Companion diagnostics are emerging as a key part of personalized medicine. Particularly in oncology, patients are being better served by drugs for which patients are selected via in-vitro diagnostic tests.
The field is promising but nascent: so far, the potential of companion diagnostics is greater than the number of drug-and-diagnostic products that is actually commercially available.
To find out why this is the case, and to learn more about the future of companion diagnostics, IVD Technology editor Richard Park spoke with Walter Koch, PhD, vice president and head of global research for Roche Molecular Systems.

IVD Technology: How would you characterize the current state of companion diagnostics for personalized medicine?
Walter Koch: A metaphor comes to my mind: cherry trees at the beginning of the spring season. I think the field is blossoming. We’re seeing the first flowers on the tree, but there are a lot of buds waiting to explode.
That is how I envision companion diagnostics right now. We have a few examples already in the marketplace. Just recently, Zelboraf and the cobas 4800 BRAF test is the first time in at least ten years that the FDA has approved a diagnostic together with a therapeutic. The field is nascent but coming on strong.

You sound optimistic. Are there other examples of products or companies that give you that impression of blossoming within the field of companion diagnostics?
Yes, specifically in the field of oncology, but not limited to that area. The state of the science has reached a point where specific aberrations in cancer cells can be targeted at a level of a therapeutic. But you need to know that that aberration is present to target it, and that’s where this model is playing out quite a bit. In fact, within a couple of weeks of our approval, the anaplastic lymphoma kinase, or ALK, inhibitor Xalkori was approved, and it also requires a diagnostic to identify the three to four percent of lung-cancer patients whose cancer is driven by an ALK fusion gene. Just within a few weeks of each other, there are two examples where this type of approach is really coming to fruition.

What are the current challenges that remain when IVD manufacturers develop companion diagnostics for personalized medicine?
The biggest challenge is the somewhat disparate development processes and timelines for diagnostics and therapeutics. In the BRAF mutation test/Zelboraf example, starting back in 2005 or 2006, we planned in preclinical stages of the therapeutic development and developed a prototype assay. We got that assay ready for Phase 1 and 1B studies so that we could obtain an investigational device exemption.
In the case of Zelboraf, there was a very strong indication that the presence of the BRAF mutation was associated with response of melanoma patients, something you rarely see with a 30-patient safety study. Importantly, it also gives the diagnostic manufacturer time to lock down the assay, the instrument platform, and the software in time for the pivotal trials, the Phase 2 and then the Phase 3 efficacy trials.
Thanks to a great collaboration, in the case of Zelboraf and our test, we managed to have both the therapeutic and the drug approved on the same day.
On the other hand, a less desirable scenario is that thedrug enters the market without any selection of patients. An example of that is the KRAS mutation test for colorectal cancer patients, where for several years now we have known that colorectal cancer patients with KRAS mutations do not respond to monoclonal antibodies to EGFR.It’s the standard of care , but there is no FDA-approved test because of the conundrum of trying to register a test that had not been used in the clinical trial. So the earlier we are engaged, the better.

How can IVD manufacturers overcome challenges such as these when developing companion diagnostics for personalized medicine?
Early engagement and involvement in the co-development project planning. In the Roche Zelboraf example, we have actually had a diagnostics team member embedded within the pharma development team. That is the ideal situation.

What sort of challenges are involved in overcoming differences in culture between the diagnostics and pharmaceutical businesses? Is culture a factor as the two groups work on a major project like this?
The first time a diagnostics group tells a pharma team what is involved in developing a companion diagnostic, jaws do drop. They are surprised at the level of rigor that is required from the FDA.
So when you tell pharma colleagues the timelines, the multiple lots that have to be made in the manufacturing scenario, that have to then be tested in multiple sites by multiple users on multiple days to validate performance, they are a bit surprised.
It is an education process as much as it is a cultural difference. I believe we understand much better now, because of our several years of experience.

What companion diagnostics products does Roche currently have on the market?
The cobas 4800 BRAF mutation test was approved in the United States and also has CE mark approval. The cobas KRAS mutation test also has CE mark approval. Ventana has the newly approved Inform HER2 Dual ISH assay to select breast cancer patients who may be eligible for treatment with Herceptin.
Ventana also has estrogen receptor and progesterone receptor immunohistochemistry tests that are used for breast cancer patients to determine if they will receive hormonal therapy.

What companion diagnostics development projects is Roche currently involved in and what is the status of those projects?
I would say across the entire Roche organization there are well over 100 codevelopment and personalized healthcare projects across all therapeutic areas.
Every drug development team considers a biomarker strategy. If the hypothesis is sound, and there are early preclinical data that suggest a certain level of evidence that it should go forward, we work on a potential companion diagnostic.
I don’t personally believe that every drug is going to have a companion diagnostic, but it is at least considered in every program.  

Among the many projects that Roche is involved in, what clinical area or disease state is most represented?
I believe oncology is clearly the area where we have the greatest presence and where there’s also the greatest amount of activity. Oncology is where we see the most intense activity, but I certainly know of several in virology as well, and in the neurology area.

