Round 2: How should the breadth of pathway inhibition and regimen intensity influence treatment selection after CDK4/6 inhibitor progression?

The discussion turns to how broadly and intensely to inhibit the PI3K/AKT/mTOR pathway after CDK4/6 inhibitor progression. Panelists debate whether broad, multi-node PAM inhibition should be prioritized regardless of PIK3CA status, weighing potential gains in disease control against added toxicity and monitoring needs, and discuss when selective pathway targeting or a less-intensive regimen may be preferred.

To hear arguments related to treating patients with an actionable mutation, see Round 1.

To hear arguments related to route of administration and patient preference, see Round 3.

To view the full series, click here.

Seth Wander, MD, PhD: All right, round 2: How should the breadth of pathway inhibition and regimen intensity influence treatment selection after CDK4/6 inhibitor progression? And so, thinking about the PI3K pathway, we know that this pathway is very important for cell survival, growth, division, metastasis, motility, and this is true not just in breast cancer, but in other tumors. We know 40% of breast cancer is ER-positive and will have a PIK3CA mutation truncal early in the course of their disease, and another 5% to 10+% will have acquired alterations in the pathway, often downstream at the level of AKT or P10 or mTOR. So, signals outside the cell turn on receptor tyrosine kinases, they activate PI3K, PI3K turns on a bunch of things, including AKT and mTOR, and lead to all that bad stuff we talked about. 

Now, we've been targeting this pathway in breast cancer for many years, if you go all the way back to the development of everolimus, and we have alpelisib and inavolisib (PI3K inhibitors), capivasertib (AKT inhibitor), everolimus (mTOR inhibitor). But the unique thing here is that those drugs are all single-node intervention. So, think of it like a rung of a ladder; you're only going to one step of the ladder, versus the geda [gedatolisib], which is really hitting upstream and downstream. As I was just alluding to, we have data from earlier development of PI3K inhibitors that P10 loss or disruption and subsequent activation of AKT and mTOR is a really important mechanism of resistance to first-generation PI3 kinase inhibitors. 

And so, that may be one of the important reasons why this drug acts differently, for example, than some of the agents that we've been used to having. So, we're going to go through this question of how should we inhibit the PAM pathway? Now, one of the things that I was thinking about for this debate as we were designing it was we have the question I gave you, but then there's another way to translate that question. So, I'm going to take the liberty of translating this a slightly different way. When geda eventually has a label that encompasses both mutant and wild-type, would you give capivasertib with a 7+ month median PFS (oral, 4 days on, 3 days off) or gedatolisib with an 11- to 12-month PFS in the mutant, (IV, 3 weeks on, 1 week off)? And this is not an abstract philosophic argument like the one we just had, this is a discussion you are going to have in clinic very, very soon. And so, as you guys are going through this, I'm curious how you're going to fall on that point. 

VK Gadi, MD, PhD: So, we just saw a version of this moments ago and I'm just going to bring it back for a moment. I think you can appreciate the complexity here, and this isn't even the full truth. There's many, many more molecules in these pathways. And we think about, and we search for oftentimes what we call backbones in therapy, something that we give everybody and then add something else too. And that does a disservice in breast cancer because it's actually the entire lumbar, it's the entire spinal column with 3 regions. And if you think about it, at the base of that, the lumbar spine is probably your endocrine therapy. And for the longest time it's been aromatase inhibitors. Keep in mind that even though these oral SERDs are being approved for patients with ESR1 mutations, they work also exceptionally well in patients who don't have mutations. So, you don't give up anything. And it's only a matter of time before those therapies are available for unselected patients, regardless of what's going on with ESR1. They're just better drugs. 

The next component of that spine is the thoracic, and we've been able to target that with CDK4/6 inhibitors for the longest time. Cyclines matter, and they seesaw back and forth between the other mitogenic pathways in these cancer cells, you must always have to deal with them. And you've seen that when we only deal with ESR1 mutations and then don't give a CDK4/6 inhibitor, the curves go straight down and then you find a group of patients who but for that ESR1 mutation would have been fine. But you give the CDK4/6 inhibitor and that line gets up a little bit more. Now, the next component of this is of course the cervical spine, let's talk about that. That's this PAM pathway. And we've been targeting pieces of it while not targeting all of it. And here we have a drug, gedatolisib, that we're going to look at in a minute, that nails all of that entire pathway in some way. 

None of these are perfect molecules, but they're what we got right now. And when you comprehensively shut down this pathway, you've really given patients a backbone on which you can then explore what other things are going on. So, this is potentially the new standard. And so, we're going to look at the other side of the house here, which is the patients who actually have a mutation detected, this is just a newer study, so I'll walk you through it just a little bit. Three arms, randomized 3 to 1. So, one of the arms is a little bit underpowered and that's a doublet here. Here, instead of giving fulvestrant alone, we're giving alpelisib standard dosing there and then the gedatolisib dosing of 3 weeks on, 1 week off. Now, funny thing about this trial, the entire VIKTORIA-1 trial, both the wild-type, or not-detected cohort, and the mutation cohort had an A1C limit of 6.4%. 

Reflect on that for a minute. That was actually requested by the FDA because alpelisib was a comparator. It had nothing to do with what we were expecting to see with gedatolisib. You'll see that borne out in the toxicity data. While we do see some patients with grade 3 hyperglycemia, it is not nearly at the rate that you see with the oral agents. Otherwise, traditional endpoints of primary progression-free survival by blinded independent central review, OS, response, safety, et cetera. So, this is the top-line data for [the wild-type cohort], 9.3 versus 7.4 months, respectively, for the triplet and the doublet, and then you see the OR rates of 31%, 28%. Fulvestrant by... Sorry, this is the wild-type cohort. 

