There are a number of things that need to be done to help lower the barriers and improve the success rate of drug development. Biomarkers—new biomarkers—are one of the most important things that we can strive for.
So over time, FDA has developed a process for the qualification, and they will say, “This is qualified for this purpose”—explains context of use and explains other things about the biomarker.
So when you think about qualifying a biomarker for a certain use, you have to think about a number of things. First of all, you have to define the use very well, so let me give some examples.
We want to select people for trials who are more likely to have the problem. So if you have more people having the problem and you’re trying to prevent a problem, then you’re more likely to find where they’re going to have an effect. That’s called an “enrichment marker.”
Another context of use might be a biomarker we want to use as an endpoint in a trial to find out whether the drug is working, all right, and so that one would be a different context of use.
For each context of use and each disease, there is going to be criteria that have to be set for how much evidence is good enough. And we call those the evidentiary criteria and we are working on developing those.
Some people wonder, “Why can’t we have one set of evidence for everything?” But it depends on the stakes of the decision. So you think about an enrichment biomarker: You want to have more events in the trial. Okay, if it doesn’t work very well, maybe you’ll just have a few more events than you would have before, right? But it’s not the end of the world. But if a biomarker you’re relying upon to keep people’s kidneys safe and it doesn’t work and their kidneys are harmed and you don’t know it, then that’s a terrible failure. And so the stakes of that biomarker being correct are very high, and so the evidence bar is higher than something where the consequences of a wrong prediction are not that serious.
The FDA has been working with a lot of outside groups, scientific groups and the general scientific community, on trying to set sort of general criteria. They are going to be different for a safety biomarker versus an enrichment marker versus a trial endpoint and so forth. But there are some common criteria, such as they all should have good analytical validity, which means the test works, basically: You don’t get a lot of false positives and false negatives and so forth. That sort of thing is pretty well worked out.