11/09/2016 // Interview // Interview: Richard Johnson / Photography: Andreas Fechner

We asked three well-known leaders in the genomic revolution to discuss the future of molecular testing and medicine. The panelists are Elaine Mardis, Ph.D., the Robert E. and Louise F. Dunn Distinguished Professor of Medicine at Washington University and co-director of the McDonnell Genome Institute; Heidi Rehm, Ph.D., FACMG, associate professor of pathology at Harvard Medical School and director of the Laboratory for Molecular Medicine at Partners Healthcare; and Tadd Lazarus, M.D., a physician who serves as chief medical officer of QIAGEN.

In five or ten years, what role do you expect genomics to play in medicine?

Dr. Mardis: We have clear indications today that genomics can add evidence to medical practice in terms of helping to better define diagnoses, treatments, prognoses, etc. I don’t think that genomics is always the answer or will provide the complete and only answer. But we are seeing evidence of where it is becoming incredibly helpful in the practice of medicine.

Dr. Lazarus: In my 18 years doing this, I have to say, what was considered wildly esoteric just a few years ago is considered ho-hum today. When I was an attending physician in a hospital in New York, it took an act of God to get authorization to run a PCR-based test. Back then, PCR occupied multiple rooms and took an entire day for the best technologists. Now, automation and plug-and-play systems make PCR routine. Today, we’re bringing in next-generation sequencing, which is radically different in the amount of data generated, the breadth and depth of insights. This has not taken all that long to emerge. So looking five to ten years down the road, what we’re talking about today will become routinized tomorrow for hospitals and physicians.

Dr. Rehm: Over the next five to ten years I anticipate that we will increasingly move to exome and genome sequencing as a routine aspect of care both for diagnostics and preventative care. But we still have a lot of work to do to understand the predictive capabilities of our genomes.

In the future, will a patient receiving a first diagnosis, say for cancer, be tested for genetic mutations as a routine procedure?

Dr. Rehm: It’s already happening. We have access to multiple tests, and different genes and mutations are being analyzed across many tumor types. So, yes, genetic testing will become more commonplace, even in community practices where it isn’t as common today. Over time, every tumor patient will have a genetic test to help inform both the treatments as well as prognostic information.

Dr. Mardis: Of course, it’s going to be a different answer for different people, types of tumors and access to a biopsy specimen. The paradigm here that’s quite interesting is lung cancer, because there are a large number of targeted therapies linked to genetic alterations that indicate a likelihood of response. It’s taken just 10 years from when mutations in the EGFR gene were first linked to responses to tyrosine kinase inhibitors to where we are offering these drugs as first-line therapy to patients with EGFR mutations. Just 10 years. With next-generation sequencing, and the increasing numbers of genes indicating targeted agents, I expect healthcare providers to turn to testing earlier in the process for more cancers.

Dr. Lazarus: If I could offer a third scenario – maybe looking out 10 years – I wonder if we’re going to get to the point where we’re not even going to wait until we see solid tumors but will be able to have a degree of surveillance that we don’t have presently. With the emergence of liquid biopsies, which are much less invasive, it is very much in the realm of possibility.

Dr. Mardis: That’s future-forward, but definitely a place where you’d like to be – cancer prevention as opposed to cancer treatment.

Speaking of prevention, what about testing for genetic factors that predispose people to a particular disease?

Dr. Rehm: It’s being done commonly now in families where an individual has a disease and is identified as having a particular mutation. That leads to other family members being screened genetically before they develop symptoms. That’s been shown to be very effective at preventing adverse outcomes, whether it’s cardiomyopathy or breast cancer. In some cases it may make sense even in the absence of a family history. Over the next five years, genome or exome sequencing may emerge as a common part of preventative health plans.

Dr. Mardis: One limiting factor is that there aren’t enough genetic counselors to go around and deal with all of the sequencing data. When you do these tests, you can’t enter into them lightly because you don’t know what you’re going to find. What you find may be very sobering and change the patient’s life in a dramatic way. We have to be prepared in ways other than simply being able to offer the tests.