Wander: Wild-type. 

Gadi: Sorry, wild-type cohort. You're looking at the triplet, doublet, et cetera, against fulvestrant by itself for the efficacy. Now, here's the mutant cohort. Sorry about that. And then, you're looking at alpelisib/ fulvestrant, about 5.6 months, we're doubling that with both the doublet and the triplet here. Interestingly, the doublet really does look like it's performing well, but it is an underpowered group, so the estimates around that are a little bit bigger. It's possible that the triplet is still the superior combination here. But either way, it gives you some optionality if this regimen is approved for patients who may not be tolerating that CDK4/6 inhibitor, you have an option not to give it. 

So, really, yes, this is a more intense regimen in the sense that there's many drugs here involved, and it's complicated. You've got a poke in the butt, you got a pill in the mouth, and you've got an infusion in the vein, that is not ideal, and it's acknowledged. But it may be the best therapy we've got right now. And it's a conversation that I think we can have with all our patients, saying that this triplet is likely to get you the response you need for the longest amount of time. And really you have some optionality to drop some of it or play around with it, if patients can't tolerate it or just have compliance issues around trying to stick with it. So, really the last case here is we are oddly enough going old school here. In general, we've been sequencing and then giving therapies based on what we learn, but here you have something that I've been terming biomarker agnostic. 

You don't got to know, you can just start, and then you can kind of keep this as prognostic as to whether or not a patient's going to respond for a long period of time or not based on things like the presence of a PIK3CA mutation. But it makes things easier in some ways to just not have to think about it. And when you're a primary oncologist, 95% of patients are taking care of by such doctors, it's kind of nice to not have to go back to the sequencing and try to figure out what's going on. 

Paolo Tarantino, MD, PhD: All right. So, I like going back in time, I have this time machine that I always for every round like to take, and this time the time machine is to Oncology 101. Why we actually do what we do, why we discuss with patients drugs when they're in clinic, we do that for two main reasons. The first one is to improve overall survival, to make them live longer. The second one is to improve quality of life or maintain quality of life, to make them live a good life, and make them do the things they like in life. Things we do not try to do in clinic is to shrink the tumor, eradicate the tumor, hit the disease hard, try to block every pathway. These are just collateral things, surrogate things that hopefully lead to what we want, improving overall survival, maintaining or improving quality of life. 

And so, whenever we try to achieve any of these goals, we have to check what is the minimal burden of treatment we are willing to offer to the patient, and to have the patient accept to achieve these goals because increasing that burden may actually take us away from what we want, which is maintaining quality of life and extending their lives. And sometimes that just takes monitoring a mutation and switching therapy, like in the SERENA-6, which was shown to improve quality of life, of course, but also sometimes it's just looking at the results of a trial and being critical about them. Because you can look at the results of VIKTORIA-1 and think, okay, there's 3 numbers here. There's 2 months, 7 months, and 9 months. Well, 9 is better than 7, let's go for the triplet. Well, turns out that oncology is a bit more difficult than this. 

There's so many moving parts, there's statistical assumptions, for instance. This trial was not powered to show that the triplet was better than the doublet, they were just tested in parallel. Both of them were better than fulvestrant, as you can obviously see, but the confidence intervals of that progression-free survival are overlapping. And so, that 7 could be 7 in this experiment, but if you do the trial again, it becomes 11. It's just like, there is a certain uncertainty that we have in oncology, but what is not uncertain is the burden of toxicity. Once you add a third drug, for sure you add toxicity from that drug. You add neutropenia for palbociclib, you add complications in the logistics. The patient has to remember to do 3 weeks on, 1 week off, and then come back, and the labs show that there is great for neutropenia, you have to hold, but continue the other drugs... It's just more complicated. It adds costs because let's not forget, each drug comes with its own costs, and inevitably it has complication. 

And so, are we really sure that these data warrant that additional complication? I'm not sure. I really feel that we are sure here that gedatolisib adds a lot, because both arms that include gedatolisib show a major improvement in progression-free survival, but in this cohort and in the PIK3CA mutant cohort, even better data as we have seen. But the palbociclib, I think it's still totally unclear. And so, if you go back and you look at the [inaudible 00:11:10] data, you will see that in truth, here, 11 months, 11 months, because it's another experiment and suddenly the triplet doesn't add much. And so, I do feel right now, palbociclib in the second line, we've never considered it that much because it has negative data in the second line. In the PACE trial, in the PALMIRA trial, in several trials, BioPER, palbo didn't add anything in the second line. 

I struggle to think that suddenly, if you couple it with another drug, suddenly palbo shines in the second line. I feel there is some uncertainty in the statistics of trials. Here you have a 2-month improvement, which may be there or not, but based on the burden of side effects, I think I want to actually make gedatolisib here become as easy as possible to give to the patient so that I can extract all of the activity I can from that drug, and avoid adding the palbo, both in patients with PIK3CA wild-type and mutant disease. So, I'm definitely for the doublet, and this is fortunately now available, and both the triplet and doublet by the FDA are available, but I definitely vote for the doublet. 

Wander: Great. Yeah, it's an interesting question too, that we don't have an answer for, to Paolo's point, looking at that data with all the caveats you just mentioned, the inclusion of the CDK inhibitor seems to mean more in the wild-type than it does in the mutant. And biologically, we need to do some more work to understand that. I mean, to me in the wild-type, I agree with the point you're making about added toxicity, cost, et cetera. If you have a patient who's losing response to a CDK inhibitor because of activation of AKT/mTOR, and then you block the AKT/mTOR, maybe that's where you're seeing reintroduction of the CDK helping you, in theory. Why is that not the case for a PIK3CA mutant patient is another layer of complexity that I think we need to do more work in the lab to figure out.