The Medical Matchmaking Machine

24:49And we got Michael on this medicine. And April of this year marked nine years that he's been in remission from his angiosarcoma. Oh, wow. Other patients have been treated with this. Other doctors learned about this and started treating their patients. And it turns out about a third of people with this horrible cancer, previously uniformly fatal cancer, will respond really well to pembrolizumab, to this medicine. and it's now standard of care for his former cancer. It's now standard of care? It's now standard of care without ever doing a clinical trial. And that goes to show you when you have a disease that's this bad and you find a drug that works this well, you can change the paradigm for the disease for relatively, I mean, as close to$0 as humanly possible.

25:34But like, why is it that everything is so, like it's like we almost have this idea of like lock and key, Like this medicine does this thing for this drug and da, da, da, da. But then it's like, and then you're like, no, no, no, but this key works in this lock over here. Like, why is it that that is that? Anyway, yeah. Tell me there. So the reason I think that our system is like this drug works for this disease is because in order to get a drug approved, a drug company has to develop a drug for a specific disease and submit it to the FDA for that disease. The FDA approves it for that disease. And if that drug company mentions a single word about that drug working in another disease, they will get fined billions of dollars for what's called off-label promotion.

26:16So when the FDA approves a drug, what they're really doing is they're approving a drug company to market a compound for a specific disease. And that company cannot market that compound for any other diseases until they come back to the FDA to get that change made. But every time a drug company does that, it costs lots and lots and lots of money, so they don't go after all the opportunities they have. But insurance companies and payers realized, well, if this drug that's approved for this one thing could also be useful in this other thing, and it would be good for patients, shouldn't we allow doctors to prescribe things off-label?

26:49And so that's something that happens very commonly. About a quarter of all prescriptions in the U.S. are off-label. No. Yeah. So it's somewhere between 20 and 30 percent of all prescriptions written every day in the U.S. are off-label. Are not for the reason they're supposed to be. Yeah, exactly. So that includes examples like doxycycline for Lyme disease, where every doctor in the world would be like, yes, use doxycycline for Lyme disease. But doxycycline is a cheap old generic antibiotic. So whoever made doxycycline 100 years ago, 30 years ago when people figured out it worked for Lyme disease, they aren't going to submit for a label change.

27:19And that gets into the other factor here, which is that once a drug becomes generic, whoever originally made the drug, they stopped making money off of the drug because you have generic competition. You have multiple companies that make the identical drug and the price plummets per pill. And so no one in our system makes any money off finding a new disease for that drug. But why wouldn't like like you would think from the drug company's perspective, like that they would want to go to the FDA, get as many uses approved as possible because then then they could go out and say this helps this. This helps that.

27:53This helps this. This helps that. Like you they would make their market as big as possible. Except it's more complicated than that because you can only sell a drug for one price, regardless of what disease you sell it for. It always has to be the same price. So what that means is that you have to pick the first disease that you get your drug approved in. You have to pick the optimal market for that drug for the optimal price because pricing is actually not based on the cost of the medicine. Pricing is based on the value for that disease. So the fewer competitors there are for a disease, the more expensive the drug.

28:27The rarer the disease, the more expensive the drug. There's all these factors that affect how expensive the drug is going to be. And if you're a drug company, you have to maximize your profit. So you need to come up with the highest price for the highest number of people. But it might be that a low number of people at a higher price is better than a high number of people at a lower price. And so you can imagine it gets really complicated really quickly. And it's all about the first disease you get your approval in. And so companies have to be really thoughtful and strategic to maximize their profits about what their first approval is.

28:58Once they get that first approval, now they have to remember that they can't change the price for the next disease. And so this is this horrible economic issue, which is just so depressing because like on the other side of these economic issues are people suffering. Okay. I want to hear more about what you're doing now. You're like, amazing. This works for me. How does that then go to, oh, no, wait a second. I'm not just doing this for my family. Like I can big this up. Yeah, the next big milestone was early in the pandemic. I was actually driving down to Raleigh, North Carolina, had my wife in the car, and I'm listening to the radio about this pandemic.