»There will come a day when an entire set of genetic variants found in a patient can be reliably put in electronic health records and a physician will be able to ask questions of that data in real time.«
Dr. Heidi Rehm, FACMG Associate Professor of Pathology at Harvard Medical School and Director of the Laboratory for Molecular Medicine at Partners Healthcare, Boston, Massachusetts.

Let’s talk about the people side. Are we ready for this genomic revolution, in terms of MDs and genetic counselors? 

Dr. Mardis: There is an education issue for MDs. For genetic counselors, there just aren’t enough of them, quite frankly. It’s not a profession that attracts a lot of individuals, so there aren’t enough people to go around for all the genetic testing we’d like to do. Dr. Lazarus: From the point of view of someone who was teaching medical students and house staff for 15 years, I do wonder if we’re giving the next generation of clinicians the cognitive and linguistic skills they need. The majority of cancer care, for example, is in community-based centers, so those physicians have to be able to explain genetic information to patients. They can’t view the testing component as a black box. Many still view PCR that way, so with NGS as a more powerful and more mysterious tool, we need to provide insights in medical education.

Dr. Rehm: I do think genetics is becoming more commonplace in medical schools. We haven’t done it perfectly, but there is a lot of effort to understand and support the introduction of genetics into the routine practice of medicine, as opposed to only in medical genetics clinics.

Two of you are employed in research hospitals and great universities, doing cutting-edge work. Do you expect to see a democratization of these genomic technologies? 

Dr. Mardis: Clearly the goal is to democratize. If you look at current panel testing for cancer patients, for example, most people today are not getting access to these types of tests. To change this we’ve got to secure insurance reimbursement. And we need technologies that are easy to operate. For example, I think devices like the GeneReader NGS System recently introduced by QIAGEN will aid that democratization process because it offers a fairly straightforward workflow, doesn’t require a lot of technical expertise, and delivers the back-end analysis that’s so critically needed for community-based care centers.

Dr. Rehm: We’ll have to see whether these tests will be done on-site in every hospital or sent to large reference laboratories. In the beginning, academic labs are the ones who typically pilot the cutting-edge technologies. They get them validated, demonstrate the utility and prove it to the market, then the larger reference labs pick them up for broader clinical use. At some point, a transition to more robust approaches happens with lower cost investments for set up. You see that starting to happen now with panel-based tests in cancer, but exome and genome sequencing are still in cutting-edge academic labs and a small number of commercial reference labs.

How do you see bioinformatics aiding doctors for diagnoses in the future?

Dr. Rehm: There will come a day when an entire set of genetic variants found in a patient can be reliably put into electronic health records and a physician will be able to ask questions on that data in real time. You could support a physician with a tool that allows him to ask questions the moment he’s interacting with a patient, so a patient says, “Well, Doc, I’ve had this chest pain and I’ve also had shortness of breath and I’ve been very tired.” A physician could use a bioinformatics tool to query a person’s genome at that moment – and see if genetic information might help explain the patient’s condition. But much higher quality genomic knowledge bases will be needed to enable this type of direct physician use.

Dr. Mardis: In a tool like QIAGEN Clinical Insight, the interface puts the resources there to support the clinician. It’s not just a black and white text, it’s an ability to link out to published peer-reviewed literature that supports that variant and the therapy or set of therapies. The physician can click through to read additional information, connect to clinical trials, and see the information that’s out there. It’s a supporting infrastructure for decision making rather than a directive that says, “Mutation X in Gene Y equals treatment A.” And that’s what physicians want.

»Just having the ability to sequence a lot, in and of itself, does not solve the world’s problems. We have to be able to interpret the data and interpret it correctly. I see the next big challenge, especially in the clinical environment, coming from real data integration.«
Dr. Elaine Mardis, Professor of Genetics at Washington University and Director of Technology Development at the McDonnell Genome Institute, St. Louis, Missouri.

What is your view on the privacy of patients’ genomic information?

