Anwar Chahal & Calum MacRae discuss Clinical Genomics training

Getting Personal: Omics of the Heart

Jane Ferguson: Hi everyone. Welcome to episode seven of Getting Personal, -Omics of the Heart. I'm Jane Ferguson, an assistant professor of medicine at Vanderbilt University Medical Center and the chair of the Publications and Professional Education Committee of the Functional Genomics and Translational Biology Council of the American Heart Association.

This month I'm particularly excited to announce a new venture. We have teamed up with the journal, Circulation: Cardiovascular Genetics to bring you this and future podcasts. CircGen publishes a lot of the most compelling research in cardiovascular genetics and genomics and precision medicine in cardiometabolic disease. We've already featured a lot of the research in previous episodes of the podcast. With this new collaboration, you can look forward to even more in-depth features of the newest research published in Circulation: Cardiovascular Genetics.

This month, Anwar Chahal, a cardiology fellow at the Mayo Clinic, talked to Calum MacRae, the Chief of Cardiology at the Brigham Women's Hospital in Boston. The interview covered a variety of topics and we couldn't fit everything into this half-hour podcast. What you will hear in this episode is a discussion related to a recent publication by Dr. MacRae and colleague Aaron Aday in Circulation published in July 2017 entitled Genomic Medicine in Cardiovascular Fellowship Training. Dr. MacRae expands on the topic, How to Train Clinicians to Deal with Advances in Genomic Medicine and What we Can Do to Improve Implementation of Knowledge from Genetics and Genomics to Helping Patients with Cardiovascular Diseases. If you'd like to hear more from the interview, including discussion of Dr. MacRae's bold One Brave Idea project, and to hear more pearls of wisdom and career advice from Dr. MacRae, you can download the full interview in the special hour-long podcast.

First, I do want to highlight one recent paper from the Functional Genomics and Translational Biology Council which was published in the June 2017 issue of Circulation: Cardiovascular Genetics. This clinical genomics paper was published by Laura Zahavich, Sarah Bowdin and Seema Mital, all from the Hospital for Sick Children in Toronto. The paper is entitled Use of Clinical Exome Sequencing in Isolated Congenital Heart Disease and describes the case of an infant with congenital heart disease where a pathogenic mutation in the notch one gene was identified through whole exome sequencing. The paper highlights the utility of whole exome sequencing when candidate gene panels are negative, allowing for increased understanding of causality and the ability to make risk predictions for future offspring. At the same time, this approach reinforces the importance of well-trained clinical personnel including genetic counseling, to appropriately interpret and disseminate findings from whole exome sequencing.

A little aside, in case you are not aware of this, Circulation: Cardiovascular Genetics has a really nice feature on their website where you can apply filters to see specific types of publications. From the toolbar along the top of the homepage, you can go to Browse Features and then you can select Options from the dropdown menu, so if you click on Clinical Genomics Cases, for example, you will see all of the genomic case reports, which may be of particular interest to this audience.

Dr Anwar Chahal: My name's Dr. Anwar Chahal. I am a cardiology fellow in training from London, UK, and I am doing my research fellowship here at the Mayo Clinic. I'm very honored and delighted to have our guest, Dr. Calum MacRae, so you are the Chief of Cardiovascular Medicine. You are a MD PhD by training and you are associate professor at Harvard Medical School and your expertise amongst many other things, internal medicine, cardiovascular diseases, but in particular inherited cardiovascular conditions. Is there anything else that you would add to that?

Dr Calum MacRae: I'm a big fan of generalism and I'm quite interested in cardiovascular involvement in systemic disease as well, largely as a means of keeping myself abreast with the biological mechanisms in every system that seems to be relevant to cardiovascular disease.

