Optimal Duration of Immunotherapy

Transcript:Mark A. Socinski, MD: One of the issues that I struggle with in my practice is this issue of duration of therapy. What we know about immunotherapy agents is, in those patients that respond, they can often have durable responses that may actually go on beyond cessation of the immunotherapy. We struggle with how often you have to treat people in this particular setting. And then there’s the issue of pseudoprogression, which, I think, is probably overplayed in lung cancer. The question I’ll ask Dean about is, treating beyond progression and what your thoughts about that are. So, we’ll hear about that first. But I want to get people’s thoughts about duration of therapy and is there a role for it. For instance, with durvalumab, many of my patients are asking, “Do I really have to come every 2 weeks for forever if I’m responding?” And that’s a great question. But I don’t know that we have any trials that are addressing these issues. Treatment beyond progression?

Dean F. Bajorin, MD: That’s an easier question than how long we treat. Actually, we struggle with this in terms of our studies. If we take a look at response, we’re all using RECIST criteria for response. It’s 30% reduction in the sum of the diameter and then progression is 20% or more from the nadir. And so, some patients in whom you have a profound drop in your disease status or size, it doesn’t take much to get to the 20%. RECIST criteria were not developed for this, in terms of duration of therapy. In the renal cell CheckMate-025 study, patients who were treated with nivolumab and then had progression based on strict RECIST criteria were treated at least 6 weeks beyond that. It’s hard to do that type of analysis because you are selecting a subpopulation of patients. But what was interesting is that there are patients who went on and had regression later on because their bump in disease size was really not that profound. We’re really not talking about pseudoprogression. We’re really talking about what I think are minor changes in total disease size, which may be clinically insignificant. But, in terms of our clinical trials, disease size may represent progressive disease.

We’re seeing the same thing in the bladder cancer studies. We’ve actually seen some late responses at 8 and 9 months in terms of stable disease that fluctuated in between and could be interpreted as progressive disease. So, I think that it’s important not to withhold therapy too early and that using strict RECIST criteria may not be in the best interest of every patient. On the other hand, I don’t think that treating beyond clinical benefit is worthwhile. That’s going to really need clinical judgment.

Mark A. Socinski, MD: Yes. So, Jared, I think this is a huge issue in lung cancer, where many times I get calls about community oncologists wanting to hang on to the immunotherapy even though there may be progression, thinking that it’s pseudoprogression. But, at least in my read of the data, it’s more often progression than not. What are your thoughts about pseudoprogression, specifically in lung cancer?

Jared M. Weiss, MD: I think we need three major categories of apparent progression in lung cancer. The thing being discussed most often at the podium is this phenomenon of pseudoprogression; the idea that the cancer appears to be growing, and then later, wonderful things happen. This is much, much less common in lung or head and neck cancer than it is in melanoma. The vast majority of a clear apparent progression up front is real in lung and head and neck cancer.

The second pattern that we might see is mixed progression. So, you have a patient on trial. This actually isn’t new to us. We know this from the EGFR story. We know this from the ALK story, that you can have some lesions that are responding very nicely while others can be growing. And there, in the real world, we apply clinical judgment. If we have a central-growing lung nodule that’s pressing on a major airway, pressing on a major vessel where that’s really threatening to hurt the person, you have to switch therapy. You don’t just keep going with that. Whereas, if you have a small peripheral nodule that doesn’t have a lot of threat, you might keep going. And I think that pattern, that need to apply to clinical judgment, very much applies to the PD-1 and PD-L1 inhibitors. Whereas on the trials, and now in the clinical setting, we’re certainly seeing that pattern of mixed progression more often. And there are systems for trials that allow you to account for that.

The final pattern—we can’t ignore that this happens—is progression. The drug simply is not working, and that needs to be acknowledged, that quite a large amount of apparent progression is, in fact, real in lung cancer.

Mark A. Socinski, MD: Back to you, John, this issue of duration. I’m very tempted, in my practice, to give people breaks because they’ve had a nice response. They may not need it as frequently. I don’t know that any trials are addressing that issue. Certainly, in lung, the paradigm of four cycles, or so-called induction therapy, and then a maintenance strategy is very appealing. We do it in lots of situations. So, your thoughts on that?

John V. Heymach, MD, PhD: I think you’re raising a question that’s even more of a question than with chemotherapy. So, I don’t think we really think of maintenance chemotherapy, like pemetrexed, for example, as having residual benefit after you stop. But we know for immunotherapy, we’re inducing the adaptive immune system. And those T cells don’t disappear, they hang around. It is quite possible that we’re going to be over treating by treating for years, or you may be able to give a break. Then, if it wakes up, you can start treatment again.

I’ll give you an example from a patient of mine. He’s described his story in some newspaper articles in the Houston Chronicle and elsewhere; he’s happy about his story being told. He was a gentleman I treated in the KEYNOTE-001 study with pembrolizumab, and he had metastatic squamous cell cancer. He had been told previously to work out his affairs and get things in order. Because with metastatic squamous cancer after he had had first-line chemotherapy, he should expect a median survival of 3 to 6 months, which was not inaccurate at the time. And so, he went on the study and his tumor slowly shrunk, and shrunk, and shrunk. He’s an engineer, he actually was involved in the Navy, and so he would track quantitatively every single lesion, and the volume, and so forth. After 2 years, the lesions had just gotten down to tiny residual things, and he said, “Do I have to continue this?”

So, in the KEYNOTE-001 study, patients stop after 2 years and are given the option of continuing. But when you talk to people from the company and investigators, they said, “Well, we didn’t think that was something that was really going to be happening. We just sort of put that in there, and it is happening.” After 2 years we said, “Well, you’ve got a couple little lesions. Why don’t we just radiate them and see what happens?”

Mark A. Socinski, MD: Call it a day, yes?

John V. Heymach, MD, PhD: So, that’s what we did. A number of months ago, he just got back from a trip to Hawaii. He stops in and sees us, gets his scans, and just goes about his business now. He’s not getting any active treatment, and he’s now coming up on 4 or 5 years after he started his treatment overall.

Mark A. Socinski, MD: This is exactly the type of patient that raises this issue. And I think at some point, we have to deal with the duration-of-therapy question.

Howard L. Kaufman, MD: Well, we’ve actually had a lot of experience with IL-2 patients in the past. And there’s never been a very strong correlation with the total number of doses and the response. I recall a few patients that I had where they got one or two doses and said, “This is not for me.” They’ve gone on and done really, really well. So, I think the concept with the immune system is that if you’re going to get a response, you may get it regardless of how many doses we give. This is why finding a biomarker—particularly predictive biomarkers, which are a way to measure response—is going to be really important. One of the things that we’ve been using quite a bit has been PET scanning. And what we’ll often see, at least in the melanoma population, is the PETs will begin to show a decrease in the hyper metabolism long before the CT scan will clear. We’re finding this particularly useful in those kind of equivocal cases, where we’re not really sure whether we might be dealing with a pseudoprogression or a real progression.

Transcript Edited for Clarity