John L. Marshall, MD: This is a new technology. We’re learning how to apply it, [it is] not always approved and not always on pathway. Using these technologies in these clinical scenarios, they need to show their value and worth. Because these are not inexpensive tests to run and they have turnaround time. So can they make us better doctors, basically, can they make us more efficient with our therapies? One new poll is looking at using ctDNA [circulating tumor DNA] levels, if you will, to tell us whether a patient’s responding or beginning to progress. You know what we do, we do a baseline CT, or some kind of picture. Usually 2 to 3 weeks later, the patient gets their first treatment. We then treat them for some period of time, sometimes it’s 2 months, sometimes it’s 3, sometimes it’s 4, using different measures like tumor markers, patient’s symptoms, and adverse effects to decide when to do another test. Sometimes it’s the insurance company that’s telling us when we can do the next test. Then we measure, did the tumor get better or not. Now, you could envision, let’s say in 3 months that the cancer got better. But by the time 3 months came around, it was starting to go back up and get worse, but compared to the baseline, it looked better. So we all smiled at each other and kept going for another 3 months, and in the next 3 months that treatment didn’t work. We could also see a time when we’ve given treatment for a month or two, and the patient seems to be doing OK. But we don’t really have any measure to prove if this has been worth it. Should I give a third month of treatment or not?

Using these kinds of molecular and more sensitive blood tests that are reflective of the patient’s tumor DNA, could we in fact predict better efficiency? This analysis in this case, using a Guardant assay, showed that we could be more efficient, that we could show a longer benefit of treatment, that we could turn off treatment more quickly when it wasn’t working. I do believe that we will increasingly be moving to these kinds of tests, getting away from our traditional CT scans, so that we have a much more real-time impact on what we’re doing. Now, the tests are evolving, as I said, but I think you’ll see more and more of this as a more efficient way to monitor the benefit of our treatment compared to scanning, etc.

One of the other major approaches that we talked about earlier is the concept of minimal residual disease. Is there still cancer there or not? We’ve all been seeing an explosion of data, a lot of it out of Japan, some of it out of other places as well, looking at postoperative analysis, that yes or no, you still have tumor DNA floating around. If yes, you should receive adjuvant therapy, if not, maybe one day we’ll get comfortable enough to say no, you don’t need adjuvant therapy. One of the problems with this test is the long turnaround time, when should you draw the baseline blood, etc. Most of us didn’t learn this in medical school, but right after surgery with all the healing going on, there’s a lot of human DNA in our bloodstreams. So the question in one of the abstracts that was presented is, can I find the small amount of tumor DNA in what is now a much bigger haystack of DNA right after a surgery? They looked at whether you drew it at 1 week, 2 weeks, 3 weeks, etc. Most of us have been drawing these assays between 3 and 6 weeks after surgery, but before starting adjuvant therapy. This basically said, yes, we could find it even at 2 weeks after surgery. But the recommendation was still to let some of that big haystack clear out before drawing the blood tests. More and more of these assays are being done, led by colorectal cancer, but are being applied to many other disease sites as well. So stay tuned for more data. And as I said, my prediction will be that this will quickly become incorporated into the standard of care, as we get more efficient and understand how to use it.

First off, really, nobody has an in-house laboratory for this. And those that do are probably not performing up to the level that some of these private companies are doing. It’s how they’re staking their claim. It’s like, do you have a homemade CT scanner, or do you buy one from Siemens? You buy the one from Siemens because they know how to make CT scanners. This is where we are with this technology. We’re using it a lot. We were early adopters here, particularly in the colorectal space. But we’ve begun to do this also in other GI [gastrointestinal] tract cancers. One of the criticisms that’s been levied at us at our site and others that are doing it—we’re not alone, lots of people are doing it—is what are you going to do with the data? And it’s just like knowing bad news even sooner, can you do anything about it?

So we understand that right now, we don’t have any action to be taken if you’ve already been given adjuvant therapy, and the ctDNA is still positive. One of the things we are piloting at Georgetown [University] is a clinical trial looking at an immunotherapy combination in patients with GI cancers who have had curative therapy, they’ve had all the adjuvant therapy we think we should give, and yet the ctDNA is still positive. We’re just looking to see if these novel approaches using novel therapies can flip the ctDNA to negative. We’re preparing for what we think will be the future of tailored adjuvant therapy, so that you won’t give the same thing to everybody. There’ll be a group of patients you don’t treat at all. There’ll be a group of patients you give today’s standard, because we know we’re already curing those people. And there’ll be a group of people we will be giving novel therapies to, and so we’re hoping to be out in front of what those novel therapies might be.

Transcript edited for clarity.