29:33And I'm sitting there thinking, you know, gosh, this involves the immune becoming activated and causing all these problems. And gosh, it's going to take us months or years to come up with new drugs. I really wish there was a lab somewhere out there that was really good with inflammatory stuff and could repurpose drugs and could direct drugs at this thing. And then I was like, oh, maybe we should do that. And so we decided to create a program called The Corona Project where basically we redirected my 15-member lab to focus specifically on COVID. And early on, as you'll remember, there was a lot of drugs that were repurposed.

30:08Some worked, some didn't work, but there was a lot of repurposing. This is the first time we did like a very concerted effort to be like, what else is out there for this one disease? Very much informed by what we'd done previously. And COVID, of course, there's lots of controversy about what worked and what didn't. But the two drugs that unquestionably worked incredibly well were dexamethasone and tocilizumab. They saved millions of lives. And they were, you know, old drugs have been around for a long time. And so that further got my wheels turning on like, what if we could create a system to automate what my little lab was doing for one disease?

30:41But we did it for all diseases and all drugs simultaneously. And thankfully, in parallel to those dreams, the field of machine learning, artificial intelligence has matured so much that we can actually do that. OK, so tell me about AI. How are you using AI to like to matchmake here? How did you think like, OK, this part of it can be done by AI or this part of it or whatever that kind of thing? So in my case, you know, you can think about this. We use what are called biomedical knowledge graphs, which are just sort of mapping out like every medical concept on a map. So you can imagine like if you have this giant wall and every single gene, every disease, every protein, every pathway was put against the wall.

31:21So if we were to start with that concept and say, well, what do we do for me? Well, you'd find Castleman's on that wall. It would only be there in one place. You'd find Castleman's. And what you'd find is you'd find an edge or a line between Castleman's and activated T-cells, because I discovered that T-cells were activated in my disease. You'd find another line to mTOR activation, because I discovered that mTOR activation was really up in my particular immune cells. And then you would find a drug from T-cell activation and mTOR activation to serolimus. Because serolimus is able to inhibit mTOR activations and able to inhibit these activated T cells.

31:55And so now within this giant graph of every disease, every gene, every protein, you would find Castleman's with lines or edges to these two concepts, and then lines or edges to serolimus. And you would see a connection between them. And so now imagine doing that for every disease, every gene, every protein, basically what the world knows about all of medicine. Well, it's almost like mechanical what you're doing. It's like, it is like you're trying to To make like the mechanical blueprint of what is going on. That's exactly right. This leads to this, leads to this. And this reverses this, which reverses this, reverses that.

32:30Everything is there. It's everything. Well, now what we do is we train machine learning algorithms on all of those known treatments. So like the sirolimus for Castleman's, sildenafil for pulmonary arterial hypertension, insulin for diabetes. Imagine training this algorithm because machine learning algorithms are so good at finding patterns. And so we're giving the machine learning algorithm lots of information about known treatments. And we're saying, this is an example of when a drug works for a disease. And we do it thousands of times with all of the treatments that are out there for all the diseases that are out there.

33:03And then we say, okay, algorithm, now go and score how close of a pattern the connection is between a known treats relationship for every other drug versus every disease. So if a toenail fungus drug looks like there's no way it could work for pancreatic cancer, You need to give it as close to a zero as possible, 0, 0, 0, 0, 0, 0, 0, 1, right? Right. But if leucovorin looks really promising for a subtype of autism because the pattern of connections are there and there's a clear intermediary between that subtype of autism and that metabolite, give it a high score. So you get a 0.99. And so now what we do, we do all 4 ,000 drugs, all 18 ,000 diseases.

33:43So it's about 75 million scores that we generate, that our machine learning algorithms generate. And then that gives us a list in rank order from the things that are 0.99 all the way down to things that are 0.00 of every drug versus every disease. And so we come across matches that are incredible that we never could have imagined that now the algorithm is saying you should really look at this. It's like the body is so complicated. These drugs are so versatile. It's like our minds can't even comprehend that. It's like that's why you need to go to AI. you would have to, as humans, think about 75 million possibilities.

