Medical Device Link.

Originally Published IVD Technology November/December 2003

Rob Lipshutz, PhD, is senior vice president of molecular diagnostics and emerging markets at Affymetrix (Santa Clara, CA). He can be contacted at rob_lipshutz@affymetrix.com.

In recent years, it has seemed to be an elusive goal: revolutionizing the healthcare arena through the advent of molecular diagnostics. Yet there is now light at the end of the tunnel. With the launch of the GeneChip System (GCS) 3000Dx by Affymetrix, both IVD manufacturers and the bodies that regulate them have a model to follow when bringing their molecular diagnostic tests and systems to market.

Yet many IVD manufacturers know not to be too optimistic just yet. More advances from the discovery arena must take place before the market will make full use of this technology for molecular diagnostics.

To learn about the experience of developing a molecular diagnostic system and then seeing it through a regulatory review, IVD Technology editor Richard Park spoke with Rob Lipshutz, PhD, senior vice president of molecular diagnostics and emerging markets at Affymetrix (Santa Clara, CA). In this interview, Lipshutz discusses bringing molecular technology from the research lab to the diagnostic market, steps that will enhance market integration and the future focus of the molecular diagnostics industry. The full text can be accessed on-line via the IVD Technology Web site at www.ivdtechnology.com.

IVD Technology: What have been the biggest developments and advances in the molecular diagnostic sector during the past few years?

Rob Lipshutz: From our perspective, the biggest development in the molecular diagnostic sector has been the ability to apply new genomic-level technologies to the diagnostic-discovery arena. This ability has led toward the development of signatures that can be hosted on new platforms like the Affymetrix microarray and, in some cases, more traditional platforms.

So, from a microarray company’s perspective, this has meant large-scale expression profiling and genotyping analysis. Research groups are continuing to use our tools to identify expression profiles for cancer to better classify disease, and to better classify and predict drug response and overall patient outcomes.

One example is the work by Todd Golub, now at the Broad Institute (Cambridge, MA). Initially, Golub used Genechip microarray expression profiling to distinguish between two clinical forms of leukemia–acute myelogenous leukemia (AML) and acute lymphocytic leukemia (ALL). His research team then went on to identify additional subtypes of the disease, proving that they could use expression profiles to go down to even lower levels of classification. That research was recapitulated in three different studies, namely those at Erasmus University Medical Center (Rotterdam, The Netherlands), St. Jude Children’s Research Hospital (Memphis, TN), and Torsten Haferlach’s lab in Germany.

St. Jude’s is looking at the same disease but focusing on ALL. The group is demonstrating that gene expression profiles can distinguish ALL subtypes according to their response to different therapeutics. Being able to determine whether a disease is resistant or sensitive to a certain treatment ahead of time is going to completely change the way people think about using molecular diagnostics for cancer.

As far as the discovery side of molecular diagnostics, people are making more extensive use of microarray-based genotyping. For example, researchers recently used Genechip arrays to identify the first gene liked to a form of Sudden Infant Death Syndrome (SIDS) that has tragically affected the Amish population. The amazing thing about this stuff is that it took the researchers less than a week to identify a gene that had completely eluded them using other technologies.

More and more people are using microarray-based technology for large-scale whole-genome scans, which are enabling genetic-level marker discovery on an unprecedented basis. I think one indication of this is the work that has been going on at Perlegen Sciences Inc. (Mountain View, CA). The company is using our technology in collaboration with a number of not-for-profit organizations to do marker discovery for diseases. Perlegen is also working with pharmaceutical companies like Glaxosmithkline (London) and Eli Lilly (Indianapolis) to look at drug efficacy and pharmaceutical-associated adverse events. In these studies, researchers are stratifying populations on the basis of genetic markers to identify patients who are susceptible to an adverse response and those who will respond in the most positive way to a given compound.

When covering what is new in terms of getting molecular diagnostic tests out to patients, I am going to speak more from our experience here at Affymetrix (Santa Clara, CA).

It was not that long ago that we talked to people in academia, in the clinical labs, and at large diagnostic companies about getting this technology out to patients. Some of them said that they didn’t believe the technology could be certified and manufactured for diagnostics. However, the fact that we have received the in vitro diagnostic CE mark for the GCS 3000Dx and that Roche Diagnostics (Indianapolis) has it for the Amplichip CYP450 Test is a testament to the fact that this technology can be made ready for diagnostics. A big part of developing diagnostic quality products lies in our commitment to quality manufacturing practices. In fact, this past year, we achieved ISO certification on our manufacturing facilities, both for instrumentation and the microarrays themselves.

This is the first new broad-based molecular technology that has moved into the market over the last couple of years. I think there is a big opportunity to integrate molecular testing now more than ever.

