On the Road to Personalized Genomic Medicine in Oncology

This presentation will address the validation of prognostic and predictive biomarkers for identifying which patients require treatment and which benefit from specific regimens. I will distinguish analytical validation, clinical validation and medical utility. I will discuss key structural limitations of the vast majority of prognostic marker studies that render them therapeutically irrelevant and how this can be avoided. I will discuss how it is necessary to shift from a retrospective data-dredging “correlative science” posture to a prospective predictive medicine approach. I will present prospective clinical trial designs for co-development of new cancer drugs and companion diagnostics. I will discuss the prospective-retrospective approach to establishing the medical utility of prognostic biomarkers using archived tumor specimens from key clinical trials.

The opportunities for rapid progress to personalized predictive oncology are enormous today. The greatest challenges are only partly scientific. Maximum progress does require better understanding of the key oncogenic mutations and de-regulated pathways. But misunderstanding of the role and nature of properly designed clinical trials in establishing medical utility of diagnostics is also limiting progress.