34:19Like my lab's really good at looking at like dozens of possibilities for like one disease. Like my lab can spend like a year and we get through a few dozen for one disease, right? But like we could never think about like 75 million possibilities and then compare them. And I'm not saying AI is perfect, but directionally it's really good. The things that are the 0.99s are way better than things that are the 0.5s. How likely do you think it is that like if there's an ordinary person with an ordinary disease that existing treatments don't work for, that there is something in there for them? Yeah, I guess there's two probabilities here.

34:55I think that one is that what is the likelihood that there is a drug out of those 4 ,000 that could work for that disease? And then what's the likelihood that you or anyone else is going to find it, right? Because it's just like, A, does it exist? And B, can you find it? I think the A, does it exist? This is obviously a really hard thing to guesstimate on, but like, I'm going to say somewhere in the realm for any given disease, somewhere in the realm of maybe 10 to 20 % that there's something out there. And then in the realm of, are you, or is a team going to find it in time for you? It becomes much lower than 10 to 20 % likelihood, right?

35:31Just because the steps that have to happen. And so for us, we're going to be the organization that is going to identify and unlock as many of these drugs as possible. So that way we don't have to be throwing Hail Marys so that when you get diagnosed, it's, oh, wow, you have pulmonary artery hypertension. You should just take this medicine. Oh, wow, you have glioblastoma. You should take this medicine. And so we've intentionally taken the approach of let's use AI and data to find the best uses for the best drugs so that we can move them forward so that way we aren't doing Hail Marys. But the reality is, Latif, is that like people are suffering all the time and we are contacted all the time and we want to help any way we can.

36:10And we're going to be making our algorithms publicly available in about nine months time. Interesting. But until then, we want to continue to improve them. We feel this tremendous responsibility that once we share it, that, you know, it's out. Right. And so we're going to continue to improve over the next nine to 12 months. But then we will share it. And are you imagining patients would use that or are you imagining doctors would do that? So the intention will be for doctors and researchers to use it. So that way they can come up with new areas for research. They can think about it for their patients.

36:35But the reality, I think, is that patients will also use it. I'm trying to imagine. So that is really interesting that you're making that public. And I think it's also like there's something beautiful and hopeful. So the conversation went on for a while after this. But I was honestly surprised and a little taken aback that this algorithm that David had made, that he was going to take it public. and it took me a while to kind of process that and figure out like what I thought about that or what I even wanted to ask him about that. That part of the conversation, which actually felt a little bit trickier to me, is coming up right after the break.

37:15Stick around.

37:23Okay, so I did that interview with David Fagenbaum And it's funny, he genuinely surprised me when he said he was taking this thing public. Like I had not heard or seen that before. It is since being reported that he's doing that. But when we did the interview, like I didn't know he was going to say that. And so when he said it, I was like kind of shocked that he would do that. It like re-slotted the story in my brain from being like a slam dunk, like a no-brainer, best possible use case for AI to something that was like, wait, wait, wait, what do I think about this? Should ordinary people be able to look up what drugs AI thinks will help them?

38:04Is that helpful? Is that reckless? So I really didn't know what to think. I called up a bunch of my doctor friends. Some thought it was like so exciting, especially for people with rare diseases, you know, where there's not a lot of research money. They were like, yeah, this tool is going to be so useful for so many people. But then there were other doctor friends of mine who said, no, no, no, this is going to make my job harder and it's going to hurt people, which you will hear more about in a minute. Anyway, so I called David back and I had way more questions and I was just like, okay, just tell me the specifics.

38:40What are you putting out there exactly? And why are you so sure this is a good idea? Sure. I don't even know what to call it. Like, okay, your AI matchmaking tool. What do you call it? How do you use it? Yeah, so we call it the matrix. So it's an acronym. Everything has to have an acronym in my life. So it's ML-aided therapeutic repurposing and extended uses, matrix. Okay, so ML-aided, machine learning-aided, therapeutic, T for therapeutic. Repurposing, R. Yeah, okay. I, N. and then it gets a little sloppy then extended we're using the x in extended okay for matrix uses oh okay then uses there's no uses doesn't get a letter it's not matrix to you it's just it's just it's just matrix okay because you're a fan of the movie or something it actually is a matrix in that it's 4 000 drugs it's 18 000 diseases so it's actually we're building actually a matrix of drugs versus diseases uh and a fan of the movie okay so how does it work like say i'm just somebody, I'm a random patient with random disease and I want to use it.

