 Our next session is clinical evidence in genomic medicine sustainability. We'll have talks on the state of science and gaps from Jonathan Berg and Roger Klein. First will be Jonathan Berg from the University of North Carolina at Chapel Hill. Thank you. All right. Well, thank you for inviting me to speak. I've really enjoyed hearing about Ignite and everything that's going on. And, you know, trying to come up with a talk about sustainability and what the gaps are. And there's quite a lot to talk about. So I'm going to try to focus on just a few things here. You know, I think this is probably the key issue that we face in our society, which is the health care spending. And I think that largely this is why genetics is getting picked on as the new kid on the block, as Mark Williams told me this morning. You know, the technology that we have to use is being increasingly scrutinized. And much of that is probably because of the perception that it is technology that drives health care costs. And so that's natural for third-party payers to want to get a handle on, right? What are we doing? Why are we doing it? What are we hoping to accomplish by it? And I think we need to meet the third-party payers and other stakeholders in that space. And we can't expect that we're going to be exempt from evidence of utility. So this graph, I've sort of modified from something that I saw David Bienstra present at an IOM meeting, a workshop a few years ago now. And I've seen versions of this in different places, and you all probably have. And the idea is that if you just map things in terms of whether they cost you more or less and whether you get better outcomes or worse, then you can sort of map things out and say, well, these are the things that cost more and give us worse outcomes. We shouldn't be doing that. These are the things that give us better outcomes and cost less. And we all hope that genomic medicine will give us those types of outcomes. But the question is, really, can we achieve it? There's going to be some things that actually will give us lower costs but worse outcomes. And we could think about whether we'd want to use those if resources are scarce. If it's not that much worse and it's a whole lot less. So we might want to think about those things. But probably most of what we do is going to end up in the costing more and maybe giving us some better outcomes. And we need to address the fact that we do want to use the technology. And the question is, can we afford it? And does it give us the cost effectiveness we need? Obviously, this is the real no-brainer. These ones up here are actually difficult to change in medicine once you've got physicians established with practices. They don't want to change their practices even if it does turn out that they're worse outcomes and more expensive. It's pretty difficult to convince the public that they should have less, particularly if the outcomes are going to be a little bit less. And this one over here really depends on the stakeholder and what they value in this overall value proposition. So I think we're so accustomed to seeing this ubiquitous graph down now to about 1,000K per genome that it really seems intuitively obvious that we should be able to use whole genome sequencing clinically. That is why we're all in this field. But we still have to address the cost benefit, right? There's the cost of interpretation that's been mentioned. It's not the million-dollar interpretation as some sort of hyperbolic statements have made, but it's definitely more than zero. And the cost of downstream interventions are definitely going to be more than zero. So somehow we need to address the overall cost of genome sequencing and other genomic approaches. And we have to understand the net outcomes and sort of how effectively we can achieve those outcomes with sequencing. So what the evidence is for genomic medicine depends on who you ask, which I think is really important that you all have engaged stakeholders and that's going to be a main theme that I talked about. But it's also understanding what their perspectives are and really what they value in this process. And how much certainty is required to be able to make a decision about whether something should be used clinically. So these implementation studies are clearly necessary. We have to understand how to put things into a clinical space when there are all the barriers that we've identified. It is important that these need to be proven to be useful in order for the uptake to be maximal. But really the sustainability to me depends on convincing other parties involved, whether that's a healthcare system, insurance providers, FDA, etc., that genomics is effective in what it's trying to accomplish. So there have been some efforts to sort of engage stakeholders and think about what their perspectives are. And I like this way of looking at it. Again, Dave Vistra's on this, so you might have probably already invited him. He probably couldn't come, which is why you invited me. But the idea being that if you have some information about the risk and benefit of a particular application, and you have some level of uncertainty or certainty about whether it is favorable or not. That you could actually sort of get some consensus around how to use these things, how to develop them, whether you could use some evidence development practices to get additional data. So I think that this is the sort of work that probably needs to be done more to really understand at many different levels with stakeholders what kinds of evidence they need and how they could help us to generate it. So which outcomes matter? So ideally you want to get direct health outcomes, right? That's what we're in this business for. We are in a medicine field and these are the things that matter. But for physicians you could also imagine that just simply being able to tailor and improve management might be a reasonable expectation. And especially for our patients with very rare disorders, even if there are no evidence based guidelines or other sort of clear things to do, we can at least use our best clinical judgment to treat those patients as well as possible. And the idea that for patients and families there may be some intrinsic value to understanding the diagnosis, which is clearly also including what the recurrence risk is in the family. So I recently came across the hierarchical model presented by Frybeck and Thornberry, which I think originally was designed for imaging tests. I think is highly relevant to what we do. And they describe several layers of efficacy, if you will. So there's a technical efficacy to a test that really deals with its analytic validity, does it actually tell you, give you a correct assay in terms of the genotype, whether it gives you an accurate diagnosis and coupled to that, whether