Dr Anwar Chahal: I think training across the world has increasingly recognized the importance of genetics and genomics, but I just want to share one little anecdote. My wife's a primary care physician and I was visiting the GP practice where she works and she'd mentioned that I had an interest in genetics and genomics and one of the partners came out with one of these reports that a patient had sent their sample to a private company, got this analyzed, brought it into the clinic appointment, and asked for an interpretation. The GP partner said to me, "I've absolutely no idea what any of these numbers, values, et cetera mean. I actually am looking forward to my retirement because I really don't want to have to cover all of this. Can you help me with it?"

I sort of remember hearing Dr. [inaudible 00:05:48] talk here at Mayo Clinic, who's really pushed forward pharmacogenomics and he's been arguing for quite some time, as I've heard you say as well, that genomics and genetics is just going to be a part of the medical record in the same way that hemoglobin or a chest x-ray is and people better catch on because it's here, it's available, commercially people can send their samples directly without their doctor's involvement and then it's trying to make sense of all of that. I think as a community research and clinical we have to take this very seriously and I'd be grateful for your insights on that and then if you could then tell us what would be the best way for the up and coming generation and for programs to incorporate that into their training.

Dr Calum MacRae: I think you're right, there is a general tendency in the public domain to test a variety of different genotypes, and in many instances I think the key elements are how do we as a profession conceive of these tests? I think one of the things that we forget perhaps at our peril is that many of these things are problems that we've encountered before. There's a natural cycle of different tests in medicine where they start off in the academic medical centers, they propagate into the periphery, and then eventually they're assimilated as part of internal medicine.

I think the scale of genomics is obviously somewhat broader than many individuals have seen in the types of data that they deal with on a day-to-day basis, but I think that's something that's happening in everybody's life, and every aspect of your life, you have many more channels to deal with, you have many more choices in the supermarket to deal with. I don't see this as a sort of existential challenge to medicine; quite the opposite. In my experience, the core things that we need to remember is that DNA is no different from any other assay, except for the fact that it's relatively straightforward to do DNA diagnostics. It's technically not as sensitive a set of biochemical issues as are many other assays that we use in day-to-day clinical practice.

The other thing that I think is perhaps a key element is that I said a few minutes ago, it's a [inaudible 00:08:35] dataset and it's stable for your whole life. You only need to have it tested once. It's sort of invert to the typical diagnostic paradigm, so instead of a primary test being interpreted in the context of an ongoing clinical event, the test may have been present for four decades and the results will evolve over time in light of the changing phenotype or some new information in respect to that genotype.

What I've actually looked on genomics as is almost an organizing principle for the way that you build care. In fact, I see quite frequently, we probably now have an average one or two new patients a month in my clinic who bring their entire full genome with them, either an exome or a whole genome. We've begun to really get to know quite well how to manage patients. Obviously they're a select group of patients but one of the things that I find is that patients are really quite astute in understanding that genotype and phenotype are not deterministic relationships. What you have to do is always interpret these things in context of a probabilistic understanding. Most patients I think when they're told this understand that we're going to learn much more about genomics going forward than we will ever imagine that we could know at the present.

That will involve lots of different things. It will involve new ways of displaying data, new ways of thinking about the data in the clinical context. I actually think one of the most interesting things about genomics and to be honest any assay is that they rarely reach any form of maturity until they are used in the clinic, until they are actually used in implementation. For example, many genetic tests at the moment don't change therapy and they don't change outcomes, but partly that's because they've never been studied in that context, and one of the things that I think [inaudible 00:10:45] has to be really congratulated for is his focus on pharmacogenomics as being one of the early areas in which this will really move forward.

I believe that by immersing ourselves in it, by actually trying it within the clinic where we're going to learn much more, and part of that gets back to the original topic that we spoke about, which is phenotype. The only way to really begin to understand collection of phenotype is if you do it in the context of existing genotype I think. As we move into new phenotypic areas, we're not going to be able to test everything and everybody. I think there the genome will end up being an important framework, lifelong framework for the management of a patient's diagnosis, prognostication and then therapy, potentially in that order.