39:47What do I do? Yeah. So we're still working on some of these things. We're actually like literally like talking about prototypes and processes. But I can tell you that there'll be the ability to type in the name of your disease or maybe the drug that you care about, probably more likely the disease that you care about because most people care about diseases as opposed to drugs. But then actually go to look at a rank order list for that disease. So say like, oh, wow, I care about ALS. These are the 4 ,000 drugs ranked in order for ALS, according to this AI platform. That I'm almost certain will be available in a format like that.

40:19But the bells and whistles, we still have to work out. Is it like a chat GPT style thing? I'm imagining you put your disease name into the thing, and then it'll spit back out at you a bunch of names of drugs, and it'll give you the percentage. Yes. But then you can also be like, hey, by the way, I'm also a smoker and I have diabetes or whatever other conditions you have. And then it'll like like like how much information would you put in like as a user? I mean, what you're describing would be sort of like a holistic patient support treatment tool. And we're really not building that. I hope someone does.

41:03Someone should build that, but we're not building a tool that is going to be that sort of treatment co-pilot, though I would love for someone else to do it. Okay, so this is very much just insert name of disease, and what's going to come out is a list of names of drugs that may or may not work, and here's the odds that they will. Yeah, and these are in the same list that we're getting on our medical team and our research and development team. we're giving you the same results and same scores that we're getting because we feel this obligation or this responsibility that if we're going to put our eyes on them, the world should be able to put their eyes on them.

41:37You know, here are the tools that we use. Like our medical team uses these same machine learning algorithms. You can use them too, but it's important to remind them that when our medical team uses those machine learning algorithms and they come up with something like lidocaine for breast cancer, we still then go on to do a bunch of laboratory work of lidocaine and breast cancer. And then we think about doing the right clinical trial of lidocaine and breast cancer. So it's not like we use the algorithm to immediately move forward into action. We use it to then plan out what to do next. Okay. So why did you make the decision to make it public?

42:10So our feeling is that we as a nonprofit at EveryCure, we're only going to be able to go through like dozens. I mean, if we can get into the hundreds, I would be over the moon about it. But like there are still thousands of diseases that like could potentially benefit from our scores that we'll just never be able to get to unless it's like because the list that matrix is spitting out is just so big is that it it's so big and it's so powerful the things like when we look at the top we are blown away by the number of promising drugs we're like wow and actually let's have some of the cases there's actually been clinical trials that have shown the drug works but someone stopped after the small trial because there was no way to commercialize it so one part is let's make it available to the world so that other people can pursue these things that we're not able to go after.

42:57And the other is sort of probably a little bit inspired by maybe we shouldn't be so paternalistic in medicine and maybe we should like, you know, allow this information to be out there. Of course, when I say that, I do like cringe just a little bit because like I don't want us to create problems by putting this out there. But it feels like the responsible thing is to share the scores, but to appropriately caveat them and disclaim them to say like, these are for research purposes. At every cure or nonprofit, we don't take a score and then put that drug into a person. We take a score, we evaluate it by MDs, PhDs, and MD-PhDs.

43:34We spend months on it. And then we do laboratory studies. We do clinical trials. We work with experts to get in the guidelines. That's our process. So we want other people to take a similar process. Have you thought about what could go wrong by making this public? Yeah, a couple of things immediately come to mind. I mean, number one is patient harm. A patient taking a medicine that causes harm to them that had not undergone the studies necessary to evaluate it in that disease. Now, the good news is that every drug we score is already FDA approved for something. So we're not there's no drugs on there that like someone like, oh, my gosh, just never got a regulatory approval.

44:10They all have been approved. You're not recommending cyanide to people. Yes, exactly. Yeah, cyanide is not on our list. Right. Actually, I will tell you, there is a drug repurposing platform that I will not name on this podcast where literally one of the top five predicted drugs for Castleman disease or predicted treatment for Castleman disease is car fume exhaust as a treatment. No. What? This was the top predictions that, like, I guess inhaling car fumes, like fumes. And this is just like an AI hallucination kind of thing? It was like AI. It was like AI made the connection. You'd be like, you know what?

