Adaptive Cancer Therapy: Managing Drug Resistance & Treatment Sequencing

•Low dose

•Drug combinations

•Sequencing (not a combination cocktail)

•Mathematical simulation models Please see the notes from the last meeting for details. Discussion

•Drug combinations : Everyone agreed that drug combinations that have a greater fit with identified biomarkers are preferred, but it is difficult to get oncologists to prescribe them. There should be trials that offer a couple of drugs, including investigational drugs. However, even if you want a combination of two approved drugs, they're considered a new drug (thereby lacking evidence) if you offer them together.

•Sequencing: While sequencing drugs makes sense in theory, there is little evidence to support it from traditional clinical trials, while there is evidence that combination drug cocktails provide better patient outcomes.

•Mathematical simulation models : Saed Sayad pointed out that the models that Bob Gatenby was using were very simple, and that models today are taking more variables into account, such as DNA, mRNA, and proteomics.

•Obstacles: Emma Shtivelman noted that physicians won’t prescribe drug combinations because there are few trials that have tested drug combinations. Even if each of two indicated approved drugs would provide a better outcome, physicians won’t prescribe the combination without clinical trial evidence. There should be clinical trials of drug combinations, including investigational drugs, but the obstacles are almost insurmountable. The combinatorics of a personalized treatment using multiple drugs at multiple doses and different sequencing choices make it nearly impossible to use a randomized clinical trial to derive supporting evidence. Requests

•Do you have any comments on adaptive therapy?