I think you need a whole different set of skills, you need a whole different set of technologies, but most importantly you need information that you can interpret in the context of the person in front of you. Until you can make mechanistically important insights with one person, it's going to be very difficult for genomics to really change medical care. That's something I think we should be focusing on. I think we've tended to have an associate of strategy for genetics. We haven't driven it into the clinic. As we drive tests into the clinic, whether it's troponin T or whatever, you begin to understand much better how to use them, although sometimes that can also go in quite extreme directions that you may not necessarily anticipate. I mean, troponin originally was a stratification tool for acute coronary symptoms, and now it's virtually a diagnosis in its own right.

I think you'll see that tendency revert over time as people begin to understand the biology of troponin, of isoform switching in peripheral tissues of the way in which troponin may represent very different disease biologies. At the moment, it seems like it's a very simple and straightforward yes/no type of test. There's no such thing in medicine, and I think that's what we're learning about genomics and so instead of conceiving it as a series of 10 to the nine yes/no tests, we're going to end up with a very different vision and view of how it can be implemented to clinical practice. That can only come from having clinicians and geneticists work together on this.

In fact, one of the things that we've been doing in the partners environment with some of our colleagues, and I have NIH funding to do this with Heidi [Ream 00:13:31], with Sandy [Aronson 00:13:34] and with Sean Murphy is to think about how we display data, but also how we collect information in light of that genomic data that helps in an iterative way in the learning fashion inform genotype phenotype relationships in a much more probabilistic manner than we have done to date. There are lots of efforts and that's great, that just happens to be one that I'm involved in, but I think it's a generalizable approach that you're going to see moving into the clinic in the next few years.

From the standpoint of training, I think what you want to do is get exposure to all types of genetic information so you understand common [alleles 00:14:15], rare alleles, genomics and individual panels and I think the best way of doing that is to have it be part of training programs. In fact, with one of my junior colleagues, Dr. Aaron Aday, we recently wrote a short piece highlighting how important it will be for all of us to come together to think about how do we start to introduce the concepts of genomics into standard clinical training programs, and that's something we're working on fairly avidly at the Brigham and I'm sure there are, I know there are efforts at many other institutions to do similar things.

Dr Anwar Chahal: That article is published in Circulation in July of this year if anybody wants to download that. I think if we talked to clinical trainees and asked them what are their concerns about training, as you know training can be very long in cardiology, which is a procedurally based specialty whether or not you become an invasive proceduralist at the end of it, there is that component at the beginning, and do you think a standard, in the US, a standard three-year program with two years of clinical and one year of research can incorporate that at a sound enough level to allow somebody to practice or do you think we're going to look at increasingly a one-year or a six-month sort of add-on fellowship for those interested more on the inherited side or more on the genomic side?

I, like yourself, trained in London and the training programs are longer in the UK. It was probably six years when you were there. It shortened to five, and now increasingly it's going to become six and maybe even more with a general fellowship for five years and then a super advanced fellowship and inherited cardiovascular conditions certainly there has become a module that is encouraged for people to take and then become somewhat certified in inherited cardiovascular conditions. What do you think there in terms of incorporating all of that as well as learning basics of eco and device therapy and catheterization? What are your thoughts there?

Dr Calum MacRae: Again, I look at this as a spectrum. I think there's a trajectory for all of these types of innovation and knowledge and it starts off being super specialized, it goes into a more general location, and then eventually it's an integral part of everybody's clinical practice. I do think that what you're going to see is rather than, and this is already I think the case in many elements of medicine, medicine has already exceeded the knowledge base even when I was training by probably a long order in terms of the complexity and extent of content, not that I trained that long ago.

One of the core elements I think that we're seeing is that we need to move medicine from what I believe has become somewhat de-professionalized, say, to one where you're focusing on, not on the actual core knowledge that you bring with you to the table, but actually the way in which you integrate knowledge. I think the focus of training is going to change somewhat. It has had to change in other fields. Medicine I think for a long time favored that sort of single, comprehensive approach in one mind. Medicine is going to become more of a team sport and it's also going to become more of a knowledge integrator profession than it has been for some time.