it helps the clinician with their diagnostic thinking, which I think relates to clinical validity. So in other words, if the physician is faced with a number of possible differential diagnosis questions, does the test help them narrow that down, exclude things, or include things? Does the test lead to therapeutic efficacy? And this is not necessarily a utility, but does it actually allow the physician to make a decision about treatment, whether or not that actually has any benefit to the patient? You also have a patient outcome efficacy, which is a bit of a higher level there, which I think falls into the clinical utility category. And then at the very top, on this case in the bottom, the societal efficacy really does the cost of healthcare change. Do you get a more cost-effective outcome with this test? So I think these are useful ways to think about what we do in genomic medicine. And just as a simple example, I sort of was thinking about cystic fibrosis. So you have a child with a diagnosis. We know what it is. So what chloride test was positive, we know the diagnosis is cystic fibrosis. So really is genetic testing needed? And you could argue that it's not if you have a clinical diagnosis. But if it is, what test would you do? And should that be covered by insurance? And so thinking about the hierarchical efficacy, clearly the genetic testing modalities we have for genotyping and sequencing of the cystic fibrosis gene are quite reliable, quite accurate. The establishing a specific molecular lesion in this particular patient might not actually help your diagnostic thinking if you just find out the patient is homozygous for the common Delta 508. Doesn't really change the clinical diagnosis any. So you could argue that, well, that didn't really help you very much because you already knew the clinical diagnosis. But if there was something that you could think about in terms of enabling different decision-making about treatment, for example, using some of the new drugs like Lumicaftor and Ivacaftor, which arguably are coming on board and are facing physicians have to face this decision of, do I try this drug? Well, we know that these drugs only work with certain types of genetic variants in the cystic fibrosis gene. So in fact, doing the sequencing and identifying the specific molecular lesion would alter therapeutic efficacy and it might even improve patient outcomes if you consider the data, the early data on these drugs. And so conceivably, this might alter the cost of carrying for cystic fibrosis patients. So going back to my questions, is genetic testing needed? I would argue yes, because we really want to know whether this drug combination is a possible therapy. And if so, what would we do? You know, and I put out here the exome is clearly an incorrect answer because you don't need to do a whole exome or a whole genome sequence to understand what the CFTR mutations are. But you could imagine a different set of conditions where there was a higher range of genetic heterogeneity and you really needed to specifically identify the gene. Should it be covered by insurance? You know, this is where it comes to the value in the economics, the cost of the test, the cost of the drug, the nature of the health outcomes. You know, the early data suggests that you can reduce hospitalizations by one third in these patients. Do we consider the cost of the drug at $259,000 a year to be affordable? It's a societal question, but these are the types of things that providers, the third party stakeholders, are going to have to make decisions about. Just as a brief thing, what about the reproductive implications we talked about, personal utility, being able to know for family reproductive planning? I think the real problem with this is that there's no defined measures of utility in that context, and every family is going to have different uses for the information. And so stakeholders are going to arguably view these differently. And I don't think this is an incorrect assessment by payers that this type of information may not be worth them paying for. But these are the sorts of things that we need to be able to communicate with stakeholders about. So getting to the gaps, I think stakeholder engagement, and as I was putting my talk together, I learned about the Ignite stakeholder engagement. Workshop wasn't able to attend, but had someone here told me a little bit about it. And it sounds like this is a very, very good first start. I think we need to conduct our engagement with respect to defined clinical scenarios. We need to be clear in communicating the population that we're interested in, the tests that we're interested in, what we're doing it for, so that the stakeholders understand the clinical context. We need to understand what they value. And this could be anywhere from patients to providers to health care systems to payers about different outcomes of genetic testing. We need to understand the evidence that they need to make decisions, understanding that this may be difficult for them to tell us in advance. But if we understand those things, the studies can be designed with very specific outcomes in mind that provide that evidence that really directly addresses those requirements. I don't think that this means that we should completely ignore the other important scientific questions in our field. One example would be how to deal with secondary findings or how to understand the penetrance of these conditions and these sorts of things. But we have to frame our studies to convey our results to these different stakeholders. I think evidence synthesis is a big gap. Just as we need the databases of genes and variants at ClinVar, and that's a somewhat self-serving statement since I'm involved in that project, I think we also need to have systematic collection, curation, and evaluation of genomic studies. And I think I heard that mentioned in the stakeholder summary that somehow understanding how to organize that information, design something in collaboration with different stakeholders so that we structure information that facilitates the review of all of these studies that we're doing, make it publicly available, have it curated by experts, make it transparent so that everyone knows what evidence the stakeholders are looking at, and then we can also design our studies for maximal impact in that area. So just to kind of summarize briefly my take-homes, technology and health care costs are a big issue for us. We just have to tackle it head-on. We definitely need to understand what stakeholders value, what evidence they need to see to support use of genetic testing, genomic testing, in different contexts, and be able to really produce the relevant evidence that'll tackle those topics. So, thank you. Thank you, Jonathan. Next we have Roger Klein from the...