Dr. Rehm: This is a really important topic, privacy and data sharing. It is challenging in research, because in the past we had de-identified access to patient data but then had to worry about whether we could share important findings. I think the paradigm going forward is to consider the individual an active participant in research, to help him understand the benefits of sharing data for research, and enable the patient to drive his own data sharing, which allows a better balancing of risks and benefits that aligns with the patient’s own wishes, which are usually to increase sharing of this data.

Dr. Mardis: Right now, we spend an extraordinary amount of energy, time and caution trying to protect data and keep it private. If we could change the emphasis to criminalizing the misuse of data, such as for discrimination or another illegal purpose, that would free us to do things for medical research that we’d dearly love to do.

Dr. Lazarus: At least we have established, with the Affordable Care Act in the United States, some provisions that lend themselves to protecting the patients. The abolition of prior existing condition exclusionary periods and portability of coverage, for example, are terribly important.  

What roadblocks do you see to the adoption of genomic medicine?

Dr. Rehm: Right now, one of the things we’re desperately struggling with is a lack of insurance coverage for genetic and genomic testing. Part of the problem is that amassing enough information on each medical indication is challenging, especially demonstrating improved outcomes and cost savings. The other aspect is our inability to effectively understand much of the information we’re getting and interpret it for the patient and a physician in an accurate and understandable way.

Dr. Mardis: I absolutely agree. The educational component, I think, is critically important. And to some extent, we also have to up the ante in helping the general public to understand the importance of genomics in medicine. That’s harder, but I think it will happen over time.

Dr. Lazarus: It’s also difficult for regulatory agencies to keep up with innovation. Next-generation sequencing is so radically different, it’s going to be an uphill trek. And for reimbursement, the industry needs to implement those difficult trials to show clinical utility and outcomes.

»Autoimmune diseases are a field where clinicians very much need more insight. Diagnosis now is only at a descriptive level, so the application of NGS to autoimmune diseases is potentially very powerful.«
Dr. Tadd Lazarus, QIAGEN Chief Medical Officer Germantown, Maryland.

Do you foresee a pivotal “next big thing“ in genomics, or more incremental expansion? 

Dr. Mardis: Well, the major transformation has occurred around next-generation sequencing, but it also relies heavily on appropriate computational analysis of NGS data. Just having the ability to sequence a lot, in and of itself, does not solve the world’s problems. We have to be able to interpret the data and interpret it correctly. I see the next big challenge, especially in the clinical environment, coming from real data integration.

Dr. Rehm: I wholeheartedly agree. We had a major advance with next-generation sequencing, and the incremental advances will be in the interpretive process. There was this belief that the DNA is black and white, with four letters – A, G, C, and T – and therefore whatever it tells you is true. What we know now is that, yes, it’s useful information, but it must be contextualized and combined with other information from exposures and other omic data. Developing the tools to integrate all of these pieces of information will be a critical step to informing health.

Dr. Lazarus: For me, one of the next big things is the microbiome, both external and internal. Our knowledge of the interactions between microbial organisms and the immune function, as well as surveillance activities of the human gut, is really at the most rudimentary level. The microbiome is going to be extraordinary in what it reveals.  

After cancer, what is the next disease that will yield to genomically-guided treatment? 

Dr. Rehm: Germline genetics is an area where genetic information already is being used routinely. Collectively, if you add up the rare hereditary genetic disorders, they account for an enormous burden of disease. A lot of these cases are mutations where there is no family history, so just because it’s not in your parents or your relatives doesn’t mean it’s not genetic. Cardiomyopathy, the most common cause of sudden cardiac death under age 35, is one example. It is found in only one in 500 individuals but affects thousands of individuals.

Dr. Lazarus: Autoimmune diseases are a field where clinicians very much need more insight. Diagnosis now is only at a descriptive level, so the application of NGS to autoimmune diseases is potentially very powerful. Near-term, infectious disease is another area where NGS can be of great assistance. This is a very practical need we can address, the clinician’s dilemma of diagnosing sepsis and poly-microbial infections, as well as the physician’s issue, “I have good antibacterial coverage but should I be covering for fungi?” That is sometimes the great question in severe infections and, tragically, for many patients the diagnosis only occurs post mortem.