It's interesting, when medicine started there was so little knowledge that you really had to have almost every physician be an experimentalist using [inaudible 00:18:37] experiments in front of them. I think the way that I see medicine evolving is that as the knowledge base and the rigor of that knowledge base improves, many of the things that we think of as professional activity today will actually devolve to primary care, and to be honest into the community. There are many things where the rigor of the underlying data are such that there's no reason for a provider to be involved, for a licensed provider to be involved. We allow our patients to install their own wireless networks without a technician. I'm sure most of them can look after their own lipids pretty effectively if they were given the right information.

A lot of stuff will begin to move in that direction. As that happens, I think the way in which information is displayed, the way in which data are collected and the workflow around integrating information will change. That doesn't get past the point that you brought up, which is that that will probably take a couple decades and in the interim, I think people are going to end up training in modules of sub-specialties, but I think one of the things that I sometimes like to ask myself is, "What's the end game? Where is this going to end up? Can we build systems that train directly for that end game rather than going through these intermediate steps?"

I think that's something where I think we tried in the short piece that we wrote in Circulation to argue that everybody should have some exposure and that that exposure can change over time. We should be equipping people not to know genomics but to be able to learn how genomics is impacting their patients for the next 50 years. That model of professional training is actually the one that really was the dominant model until maybe a hundred years ago, and then the reasons for it don't quite seem obvious to me, at least at the moment. We sort of tended to slowly move to more of a learned knowledge base that was then applied. Physicians sort of steadily got to the point where we're now data entry clerks. The actual amount of professional and intellectual engagement has I think slowly diminished in many medical sub-specialties and medical specialties.

The opportunity that genomics and other advancements in technology in medicine bring is the chance to I think re-professionalize ourselves to move from just simply defining ourselves in terms of the knowledge base that we each bring to the table, but defining ourselves rather in terms of how we put the knowledge together around individual problems and individual patients, a very much more patient-centered, biological approach than perhaps we've had over the last couple of decades. I think these are, I'm obviously stating a lot of this in somewhat in extremes, but I think these are general trends that you see in medicine. They've happened in other fields as well and people have overcome them. It's usually a function of changing the workflow itself, of changing the way in which the information ends up in the professional's hands and how you collect the data that you use then to interpret the existing knowledge.

That I believe we haven't really reworked probably since Osler's time. It is amazing that we still have workflow, I mean it's amazing in lots of ways. It's an amazing tradition. It is quite interesting that we still have workflow that's probably largely dependent on what Osler liked to do when he was growing up, in terms of the times of day that he got up and his workflow. That's sort of instantiated in many ways in everything that we do. Nothing entirely wrong with it, but there's a lot happened since then that we haven't really changed. Medicine is not yet in many instances a 24/7 profession, and yet most other things that have much less in the way of impact in society are already 24/7 professions in many settings.

I think you're going to see a lot of demographic changes in medicine come from the advent of technology in other industries. I think those will all transform the way that we imagine training in medicine. Along the same sort of timeline as some of the traditional approaches that you described, building out a training module and then having a subgroup of people do six months or a year of extra training. I see that as a short-term solution. I think ultimately longer term solutions are changing the whole workflow of medicine.

Dr Anwar Chahal: What have you done in your own program at the Brigham to introduce genomic medicine training for fellows?

Dr Calum MacRae: We are building out, obviously we have a fairly large cardiovascular genetics clinic, I think probably the largest in the world. We have now seven, soon to be eight providers working only and wholly in cardiovascular genetics. We therefore have the ability to have our fellows rotate through our genetics clinic. We have in-patient and out-patient genetic services and we also obviously involve our fellows in a lot of the academic pursuits going on in both our genetics and genomics programs in the cardiovascular clinics, as we do our colleagues who are no longer in training. We have regular, in our clinical conference slot we have several times a year, we have a genetics component, and then what we have also is an integrated training program with clinicians and pathologists that is really bringing the individuals who are understanding the technical aspects of the genetic testing with the individual sort of learning and understanding the clinical aspects of that testing.