44:42And I was like, oh my gosh, that's the problem. I just haven't been inhaling enough car fumes. Like that's why my castle was as out of control. Right. I just needed to inhale car fumes. Just every morning after breakfast. So I say that not just to say that, oh my gosh, this one platform had this bad prediction, but it's to say that AI is going to make silly predictions that make no sense. Not silly, like harmful. Harmful, yes. Harmful, yes. Right. That's why humans have to be a part of this. And humans who can critically evaluate this and say like, this is not good. So I think the most important thing is going to be, I think, how we communicate around these scores when they are made publicly available, that these scores are intended for our research team to find things for us to research more.

45:23These were not ever intended to be, you know, scores to say this thing that's number one is what I should get because that's going to save me. And so I think it's going to be context is going to be really important. But like, what's the thing you say? Like, how do you expectation set in a really clear way to make it super distinct? Like this is a machine that generates research ideas versus something that tells you what drug to take now. I see what you're saying. Your point is, is that you can say it all you want, but for that, not that like just may not stick. Right. It may not stick. And also people are desperate.

46:01Like a lot of these people, I mean, you, you know, you know better than I do. Yes, they will do anything. It's a great point. I think that the way to make it stick, I think it's trying to explain that we don't use these predictions to decide how to treat someone. This is not, and maybe even to use the terminology, this is not a solution engine. This is an idea generator. And we at EveryCure do a lot of further work. And so we hope you, if you're going to use these scores will also do further work. And so if you're a patient, that might mean working with a research lab to do the work to figure this out.

46:39What we recommend is talking to the disease organization that you're a part of, whether it's the ALS Association or you name it. It's talking to a lab to see if there's further work you can support. But none of those options are go take this medicine. I'm trying to imagine like so that is like it's there's something beautiful and hopeful and democratic about that. I can also see though, there's a big danger here. Like, so I talked to a friend of mine, um, about what you're doing. And she was like, I've already seen this play out. Like I, I, she's like, I know exactly what's going to happen here.

47:13I saw this play out with ivermectin during COVID. So all my patients were coming in asking me for ivermectin, even though like I knew that that wasn't going to help. She was like, last week, someone asked me for ivermectin for cancer. It's an antiparasitic drug. It's not going to help you for your cancer. So I said, no, like, no, I'm not going to sit here and prescribe you a thing that there's no evidence for. But what then this is doing is it's like, it's like driving a wedge between the patient and the doctor, because now once the doctor says no to the patient, then the patient now doesn't trust the doctor.

47:53Now the patient is either going to probably, you know, go doctor shopping until they find a doctor who will prescribe it to them, or they'll wind up at the black market or do medical tourism or some other, you know, non ideal situation. But anyway, like, like the her point was that the doctor-patient relationship is already in a bad place. Like it's already in a really bad place. And now if patients come in with these drug recommendations, like who knows, maybe they'll be great, but also maybe they won't. And then she has to be the doorstop. Like she has to be the one who crushes the hopes and dreams and has to hold the line.

48:38And that's like - I agree. And I totally empathize and can really see the concerns. I think that where I go when I think about COVID is not so much ivermectin, but I go to dexamethasone. So dexamethasone saved millions of lives during the pandemic. It was the only drug, Latif, that was recommended against when the pandemic started. It was whatever you do, don't give people a corticosteroid. Corticosteroids weaken your immune system. Don't take dexamethasone. Literally, there was no recommendation for what to take. There was only recommendation for what not to take. Well, some amazing pioneering doctor in the UK still decided to do the trial of dexamethasone.

49:19And it worked. It reduced mortality by 35%. But the prevailing medical system believed that it would actually be harmful. So we didn't know what to do. We just don't do dexamethasone. Turns out dexamethasone actually reduces risk of death by 35%. So I'm so glad that someone asked the question, are we sure dexamethasone shouldn't be used? So I love that. I'm glad that DEX saved like millions of lives. And I'm glad that like, that there was a sort of a one doctor who was willing to go against, you know, everyone else. And so I think that the whole point of this is to find dexamethasone, not ivermectins.

