TCS Daily

First, Do No Harm

By Scott Gottlieb - July 1, 2002 12:00 AM

Eight years ago, working on the public offering of the Salt Lake City-based biotechnology company Myriad Genetics, I thought I found the next Microsoft.

Myriad was the discoverer of BRCA1, the so-called breast cancer gene, which left women who harbored it as much as ten times more likely to develop the disease. The company's product probed patients' DNA to see if they had the mutant gene and was the leading edge of a new medical paradigm.

Eight years later, Myriad's stock has traded far behind Microsoft's meteoric rise, not because the tests didn't work, but doctors didn't know how to use the information they were generating. The technology to diagnose the problem was far ahead of the ability to treat it. That's why most of my patients pass on Myriad's test. While they're curious to know if they have the gene, nobody knows what a positive result really means.

BRCA1 is a molecular diagnostic, a term used broadly to refer to the full spectrum of gene-based testing. As an investment vehicle, diagnostic tools have some advantages over new drugs. First, they're cheaper and easier to develop than drugs. About 90 percent of new drugs that make it into clinical trials fail, while more than half of all diagnostics eventually get approved. And bringing a new drug to market takes between five to 15 years and can cost as much as $800 million, while diagnostics take from 18 to 48 months, and cost only $5 million to $20 million.

If the diagnostics sound like the smarter investment, there's a rub. Molecular tests are the low-hanging fruit from the human genome and have been developed first, before the targeted treatments that are sure to follow. So in Myriad's case, that means the company knows BRCA1 is linked to breast cancer, but doesn't fully understand how. Unlocking that second mystery is the key to developing treatments that can intervene in the gene's malicious machinery and reveal the true value of the tests that probe for it. Until corresponding drugs are available, diagnostic tests are often little more than a curiosity.

This reality is lost on an increasing number of patients, pundits, and regulators, as well as journalists like the New York Times' techno phobic science writer Gina Kolata, who predicts recently in The Science Times that the future of gene-based diagnostic tests for diseases like cancer may be dead.

Her peeve is with a series of articles published recently in the New England Journal of Medicine, examining the benefits of an early, gene-based test for a particularly vexing childhood cancer: Neuroblastoma. The disease attacks the nervous system, and is one of the most common, and most lethal, tumors in children. When Japanese researchers found a diagnostic test that could scan children's urine for signs of the cancers long before symptoms appeared they were rightly relieved. A large screening program started in Japan, while studies of the test began in Quebec and Germany.

At first, the results looked spectacular. Many more cancers were found, and they were found early. The children underwent surgery, usually of the adrenal gland, where the tumors tend to lodge, and their cancers went away. But in the articles recently published in the New England Journal, researchers reveal there was no decline in the number of toddlers who developed advanced cancers, and the death rates from the disease stayed unchanged.

While some would use these anecdotes of human suffering to wrap molecular diagnostics in tighter regulations, these results didn't turn on faulty technology but the lightening pace of that development. One reason the diagnostic tests have been developed ahead of the treatments is scientific: It's easier to find a gene and develop a way to test for its presence than develop a drug that targets it. All you need for the test is the gene. To make the drug, you have to learn how the gene works, a far more difficult scientific task. And then find chemicals that will safely block its cellular machinery in humans.

There's another reason the tests have come first: Relaxed regulations. FDA rules regulating the development of genetic tests are a lot less restrictive than those governing the development of new drugs.

Right now, the majority of new genetic tests are being developed by laboratories for their own use, referred to as "in-house tests" or "home brews." The FDA says it has the legal prerogative to regulate home-brew tests, but so far the agency has exercised some discretion and refrained from flexing that authority.

As a result, new genetic tests can be quickly used by doctors even though they don't have FDA approval. Of course, there are still some restrictions. Labs that sell them must instruct doctors to tell their patients that the results are "experimental" and shouldn't be used for therapeutic decision making. And so long as the tests aren't FDA approved only a small number of "high complexity laboratories" are permitted to perform these tests, and the labs that do so can't get full reimbursement from Medicare.

If that sounds like a lot of red tape, it's far better than the system for regulating new drugs. In the case of genetic tests, the regulatory machinery, as it's currently practiced, allows companies to use tests on real patients, generating revenue at the same time that they collect data that can be used to apply for full FDA approval.

It also gives doctors and their patients a chance to use some of the best tests as soon as they're available. For example, new genetic tests were first used to augment an earlier generation of tests to screen the blood supply for Hepatitis C and AIDS, long before they gained FDA approval. The genetic probes were far better than the earlier generation of tests, and caught many tainted blood units that the old tests missed. If labs had to wait for FDA approval to introduce the new tests, many more people would be infected today.

What a difference a similar mechanism for drugs would make. Most people think the FDA exists to protect them from drug dangers. But safety is resolved in the first of three required phases for drug approval, and it's relatively cheap to prove. Of the $800 million spent on drug development, $500 million is spent solely on clinical trials, and most of that is spent proving "efficacy" - that a drug will perform as claimed, using massive, placebo-controlled clinical trials.

Prior to 1962 companies didn't have to prove efficacy to market a new drug; just that it was safe. Demonstrating efficacy was left up to doctors and patients, who decided just how well a new drug worked. It was decided through peer reviewed studies published in medical journals. Year's ago, publication of a drug advertisement in the Journal of the American Medical Association was tantamount to approval. It signaled to doctors that the Journal's esteemed editors reviewed studies examining a new drug and agreed it worked as advertised.

We've already gone back to this sort of model for AIDS medications in the 1980s, where the FDA granted preliminary approval to drugs right out of phase II trials once safety was demonstrated, and then used post marketing studies to grant full approval. This is how diagnostic tests get regulated, and it's one reason why the tests make it onto the market in a fraction of the time it takes to approve a new drug, with no apparent harm to patients.

Far from expanding this practice to drugs, however, the FDA might take it away from molecular diagnostics. Among the moves the FDA is considering is to establish a measure of regulation of home brew tests by instituting controls over the active ingredients (analyte-specific reagents) used by laboratories to perform genetic tests. This regulation will have the effect of granting the FDA control of every phase of development for a new genetic test - the same kinds of crippling controls they now exercise over drug development. With the same effects: To slow introduction of lifesaving technology.

The regulatory quagmire is coming to a head amid introduction of the first genetic tests with real commercial potential. Labs are reporting a big increase in demand for cystic fibrosis carrier screening, the result of recommendations made this summer by two influential professional organizations, the American College of Obstetricians and Gynecologists and the American College of Medical Genetics, that all new moms get screened.

As technology continues to improve, so will the diagnostic tests. In the case of neuroblastoma, the test was so sensitive that it wasn't finding true cancers but tumors that would regress on their own without treatment. A future test for neuroblastoma might now look for multiple markers capable of predicting a tumors propensity to metastasize.

Gene based tests are already transforming medicine from a surface science based on imprecise physical findings, to an information-driven practice that turns on genetic certainties. These first-generation diagnostic tests grew up in academic laboratories largely as research tools and quickly made their way to the clinic where they dramatically enhanced the practice of medicine, allowing us to diagnose difficult-to-find viruses such as HIV, Hepatitis, and Chlamydia. While this early wave of technology wasn't perfect, it's rapidly evolving into a mature industry, with newer tests more tightly focused on issues of accuracy and ease of use. The pace of development has been spectacular precisely because the regulatory authorities have kept their hands off the industry.

Not anymore. Bumps in the technology, like the case with neuroblastoma, give pundits and regulators who misread them impetus for greater oversight, costing us all the next generation of molecular breakthroughs.



TCS Daily Archives