By Tom Ulrich
We all remember how the genetic testing firm 23andMe roused the FDA’s ire in 2013, earning itself a warning letter to stop marketing its direct-to-consumer Personal Genome Service. The kerfuffle, though partially resolved, remains at the center of an ongoing debate in diagnostic and regulatory circles over laboratory-developed tests (LDTs) offered directly to the public, and the agency’s role in regulating those tests.
But like nature, business abhors a vacuum, and longs to fill it. Many companies and institutions have already jumped into the LDT ring, offering up genomic or pharmacologic services that they say would help guide patients’ and doctors’ treatment decisions and improve outcomes. Especially for patients with cancer.
How solid is the science behind these claims? And do vendors do a good job disclosing the strengths and weaknesses of personalized medicine? Those questions form the core of a study published this week in the Journal of the National Cancer Institute.
Dana-Farber Cancer Institute’s Stacy Gray, MD, and Dana-Farber/Boston Children’s Cancer and Blood Disorders Center’s Katherine Janeway, MD, started the study after realizing they were both interested in understanding personalized cancer medicine products being offered on the web. That interest grew after reports a few years ago of misleading websites offering nutrigenetic testing (tests that purport to use people’s genetic data to provide tailored nutritional advice).
“We both have patients come into the clinic asking about personalized medicine, and a lot of times people will bring in things they get off the Internet,” says Gray, a lung cancer specialist and population scientist. “The question is, as more and more companies, academic cancer centers and physicians engage in cancer genetic testing and advertise personalized medicine, what’s out there for patients to see?”
Gray and Janeway selected 55 websites marketing personalized medicine services for cancer, such as:
- somatic cancer gene testing (for single genes or gene panels)
- germline tests for cancer susceptibility
- genetic test interpretation
- chemotherapy susceptibility testing
A focus on the upside of personalized cancer medicine
The overall view of personalized medicine painted by the sites tilted heavily towards the positive. Forty-seven spent a lot of virtual ink touting the benefits of personalized cancer medicine, such as tailored therapy, improved outcomes or matching to clinical trials. Only 15, however, offered a nod to personalized medicine’s limitations, such as challenges of data interpretation, lack of proven benefits of personalized therapies or risks of false-positive and false-negative results.
The results complement those of other studies that have looked at the accuracy of claims made by personalized medicine providers. “A lot of times claims are rather hyped,” Gray says.
Do they do any good?
But do the services offered actually provide any benefit to doctors or patients? It’s an important question, but isn’t one labs currently have to answer.
“All labs that return results to patients have to be CLIA (Clinical Laboratory Improvement Amendments)-certified,” explains Janeway, a pediatric solid tumor oncologist who studies personalized medicine and tumor genomics. “To get CLIA certification, a lab needs to show that a test is reproducible and that it works. But it doesn’t have to prove that a test is of clinical benefit to a patient, or is appropriate for a particular situation.”
To probe that question, Gray and Janeway had a panel of oncology experts look at the tests offered by the surveyed sites. The panel then decided whether each test was clinically useful – whether it offered information that could be used to guide a patient’s care – in the context of cancer.
The panel’s decision? Only 28 percent of the sites offered tests that could actually guide care. Meanwhile, 88 percent of the sites offered tests that, with today’s knowledge, provide no value to the patient or doctor. And that was based on the sites that actually listed which gene tests they offered – more than half did not.
Mind you, we’re not talking small change here. While only about half of the sites surveyed disclosed their price list, the costs Gray and Janeway could find ranged between $99 and $13,000 per test.
Informing the regulatory gap
The whole debate around LDTs represents FDA’s efforts to catch up to a market that’s grown in ways regulators haven’t anticipated.
“We’re dealing with a set of policies that were introduced before genomics had come to the level of use we see today, and when there wasn’t much happening in the genomics landscape,” Gray explains.
It’s also, Janeway notes, a matter of the FDA clarifying what falls under its authority and what doesn’t, and using that authority. “The FDA has always had the ability to regulate LDTs, but they haven’t because up to now there haven’t been that many of them,” she says. “So CLIA certification was felt to be adequate.”
Now that companies offering LDTs have proliferated, the FDA is flexing its muscles, recently soliciting public comments on a risk tier-based policy for triaging tests for different levels of review and approval. That policy includes language about clinical utility.
“The FDA’s decisions could have profound implications for commercial and academic laboratories,” Gray says. “We hope these results will help inform the policy debate about genomic and genetic testing.”