ctDNA as a Predictive Biomarker in Patients Treated with Pembrolizumab


Two medical oncologists summarize data on use of circulating tumor DNA as a predictive biomarker in solid tumor patients treated with pembrolizumab.

George Ansstas, MD: Welcome, everyone. I’m Dr. George Ansstas. I’m a medical oncologist at Washington University, and I’m the clinical director of melanoma. It’s my pleasure to present this article through CancerNetwork® Between the Lines. Today, our featured article is about personalized circulating tumor DNA analysis as a predictive biomarker in solid-tumor patients treated with pembrolizumab. It’s my pleasure to have Dr. Firas Badin from Baptist Health in Kentucky. Dr. Badin is a medical oncologist, and he can have the stage for him to introduce himself further. Please, Dr. Badin.

Firas Badin, MD: Thank you very much, Dr. Ansstas, and it’s also my pleasure to be here with you and our audience. I’m the medical director for oncology research at Baptist Health System in Kentucky and Indiana. My focus is thoracic oncology. And I’m very happy to be talking about 2 topics dear to my heart: immunotherapy and circulating tumor DNA. Why is that? Like most of us in the oncology clinic, we use immunotherapy and immune checkpoint inhibitors. They have a wide range of indications and different tumor types in my thoracic clinic, your dermatology or skin cancer clinic, GI [gastrointestinal], and GU [genitourinary].

So we’re using checkpoint inhibitors in wide tumor types. We’re using them in different settings: first line, second line. We’re using them as a single modality or in addition to chemotherapy. However, it’s not always easy to assess response or evaluate outcomes for patients on I/O [immuno-oncology] or checkpoint inhibitors. Sometimes, tumor markers are not secreted. A lot of times, we follow a CEA in colon cancer, a CA [carbohydrate antigen] 19-9, or different tumor markers. But sometimes we know the tumor may not secrete a cancer antigen. We know there are other reasons why the cancer antigen might be high. It’s the same thing for radiology; a lot of times, we count on CT and PET [positron emission tomography] scans. But in my thoracic clinic, for example, the other findings may not mean, necessarily, progression. Pneumonitis, scar tissue from prior radiation, pneumonia—they really might mimic disease progression. We’re all familiar with pseudo progression when it comes to I/O response.

The bottom line is that it’s not always clear what’s happening. That’s why I’m excited to talk about this study, the INSPIRE trial, published in the Nature journal. It has taken it to a completely different level. Here, we’re not only counting on imaging or serology but taking it to a personalized circulating tumor DNA. Personalized is important to highlight. There are too many tumor assays that look over circulating tumor DNA or fragments of cancer cells and blood. But we’re using a personalized assay based on matching a tissue, as well as with the patient’s blood, and looking with that, guiding our approach that reliably can evaluate response or correlate with survival.

It’s a very simple phase 2 study, about 100 patients. It’s a basket trial, so there are different tumor types. We included 25 tumor types in the study. As you can see here on the slide, we have different cohorts. Those are patients who progressed on their treatment, and they don’t have any further treatment options, or they’re not candidates for further treatment. They were enrolled and received Keytruda [pembrolizumab] at the FDA-approved dose, at 200-mg IV [intravenous] every 3 weeks. Those patients were followed, as we usually do, with pseudo imaging. What’s interesting is that the study captured their circulating tumor DNA levels. They did that at baseline, as well as every 3 weeks, up to 6 months. Then they looked at outcomes. They looked at response rate, progression-free survival [PFS], and overall survival. As you can see on the slide, there’s a very clear correlation between outcomes, as well as circulating tumor DNA.

I’m going to present data here at 2 time points. We can look at circulating tumor DNA level at baseline and at 6 weeks. On this slide, as you can see, we have overall survival as well as progression-free survival. You have 2 groups of patients. The patients who had a higher level of circulating tumor DNA at baseline did clearly worse than patients who had a lower level than median of circulating tumor DNA at baseline. When it comes to overall survival, as well as when it comes to progression-free survival, you see the wide split between the 2 groups of patients, so it’s very interesting. What’s also very interesting, if you look at the 6-week time point and the change in circulating tumor DNA, is we’ve learned a lot from this study. If patients didn’t decrease their circulating tumor DNA very early, by only 6 weeks, those patients 98% of the time aren’t going to respond to Keytruda.

Furthermore, if you look at the survival data, as you can see on the slide here, if you look at the left side, we have the overall survival. On the right, we have the PFS. If patients didn’t also decrease their circulating tumor DNA, their outcomes are worse. For patients who have a decrease in their circulating tumor DNA, they did significantly better. Their overall survival curves, as well as PFS curves, are better than patients that their circulating tumor DNA did increase by 6 weeks. Also, there are more interesting data that we’ve learned from the study, especially when it comes to correlation between circulating tumor DNA and imaging, and evaluating response by imaging. Dr. Ansstas, what do you think about patients who cleared their circulating tumor DNA? How did that correlate to their survival and their outcomes?

