Advances in molecular diagnostics will ripple through the landscape of medicine. As these developments influence drug development, many believe the FDA will be forced to hold developers to higher standards.
Abstract
The dynamics in the field of molecular diagnostics are likely to ripple through the landscape of medicine, creating fundamental changes. Some believe that the U.S. Food and Drug Administration will play an increasing role, as companies are held to the new standards being created by the molecular diagnostics industry.
By the end of this year, about 100,000 PathVysion tests from Abbott will be performed. These are used to examine the cells of women diagnosed with breast cancer. In most cases, tiny fluorescent probes will bond and light up the presence of two copies of the HER-2 gene inside a cell. Once in every 4 or 5 tests, however, those twin copies will be replaced by a microscopic array of red-orange genes, an unmistakable alert that those women are prime candidates for trastuzumab (Herceptin) — a therapy that has a demonstrated high response rate among targeted cancer patients.
“There won’t be many, if any, novel oncology drugs to come to market without some form of diagnostic” in 5 to 15 years, predicts Oliver Schacht, PhD, the U.S. CEO of Epigenomics.
PHOTOGRAPH BY ALAN ABRAMOWITZ
Trastuzumab is held out as a therapy on the cutting edge of a new generation of drugs that will rely on molecular diagnostic tests to identify who is likely to benefit from biotech drugs now in the pipeline. Diagnostics groups have inked a raft of collaborations with drug discoverers to develop diagnostic tests. There are plenty of hurdles along the way, but diagnostics researchers believe they are at the creation of a new model of drug discovery that will play a leading role in speeding the emergence of personalized medicine in the next decade.
“I think it’s a really big deal, but I’m biased,” says Steven Projan, PhD, vice president, biological technologies for Wyeth Research. “There are two schools of thought: One, go after the biggest market possible and don’t ask dumb questions that might stratify your population; or, two, go after 100 percent of a smaller market rather than 5 percent of a bigger market.”
If a patient has a 20 percent chance of responding to a drug in category one or a 90 percent chance in category two, Projan wants to go with strategy number two. In the process, drug developers would identify a new set of drugs that will prove impossible for doctors and payers to ignore. Leaders in the diagnostics field say their products will help engineer a drug-discovery process that will significantly scale back the number of patients needed in costly human trials while delivering compelling data on safety and efficacy. And they’ll be able to prove to regulators the utility of some drugs in trials that earlier would have been junked by adverse results.
“This is an extremely nascent segment of the market,” says Frost & Sullivan clinical diagnostics analyst Martin Nejat. Yet, it will grow. Looking at the U.S. market for pharmacogenomics in terms of reagents, the $95 million business Nejat saw in 2004 should barrel along at a compounded annual growth rate of 26 percent through 2011.
“It’s exciting,” says Projan. “Many people are concluding that this is the future of drug development, as well as the marketing of drugs. In 10 years, probably 30 percent of new drug approvals will be linked to a molecular diagnostic test that will predict the drug’s efficacy or safety. Right now, it’s 1 or 2 percent.”
A LOT OF MONEY
“The best place for diagnostics players to participate is in the actual discovery work, when they think they have an idea,” says Steven Seelig, MD, PhD, divisional vice president of research and development at Abbott Molecular, and a leading figure in the development of the HER-2 test. “As a diagnostic partner, we can provide ideas about what to measure, how to phrase the question about utility, how to get regulatory approval, and how to get datasets for approval.”
That advantage, he believes, will grow, as new diagnostics push drugs to faster approval.
“I believe a model will evolve in which drugs may get approved at phase 2 — after a relatively small study with a highly selected population — for a very tight indication. Then you would have a drug on the market and that would allow you to expand its utility more rapidly.”
That early approval may not add up to anything close to a block-buster, but once a drug gains its first regulatory green light, developers will be able to start tapping revenue while they test it for new uses.
There are other paybacks as well.
“Looking at the selection of the patient population for the trastuzumab trials, they were able to save $35 million in clinical trial costs,” says Seelig. “And they got their drug on the market roughly 4 to 5 years sooner, but only for patients with metastatic breast cancer. Add up the total revenue that Genentech appreciated with those five years of earlier approval: $1.3 to $1.5 billion, and that’s just for trastuzumab.”
Genentech also got adjuvant approval at a quickened pace, adding even more accelerated revenue as a result of the diagnostic link.
