Panel Reacts to CONTACT-03 Data Presented at ASCO 2023, Thoughts on Adjuvant Therapy

A roundtable discussion, moderated by Brian Rini, MD, of the Vanderbilt-Ingram Cancer Center, focused on treatment considerations for frontline and refractory renal cell carcinoma, including a discussion of new data presented at ASCO 2023. Dr. Rini was joined by a panel that included Tian Zhang, MD; Pedro Barata, MD; and Michael Atkins, MD.

In the next segment of the roundtable series, the panel discusses treatment options in the refractory RCC setting, including a review of CONTACT-03 data and other studies.

Watch the next segment in this roundtable series.

—

Dr. Rini: Let’s move on to the refractory setting. We’ve started to touch on sequencing. Tian, maybe I’ll talk with you. At ASCO, we saw results from CONTACT-03. There was a Lancet publication the same day, and this was basically cabozantinib/atezolizumab versus cabozantinib in an immunotherapy (IO)-refractory setting. Patients had to have IO as their most recent therapy. I think about half had ipilimumab/nivolumab if I remember. It was pretty typical for everything we’ve been talking about. A lot of ipilimumab/nivolumab, a lot of IO/tyrosine kinase inhibitor (TKI). I think fairly typical.

Very few adjuvant patients, which we can talk about; I think 1 in each arm. But it was really built on can patients get sequential immune therapy? Is it a benefit? Do they have what I would consider as another shot of cure with immune-based therapy? There was some data with lenvatinib/pembrolizumab in a single-arm study that looked pretty good. Response rates in the 50%-plus; PFS [progression-free survival] was 11, 12 months. When I saw that data, I was like, “Well, pretty good for immune-refractory patients.” Then the lenvatinib/pembrolizumab frontline data came out and maybe it’s just a more potent regimen from a tumor shrinkage standpoint, but be that as it may, and this was happening in practice. I’m sure we were seeing this and in roundtables like this, people would give IO/TKI after ipilimumab/nivolumab, etc. This was happening based on that lenvatinib/pembrolizumab data or other data. But now we have a big randomized trial that shows no difference. The cabozantinib monotherapy arm did tremendously well; we can talk about that, a 10-plus month PFS response rate of what 40%. All the curves are overlapping. The data is absolutely identical. It couldn’t be more negative. There’s not even a hint of some advantage. Importantly, there were 3 toxic deaths in the IO/TKI arm. The 2 drugs and the IO is not without consequence. Pedro, maybe I’ll start with you this time. Reaction to the data. What were you doing before and does it impact your practice?

Dr. Barata: I think that data is important. It’s not a coincidence that we see it out at Lancet Oncology as you know.

Dr. Rini: Lancet.

Dr. Barata:  Lancet, excuse me. It’s a practice changing study. I think that information is relevant. To your point, a lot of us were doing IO with a PD-1. Mainly with a PD-1 I think, until the data was out. Actually, it’s interesting because you go back when we start trying to figure out what was the role for CTLA-4 inhibitor, a lot of us were giving a chance at different lines of therapy, very late in the game. We quickly realized, well, not that quickly, but we realized that perhaps the right role for a CTLA-4 inhibitor is upfront. I think we understand that, but we didn’t quite know the same with the PD-1.

Now patients progressed, to your point, they progressed on a prior checkpoint and CONTACT-03 couldn’t be more negative. I think there’s a couple of comments that we’ll have from folks out there. The first is right now we have no data that supports the use of salvage checkpoint inhibitor if you’ve progressed on a prior checkpoint inhibitor. I’m not sure if you can apply that necessarily to patients who will be getting adjuvant more and more adjuvant therapy. They complete 1 year of immunotherapy with pembrolizumab and then you recur a year or 2 later. I don’t think we know that for a fact. But at least for patients who progress on frontline IO, I think right now we have no data to keep supporting that as moving forward. At least until a lot of trials like TiNivo will read out.

Now, there’s a couple of comments that we can make about the design. First of all, I think it was the first design to keep the same control, the backbone, in both cohorts. Other trials out there exploring salvage IO are comparing it against something else. This one compared cabozantinib in both arms. I think that’s important because it puts it into context. I think one point you just mentioned quite well, cabozantinib did perform quite well in the control arm. The other point is for whatever it may be, I mean we have data that perhaps asks the question, do we really believe in PD-L1 the same way we believe in PD-1s? I mean PD-1s are different from CTLA-4. I’m not quite sure PD-L1s are the same as PD-1. I guess if we were back, if were to redo that, we probably would’ve chosen a PD-1 instead of a PD-L1.

We don’t know that. Maybe the TiNivo [study] will help to answer that. That will always have that comment. My other editorial comment about that data would be, we haven’t seen that data. I believe there were some thoughts about it, is there’s a heterogeneous population. Some are primary progressors on IO. We did not exclude those patients on the trial. Those other patients can stay on therapy for 2 years and then go on trial. To my understanding, I have not seen a pure breakdown of flat negative for everybody or are we for the way we enroll in this particular trial, patients who went on second line CONTACT-03, they progressed within a year, let’s say, 6 months. Then you don’t see that because those patients are less likely to benefit from IO in the salvage setting. I’m not saying that’s the case.