Have you seen an increase in people using whole-genome technologies for biomarker discoveries that will eventually lead to a clinical diagnostic?

We’re now at the point where we’re beginning to see clinical research turned clinical diagnostics. In the last couple of years, whole-genome studies have turned out discoveries that some of our partners have picked up, with the interest of deleveloping into a diagnostic.

Increasingly, pharmaceutical companies are beginning to look at discoveries coming out of whole-genome studies for potential ways to improve pharmaceutical success. Where it goes in terms of the diagnostics, we have yet to see. Already, the pharmaceutical industry is using microarray technology in trials for biomarker discovery. Whether or not they decide to use microarrays as a diagnostic becomes a question of where the company sees its product going at the end of the trial. This is the whole pharmacogenomics question: which way to go with a compound geared toward molecular targets.

While it’s still early, I can say that we have had substantive conversations with people at pharmaceutical companies who are definitely interested in understanding the availability of our technology for a companion diagnostic to go with a compound.

Obstacles to Commercialization

What obstacles are IVD manufacturers encountering when developing their molecular diagnostic technologies and when trying to do business in molecular diagnostics?

Particularly when you look at new technologies, it is really dependent on the levels of quality assurance that are acceptable for a regulated product. Making the jump from a research tool to a diagnostic product can be difficult. This is especially true for molecular tests because most new types of molecular diagnostics have their roots in research tools.

A lot of these technologies are coming out of relatively small, young companies, and the transition to an approved IVD is a difficult one—particularly for companies of that size. Typically, start-up companies don’t have the necessary experience and the background in GMP, QSR, or ISO, for example. That said, it can be done. But it takes a significant amount of dedication and commitment—organizational commitment—to really get a company to that place.

For a successful molecular diagnostic, a company must have a scalable technology that allows the manufacturer to get into the price points that make it attractive to the diagnostics business. The company must also be able to get good, strong, validated molecular signatures that meet an unmet medical need, and then to be able to translate that into patient education and physician education to create the demand for these products out in the marketplace.

A few years back, many in the IVD industry were saying that molecular diagnostics was going to revolutionize the industry. Unfortunately, it really has not. Do you think that it eventually will? And if so, when will molecular diagnostics play a larger role in the clinical diagnostics area?

The limiting factor has been the availability of good, validated molecular signatures that meet an unmet medical need that ultimately provides the value for a company to bring a product forward. However, the right set of research tools that allow scientists to exploit all the genetic information made available from the Human Genome Project has been lacking. That said, it’s been difficult to make the correlation between genetic and genomic signatures and physiological conditions.

For the great number of markers that people find, only a modest number will have enough value to take them forward and actually turn them into diagnostic tests. Now, using tools like microarrays for broad-scale genetic and genomic studies, enough research is being done and enough new markers are being identified, that we can begin to elucidate the complex signatures that may be required to develop new diagnostics for some diseases.

The work that was done in the first 15 or 20 years of modern genetics was largely focused on Mendelian traits (i.e., highly penetrant traits that show up in a relatively modest subset of a population). Well, the real unmet medical needs in the population, those that affect large numbers of people, tend to involve fairly complex diseases when you look at them on a genetic basis.

If you think about heart disease, diabetes, various kinds of neurological or neurodegenerative diseases, they all have complex genetics and it has taken a while to obtain the right set of tools to work on these diseases. Over the last three or four years, people have started to elucidate these more complex diseases and create the kinds of signatures that could be taken forward to diagnostics.

Gene expression analysis, particularly in cancer research, is moving very quickly. I see enough irons in the fire in different places that the future of molecular medicine will be realized, and it is now starting to grow at an ever-increasing rate.

In the past, people probably underestimated the challenges on the discovery side and did not take into full account the challenges on the implementation side of getting these tests to market.

Market Integration

Why did Affymetrix decide to launch the GCS 3000Dx system in Europe? What hurdles and obstacles did the company encounter in acquiring the CE mark?

The quality systems to which products and manufacturing processes have to conform are not significantly different between the United States and Europe. In our efforts to meet both the European requirements for the IVD CE mark, and the FDA requirements to be able to submit a 510(k), we found that we had fulfilled the requirement to register for a CE mark first, and so we did.

At the same time, we have always been committed to submitting these products for a 510(k) and we stand by our commitment there. It just happens that the process of applying for a CE mark involves self-evaluation and registration combined with the appropriate audits. Having gone through those audits and passed them, we decided to take the European registration so that we could start moving into that market.

The hurdles have certainly been there, but they haven’t been insurmountable. It was really a matter of saying, “What is it that we need to do?” Then putting the planning in place to do it. And that required placing the right kinds of people in control of quality assurance and regulatory requirements.

