For more than a decade, the marching order for drug developers has been that every drug candidate proceeding beyond a proof-of-concept study should be accompanied by a biomarker. Immuno-oncology (IO) is turning that notion of biomarker development on its head. In IO, the role of a biomarker goes well beyond identifying whether the drug target (usually a genetic mutation or rearrangement, in the case of a targeted therapy) is present in a given patient and whether the drug candidate can be delivered at a dose that allows for effective modulation of that target. The challenge in IO – especially for the current wave of drug development programs that look to combine PD-1/PD-L1 targeting agents with drugs that modulate additional targets, IO or otherwise – is in how to orchestrate a program efficiently in a highly competitive landscape without the benefit of the kind of binary measurement that is used prospectively to define and select a patient population for a targeted therapy. In IO, predictive tests will only be identified retrospectively after much guesswork around choosing the combinations to test: the tail will be wagging the dog. To paraphrase from the realm of political investigation, discussions about a successful biomarker development path in IO will turn on questions of what did you know and when did you know it.
As the initial wave of checkpoint blockade drugs establish themselves, led by the first PD-1/PD-L1’s Opdivo (nivolumab, from Bristol-Myers...