49:52If there's a drug that someone thinks might work for a disease and it's snake oil and it's not working, we want to study it. We want to prove that it doesn't work. If there's a drug that looks kind of promising, but no one's studying it, we want to study it and prove that it does work. We just want to prove that they work or they don't work. So do you feel like, like what you're doing, do you feel like, oh, this existing medical system needs to be respected, it needs to be shored up, and it's like you're trying to like fix a part of it that's wrong? Or do you feel like it's like, no, no, no, no, no, the existing systems are failing us in these key ways, we can't be bogged down by them, like I'm building a whole new thing here.

50:38Yeah, I think where my mind is, is that I think I'm still so appreciative for what doctors do for patients and that doctors bring just this laser focus on helping the person in front of them. I'm so grateful for that. And I'm so grateful for all the things that our biomedical system has figured out are true. Like this drug works for this disease. Like I'm so grateful for that. So I don't want to break down any of that. I want everything about our doctors caring about patients and the relationship. And I want everything about all the known knowns, like where we know this drug works for this disease.

51:16I think what I really, really want to bring forward is uncovering the unknowns. So that way those doctors can use, that unknown can become a known so it can be easy for them to use it. I don't want to create some crazy new system where patients are picking drugs off AI. Like, I want to use AI so we can find out what we can elevate to the level that a doctor feels comfortable. Like, wow, that steroid actually could be useful for this thing. Huh. Never would have thought about it. But you know what? They did a clinical trial and it works. So, like, I'm going to do that. So, in my opinion, it's not about breaking down the system.

51:51It's about enabling the system to do exactly what it's trying to do, but that we're caught up in these assumptions that I think we have around what we know and what we don't know. And I think we're really, we're certain we're very good at what we know. Like I totally, I believe everything we know in the system is rock solid. I think we're just not as good at understanding what we know to be not the case versus what we just don't know at all. Like the amount of our, and actually there's a term for it that the computer scientists use, and that's the ignorome, which is basically the things we don't know about medicine.

52:23Like the ignorome, I think, is a lot bigger than most of us in medicine want to appreciate. And I think if we can be a part of uncovering the ignorome and making it less ignoromous or whatever it is, then I think that that's where we can serve medicine, doctors, patients, and not to try to break it down. And it's really about lifting things up. Yeah, I mean, look, you've already been doing miraculous work. At the same time, I can't help thinking about this other Radiolab story that we did. I don't know, it was like a year or two ago, about this ICU doctor named Blair Bigum, who was kind of like you, was like the annoying patient relative.

53:11And the story is kind of about him watching his dad basically contract cancer and die. and his takeaway from that experience was like we we don't need more hope we don't need more like hail marys at the end we need more dignity like we just we glamorize doing everything we can you know because we because we want to like keep the people we love alive but actually the reality is like those last ditch efforts like tend to make things worse for the patient for the family for the hospital. And now it's funny because I'm talking to you and I feel like you have the exact opposite, 180 degrees, different takeaway.

53:56Fight, fight, fight, never say die. Like try every drug in every pharmacy. So like, like, I just don't even know, like, how do you square those two things? So I actually think that during this, this discussion, I think you've actually sort of opened my eyes a little bit. Just because you sort of like highlighted to me, I just told you the three most special months of my life were the last three months of my mom's life and we weren't fighting for a treatment. Yet I just try to extend other people's lives with the drugs we already have in my own. And so I think everything's context dependent. And I mean, if I think that what he said is conceptually correct, but I think that when you feel it, when you experience it, especially when you experience like the positive side when you do make it.

54:44It creates a new sort of value that you put on if you do get extra time. But of course, at the end of the day, this is like philosophical around like individual versus collective societal. Like, you know, there are people that will say like, I want to die at 70 years old because I don't want society to have to take on my burden. Like even if I'm not sick, like I just should die at 70 years old. So society doesn't have to pay for my costs. And then there's other people who are like, we're only on this earth once. So like, like I'm going to like squeeze out all the time I can get, you know, I'm going to live as long as I can get.