•This adaptive strategy seems intuitive, yet it’s not widely practiced. What are the barriers or objections to it? The information and opinions expressed on this website or platform, or during discussions and presentations (both verbal and written) are not intended as health care recommendations or medical advice by Cancer Patient Lab/Prostate Cancer Lab, its principals, presenters, participants, or representatives for any medical treatment, product, or course of action. You should always consult a doctor about your specific situation before pursuing any health care program, treatment, product or other course of action that might affect your health. — Meeting Notes Brad Power: Today we're going to be talking about the presentation that Bob Gatenby made a week ago on adaptive therapy – using evolutionary and game theory to strategize about how to manage treatments. For those of you who weren't there, or didn't see the notes of the meeting, there are basically four pillars: drug combinations, low doses, sequencing, and mathematical models. Saed Sayad: We were talking about old fashioned modeling, which is using mathematical equations to measure the change of the concentration of drugs in our system, in our blood. Based on that model, we can decide that when we reach 50% of the concentration, we can increase the dosage. This is kinetic modeling. This type of model is very limited because we now have huge amounts of data which are mixed data. We should be able to use those data and find the interaction between proteins and genes. And then instead of the simple mechanical mathematical model, we can use machine learning, data science, and predictive modeling to include many more factors in our equations. Brad Power: In “The Signal and the Noise ”, they talk about weather forecasting, and how the models get more sophisticated and the data gets more sophisticated. What I hear you're saying is that the standard of modeling today is to have many more variables in the model, instead of just a handful? Saed Sayad: It’s the variety of elements in the model. It means it's not just a simple concentration of a drug; rather, it is about the interaction between different components of the cell, like the DNA, mRNA, and proteins. Brad Power: Integrating all the different kinds of information you could be bringing together, such as different kinds of -omics, as well as medical history, has been a big theme for Brian. Emma Shtivelman, you put something in the chat last week about bipolar? Emma Shtivelman: I asked about bipolar androgen treatment. It's sort of taking the evolutionary approach that has been proposed to an extreme. Not only do you do short time treatment in this case with androgen deprivation, you then flood the patient and the tumor cells with testosterone, hoping that this will prevent the development of resistance. The cells that were kind of responding to androgen deprivation will now flourish and the resistant clones will be pushed out of the picture. Dr. Emmanuel Antonarakis is a big proponent of this approach. He reported some successes, particularly in patients who have mutations in the DNA damage repair pathway. It's literally case reports, not big studies, and after bad deprivation. He reported that patients in several cases responded well to immune checkpoint drugs. The first report was maybe nine years ago or so. Brad Power: Pete Kane, do you have any resources or people you've run into that might be able to help us with simulation models? Pete Kane: Possibly. I'll give it some thought and see if I can make some introductions. Certainly there's been a wide cast of characters that we've encountered. Brad Power: Ally Perlina, we have been having a conversation on drug combinations. We agree on that pillar of Bob Gatenby’s approach. But you challenged the notion of administering the drugs in sequence, rather than as a cocktail combination all at once. That's a very valid question. I can represent Bob Gatenby's argument: He's saying that if you knock the population down successively, that's a way to get an extinction event. But if you hit the heterogeneous population all at once, up front, when there's a large population, you're going to get some resistant strains that are not going to respond. If you hit it with a combination, you'll get a better response. Progression free survival and the other metrics will look good, but you won't actually kill it off. There will be some resistant strains that are lurking there in the background, and they will eventually come back. When he talked about Brian's case, he said, “Brian, you have a low tumor burden. There's a low tumor population. You can go for an extinction, knockout blow if you take three drugs in succession.” If Dr. Gatenby looked at the CureMatch drug combinations, I think he would want to do them in succession, not as a cocktail. I may be misrepresenting him, but I think that's the argument. Ally, you were saying there isn't any evidence to support that. Ally Perlina: I didn't state that. I was asking because at a quick glance, we couldn't find any papers, but we may not have been thorough enough. I was asking if there's any clinical evidence to validate this theory. Saed Sayad: I read the paper. They said that not only is sequencing the drugs important, but also the order. They were using doxorubicin and other CTL immunotherapy drugs. They showed that if you change the order, the effect is going to change. In many cases you can make a cancer cell uncomfortable, instead of killing it. Based on my research, this is a very new field. There are many questions, and few answers. Brad Power: There's a specific example we've talked about. Rick schooled me on the notion of radiation turning a cold tumor hot, and then being responsive to immunotherapy. That would be a sequence of radiation followed by immunotherapy. Rick Stanton: That's our hope. We looked at our immune deconvolution as assessed by Tempus RNA seq data, and we have very cold tumors. We have some evidence of cytotoxic T- cells in perforin and granzyme A (perforin and granzyme cooperatively induce target-cell death) that contradict that there's no CD8 T-cells. Nevertheless, it looks like we have pretty much no CD8 T-cells. Immune modulation without getting TILs into the tumor microenvironment is probably not going to work. We need a prerequisite, which in our current thinking is radiation. It might be Pluvicto, which would come in and basically carpet bomb anything that has PSMA (prostate-specific membrane antigen) on it, and that would awaken or heat up the cold tumor. That would be perhaps the first step in what would be a great sequence. The second step would be immune modulators. I'm very curious about what CureMatch and Ally bring to the table. When I first saw the Tempus report with a super high BRAF expression for Brian, I took it with a grain of salt because I didn't get that section on overexpression in my Tempus report. Looking at the transcripts per million data that's coming out of Tempus for Brian, it was really apparent that his BRAF is off the charts. Either they made a mistake or that's very reportable. That brings me into the whole concept of BRAF. Nobody is really hitting BRAF for prostate cancer. Why would this make sense? BRAF is a signaling molecule, or protein inside a cell cascade. It doesn't mean that you have activated or phosphorylated BRAF, it just means you have a ton of BRAF. I make the analogy to a soccer game. At the cell surface an EGFR receptor kicks a soccer ball from the defense, to the midfield, to the strikers, down into the nucleus. Having a bunch of BRAF is like having a thousand midfielders, but it doesn't mean they have the ball. It doesn't mean that they're being phosphorylated and actively signaling. But Brian’s BRAF was not just a little high. It is like 10 