“If we hadn’t selected patients for the pivotal phase 3 trial with a diagnostic test, we still wouldn’t be finished developing trastuzumab,” says Tom Metcalfe, head of the Roche Biomarker Program. “We would have had too few responders in the trial, and that would have diluted the therapeutic effect.”
“I hesitate to say breakthrough, but there will be studies that will allow us to choose a smaller number of patients,” says Projan. “This is going to happen.”
And probably, it will become run of the mill for many biotech drugs.
Metcalfe is looking for opportunities to develop diagnostic tests for Roche’s experimental drugs or in collaboration with other companies. And Roche Diagnostics is also developing tests to target existing therapies better.
One example, says Metcalfe, “ is a drug that we are developing to treat solid tumors — pertuzumab (Omnitarg). Our work to date has shown that not all patients and not all tumors respond in the same way to this drug, and we believe that we will need a test to identify those patients most likely to benefit.” So researchers will work to find bio-markers that predict a response to the drug and will examine how the diagnostics work as the drug goes through a midstage trial.
“We’ve got some fairly good hypotheses about the direction we should move in,” says Metcalfe. “Patients will be treated for the drug, and we will monitor their response at the same time we look at these markers. Some are basically blood tests. We’ll do some tests on the tumor and then, afterward, look at the correlation of the presence or absence [or level] of the marker and the response to the drug.”
That may not be easy to accomplish. Before this approach to looking for a novel diagnostic came along, researchers would have abandoned pertuzumab long ago.
Gaining approval from the U.S. Food and Drug Administration for a diagnostic test may be difficult and expensive, but not if you compare it to the cost of drug development. The total cost for a new, DNA-methylation-based cancer molecular classification test being worked on by Epigenomics could be some tens of millions of dollars — far less than the hundreds of millions of dollars that go into many new drugs, says Oliver Schacht, PhD, the company’s U.S. CEO.
BIGGEST OPPORTUNITIES
The groundbreaking development of the HER-2 test for breast cancer — or other tests being developed now around other cancer drugs — is no accident.
“There won’t be many, if any, novel oncology drugs that come to market without some form of diagnostic, 5 to 15 years down the line,” says Schacht. “Oncology is immediately life threatening, and most drugs don’t work for all patients.”
“Oncology is the area with the most potential, ” agrees Metcalfe. “The reason for oncology is the extremely large unmet medical needs, the fatal nature of the disease, the work that’s being invested in terms of many novel therapies.”
Yet there are several other areas where there is significant potential.
“If we had a reliable marker for diagnosing patients with early onset Alzheimer’s,” says Metcalfe, “and distinguishing it from mild cognitive impairment, this would speed up our ability to develop drugs for Alzheimer’s and how we go about targeting the right patients for trials.”
“If we hadn’t selected patients for the phase 3 trial with a diagnostic test, we still wouldn’t be finished developing trastuzumab.”
— Tom Metcalfe, Roche Biomarker Program
Neuropsychiatric diseases like clinical depression, for which psychiatrists are typically limited to drugs that work in 1 of 5 people, would greatly benefit from molecular diagnostics, says Seelig. “A lot of people have behavioral disorders. This is a huge drain on our society.”
Another expensive area with hit-or-miss therapeutics comprises hypermetabolic syndromes, such as hypertension or type 2 diabetes.
Nevertheless, there are limits to what diagnostics will be able to do. They are unlikely to pick up on rare adverse reactions, which could affect the life or death of 1 in, say, 20,000 patients. To become more readily detectable, the signal would have to be demonstrated in 0.5 to 1 percent of the population studied.
“I wish I could tell you we had a battery of tests to pick patients for trials,” says Projan, “but we’re just getting started. It will happen only if there is access to pharmacogenomic data from the patients.”
That won’t be easy. Just asking for patients’ permission to include them in a pharmacogenomic study can lead quickly to a dead end. “We live in an environment where pre-existing conditions might prevent people from changing jobs or even could restrict the benefit packages they receive.” says Projan.
The genetic signatures for predisposition to heart disease or Alzheimer’s are well defined. And people dread the idea that if they fall into one of those groups, that information could blight their lives if it reached their employers.
