Lack of predictive tools for conventional and targeted cancer therapy: Barriers to biomarker development and clinical translation

Predictive tools, utilising biomarkers, aim to objectively assess the potential response to a particular clinical intervention in order to direct treatment. Conventional cancer therapy remains poorly served by predictive biomarkers, despite being the mainstay of treatment for most patients. In contrast, targeted therapy benefits from a clearly defined protein target for potential biomarker assessment. We discuss potential data sources of predictive biomarkers for conventional and targeted therapy, including patient clinical data and multi-omic biomarkers (genomic, transcriptomic and protein expression). Key examples, either clinically adopted or demonstrating promise for clinical translation, are highlighted. Following this, we provide an outline of potential barriers to predictive biomarker development; broadly discussing themes of approaches to translational research and study/trial design, and the impact of cellular and molecular tumor heterogeneity. Future avenues of research are also highlighted. Crown Copyright CO 2021 Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).