Biomarker Selection and Validation

Chapter 1 - Definition: Translational and Personalised Medicine, Biomarkers, Pharmacodynamics

Most prognostic biomarkers are not routinely used in clinical practice, as they are generally developed in unfocused clinical studies composed of samples of heterogeneous patients with available tissues. These studies, which are not specifically designed to address the clinical significance of a prognostic biomarker, do not result in validated biomarkers with clinical utility. In order for a biomarker to be clinically useful, it should correlate with tumour behaviour and/or treatment outcomes. Its use in guiding treatment decisions should lead to improved clinical outcomes. Since a particular biomarker may be useful in one disease but not in another, it is also necessary to define the disease in which it should be used. Selected biomarkers used in routine clinical decision-making for advanced solid tumours are listed in the table below.

Selected Biomarker Tests Routinely Used in Clinical Decision-Making for Advanced Solid Tumours

Modified from Bedard PL, Hansen AR, Ratain MJ, et al. Tumor heterogeneity in the clinic. Nature 2013; 501:355-364.

Tumour type Biomarker Prognostic or predictive biomarker
Oligodendroglioma 1p and 19q co-deletion Prognostic/predictive
Oligodendroglioma MGMT promoter methylation Prognostic/predictive
Medullary thyroid RET mutation Prognostic
Breast ER expression Prognostic/predictive
Breast PR expression Prognostic
Breast HER2 amplification Prognostic/predictive
Lung EGFR mutation Prognostic/predictive
Lung EML4-ALK translocation Prognostic/predictive
Gastric HER2 amplification Prognostic/predictive
Colorectal KRAS mutation Predictive
Melanoma BRAF mutation Prognostic/predictive
Gastrointestinal stromal KIT mutation Predictive
Gastrointestinal stromal PDGFRA mutation Predictive

In order to be clinically useful, a biomarker should be assayed in a specimen that is easily accessible in the clinical setting and can be readily obtained with standardised collection and processing protocols. Biomarkers that require serial assessments over a long period of time should be able to be collected in a minimally-invasive manner to be clinically feasible. The biomarker assay itself should be specific to the disease type tested and be reproducible, with appropriate cutoffs indicating the presence or absence of the biomarker. It is important for biomarkers to be validated in an independent clinical population that differs from the population used to develop the biomarker. Use of a particular biomarker should lead to a clinical decision that is linked to clinically meaningful outcomes, such as improvement in survival or quality of life or a decreased toxicity.

« Previous Page Next Page »

Last update: 03 September 2015