Designs for evaluating precision medicine trials - Archive ouverte HAL Accéder directement au contenu
Chapitre D'ouvrage Année : 2015

Designs for evaluating precision medicine trials

Résumé

Most new agents are being designed to target molecular alterations involved in carcinogenesis. They are expected to produce antitumor activity only in the presence of the matching molecular alteration or companion biomarker. Prevalence of the molecular alteration is often low. Specific tumor types with the molecular alteration are often rare diseases, and testing the treatment effect independently in each of them is almost impossible. Therefore, several histologic-agnostic trials have been initiated to test several treatments directed against several molecular alterations in various tumor types. The objective of such “precision medicine” trials is then to investigate the added value of a treatment algorithm to select the best treatment based on molecular abnormalities. The design of these trials raises numerous statistical challenges that we review in this chapter. We use the SHIVA trial as a running example to illustrate the various methodological aspects of these trials, to highlight the benefit of randomization, and to review the answers that can be expected from these complex trials as well as the pitfalls. In particular, we explore the power of randomized trials in case only part of the algorithm would be efficient, that is if only some targeted agents actually work in the presence of the selected target, while others do not. Finally, we present alternative designs and discuss their main features.
Fichier non déposé

Dates et versions

hal-02560954 , version 1 (02-05-2020)

Identifiants

Citer

Xavier Paoletti, Bernard Asselain, Christophe Le Tourneau. Designs for evaluating precision medicine trials. Pan-cancer Integrative Molecular Portrait Towards a New Paradigm in Precision Medicine, pp.113-131, 2015, 9783319221892. ⟨10.1007/978-3-319-22189-2_8⟩. ⟨hal-02560954⟩
30 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More