Is developing a companion diagnostic for a particular drug something that is always taken into consideration and discussed at the beginning of the drug-development process?
Absolutely. Regardless of how early it is, once a drug development team has a molecule that looks like it might have the right kind of properties to hit a target based on cell culture data, in vitro systems, and so on, the team starts to contemplate whether there would be a biomarker for that. That is probably the best way to deliver effective medicines.
No drug works for everyone. If you’ve got a headache, there are some people that respond better to aspirin, while some respond better to acetaminophen and others to ibuprofen. We don’t always know why, but it clearly reflects differences in humans that manifest in many other ways as well. There are drugs that don’t provide much benefit for a very significant number of people, and if you could avoid treating them with something that doesn’t work, not only are you sparing them unsuccessfully treating their malady, but also potential side effects as well.
It makes sense from every perspective that you would want to try to develop a companion diagnostic test if a good predictive biomarker exists.

Roche’s COO, Daniel O’Day, was recently quoted as saying, “Efforts by government to curb healthcare spending will be beneficial to growth in companion diagnostics for personalized medicine.” Do you agree?
I agree. It is a logical extension that with the aging demographics of the planet, and every healthcare system in the world under some duress to deliver quality healthcare with a finite amount of resources, anything you can do to improve the efficacy of how you deliver healthcare is a good thing. You can quantify it in a variety of different ways.
If we can make new therapeutic agents more efficacious and focus more on the patients who will benefit versus those who will not, by using companion diagnostics, that is simply a dollars-and-cents equation.
Diagnostics can increasingly improve the value of the therapeutics not only to the individual patients, which is the first thing we care about, but also to the healthcare systems and the payers.
It is clear to see that there are many opportunities for diagnostics to contribute to improving outcomes with medicines and better healthcare delivery. And we certainly plan to be a part of that, with oncology leading the way.

FDA recently released its draft guidance on companion diagnostics. Do you think the document adequately addresses the various regulatory issues and challenges IVD manufacturers encounter when developing companion diagnostics for personalized medicine?
It is good that the agency is thinking about this and that they’re soliciting feedback from industry and professional organizations.

What is currently the level of clinical use of companion diagnostics by doctors, physicians, and other clinicians?
There is a lot going on, some of which is actually standard of care, mainly for oncology. For breast cancer patients there are three biomarkers that are going to be evaluated: ER, PR, and HER2, which covers about seventy-five percent of breast cancer patients who, based on test results, are now going to be candidates to receive a targeted therapy-either an aromatase inhibitor, or tamoxifen, or Herceptin.
In colorectal cancer, a patient with metastatic disease who is a candidate for the EGFR monoclonal antibody therapies will get a KRAS test today, despite the fact that there’s not an FDA-approved test. In the United States, Europe, and throughout the world, this is the standard of care. For melanoma, patients will be tested for the presence of the BRAF V600 mutation.

So from what you understand, have physicians, doctors, and clinicians accepted the usage of companion diagnostics as part of the standard of care, without much resistance?
When a companion diagnostic has been validated in a rigorous clinical trial, there appears to be rapid uptake.

As long as it’s validated, they essentially are on board.
Yes. And particularly if it ends up in the professional guidelines, clinicians will do it, because if you don’t, you’re really not practicing modern medicine. Those professional guidelines are a critical element.

What has been the reception to and usage companion diagnostics outside the United States?
I don’t think there is any fundamental difference. There are differences in regulatory paths for getting into the market. Zelboraf is not yet approved in Europe, but it is approved in the United States.
Interestingly, it is a bit easier for us to get diagnostics approved in Europe. The CE-mark process, with some exceptions, is a self-certification process. But that is the only fundamental difference. Drug and diagnostic development companies are global.

What are the future prospects for companion diagnostics?
I think they are really, really good. I believe we’ll see many more examples in the coming years where diagnostic tests are developed together clinically in the same clinical trial with the therapeutics, because it enables the personalized medicine approach and can specifically improve the outcomes for patients over the one-size-fits-all approach.
The continued focus on companion diagnostics and personalized medicine is good for the industry. It’s good for patients. It’s good for healthcare systems and payers.

What future challenges will IVD manufacturers encounter when developing companion diagnostics for personalized medicine?
From where I sit in Research, science and technology are moving at such an extremely rapid pace. And with ever-decreasing costs, there are daily announcements of new biologically relevant biomarker discoveries. In academic centers that are on the cutting edge, we’re seeing some of those discoveries converted into laboratory-developed tests very rapidly, with the results used long before they’ve gone through the more typical rigor of a randomized clinical trial that would be required for an FDA approval.
Once the FDA has approved a new therapeutic with a specific clinically validated companion IVD, there may already be several laboratory-developed tests on the market, and the clinical laboratories will market their tests for use with that therapeutic, even though that use is not supported by clinical data, nor has it been subject to the FDA’s rigorous review of safety and effectiveness.  
We believe there needs to be a more consistent, risk-based regulation of diagnostic tests regardless of the originator, and that without it, patient safety may suffer.

Do you have any final comments on the state of personalized medicine and companion diagnostics?
It is exciting and exhilarating for those of us who are able to participate in this new way of developing medicines. And I am convinced we’re making a fundamental difference for patients.
We’ve had patients come to our site and tell us what a difference our tests made for them and how their disease was managed. People who had been diagnosed with malignant melanoma and, frankly, didn’t have any treatment options left, have a new lease on life with Zelboraf following BRAF testing-based eligibility.
So it’s very gratifying to see that we can help patients. It is very exciting to be a part of this.

Walter H. Koch, PhD is vice president and head of global research for Roche Molecular Systems. He has held this position since 2005. As a member of the Executive Leadership Team, he sits on the Life Cycle and Business Development committees and chairs the Research Portfolio Committee. He can be reached via Jacqueline Wallach at jacqueline.wallach@roche.com.