 As Terry mentioned, I'm one of at least three presentations on pharmacogenomics, and they're almost every center that put forward one of the forms that got came is doing something in pharmacogenomics. So I think there will be some issues there in terms of whether the room is big enough for the breakout session. But hopefully people will go to all the breakout sessions or go to other ones as well. I also think in terms of one of the things that Rex and Terry charged us with was trying to find some projects that could go forward. And because of the commonality of pharmacogenomics, I think there will be one or two there that will come out of maybe Mark's presentation, Maya and Alan's, et cetera, that could be taken forward. I think that's it. We'll find out. You could always give someone else's. Yeah. I like yours. Yeah. At the Christmas party at University of Aberdeen, back when you used slides, they would invite four faculty to bring four slides, and then they'd mix them all in a bowl, and you had to pick out four slides and give a talk on those four slides. So it was, of course, this was in Scotland, so ethanol was involved, so there it was. But they were either, it could be awesome. So I'm going to talk a little bit about pharmacogenomics, but mainly hit on some of the end points that I think we need to be focusing on, because really what's represented in this room are a bunch of people who have been doing things one way, and a receiving end point. So we heard from a couple of CEOs that really look at the world a very different way. Now, St. Jude's is still pretty research-intensive, but as we got from the presentation from Geisinger, the key end points for many of these hospitals have nothing to do with what study section recognizes. And so I'll come back to that point. I think we need to start tackling those, because right now we're building things for an audience that's different from the actual application audience in many cases. So pharmacogenomics can be a lot of things to different people, and also hitting on that same point, we've spent a lot of time discovering things and hunting down the rare aspects, and there's been relatively little time on the clinical aspects. And our institute has, Terry wanted me to put one slide at least on my institute, is made up of a bunch of different people trying to go across this divide. But as you, as we've already heard about, man, tired of pushing that But as we already heard about, most everyone is down here on the discovery end, because that's where the funding is, and there's very few people really pushing all the way across the divide in its truest sense. Now we have these examples, I trimmed down my usual list to just include germline examples, and there are at least one FDA-approved drug that I haven't put on here yet. But we do have a number of examples that could be used. And that group of people I showed you on the previous slide, and Peter, you can accente formally of Mars Group now at UNC, is part of this effort. There are studies in all of these different areas that are going on, prospective studies where we're doing interventions based on genetic markers for that. But they're very focused, one question, one genetic test, one or a very small number of endpoint types of questions. Think R01, R21 types of questions. And that's fine. There certainly is a need for a lot more of those. But I think that genomic medicine, if it's to go forward a little bit faster, is going to need to be asking broader questions, and I'll come back to that point in just a second. So one example, and this is data from Matt Getzde at Dick Winsher Bombs Place at Mayo Clinic, who saw differences in outcome from breast cancer with treatment from Tamax has been based on the SIP-2D6 endpoint. And the poor metabolizers here, if I can get this to work, need some other drug. Tamaxfin works okay in these green folks. But the intermediate folks, which is about 40% of women, was the part that was really neglected. And so many of us in the room got a lot of phone calls from a lot of oncologists who were ordering this test and asking the question, what do we do with the 40% of our patients? And so we did a very simple study. It was just published in the Journal of Clinical Oncology a couple of months ago. And we asked the question, we looked at active metabolite levels after Tamaxfin, and initially in a pilot study of 119 patients. And we found that the intermediate metabolizers indeed had lower blood levels than the rest of the group. And so we did a simple thing. We took the extensive metabolizers and left them on the normal dose, 20 milligrams. We took the intermediate metabolizers, if I can get the arrow to move. Anyway, you can see the orange ones. And we doubled their dose. So we did a test in a clean environment, and we either left them on the same dose or doubled their dose. And what we were able to demonstrate is after that simple change, we normalized blood levels. So active metabolite levels were now no longer statistically significant different from each other. And yes, I don't know why that purple group went down a little bit. It wasn't significant, but it sure was irritating. But there was no longer a statistical difference. Now blood levels do not equal outcome. We haven't proven that we're curing women with breast cancer. But we are demonstrating that genetics in a prospective setting can lead to an intervention that can be simple and easy to conduct without needing even an algorithm. This study was done. We've now expanded it to over 500 patients. And this study was completed in sites across 64 of the 100 North Carolina counties. So coming back to the point that Jeff made, wherever he is, that we did this mainly outside the mothership, partly because academic centers are so painful to work in, and partly because most patients are not seen at academic centers. So if you're asking a practical question, the academic center may be the wrong place to ask that question in some cases. And at least we need to be including the real world, if you want to call