Dr. Rini: Thinking about the forest plot, I don’t think there is any signal there.

Dr. Barata: I don’t think there’s a signal there, but at least hypothetically. Because that opens the question, are you done when you try checkpoint inhibition once or do you need a break from that MOA [mechanism of action] and to revisit that in the future? I don’t think we know that.

Dr. Rini: This trial doesn’t answer that.

Dr. Barata: That trial does not answer that. I think those remain unanswered questions for the future. But I will finish as I started answering your question, I do believe it’s a practice-changing trial because I do believe that’s going to give a lot of pause to folks out there to stop using an IO in the salvage setting for patients progressing on frontline IO.

Dr. Rini: Mike, what do you think?

Dr. Atkins: I agree. Very important trial for the reasons that Pedro said. I think a couple of points just so that people don’t over interpret it or misinterpret it. I think that it doesn’t mean that immune therapy won’t work in the second line. But we’ve always taken the approach that if you want to have immune therapy work in the second line, you have to add something that’s another immune therapy to it that’s effective. Then it’s a good idea to keep anti-PD-1 around. Because if you’re activating the immune system in another way, you want to make sure it doesn’t get shut off by having PD-1 expressed or PD-L1 expressed by that tumor. I don’t think the interpretation should be you only get 1 shot at immune therapy; you could say cabozantinib is not an immune therapy.

The second point is that the cabozantinib activity was better. I’ve heard people say maybe that was because there was leftover immune therapy and so you were actually giving IO/TKI even in the cabozantinib arm. But I don’t think that’s the case. I think there’s not enough consideration of how the second-line patients are different than the first-line patients. The second-line patients who’ve had prior IO, are relatively depleted in patients with aggressive disease who might have potentially responded in a major way to the frontline. If they didn’t respond, they never got to this trial. You have a second-line patient population that is more TKI responsive and that’s why the cabozantinib response rates are, you would call, surprising and better even than what you saw in METEOR and a little better even than what you saw in CANTATA.

Dr. Rini: You think the immune responsive subsets depleted in the angiogenic responsive therefore is-

Dr. Atkins: Yes.

Dr. Rini: …enriched?

Dr. Atkins: Yes.

Dr. Rini: That’s leading to the results.

Dr. Zhang: Well, to that point, I mean there was that cohort. This second-line therapy patient cohort, to your point, is quite heterogeneous. What they received in the frontline setting and how they responded in the frontline setting and how they’re progressing, timing of progression, where they’re progressing, those are all patient characteristics that as clinicians, we think about when we’re think selecting subsequent therapy. But as they’re enrolling into a trial that was designed this way, everybody is sort of randomized and treated the same way, which is great for trial design. But suffice to say that this cohort is very different. The ipilimumab/nivolumab-treated frontline patients, they were never exposed to a VEGF TKI in the frontline. This patient, that 30%, 40% of patients are not the same as the METEOR patients who had VEGF TKIs in the frontline and then went on to cabozantinib.

That is a TKI-naïve cohort that I bet if you looked at the stratification factors in the cohort, those were probably the patients who did really well on the cabozantinib control cohort. That drives the difference a lot less than we possibly could have expected if those patients were taken out of that comparison. I don’t believe we see that in the current data that’s public. But I think it’s an important trial. I was one of the proponents of using lenvatinib/pembrolizumab from the Memorial data and in the refractory setting.

But now that CONTACT-03 is out, I think thinking about the change out of mechanisms of action, I think is appropriate. If people have had ipilimumab/nivolumab upfront, they can do very well on cabozantinib monotherapy. I won’t add atezolizumab. Certainly, want to wait until TiNivo-2 reads out to think about that again. But that sequencing is an important one. One way, it’s a part of PDIGREE too, is thinking about that sequence of ipilimumab/nivolumab followed by cabozantinib. But then also if you take a break, so I wonder if you take a break, say they received axitinib/pembrolizumab or they received ipilimumab/nivolumab upfront, then they get cabozantinib alone, in that third line, could people, and they’ve had 9 months or 12 months for their T cells to repopulate and the exhausted T-cells to go away, could they be again responsive to PD-1? I wonder what you think about that.

Dr. Rini: I guess my opinion is it’s possible. I wouldn’t do it in practice because there’s no data to support it. I don’t know that that trial’s going to get done.

Dr. Zhang: How is that trial going to be done?

Dr. Rini: I don’t know that that trials’ going to get done. I’d love to see it.

Dr. Zhang: Yeah, multiple years.

Dr. Rini: We’ve been talking about TiNivo, which is tivozanib/nivolumab versus tivozanib. Nivolumab is a PD-1 inhibitor versus PD-L1. That may be a mechanistic difference. I know that did allow prior adjuvant therapy and there will be more. Because I know we put several adjuvant; I don’t remember if they had to be within 6 months. I don’t remember quite the parameter, but it will give us a little insight into that question, which we’ve been alluding to is well what if 2 years pass since your adjuvant therapy, we all think, “Well, we probably would start over with that patient.” But we don’t know.