We built our manufacturing facility in Sacramento knowing that we were going to go for GMP, QSR, and ISO certification. So we anticipated all of the things that we knew we were going to have to put into the processes—things about the building, what you do about the walls and the ceiling— at that time. Quality is something you build in. You don’t just add it on at the end.

How does FDA and the U.S. market compare with the European regulatory bodies and the European market when it comes to molecular diagnostics and using them in the clinical setting?

The way I see it, both groups ask that products meet very similar requirements in terms of quality and safety, as well as efficacy for patients. Europe uses a process whereby certain inspections of quality systems are performed through a group that serves as a competent authority—the group that performs ISO inspections. But then parts of the inspection process that are specific to a given product are performed as a self-registration process that is then subject to a subsequent additional audit.

In the United States, FDA takes a different path. The agency says, “We would like you to meet many of the same requirements as needed in Europe. Then we would like to review what you are doing on a product-by-product basis prior to giving you approval to sell the product as an IVD, let us say, under a 510(k) or a PMA designation.”

Now, that said, certain levels of products in Europe have higher bars for regulatory approval and have more regulatory requirements in common with the U.S. system. But I want to reiterate that there is a great deal of similarity, as far as the quality and design systems you have to put in place, and the kinds of validations you have to do for approval in Europe and the United States.

What impact will the launch of the GCS 3000Dx system have on the IVD market and IVD manufacturers?

Well, this is the first system out there to run DNA microarrays that has met IVD requirements. This launch sets the stage for future molecular diagnostic tests that other companies could develop much as Roche has. We have a partnership with bioMerieux (Marcy l’Etoile, France) that has been going on for quite a while.
By putting out a regulatory-ready microarray platform on which different IVD manufacturers could create tests, we have changed the way IVDs can be brought to market. New molecular tests could be hosted on a common platform, which would create a lot more simplicity for the labs. It also provides an easier route to market.

Do you believe that other smaller to midsize companies should consider partnering with major companies such as Roche and bioMerieux when approaching the molecular diagnostics market?

Well I cannot speak for whether or not other companies should take this route. Given the nature of our technology, we decided that we are very good at making our microarrays.

When we thought about going into the diagnostics area, we determined the kinds of expertise we would need: the regulatory expertise, the ability to do all the commercial things you need to do in the diagnostic marketplace, and knowledge of the whole developmental life cycle. We concluded that it was a pretty big step up.

In order to really get the broadest distribution and broadest adoption of our technology, we decided to find some partners to go forward. So we developed what we call the “Powered by Affymetrix” model, in which Affymetrix manufactures its microarrays and instruments, while the diagnostic partner develops and commercializes the test. The resulting microarray products enable users to examine genomic sequences in more detail than ever before, providing a more efficient and complete method to diagnose a wide range of conditions and offer the possibility for improved patient care.

We see the Powered by Affymetrix program as a viable model that will allow us to go into diagnostic markets more broadly than if we tried to do it ourselves. The larger companies have the wherewithal, the experience, and the knowledge of markets that is well beyond what we could have had alone.

What is preventing genetic disposition tests from becoming a more substantial part of the market?

There are not many well-developed signatures that are both really informative and for which there are good choices that patients can make based on the outcomes of those tests. You don’t want to have a test and then not know what to do once you get the results.

Clearly, if you are going to do this kind of testing for genetic predisposition, there should be a change, either in lifestyle, in monitoring of a patient, or in diet that might help counter a certain disease that the person is predisposed to developing.

Another substantial component that is hindering integration of the tests is both patient and physician education. Patients and physicians will need to better understand why these tests are useful and how they can improve physicians’ ability to care for their patients.

In medical school, the use of genetics is going to have to move from being one course to being an integrated part of learning about treatment across the board in multiple disease areas.

Also, getting effective legislation passed on state and federal levels for protecting privacy and enforcing nondiscrimination on a genetic basis will be of fundamental importance to making genetic predisposition tests a part of routine clinical care.

Is there anything that IVD manufacturers can do to help facilitate education, maybe not so much in medical schools, but more for working with physicians, end-users, and your customers?

We are just starting to get our products into clinical laboratories, so we are not doing too much broad-based education there, but we will be starting to work toward educating them. Our diagnostic partners are going to be significant proponents of this kind of education. We would expect them to spend a fair bit of time, especially at the level of laboratory and physician education.

Relative to the general population, we have taken the initiative over the last couple of years to run an annual series of seminars that are open to the public and that address some of the areas around genetic technology, genetic testing, and genomics in general. We have run these Genetic Age Symposia for the past several years now in cities including, San Francisco; Washington, DC; and London.

When we have the opportunity to speak about our successes or the successes of our research customers in the marketplace, we do what we can to help inform and educate others about how this kind of genetics can benefit people’s lives.