it that, in that. Now ironically, these 500 patients were accrued in less than a year. We planned on enrolling 20 patients a year until we went out to the community sites. We had a very simple question, easy to conduct. Anything they could do, patients demanded this. And we enrolled 500 people in less than a year. Jeff enrolled 1,000 in whatever it was. There are opportunities that can be done more efficiently in these types of settings. So a number of these examples that I'm showing you here, these are examples from the FDA package insert changes, are being conducted out in the community setting. Sometimes with no patients being enrolled from the academic center, as we try to understand how does this stuff really work from a process standpoint and other aspects out in community practice. Places where we wouldn't want our relatives to be treated, but where many patients are treated. And there's a need there. The other part, as I mentioned, is there's a huge gap that we really ignore. We're great at step one and step two here, and really have not put a lot of time and effort into these latter stages. And now it's time to do that. And I think we need to be looking at what are the drivers for step four and step three, because they're not the same as step one and step two. And it's a hard thing for those of us who are used to writing NIH-style grants that study sections like certain things, they often do not resemble the endpoints that we need to drive forward. I'll come back to that point in a few slides. And it's mainly because of how boring these aspects are that we've ignored them. And now we need to figure out a way to get excited about that. So it comes also to parts of the health system we really, I really don't like interacting with. I apologize to any of you who are health economists, but you are really boring. I mean, health, you know, having to interact with these folks is really challenging. It's totally different input. The language is very different. Health system integration, what works for the VA doesn't work for Duke and UNC, et cetera. Medical informatics, I mean the impact that our IWAR for an iPhone app had on our prescribers was dramatic. People said that they didn't believe the data was there until they had their iPhone app, and now they're using it routinely. It's, you know, trying to get to the real questions, often the real answer is not lack of data. It's lack of ability to translate. And so we need to be working on things like that. And I think certainly Scott and possibly Howard will talk about the research to clinical assay aspects, because certainly he's done a lot of work in that area. So we have a new audience. In the past, the audience was ourselves, editors, reviewers, study section. And now, often the people we're trying to convince are clinic administrators who are more interested in clinical efficiency than they are preventing bleeding in the head and the gut. And so if we can make Warfarin, the Warfarin Clinic see 5% more patients, they will mandate genetics. Whereas if we do a randomized trial showing that bleeding is decreased, they'll make us do another trial to see if we can do it twice. And so we really need to be looking at what are the real drivers? The payers themselves are certainly in an audience. And then patients are getting more involved with this. Patients are now starting to order the test themselves, starting to demand the test, even when it's not appropriate. And we need to at least address that audience even if we don't think they're right. We also have additional endpoints. And I said additional, I said new, but also additional because survival, stent, thrombosis, adverse events aren't gonna go away. Those are important endpoints. But we need to add some things that are equally important in the eyes of the health system. Things like selecting amongst equal therapies. Now you might have your favorite therapy for your favorite disease, but I bet you there's some level one, or at least level two evidence for other therapies that are equally as good. You know how to spell this one, so you use it. But there's something else that's almost as good or maybe even better. And so how does one adjudicate between equals? Is a common thing done in health systems and not a common thing done at academic centers? Again, because it's boring. And so getting at some of these things. Return on investment for the medical home. When adverse drug reactions used to be a cash cow for medical centers. I mean, if you're having a tough year, all you need is a couple Steven Johnson syndrome cases and you're flush again. Now, with the medical home concept, you're given set amount of money and adverse drug events start losing you money. We now have incentives to try to really get, to tackle that and look at that return on investment. Quality measures, something that is again in the boring category, but very important in terms of how health centers are operated. And I can't remember, oh, patient satisfaction also is part of that. Now I'm gonna skip over that in interest of time. But the concept of intervention or preemption is something that we're very comfortable with breast cancer screening, with colon cancer screening, Down syndrome screening. And they're all about the inflection point. Find an inflection point and offer the screening. Very simple types of messages that are out there. Adverse drug reactions are certainly a major problem. Most of us here are working that area and I realize that. And yet we haven't really approached it in that same way. And we have some initial data and it'll be a little smoother and better soon working with some of the colleagues at MedCo. Trying to define where is the threshold for risk of needing a medication. Or risk of needing a medication with a high level of adverse events. Or risk of needing this medication where there is a genomic test that could offer some level of benefit. And so this concept is something that we're pushing on. And something that I think this group needs to help tackle if we wanna move the field forward