NSCLC: Increasing the Value of Molecular Testing

Transcript:

Gregory J. Riely, MD, PhD: As we identify more and more very uncommon mutations that happen in 1%, 2%, or 3% of patients with non—small cell lung cancer, it becomes more challenging to develop clinical trial programs so that we can understand how best to treat these patients. The way we do this is currently conceptualized as 2 approaches. One is a basket trial, where you take all patients with the same mutation, put them on a trial, and start to separate them out based on tumor histology: for instance, lumping together patients with KRAS-mutant lung cancer and those with KRAS-mutant colorectal cancer. KRAS is probably not an example you’d use because it’s relatively frequent, but the NCI-MATCH study is an example where patients who had ALK gene amplification were, with whatever histology, enrolled in a trial to get crizotinib. We have seen a number of trials like that that really focus on 1 particular entity and try to target that entity regardless of tumor histology. These are generally called basket trials.

Another approach is to have what’s referred to as an umbrella trial, where you have 1 big trial that tries to incorporate everything that you might find. You enroll that patient in this umbrella trial and you do molecular testing. You find the patient’s molecular phenotype and then you divide those patients based on that into different arms of the trial. I think whether you talk about an umbrella trial or a basket trial, they’re both trying to address the same problem, which is, how do we deal with these very small groups of patients? How do we best test them? I don’t think there’s a right answer, but I think these novel trial designs are helpful, and hopefully we’ll be able to learn more going forward.

When we identify specific mutations as part of a broad-based mutation testing program, sometimes it’s clear what we should do. If we identify an EGFR mutation, we know that patient has a standard-of-care approach. But sometimes we identify things that are very hypothesis generating, or early in their development as a drug biomarker. The best thing to do with those patients is to try to enroll them in early-stage trials, whether that’s a phase I clinical trial program that has new drugs, or an early drug development program that more generally has specific drugs for specific mutations. I think it’s critical that we all develop a plan so that when we do broad-based mutation testing, we know what we’re going to do when we find these rare entities.

It’s become much more complicated over the years as mutation testing has become more complicated. Getting it paid for has also become more complicated. We’ve seen a variety of approaches from individual payers as well as Medicare that determine how a drug, or how a testing platform, should be paid for. I don’t know the details of all this, but I would note that recently CMS [Centers for Medicare and Medicaid Services], the Medicare provider, has identified broad-based mutation testing by a variety of platforms as something that they think is reasonable to pay for. I think we need to learn more about what those platforms look like and which platforms qualify, to help us understand which ones should be used in our patients.

Transcript Edited for Clarity.