55:21I don't think that I'm sort of on either camp. I think you can, um, you know, you know, be reasonable and decide. I mean, I should also share about a patient recently that, um, that I had helped to discover a repurposed drug for, and I mean, it got him out of the ICU. And I remember sitting with my team and jumping up and down. And like when we got the news that he was responding to this medicine, it was like literally. What was the sickness he had? He has calcium, the same exact subtype that I have. Wow. It was just, I was so excited. I remember I cried tears of joy. I was so happy that we found this drug for Paul.

55:53And the next couple of weeks went by and, you know, he kept getting better. And he got out of the hospital and I got in touch with him. And he explained to me, David, this drug got me out of the hospital. It turned everything around to save my life. but I feel horrible on it. It makes me nauseous. I'm vomiting all the time. Like it's controlling my disease, but I don't like the way that I'm living. And, um, he was a 70 year old gentleman and he decided to go off that medicine for the exact reason you mentioned a lot of is I want to spend the time that I've got with my, with my, my kids and, and with my wife.

56:29And I was like, Paul, but we already found something for you and it got you out. Like we've shown that we can do it. Like, like, let's try another one. And, and he said, he said, no, David, I don't want to. And, and then the two of us just cried, you know, tears of sadness, you know, cried tears of joy before that cried tears of sadness, but it was, but it was okay. Like this was his decision. Like he knew that if, if he put me on the case, he know, he knew that I was going to be all in and he knew that I'd, I'd been able to do it once before. Like, but he told me he didn't want to. I was very sad, but I felt that it was absolutely the right thing, and I 100 % respected it.

57:11I understood that for him at this moment in his life, that was the right decision. And he passed away a few days later.

57:26So you're saying it just basically needs to be personal, it needs to be case by case. Yeah. Yeah, maybe that's the big takeaway, is that it seems like everyone wants to tell us what everyone else's decision should be. So like, it's best for society for you to do this, or it's selfish of you for you to do this. But I think that maybe that's the real fundamental thing this comes down to. It's got to be that patient's decision.

58:02Thank you.

58:32Chloe Shasha Brooks and Helena Bowen, who introduced me to David and his work. I was right there in the wings when he did his talk on the TED stage. That talk should be on their website, the TED website, very soon. And in the meantime, for more on David's story, you can check out his book, Chasing My Cure, or for more on the work that him and his team are doing, you can go to their website, everycure.org. Special thanks to Peyton and the rest of the staff at Every Cure. If you have not had enough Radiolab, we referenced two prior episodes today, and I think they're both totally worth a listen.

59:15The first was the one I mentioned about that ICU doctor, Blair Bigham, who basically writes a book about how we should make our peace with death and we should die with dignity. And then while his book is on the bestseller list, his dad gets pancreatic cancer. And all of a sudden, everything he wants to do completely contradicts everything he wrote in his book. That one is called Death Interrupted. Second Radiolab episode that we mentioned was kind of the backstory of the drug that saved David's life, rapamycin. It is an entirely improbable backstory. It is shockingly dramatic. It's about one immigrant scientist who basically single-handedly saves this potential drug from the trash can by smuggling it across a border.

1:00:10That episode is called The Dirty Drug and the Ice Cream Tub. That is all for us today. Thank you so much for listening. Until next time, I wish you good health.

1:00:36Hi, I'm Connor, and I'm from Minneapolis, Minnesota, and here are the staff credits. Radiolab was created by Jad Abumrad and is edited by Soren Wheeler. Ulu Miller and Latif Nasser are our co-hosts. Dylan Keith is our director of sound design. Our staff includes Simon Adler, Jeremy Bloom, Becca Bressler, W. Harry Fortuna, David Gable, Maria Paz Gutierrez, Sindhu Nyanan Sambandan, Matt Kilty, Annie McEwen, Alex Neeson, Sarah Kari, Sarah Sandback, Anissa Vitsa, Ariane Wack, Pat Walters, Molly Webster, Jessica Young, with help from Rebecca Rand. Our fact checkers are Diane Kelly, Emily Krieger, Anna Pujol-Mazzini, and Natalie Middleton.

1:01:18Hi, this is Jenny from Brooks, Maine. Leadership support for Radiolab science programming is provided by the Simons Foundation and the John Templeton Foundation. Foundational support for Radiolab was provided by the Alfred P. Sloan Foundation.