We imagine over time that this will evolve into potentially the type of specialist module that you described but also into a fixture that goes all the way through our two-year clinical training program. We've sort of taken the point of view that we probably need to do a bit of both. We need to, given what I said in the last few minutes, that we need to take a thread that recognizes a short-term and intermediate term need for specialization but also recognizes that we have to equip every one of our trainees and every one of our physicians with the ability to begin to learn the underlying science of genomics and the underlying approaches to using genomics in every aspect of clinical cardiology. We're doing both of those things and have active efforts in both.

Dr Anwar Chahal: You mentioned integration with pathologists but for our colleagues who are not clinicians, what about the research angle and the scientists when they're in training, is that integrated so that we are getting this meeting of minds that is essential?

Dr Calum MacRae: Absolutely, in fact we, thanks to a variety of efforts at Brigham Women's we have now at least three separate venues in which this occurs. I mentioned cardiovascular genetic clinic. We also have a genomic medicine clinic, which I'm one of the clinical codirectors for, where we actually have cases of [inaudible 00:26:45] through routine clinical care that seems as if they would benefit from whole genome or whole exome sequencing, and then we have a weekly conference that's actually led by Dick [Mass 00:26:58] and Shamil [Sonaya 00:27:01], two of our genetics colleagues and takes in specialists from all throughout medicine as well as scientists from the entire Harvard Medical School environment, and so we bring everybody together around mechanistically solving individual clinical cases.

The third venue is one that's part of a national network, the undiagnosed diseases network. We're one of the sites on the national, the NIH-funded UDN network. There again one of the themes is identifying individuals or families who would benefit from both rigorous genomic analyses as well as much deeper phenotyping. That's been a program that I think has been very exciting and one that we again have learned a huge amount from in terms of how do you begin to build the infrastructure that brings not just the first clinician to see the patient, but somebody who, a whole team of people who understand and can evaluate all the biological aspects that are relevant in that patient. Then also brings to bear the scientific expertise that you might need in order to make a mechanistic connection between genotype and phenotype in that one individual, and some of that involves animal modeling.

In cancer for example there's a concept that has emerged over the last two to three years of what's called co-clinical modeling that once you've identified some of the genomic features it allows you to begin to model in an animal in parallel with the trajectory of a patient-

Dr Anwar Chahal: [inaudible 00:28:40] as some people call them.

Dr Calum MacRae: Exactly. Creating an avatar. In many instances that's an avatar that includes multiple different disease models. We've begun to do that in the cardiovascular space. I think obviously it's early days yet, but I think there are lessons to be learned about how you build the types of infrastructure that allow people to move beyond this state where a patient's outcome is dependent on him seeing the right doctor on the right day at the right time. There are actually systems that funnel the patients into the right venue based on objective criteria at every stage. I think that's the type of re-organization, re-imagination of the medical system that we need.

We sort of duplicate things in lots of different areas and you're still dependent on hitting the right specialist at the right day at the right time, or not seeing a specialist, seeing a generalist on the right day at the right time, who's able to put everything together, or even hitting somebody who has the time to listen to your story in a way that helps you identify the exposure or the genetic basis of your condition. If we recreate the professional environment that I talked about earlier, I think, in ways that are both traditional and novel at the same time, I think we'll do ourselves a great service and build a platform that lets all of the technologies, including genomics that we've talked about today, begin to impact patients in a real way on a regular basis.

Dr Anwar Chahal: Thank you for that Dr. Calum MacRae for giving up your precious time and sharing your thoughts and insights and experience.

Dr Calum MacRae: Thank you for your time and I've enjoyed talking to you.

Dr Anwar Chahal: Thank you Dr. MacRae.