Abstract
With a focus on imaging, this narrative review describes why biomarker research rarely culminates in anything clinically useful. We explain the difference between diagnostic and prognostic biomarkers and the different phases of their research. We describe what factors need consideration when designing research studies to generate evidence sufficient to translate a potential imaging biomarker into clinical practice. We emphasize the need to balance the number of biomarkers investigated with the number of available patient events, so that studies are not overwhelmed by false-positive results. We explain the need for rigorous evaluation, so that clinicians can be convinced the biomarker will work in their practice and benefit their patients. In particular, we explain why novel biomarkers should be evaluated alongside factors already known to be clinically useful, so that their incremental benefit, if any, can be determined.