Personalized Cancer Treatment: Challenges and Access for Patients

•Testing: We have heard from Tony Letai and Payel Chatterjee of SEngine about functional testing . Blood-based liquid biopsy using ctDNA and surrogate markers of efficacy can be used in cases where fresh tumor tissue is unavailable. (Peter Kuhn is scheduled to discuss.) Blood and other novel tests (Karin Rodland) can enable more frequent monitoring of disease progression, enabling fine-tuning of treatment and personalization.

•Predictive Models : There is a lot of investment and effort in developing models that will predict drug response by large pharmaceutical companies and academics that can be repurposed.

•Real World Evidence : Every patient should be tracked in an observational trial to share results of their unique, personalized, N-of-1 experiments. GCTA (XCELSIOR) is one such unique registry (Jeff Schrager): It allows you to create "n-of-1" arms, does not drop the patients on the floor ever, has no exclusion criteria, and understands "arm" in a dynamic way not available to any other trial model. Several other ideas were raised to address the personalization conundrum:

•Glenn Sabin proposed that the patient could consent to hold the clinician harmless, lowering liability concerns.

•An anonymous caregiver suggested that the job of patients is not to be a grateful consumer of an industry that serves itself, but rather a person with needs, and that you are seeking people who can help you, where your needs and your wishes are primary.

•Ally Perlina recommended having patients raise specific treatment options with their doctor and listen to the specific feedback.

•Saed Sayad pointed to creating a logical process that leverages existing public data. 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 have a conversation about some of the things we've learned about various aspects of making complex testing and treatment decisions for advanced cancer patients, and to get your input on some countermeasures to the “personalization conundrum”. A lot of the things that we have been talking about, such as drug combinations, have been generally accepted by most people that we've spoken to as a better option. If you can hit multiple biomarkers at the same time, that’s better, but there's a concern around toxicity when you have drug combinations. How can we balance this conundrum of, on the one hand having what appears to be better, more effective treatments, e.g., through drug combinations, and on the other hand, a concern around toxicity? This is the Prostate Cancer Lab organization. We launched in March, so we're about four months in. This is a checkpoint for reflection. This picture has changed almost every couple weeks because we've added another diagnostic company, another presenter (we have had 15 meetings with a variety of discussion leaders), or another patient, molecular biologist, bioinformatician, or physician. This is our 16th meeting, and each time we've had various people lead discussions, many of whom have raised new ideas. We have a calendar going forward of meetings, about 10 or so scheduling into September. The discussion leaders are providing insights on this decision process of identifying treatment options, targeting them based on diagnostics and analysis, and then trying to reflect those and prioritize them. This slide summarizes what we've learned and what I'm calling “the personalization conundrum”. Treatment Strategy : Ally Perlina of CureMatch has presented their recommendations for matching approved drugs indicated by the patient’s biomarkers. CureMatch’s distinctive contribution is drug combinations. Their premise is that hitting multiple biomarkers is better than hitting one at a time. Saed Sayad has also made the point that more drugs at lower dosage is