times higher than anyone. It's off the chart. Brian McCloskey: I’m something like number one out of 500 patients. Rick Stanton: Yes. Brian is by far the highest of the 512 prostate adenocarcinoma patients (PRAD) in the Cancer Genome Atlas (TCGA) for BRAF. So you start thinking, maybe there must be some reason there are a thousand soccer midfielders there. That doesn't happen by chance. It made me very curious as to CureMatch bringing in BRAF inhibitors as potential therapies that I and the medical community would've never thought of. For Brian's case there could be a BRAF therapy in the sequence, if we kind of can adhere to Bob Gatenby's approach. In Brian’s case, maybe BRAF is a good idea. Ally Perlina: Brian’s BRAF was overexpressed compared to the cancer cohort, and If it were compared to the normal tissue, it probably would be even more overexpressed. I can run another CureMatch pass with you with that data in mind. We believe in molecular matching. Your molecular profile is an overall, mechanistic, biological picture of what's driving the cancer. You want to be able to interfere with that biological profile as a whole, as comprehensively as possible. If molecular matching is possible for BRAF, we know it's an oncogene, we know how to target it, and there are drugs available, then this would be a potential match, alongside other matches. The degree of matching has shown clinical correlation in prospective clinical trials. The reason I have skepticism about delivering drugs sequentially is that the science says if you pick one target at a time from among multiple targets, you're doing a partial match. Partial matches correlate with poorer outcomes. When you just pick one, they may not live to try the next one. The sequential approach assumes that you can afford to try just one and there is no risk of giving a partial or a poor matching therapy. The data shows there is a lot of risk. Brian McCloskey: I wonder if those clinical trials are not an exact match to Bob Gatenby's proposed process, because those patients probably did not progress to another drug until they completely failed the first drug in the treatment sequence. What he's looking at is you're not completely failing the first drug, you're waiting until you're progressing, like halfway down – whatever his mathematical model would be. You're progressing, but you haven't hit the nadir with that first drug. Then you're bringing in the second drug, and you're attacking that way. And then you maybe bring in a third drug. I don't know if that's what Bob's approach would be, but that's how I interpret it. Ally Perlina: There are two points: (1) the types of therapy and (2) the number of lines of therapy. On the types of therapy, the first line of therapy is usually going to be a broad action therapy, like chemotherapy or radiation, not a targeted therapy. Brad Power: He's talking about advanced cancer, so people who are past the standard of care, so he's talking primarily about targeted therapies, not chemo. Ally Perlina: The second point is that there was a publication with our approach applied to treatment naive patients with advanced cancers and complex profiles, but they did not have any prior treatments. It was published in Genome Medicine last October. [Sicklick JK, Kato S, Okamura R, Patel H, Nikanjam M, Fanta PT, Hahn ME, De P, Williams C, Guido J, Solomon BM, McKay RR, Krie A, Boles SG, Ross JS, Lee JJ, Leyland-Jones B, Lippman SM, Kurzrock R. Molecular profiling of advanced malignancies guides first-line N-of-1 treatments in the I-PREDICT treatment-naïve study. Genome Med. 2021 Oct 4;13(1):155. doi: 10.1186/s13073-021-00969-w. PMID: 34607609; PMCID: PMC8491393.] It shows an almost linear correlation. If you only partially address the molecular profile of the tumor then people do much worse. The progression-free survival and overall survival are poorer. And the more you match the profile, the better they do. It's not black and white. If you do a partial match, there's a worse outcome. If you do a semi-decent match, it's better. If you do a perfect match, it's a lot better. That's where the unexpectedly great results come from. That's why I was asking if the sequential theory has clinical trial results. We should be comparing on the same level, apples to apples. Brian McCloskey: I don't think that data exists. I think that this would be a novel approach. I'm going to bring in the voice of Saul Priceman, who happened to attend Bob's presentation last week. Saul runs the Priceman Lab at City of Hope. He suggested a potential approach where I would take, say, Pluvicto to target PSMA, and then use some of the other data that I have about overexpression of certain genes to sequence in other targeted therapies. I know that there's a clinical trial that actually does this combination with Pluvicto and pembro. I've been on pembro, and I don't think that it is going to work for me, but potentially a combination of Pluvicto plus trametinib, which targets BRAF expression, and some of these other biomarkers. What do you think about that approach? Ally Perlina: What markers do you have? it would make a lot of sense to take drugs which would target multiple biomarkers. The more of them the better. Brian McCloskey: AR, FANCA, and BRAF showed up. Rick Stanton: If you go into the immune modulators then, B7H3 is super high, and Brian's androgen receptor is also, wonderfully high and targetable. Those would pop off the top of the list. Even if it was cold. Brian McCloskey: You suggested, for example, like Apalutamide to target that, which would make sense. Ally Perlina: Then if you combine it with trametinib, it would be very analogous to some of our matched options. Brian McCloskey: Let me push this just a little bit more. As far as I know, the only trial that Bob has done is the one that he presented last week, using abiraterone. There were only 19 patients. I thought that the outcomes were very compelling. But what we're talking about now is not monotherapy, we're talking about leveraging CureMatch’s combinatorial drug approach, looking at multiple targets. How can we build a proposal that would leverage CureMatch’s approach for targeted drugs? I would probably throw Pluvicito in there for me as well. I know that that was not part of your solution. The reason I would throw Pliuvicto in is because we do have this issue that I have a cold tumor, and potentially that would make it hot. How can we get some support to combine your approach with Bob's approach? Because intuitively both of your approaches make sense. And they seem very symbiotic. Ally Perlina: Here's how: you try one drug, and then you add another one in a week, and another one in another week. Therefore it's sequential. Problem solved. Brian McCloskey: I like the math – every seven days. Ally Perlina: We do our own mathematical modeling on the fly here. Brian McCloskey: Have you ever encountered this sequencing approach, rather than the all in one? Ally Perlina: Our approach basically doesn't agree. It's not that symbiotic to what we see as the approach that is having more reception and adoption in the field. It is a wave that is coming up. It's not easy in the medical system to get two or three drugs right away for any patient. And it takes some hoops to jump through. The drug acquisition specialist can make the molecular justification case for molecular indication of combinations of drugs, but it is doable. Our reality seems so different from the sequential approach. Brad Power: One of the examples that Bob pointed to was pediatric leukemia, where the standard of care is 12 sequential therapies over a one year period, and it's curative. That was one evidence point that he cited. Emma, since you've seen so many patients over the years, what do you think about this question of combinations and sequencing?