There is an FDA project underway that could help change the way molecular diagnostics are developed, evaluated, and regulated in the U.S. one day, should things go the way the agency intends. And Mountain View, CA-based DNAnexus is getting the first crack at building the platform at the plan’s heart.
DNAnexus, the genomics analytics startup backed by Google Ventures and others, is announcing today that it’s won a key contract from the FDA. The numbers in this deal are small—just shy of $850,000 for DNAnexus, according to usaspending.gov—but the potential long-term impact on healthcare in the U.S. could be significant. DNAnexus is getting the first chance to build what’ll be known as “precisionFDA”: an open source, cloud-based tool for researchers, the makers of next-generation sequencing diagnostics, and others to pool and share genomic information in a single online repository. This platform is meant, eventually, to be a key tool to help the FDA start evaluating DNA sequencing-based diagnostics—tests that identify key genetic variants and patterns implicated in specific diseases—in a standardized fashion.
“We’re not saying precisionFDA is going to be the regulatory system necessarily,” says Taha Kass-Hout, the FDA’s chief health informatics officer. “But it’s definitely going to be a tool that will help streamline innovation and the process of validating [tests].”
Kass-Hout stresses precisionFDA isn’t a bid by the big bad FDA to come in and tell the developers of what are known as “laboratory-developed tests” what to do to gain FDA backing. (Such tests are defined by the FDA as designed and performed by a single lab, and are often sold widely by companies, among them Myriad Genetics, Foundation Medicine, and Genomic Health.) Rather, it’s supposed to put the power in their hands, and foster a community in which everyone can pitch in and help develop the standards to determine which tests are credible and which aren’t, Kass-Hout says. The FDA can then look to these standards, and how a test compares with them, in the future when considering approval of a test.
By standardizing the process, the idea is that there’ll be more faith in these tests, and it’ll be easier for the FDA to make quick regulatory decisions on them. That would be a stark change from today, as the providers of LDTs aren’t currently subject to rigorous regulatory oversight the way the makers of new drugs and medical devices are. Most LDT providers currently provide the testing under a general Medicare certification for clinical labs, but without an FDA endorsement for the validity of specific tests. It’s a big problem for several reasons: A number of important (and expensive) healthcare decisions can be made off of an LDT, like whether a cancer patient needs chemotherapy, or should get preventative surgery like a mastectomy. And companies that actually choose to run clinical tests, accumulate data, and seek FDA backing, sometimes wind up competing with other testmakers that don’t do the same work.
For years, the FDA has talked about changing things to provide more aggressive regulatory oversight of LDTs, but efforts to do so have been fiercely opposed by members of Congress and others fearing more oversight would make the path to a new test too long and expensive to traverse. In July 2014, however, the agency announced plans to move ahead and more aggressively regulate the tests—particularly “higher-risk” LDTs that impact big healthcare decisions for a patient.
The ultimate hope, Kass-Hout says, is that precisionFDA becomes a key component of this increased regulatory effort. DNAnexus chief cloud officer Omar Serang (pictured above) explains a few important aims of the platform his company is trying to build. First, that sharing and community consensus may lead to more and better “reference” genomes—that is, highly sequenced and analyzed genomes that are used as a guide against which a patient’s genetic information is compared. Molecular diagnostics look for certain differences between a reference genome and a patient’s genome—like a specific genetic mutation known to drive cancer. The platform could also provide a way to compare and contrast the results obtained by different labs or institutions. The idea here is to show that if a certain test is run “time and time again the results are consistent,” Serang says.
Kass-Hout likens precisionFDA to