better than fewer drugs at higher dosages. Bob Gatenby has been an influential, theoretical strategic guide for us with his ideas. Cancer is a heterogeneous population. Any drug or treatment is going to have an effect on that population and generate a resistant strain. If you administer drugs at maximum tolerable dose until resistance, you will breed resistant strains. Rather, we should think about knocking down the population using game theory, evolutionary theory, and adaptive therapy. Emma Shtivelman has been a leader in giving us principles for choosing among treatment options and ideas about treatment. For example, a targeted CAR-T therapy may be in Brian’s future. It could target PSMA or a couple of other antigens. Therefore, if you have a drug that reduces the PSMA-presenting cancer cells in the population, then it might make the efficacy of that eventual CAR-T less. She also had ideas about choosing different pathways. A lot of the prostate cancer therapies are targeting androgen deprivation and androgen receptors. Can we find drugs and treatments that would be targeting different pathways, and see what they can do, rather than continuing to pound the same pathway and get increasingly diminishing returns. Treatment Options : Brian has submitted his data to Cancer Commons, xCures, CureMatch, and Massive Bio. CureMatch focuses on combinations of approved drugs. Massive Bio focuses on clinical trials. Cancer Commons is mostly Emma Shtivelman. xCures is the software engine that's running Cancer Commons, and they have come up with yet another set of options. Brian reviewed the treatment options they recommended for him last week. He had 18 options, of which four seemed to be on the top. We need to give a shout out to our inspiration Bryce Olson, who has been through nine lines of treatment, though it could be more. He's currently on bipolar androgen therapy, which is an extreme version of adaptive therapy. It's being very effective. Rick Stanton has presented previously that he's identified 10 options for his treatment. Rick's treatment options have largely come from his clinical team, which includes Tanya Dorff, Rana McKay, and doctors at UCLA. There are two little thumbnails of slides that you can't quite see off to the right. One is Rick's analysis of the NCCN guidelines. I included it to make the important point that we are beyond the standard of care. The discussion and decisions we're talking about are for patients who've exhausted the standard of care, and they've exhausted the obvious choices for drugs or treatments. Most have had a prostatectomy, radiation, and androgen deprivation, and they've failed all of those. They are in the zone of discretion, where there might be a dozen or 20 options, and a patient and his medical team needs to choose amongst them. The second thumbnail slide is called “the Stacey matrix”, which says that we're in the zone of complex decisions, not straightforward and easy decisions. Treatment Selection : The essence of this conversation is how we can make physicians and patients more confident in making personalized decisions, particularly drug combinations, but also dosing and sequencing. When Bob Gatenby reviewed Brian's case, he said that since Brian has a very low PSA, he is a candidate for an “extinction event''. He recommended that Brian choose a ladder of three drugs in rapid succession or in a cocktail, and see if he could get an extinction event, to use the evolutionary terminology. But we run into the question of toxicity when we have combinations, because we're in uncharted territory, because there aren't clinical trials that have run that personalized combination at personalized dosing. Personalization Confidence? : There are three tools that we could use to give patients and clinicians more confidence.

