Name: Anna Barker
Position: President, National Biomarker Development Alliance; Professor, Arizona State University
Background: Deputy Director, National Cancer Institute
Launched last week by the Research Collaboratory at Arizona State University, the National Biomarker Development Alliance is a non-profit organization that aims to develop standards and guidelines for aiding biomarker research and implementation.
The organization plans to focus on various classes of markers, including genomic, proteomic, imaging, and complex biomarkers, creating best practices, guidelines, and standard operating procedures that will support new models of biomarker development.
It also aims to develop resources including a national biospecimen repository and a common biomarker database.
Leading the effort is Anna Barker, NBDA president and co-founder. Formerly the deputy director of the National Cancer Institute, Barker is a professor at ASU where she directs the university's Transformative Healthcare Networks and co-directs its Complex Adaptive Systems Initiative.
ProteoMonitor spoke to Barker this week about the NBDA's launch and its goals.
Below is an edited version of the interview.
What do you see as the major issues facing biomarker development?
It's a systems problem and we tend to, depending on where you sit in the process, you tend to focus on that piece of the process. When I was back at NCI, the [Clinical Proteomic Technologies for Cancer] initiative, which is still ongoing … there, we were attempting and still are attempting to determine how you arrive at technology standards, how you arrive at sample prep standards, how you arrive at a legitimate clinical question, and then once you get to assay development, how do you set cutpoints depending on the analyte, etc. So there is such a complex of problems, and the reason we actually decided to [form the NBDA] was that we felt as though there was no one group, no one organization, looking at this problem and saying, "Who really has to be engaged here to really ultimately move biomarkers along at least to a point where we are reasonably confident that if you follow these best practice standards, guidelines, whatever we actually agree is the right way, then, ultimately, you can at least depend on the fact that you have developed a biomarker appropriately."
Now, whether or not [a biomarker] addresses the clinical question or whether the clinical question is going to change practice is a secondary issue. I think one of the things we have concluded in the last 18 months of analysis of the problem is that so many people start, sometimes naively, with either a clinical question that, A, can't be answered with the biomarker they are developing because it's not fit for purpose; or, B, even if they do develop it, it's probably not something that will find its way into practice because there may be something better, or it's ultimately irrelevant in terms of the disease process.
It's so extraordinarily complex. So we are trying to [develop], essentially, an end-to-end process so that you can make decisions at each of these intersections and feel confident that as you move from one phase of biomarker development to another, you can be pretty confident that you have done everything that should be done to reinforce the decision that you are making.
How, specifically, does the NBDA plan to address these issues? Do you see your role as primarily helping to develop guidelines and standards? Will you be doing any actual biomarker development work or building of research resources?
We have been rounding up the communities, having them sit for a couple of days together and coming up with lists of things that have to be done – understanding the problem, one, and then, two, what are the directions for solutions.
NBDA is a real organization. We have start-up funds and I think one of the most incredible teams in the world to work on this. I'm fortunate to have [as Chief Medical Officer] Carolyn Compton [former director of the NCI Office of Biorepositories and Biospecimen Research], who is probably one of the best people in the world in the biospecimen world, and [Biodesign Institute founder and the NBDA's interim Chief Science Officer] George Poste, who is a deep thinker about biomarkers, especially from the pharmaceutical perspective. We have an A-plus team here to really put on this problem. We're a non-profit that is dedicated to solving the problem. Otherwise, [despite] all of the time and effort and sweat and tears that is put into issues of personalized or precision medicine, [these fields are] going to be extraordinarily slow to develop. We still have so much failure in our phase III trials; we have so much failure in biomarkers overall. And that I think has especially hit us hard in proteomics.
So, we are doing several things. For example, what we have concluded is that an awful lot of standards work, and evidence-based standards at that, in proteomics exist out there. It's just that we have never reached consensus as a community on which of those, or [if] any of those, [standards] are actually the right ones to drive what we are trying to do in terms of getting proteins that actually are reproducibly measurable and meaningful, and ensuring that we have the best samples, the best technology, etc. What we have decided to do there is start with a couple of consensus conferences this year in the fall to actually take on that problem.
The other thing that we are doing is we are working in the diabetes space to develop, ideally, a pathway for blood-based biomarkers. We are taking different approaches for different classes of biomarkers. For genomics, we are taking a different approach. For imaging, we have a very near-term project we are working on where we need a very clear path for developing magnetic resonance imaging for certain kinds of tumors. We are working with a group that is very far along in terms of having the data to actually do that.
We're also trying to set up a couple of national resources that are sorely needed. One is a biorepository that would actually have standard samples that we would be able to make available.