I think what is going to help the most is when good tests are on the market that can make a substantive improvement in people’s lives through well-defined changes in course of treatment.

To prepare education programs, we are working with patient advocacy groups in areas with specific development programs. Once the advocacy groups let us know about the sensitivities and issues surrounding individuals who have certain diseases, we can more effectively let the groups know of the technology choices that are out there and what is coming down the pipeline. These groups are very, very powerful advocates of patient education and broad-based population education in these areas. So we are trying to make sure we help inform them and then see if there are ways that we can work together to help broadly educate the public.

Future Landscape of the Molecular Market

Is the field of molecular diagnostics more likely to be driven by traditional targets like infectious disease, or is the field going to move toward genetic mutation as a basis for diseases?

Infectious disease will certainly continue to be an important part of the molecular diagnostic field. Over time it may not be the key growth engine, but it will remain an important component. It seems likely that genetic predisposition and genetic components that are associated with disease will become a larger and larger part of molecular diagnostics.

Given the current interest, it seems only a matter of time before molecular diagnostics associated with cancer are going to become important clinical tools. Cancer has a substantial unmet need that high-complexity molecular technologies like microarrays appear to be capable of meeting. Microarrays for cancer diagnostics will likely move into the market in parallel with some of the genetic markers, and maybe even a little bit faster, because of the wealth of information that is coming out of the research community on new signatures.

So I think you will see some of the cancer work ramping up first and then, in the long term, the genetic markers will probably become the most substantive part of the market. But it will take some time before many of the informative markers get elucidated through large-scale studies and are finally implemented as a clinical diagnostic.

You mentioned cancer. What might be the role of other clinical areas in promoting the adoption and use of molecular technologies?

A number of clinical conditions may be viable areas. You imagine an area like osteoporosis, where early knowledge of susceptibility can allow a patient to alter diet, exercise, and perhaps take an earlier drug regimen. That can have a dramatic impact on improved quality of life, and reduce mortality and morbidity later in life.
People need to realize, and I think what we are going to find, is that although there are much more complex collections of markers, they get you to a much more specific decision point about what to do. A lot of the markers that people have discovered thus far are markers that might affect only a small percent of the population—maybe 3% or 7%, for example. We need to find much more broad-based markers that, in combination with others, can really help to inform what kind of treatment is most appropriate for a patient.

What future challenges will emerge in developing molecular diagnostics?

Although issues of privacy in relation to genetic tests are here already, they are going to come to the forefront as we move forward with development. Some other challenges include adequately training physicians and educating the public, getting the right genetic information into medical records, and being able to move toward electronic medical records.

Recently, the president made an effort to initiate a program that would require all personal medical records to be electronic and transportable within 10 years. We have teamed up with IBM to start to provide some of those solutions on a research basis, so that we can help with this initiative and provide more-broad ways of integrating all the genetic information about a patient to help make a clinical decision.

To make the best use of genetic and genomic information for patient care, that information can’t be used in a vacuum. It has to be integrated. Practitioners will want to take an integrative look at the patient’s current state across more-traditional clinical markers, across the patient’s history, and across anything else that is known—such as what compounds or therapeutics they have been taking. Working with IBM, we are beginning to find new ways to pull all these pieces of information together.

What new trends can we expect to see in the area of developing molecular diagnostics as companions to therapeutic compounds? Is this eventually going to happen on a wider scale?

It’s not really a question of if it’s going to happen, I think it is a question of when, and how fast. When pharmaceutical companies start to launch a compound, there’s a tradeoff between how FDA might treat the drug and how easy it might be to move things along to a companion system. That said, we’ve had a great example recently with the drug Iressa, which is used to treat lung cancer. The drug is only effective for about 10% of lung cancer patients, but for those patients, the results were absolutely astonishing. Virtually all of those patients had certain mutations in the EGFR gene, and it turns out that there is a very strong correlation between the genetic variation and the efficacy of the compound. That is an instance where the diagnostic will link up with the drug to identify those patients who stand to benefit most.

Substantial numbers of clinical trials are being run by our pharmaceutical customers, who are basically performing the kinds of studies that would lead to pharmaceutical products with companion diagnostics. We really have to wait and see what comes out. But given the relative efficacy of a lot of cancer compounds that make it to market, being able to stratify a population and offer specific treatments to specific patients with specific diseases going to be really important.

The choices that cancer patients and their practitioners have to make when deciding what kind of treatment regimes will work are fairly daunting. The information they currently have to go on could be a lot better, and we think molecular technologies, like microarrays, can go a long way toward improving that.

We are very excited to be on the cusp of a real change in the molecular diagnostics area—in helping to make that go forward in a responsible and scientifically sound way.

Copyright ©2004 IVD Technology