•Testing: We've had a discussion led by Tony Letai of Dana Farber about functional testing and what it can do. We've also had a presentation by Payal Chattergee of SEngine on functional testing. And we will have others from First Ascent Biomedical. Liquid biopsies are another source of information that could give confidence in choosing amongst options and to be more personalized. We're going to have Peter Kuhn presenting on liquid biopsies in a couple of weeks. Another testing category is monitoring. We had Karin Rodland talk about identifying biomarkers through proteomics. Obviously PSA is a form of monitoring a biomarker, but there might be others that could be identified. The big idea is that if you could have more frequent testing of your response to a drug situation, a drug that you're being administered, then you could move more quickly.

•Predictive Models : AI and machine learning is getting better at making predictions. Drug companies are throwing lots of money at figuring out whether patients will respond to drugs that can be repurposed for patients. There's been a history of looking at genetic drivers. There are models that can be developed that would predict whether a patient is going to respond to a drug. Predictive models may be off in the future, but it could give one confidence in making a personalized treatment decision.

•Real World Evidence : Nic Schork in our last session said that we could develop a protocol and put everyone into an observational trial or registry. Jeff Schrager, who works with Cancer Commons, told me about a trial called XCELSIOR, which is a continuous trial that accepts all patients and then monitors their progress longitudinally. That could be plumbed by AI and ML or patients to figure out what's working. It's using real world evidence to get some confidence to see what is working. Brian McCloskey: That was a perfect summary of where we've been. To add some color, last week I met with Rana McKay and presented my 18 options. She was amazing in that she went through all 18 of them. I spent a full hour with her. It's one of the things that makes her a different doctor, in such a great way. When we get into personalized medicine, we get into drug combinations, and the big issue for her was toxicity. How do we manage toxicity? How do we manage quality of life? How do we figure out what the right dosing is and sequencing for these drugs? I certainly subscribe to the idea that drug combinatorials are the right approach. But I'm in a zone where I don't know where to go. Dosing and toxicity are real issues. I appreciate her perspective that, as a physician, she lives by the creed that she will do no harm. You can easily get lost looking at all of these different approaches that are going to use these drugs to address these targets of opportunity that we've identified. But there are a lot of unknowns. I appreciate Rana’s guidance in putting the brakes on. We have to help her. The mission here is, how do we help? How do we help our physicians get over this hump, and do it in a way where it meets their standards of safety, meets my standard of quality of life, and helps us to get to better outcomes. For me right now, that's really the heart of the issue. Kaumudi Bhawe: Thank you for providing this platform. Your slide outlining the different approaches and the three different aspects of moving the needle forward is very informative. We need decentralized, observational trials, such as XCELSIOR, with the ability to open up arms as necessary, and are also collecting patient tissue and liquid biopsy blood samples. It needs to be written into the trial protocol that that information is captured so that it can help future patients. Currently there are multiple observational trials, but we need trials where the intervention is screening itself. We need to get these companies, such as SEngine, and all the different types of testing companies that are out there, to overcome the typical limiting factor of patient tissue. Among the three options that you've listed for a doctor or an oncologist, if there is evidence on the patient tissue that a certain combination of drugs is working, or some experimental treatment actually works on the tumor cells, then that would be a gold standard equivalent almost to having randomized clinical trial data for that combination, but in a more feasible way. Obviously it's not possible to conduct a randomized clinical trial for every single combination or new treatment. Brad Power: I don't know the universe of registries as a naive patient. This gets into data privacy issues. Of the advanced cancer patients I know who are trying to survive, privacy is not an issue. They would be quite open to making all of their data available for research and not really care if their employer or anybody else knew about it. \I believe that that would be the vast majority of advanced cancer patients. So it's not for lack of willingness of patients to have a global registry that you could then see what works and what doesn't. The “patients like me'' argument seems obvious. If 20 people have my situation, and 10 people tried this and 10 people didn't, and 10 people had something good happen and 10 people didn't, I want to do what worked for the 10 people. Therefore, more data is better in that registry. I don't know anything about the registry landscape. On the one hand, everybody should stand for a global registry that is a data vacuum cleaner. It should be very open source, like Wikipedia, a “good for the planet'' model. Yet there are competing forces and incentives that cause academics to want to hold research data close, because that's the article that they're going to write about based on their data, et cetera. It sounds like you have some knowledge of XCELSIOR. What does that landscape of registries look like? We wouldn't want to reinvent the wheel. Nic Schork offered to create a protocol, but I wouldn't want to reinvent the registry wheel. There must be something out there. Is three one out there where we could just say, “this is what everybody should be using”? Kaumudi Bhawe: There's currently no system in place that allows this to happen easily. A system needs to be created. Even something like XCELSIOR isn’t really there yet. There are multiple factors playing into this. There's the economics of it. There are different competing interests. I don't have an idea right now of who the right people would be to tap into creating a platform to make this happen. Because any interventional trials that are out there have to be led by the group that is instituting the intervention. There doesn't exist a third entity that initiates an open ended trial and then gets everyone on board. Brad Power: In the past we have talked with Eli Van Allen at the “Count Me in” project at the Broad Institute. He's also at Dana Farber. They have a metastatic prostate cancer database. However, when we spoke to him about being able to use this registry for clinical use, he said they can't do that, that it's for “research use only” purposes. If he had to make it useful for clinical purposes, he would need a lot of money to get it suitable and available for clinical use. If anybody is going to do something like that with a broad societal benefit, it might be the Broad Institute, and yet, they're doing something like it, but as our healthcare system is often skewed, it's not directly for patients, it's for research and drug discovery. Kaumudi Bhawe: Even if it's initially labeled as for research, because that's more kind of protecting the doctors and everyone. We don't want to claim that this is clinically relevant yet. It's research. Glenn [Sabin] can speak to this a little bit too with some experience that even the