in genomic medicine rather than individual genomic studies. Now the examples that are out there do lend themselves very well to this sort of preemptive action. And I'll have a slide that the mentions Dan wrote in a second because this is really the gospel that Dan has been preaching for quite some time. And that is that one can intervene. A lot of the work that he and Josh and others have done at Vanderbilt where you have examples of events that matter for health systems. Bleeding events, hypersensitivity reactions, stent thrombosis, delayed discharge from pain control, all are things that matter and matter a lot to the health system because that's where they lose money. And the idea that we've ignored the CIP 2D6 codein, oxycodone issue is really quite hilarious when you look at the percentages of patients. There are many things that we do in the hospital when the risk is one in a thousand or less. And then here is something where it's one in 10 and yet we don't think it's really that real because it's only pain and they just need to suck it up and get over it. And so there's some issues there that we need to work through maybe some of our own psychological issues. The other thing is this is not a rare issue. So what this is showing is 61 actionable variants. Yes, you're welcome. That we feel are actually as far as our PG&E project and some of our others. And then we ask the simple question across a large number of populations how often does someone have an actionable variant? And what we found is that everybody has something. Now they may not ever need that drug in their lifetime. That's a separate issue. But they have the genetic variant. So if one went in with a panel, our preferred panel you would find that everybody has at least seven variants that can be acted on. And I think the highest was like 26 in one of the African populations. And so everybody, whether you're from whatever continent has actionable variants. And so if you look at primary care setting which is why we did this analysis it's not gonna be something where do a test and sometime in your lifetime you may see a patient with this problem. But rather the results are gonna be coming through and can populate medical records for activity. And we'll probably hear similar numbers from Mark or Peter Grover's presenting from UC and from Allen. Then the last thing is coming back to the application. We have things with untoward events like pain control. We have in some times from projects is being used for insurance coverage. You can't get STUXMAP covered without a K-RAS test and with some health plans. Identifying low utility, the drug's not working well in certain tumor types. Dose selection, therapy selection, Prasagill versus Clopridigril. Preemptive prediction case of Stevens-Johnson syndrome. But then we also have these elements here in terms of bundling of care, patient safety, avoiding bounce back. Things that really we may not care about in terms of our individual academic career but are dramatically important in terms of how health systems are looking at things. And the more often a patient bounces back the less money one makes. And if you wanna ever get an administrator excited you all you need to do is go into that area. Our health system has been putting multiple millions of dollars into some of our programs specifically because of this issue. And so it really can cause activity because there's so much money to be lost. Now it's a little bit like my 12 year old daughter when we go to the store Justice. She's able to save me hundreds of dollars every time we shop. And so it's taking advantage of the health system by saving money. It's really we're losing less money I think is really the term I should have used. So in terms of the opportunity I think conducting preemptive activities is the case. And there's Elvis but there's also Rodin in terms of the cane. Dan's really pioneered this concept of preemptive testing. I know his sideburns are shorter but he'll work on that. The targeting high risk populations, hospitals already know who is more likely to have a delayed discharge to bounce back, et cetera. And there's an opportunity to go in there and tackle that. Using health system endpoints. What do our administrators care about? And I know that's painful but it can lead to faster adoption. And then lastly using panels of variants to ask cross cutting questions is something that many of us are starting to do. And I think if we work together we can do it in a way that will benefit each of our health systems, each of our careers but more importantly the patients that we're serving. So I'll stop at that point. All right, comments, questions for Howard? Dan you gotta have something to say. Come here. You do. I appreciate the call out or the shout out Howard. But I think it's self evident to anyone who thinks about the implementation of, am I allowed to use that word? Of genetic testing. I'll just echo what you said that it's just too much of a pain in the rear to do genetic testing for drug response one genetic test at a time. And that's why it's not adopted. And so the only way out of that is to do it on a platform that does a lot of testing at the same time and then file it away. And I'd echo what you said. We started to look at our data that we've been collecting prospectively and what's really striking is that there are people who, there's two parts to this. One is that everybody has at least one variant but there are people who have more than one who pop up over and over again. So people who, for example, are poor metabolizers for copitagril then turn out to be at risk for Simvastat and related myopathy. So I think that that's another part of this that we have to start taking. That's the reason to do it. Well, we've got, I mean, all of us get calls from those kind of people all the time and we think they're nuts. And they are, but they're nuts because the drugs make them nuts. Well, so could you just clarify for us because those are interesting data on your number of actionable variants? How did you go about trying