We've classified the biomarkers into what we believe are the legitimate buckets in terms of biomarkers that we know today, including complex biomarkers. And, what we have done is gone through an analysis process and analyzed every step of those various processes, and now we are getting on the ground to start to implement plans we have to arrive at solutions. We have a big agenda, and we have a lot to do. We're hiring people, and we're going to work as effectively as we can through existing networks of experts who are willing to put in some in-kind support for what we are trying to do. The people who can help us actually solve this problem are the people who are in the middle of it. Industry is involved in what we are doing. Some of the best biomarker researchers out there are involved. Patients are involved. Patient advocacy groups are involved. We have a lot of the right people, and we'll engage more as we go forward over the next couple of years.
Could you speak more about the biorepository work? Obtaining high-quality, appropriate samples seems to be one of the major bottlenecks in biomarker development right now.
It is the bottleneck. I worked on that quite a bit when I was at NCI with Dr. Compton. She set up the cancer hub, which was eventually going to be a new national biorepository. We never got that far [at NCI] but, ideally – and other countries have actually done this – our plans are that we are going to collect and actually set standards around very specific biomarker classes. A lot of this work is done. For proteins, for example, we spent an enormous amount of resources at NCI doing this for various cancer samples, and that group now has moved on to take data from The Cancer Genome Atlas and to use those samples to see what the translation is through expression and into proteins.
So, we are just getting underway, but our plans are to have representative samples and standards and standard samples for all the various biomarker classes. We're starting with genomics, but we have a lot of access to samples for proteins, as well. Over time what we would like to do is set a national set of standards that could be adopted by everybody. For biomarker science it would be great if people were working off the same set of standards. It would make everyone's life much, much easier. Having said that, it's very hard because of things like preanalytical variables that affect the quality of the sample. This has been very difficult for our community to come to grips with, by the way. I have watched this for years, and it is just that when it comes to samples it is expensive and very difficult to get them, especially for individual investigators. And when they get them, they don't [necessarily] know the history of them. So this is a very difficult problem for individual investigators and one that we hope to do something about.
Where is the field at in terms of understanding these issues? Have researchers grown more aware and more sophisticated in their approach to biomarker development in recent years?
I think it varies. I think that we should be further along in proteomics than we are in other areas, because, for instance, whole-genome sequencing has just come onto the scene of the kind of high-throughput activity, and we are just kind of getting into that, so there are lots of questions to answer there. The same is true for imaging, [but] for different reasons. The complex biomarkers — once you start combining proteomic biomarkers with imaging, now you have sets of problems you haven't faced before. This is a complex landscape that needs to be analyzed based on the question you are asking, the context of use, the type of biomarker, and, obviously, its source.
I think that we have spent a lot of resources to bring a couple of these areas, in theory, pretty close to where we should be. I think proteomics is one of those. We've decided over a couple of think tanks that our best approach [for proteomics] would be to do a consensus conference to see how close we are to being where we should be in terms of the samples, technology standards for mass spec, etc. Nothing is going to be straightforward, but I think it could be easier than some of the other areas because I believe that, especially for proteins — because they are so demanding, and we have certainly had some critical missteps in this area, especially in the oncology area — I think this is an area where we should be close to where we need to be [in terms of standards and technology].
I think the bigger question for proteins is how good is it going to be? How good are those biomarkers going to be? We've spent a lot of time looking for blood-based biomarkers and low abundance proteins, and how meaningful they are is still a question. I think for proteins I'm pretty sanguine that we might be able to make some real progress by just getting consensus on what is the best way forward on some of these things. The one area that is still troubling for proteins, though, are preanalytical variables – what happened to the patient before they got to the operating room, or the covariants here, the drugs they are taking, etc.− that could impact the biology. And we haven't gone very far down that road to answer those questions.
Does the NBDA have any role to play on the commercialization side of things? Or is your focus primarily on the discovery and validation steps?
I think we started out on the [discovery] side of things. We are very interested in getting the science right and getting an end-to-end network-based approach in place so that we can answer questions meaningfully and, hopefully, efficiently. But what has come up time and time again is the fact that the diagnostics industry overall has stalled because of economics and the lack of a transparent, predictable regulatory approach, especially for some of the sophisticated biomarkers that are coming today. And this is generally the purview of small companies, and these companies struggle to survive and many don't survive. So, we have been asked to see if we can help with this business model question.In other words, can you tie an end-to-end system into a transparent process that would make this sufficiently predictable and potentially successful enough that you could attract more capital and more investment into this field?
I don't know the answer to that, but we have agreed to hold a couple of think tanks on that to see if there are ways that we could help. One of the things that actually strikes us, and it has come up time and time again, is that many times when we start in discovery and we make a viable discovery that is particularly informative about the biology, people get very excited, and they aren't thinking about the scale they would have to have for something to be economically viable. Being able to scale something is not intuitive in many cases for people working at the bench, but that has a lot to do with the economic viability of what ultimately becomes an assay.
And then beyond that, what can you charge for it? And you know the classic story – it may cost several thousand or tens of thousands of dollars for a targeted treatment, but you may have to sell a [biomarker] assay for less than $1,000, many times less than $500. And, ultimately, they tell us that for screening, it has to be cheap because you can't get reimbursed for it. I think the reimbursement issue is a huge one.