“Confidentiality is key,” says Projan. “We go through a detailed blinding process. Deanonymizing is something we can’t do from our side. Long-term, when it becomes a diagnostic test for selection or deselection, though, that information could be used like it is for other pre-existing conditions — as far as who gets certain kinds of coverage.”
“There’s a strong emotional overlay to this whole story about using genetics to understand diseases,” says Seelig. And when patients say no to researchers, that resistance can create a major obstacle in developing new diagnostics.
“If those samples are available,” says Seelig, “we can figure things out. If we don’t have them, we can never figure it out.
“Some of it is just logistics. In cancer trials, in particular, almost all new cancer drugs have been tested first in individuals who have failed one or more therapies in the past, so it is very uncommon for a person with recurrent cancer to have tissue biopsied. If you don’t have tissue, you can’t do tests. The primary tumor was taken out before the recurrence of the disease, 3, 5, 8 years later. Finding the primary tumor tissue sample can be a challenge. Further, the tissue obtained at the initial diagnosis may no longer reflect the biology of the recurrent tumor, and thus, would not be relevant to predicting response to the new agent.
“People have gotten much smarter about informed consent. Now we’re asking for broad approval for consent. Before, we’d say we want tissue for substance A, and if that fails, and you get the results two years later, at which point the patient is dead, then you can’t measure B, C, or D. So, now we ask for broad approval to run a variety of tests.”
GETTING THE RIGHT INCENTIVES
Once a diagnostic is out on the market, though, there’s a new pitfall waiting for test manufacturers who are looking to get paid.
“This almost becomes a chicken-and-egg issue,” says Metcalfe. “If you don’t feel confident that you will get value-based reimbursements, it’s hard to invest with confidence to prove clinical utility of the diagnostic.
“Certainly, there is increasing dialogue about the value of diagnostics. There’s an understanding of the value of personalized healthcare — and that necessitates good diagnostics. There is increased dialogue about the need to have the right incentives in place to encourage people to invest in diagnostics.”
Once that’s worked out, though, diagnostics companies will have a real advantage, says Projan. After a diagnostic test is approved and out on the market, it will be used much more frequently than the drug to which it’s linked.
There are potential savings for drug companies to ponder, as well.
“The pricing model will change,” says Richard A. Silfen, president of Cangen Biotechnologies, which is working to establish a bloodbased molecular test for the early detection of lung cancer. “Once you understand how to target a drug, you also won’t need as many marketing dollars to blast the message. When a person has the test result, physicians will know exactly which therapeutic should be used.”
The United States will play a leading role in determining how swiftly these diagnostic tests are developed, Metcalfe says, because of the sheer size of the market. The FDA has demonstrated its interest in the field, especially Janet Woodcock, the FDA’s COO who is responsible for advancing the Critical Path initiative. And Medicare could well blaze the reimbursement trail for private U.S. payers.
The FDA gets a lot of credit for pushing the entire field, says Seelig. The agency’s early support contributed significantly to the fact that trastuzumab is linked to a diagnostic. “A more recent example,” says Seelig, “is Third Wave’s approval of the Invader UGT1A1 test. I believe Third Wave got approval by showing they could reliably measure the mutation, while clinical utility was established in the literature.”
Seelig too believes that the FDA will play an even bigger role, as other companies are held to the new standards being created by molecular diagnostics.
“If a company comes to us with data about toxicity and is able to assess something, and a second or third company comes along and has not measured that, the agency will send them back to assess their drug for toxicity potential,” he says.
And some believe the dynamics in this field may change medicine in fundamental ways.
“I have a hunch that we’re in the middle of a changing landscape,” says Matt Winkler, who recently sold the RNA company Ambion and is taking 100 of his employees to start the new diagnostics company Asuragen. “You’ll need sophisticated diagnostics to figure out who should be getting which drug, and therapeutics will become more fragmented. You can see this process occurring with the emergence of higher cost diagnostics from companies such as Myriad Genetics and Genomic Health.”
“It will continue to grow, but won’t be dominant for 7 to 10 years,” says Seelig. “Maybe we’ll do better in the future, maybe we’ll be smarter. But it’s going to take a while. It will take a massively concerted national effort to do it. It’s not going to be done by an individual investigator at a university or in industry. The requirements for the reliably collected patient samples, the large number of patients, and the coordination between therapeutic trials with diagnostic efforts on critical national healthcare issues far exceeds the capabilities of small groups.”