to define what you considered actionable variants? So yeah, I didn't spend a lot of time on that. For our Pharmacogenics for Every Nation Initiative, which was really more of a public health, so we work with ministries of health across the world to develop their national formularies. So we had a consensus panel on what variants are actionable at that kind of question for that at the public health level. And so we have clinicians. Drug therapy? Yeah, yeah, only drug therapy. So only pharmacogenetic markers. There is not 100% but very high overlap with CPIC, et cetera, because the literature is the literature. But the, and so there's clinical groups from about 40 different countries that are involved in coming up with consensus for that. So you didn't pay attention to like novel stop codons or something like that? No, these were all things that were on DMET plus chip type of, there was like IL-28Bs note on the chip, HLAs are on the chip, but a lot of it was things that are already known. So that's not even taking into account novel variants from a particular country, that's a different project. So I just wanted to bring up a followup issue with regard to the preemptive genotyping idea. So I think it's a great idea of course in integrated health systems where patients are gonna be seen from cradle to death and for which having that genotype information somewhere in the EMR is gonna be more likely to be useful, but you guys have to address the issue of the real world and the idea that any given patient is gonna ping pong from one healthcare system to the next. Where does that preemptive information live? How does it, how is it utilized in our current system? So you made it sound like the idea of putting data into our own health system was solved because that's not. There are some aspects of this that are stuck in the allergy section because they're kind of like an allergy and we want a red flag, not a red flag, we want a hard stop in terms of prescribing the drug. There are other parts that we just hope are not too invisible because of where they're located in the current EMRs and have them forbid you to ascend out which is uploaded as a PDF and stuck in the bowels of the EMR. So it's not solved even for the individual centers, maybe a dance place, but certainly not at ours. The other thing I would just add to it, I mean the organizing principle that we're dealing with here is the patient and we've been reluctant to engage the patient at a level that they can really be active participants and we heard the example of, hey, everybody dies of something. Yeah, had I been that patient, I would have been telling him, yes, and you're gonna die from a very pissed off patient right now, so deal with it. But we're not empowering patients with the tools that they need and granted at the present time, there's probably a relatively small group of folks who would be capable of doing it, but this is the type of thing where we need to organize around the patient at least in this country because if we don't put the patient in charge of that information so that they can literally bring it with them either in some type of a knowledge-based representation or in some type of a device that they either carry with them or is implanted or whatever. We don't have a shot because we're not gonna get to the point, I don't think where we're gonna have complete data interoperability across the entire United States healthcare system and your point is exactly right. We're transient, we bounce, we go to different places, we wanna get second opinions and even within our own healthcare system, we're wasting because we duplicate the same test over and over and over again. With the germline genetic test, you don't need to do that, that's just money that we could use for other things. It's a critically important problem. The patient has to be at the center and we have to figure out how to empower them. At the moment, the patient is responsible for bringing drug information from clinician to clinician and they do it with a very effective device called a paper bag and they come in and they open up the paper bag and spill all the bottles out, some of which they have pills they're taking, some of which they just have had in that bag forever. So it's very broke. It's right in the bag. I think this really highlights really important consideration that we've got to really tackle and actually Pearl is gonna lead a discussion tomorrow that talks about the interface between research and clinical and we come up against this all the time and we've heard now several times today that the patients demand it or the patients say, why are you asking me to consent to this again? Just do it and I think this is really a key issue and one of the ways around this is patient portals. I think that really is gonna be a driver. Great, we'll have, I think David will be the last comment and I should say the lunch is not quite ready yet so I would propose if you guys can hang on for another 20 minutes or so that we go ahead with Mark but David, please. I would say having grown up in a very integrated healthcare system that there is a germline genetic task that I have done every time I want to give blood and every time I order a blood for somebody I type and cross-match them and that's a germline genetic test which my blood group hasn't changed in the last half of many years but they order it every time. So I think one of the key issues with that why do we order a blood group every time because blood groups don't change either. The issue is that we don't trust that anyone else is actually gonna get the blood group right when it's a critical test and so I think even in the situation it's something like a DMAT even within your own hospital I could see that task being reordered when it's something that is of high significance like metabolizing drugs. So I think we do have a precedent for a genetic test that we've been doing for half a century where it gets repeatedly reordered. I'm just curious at St. Jude, do you reorder the same tests over and again? Okay.