George Ansstas, MD: It’s a very interesting observation, Firas, about what we see. If you look at the blue dashed line on the graph on the left, the A, and you see patients who cleared their circulating tumor DNA, all these patients are alive at almost 36 months. It’s the best biomarker we have to predict that kind of durable response and longevity for patients. If you look at the 2 other lines, if you clear some of your circulating tumor DNA, you definitely do much better than those who don’t clear.

Firas Badin, MD: I find that to be very interesting as well, George. I love the clear cut, the rapid and persistent decline of circulating tumor DNA vs. the stepwise, so to speak, response on imaging with CT scan. Also, as you alluded to, the pseudo progression phenomenon. I see that a lot in my thoracic oncology clinic. Sometimes, it’s just not clear. You see some progression, and you don’t know if it’s true progression or what’s going on. But then if you have the aid of circulating tumor DNA showing that it’s declining—that would be very reassuring—and how rapid and how clean it is vs. fluctuation when it comes to serology or radiology.

George Ansstas, MD: I cannot echo that more. It’s a lesson to learn for all of us. Pseudo progression is very demanding intellectually for all of us, especially those who deal with immunotherapy. Many times, we might see that progression on the scan, and we don’t know the significance. With such data, it will be relieving for all of us to know if the circulating tumor DNA is going downward or we’re having a clearance or decrease in the number. Likely, this is a pseudo progression phenomenon, and that would make all of us continue on the treatment and not change the course.

I echo what you just said on graph B, the 1 in the middle. You see how this steep drop of circulating tumor DNA occurs in patients way before what we see in imaging. That could be an early sign for all of us who deal with these patients to get a sense how their course is moving. Is it like those patients intended to respond to treatment? Are those patients in a stagnant phase? Such a big, steep drop, and quick as well, can give you a sense how your patients might do in the longer term.

Firas Badin, MD: What a wonderful segue to go over the next part of the study talking about circulating tumor DNA, as well as evaluating response by imaging. It’s easy when they go hand in hand: if the patient is responding by imaging and the circulating tumor DNA is doing the same, or if they’re not responding by imaging and circulating tumor DNA is increasing. But that may not happen all the time. What about the other patients? Those 2 things are going in different directions.

George Ansstas, MD: As you see here, the line of interest to me is the blue. If you have concurrence between increased circulating tumor DNA and progressive disease, you do the worst. But what has interest to me, if you have increased circulating tumor DNA but complete response, partial response, or stable disease, most of us—when we see such data—get more assured that these patients might do well. But if you look at the graph here, the green 1, those patients are doing similar to what’s seen with the blue line with those who’ve increased circulating tumor DNA and progressive disease. Sometimes in the clinic, we’d be falsely assured on image that somebody has stable disease, or a response that these patients might do well in the long term. However, if we see that contrast between increased circulating tumor DNA and their images, it might be a falseassurance, and those patients might not do well. For me, as a practicing oncologist, with these tools, I might be more prepared, aware of what I might be anticipating for many of these patients. Lastly, if you look at the red line here, both go hand in hand, like if you’ve decreased circulating tumor DNA and some response or stability of disease, those patients would fare very well. They do the best among all patients. This combination of imaging and blood tests can give us a better sense of how well these patients might do.

Firas Badin, MD: I’ve learned 2 things from this slide. No. 1, circulating tumor DNA might trump our conventional methods, whether it’s imaging or serology. We’ve seen clearly that circulating tumor DNA, if it does decrease, regardless of what you’re seeing on imaging, whether it’s progression or any response or stable disease, you’re going to end up having a better outcome. The second thing, the 1 you just mentioned, said very easily and very clearly that we may not deal only with pseudo progression but also pseudo response. You might think the patient is doing well, or the patient is gaining benefit and responding by imaging, or stable disease by imaging, where their tumor is progressing. That’s reflected by increased circulating tumor DNA. It’s something I’ve learned myself on my trials when I’ve also known about pseudo progression. I’ve also learned about pseudo response when it doesn’t really correlate with improvement in circulating tumor DNA.

George Ansstas, MD: Correct. Firas, on this subject, we know from many of the Natera Signatera trials, the positive predictive value for relapse, if you have positive circulating tumor DNA, is almost 98%. That goes with the same notion about this green line, or group of patients or cohort of patients, how this circulating tumor DNA trumps the imaging in terms of outcomes. Because technically, we know that patients who have positive circulating tumor DNA are likely to clearly to declare their disease shortly after. I agree with you. This is a learning lesson for many of us about the importance of knowing the status of circulating tumor DNA in the prediction of outcomes in many patients.

Firas Badin, MD: Absolutely. In summary, INSPIRE trial was a simple, straightforward study: 100 patients, different tumor types, progressive and conventional treatment, received immunotherapy with Keytruda, followed as you would usually follow in any clinic. In addition, we used personalized circulating tumor DNA and looked at outcomes, response rates, and survival and progression-free survival to see if that might correlate with what’s happening to the tumor DNA and blood.

Transcript edited for clarity.

Related Videos