/04/2014

Online publication date: 25/03/2014

Acceptance date: 02/03/2014

Date deposited: 05/01/2018

ISSN (print): 0007-0920

ISSN (electronic): 1532-1827

Publisher: Nature Publishing Group


DOI: 10.1038/bjc.2014.156

PubMed id: 24667651


Altmetrics provided by Altmetric


Funder referenceFunder name