Microarrays are wending their way into drug development — witness the recent speaker from Eli Lilly at Affymetrix’s investor day last week who revealed that the pharma giant is using Affy chips in mechanism-of-action studies for drugs. The US Food and Drug Administration, however, has not yet decided how to handle this data in the regulatory process.
To address this increasingly timely issue, the FDA invited representatives from pharma and a few microarray companies to a workshop earlier this month in Virginia.
The workshop, held May 16th and 17th, and co-chaired by Ron Salerno of Wyeth’s regulatory affairs department and Lawrence Lesko of the FDA’s Center for Drug Evaluation and Research, involved one day of presentations on microarrays and microarray data in the regulatory process, as well as a second day of “breakout sessions” in which attendees discussed the major areas of concern.
Reflecting the level of interest surrounding microarray data in the regulatory process, the FDA had to turn away about two thirds of the people who wanted to attend. “The response to this workshop was really overwhelming,” said Frank Sistare of the CDER, the co-chair of the FDA’s genomic/proteomic intercenter working group.
Those who did make the cut spent the time energetically discussing a panoply of issues, said Sistare. Among the issues were the potential to use microarray technology as a diagnostic and whether microarrays would produce data of sufficient quality to be viable as diagnostic devices; the value of gene expression data in the regulatory process; the level of association between SNPs and a clinical outcome necessary to make SNP-based screening arrays relevant; the questions about when and how microarray-based toxicology data should be submitted during the regulatory process for a new drug; as well as what role microarray data from clinical trials will play in the approval process.
On the subject of reliability of gene expression data, there was debate as to whether the data is of sufficient quality to have any role in the regulatory process. As array expression measurements are still not quantitative and not always reproducible, some do not feel it is ready for prime time as far as the FDA is concerned, while others think it is valuable nonetheless. “I can’t say there was any consensus on the issue,” Sistare said.
In fact, participants were not able to agree on many issues — a predictable state of affairs given the immaturity of the technology and the lack of precedent for clinical and regulatory use.
One theme that did emerge, however, was “that the situation is not going to be ‘one-size-fits-all,’” said Sistare. “The factors that enter into the data are going to be teased out on a petition-by-petition basis, depending on what is the product and what is the assay we need to do.”
For example, he said, the FDA would treat a test that predicts with 100 percent accuracy whether someone is going to develop a life-threatening reaction to a drug differently than one used to provide supportive data on a product that is allegedly going to reduce the severity of the common cold.
As the next step following the conference, the chairs of the breakout sessions are planning to consolidate the information gathered and publish it soon. This report, said Sistare, will not serve as any sort of formal guidance, but rather as “here are some burning questions and here are some of the viewpoints that have been espoused.” There will also be “action items” for the FDA to pursue, as it begins the process of developing guidelines for the submission of microarray data.
The agency plans to publish initial guidelines, which are subject to revision following a specified public comment period by summer 2003, said Sistare.
Meanwhile, amid this uncertainty as to the proper role for microarray data in the drug approval process, the FDA is encouraging applicants to submit microarray data along with their Investigational New Drug applications.
There is no explicit requirement that applicants submit this data at present. Companies are, however, submitting developmental plans that include the use of microarrays in clinical trials. Sistare encouraged companies who think they might want regulatory approval to “start the dialogue very early” both with FDA’s CDER and its other division, the Center for Devices.
For gene expression profiling to distinguish histologically similar tumors, as the Whitehead Institute’s Todd Golub and others have recently demonstrated, Sistare recommended confirming the initial retrospective-prospective hybrid studies with completely prospective studies.
Although the format for submission has not yet been ironed out, and there are questions as to when the data are relevant, the best course is to try to follow common sense and maximize the quality of the data, as well as to include as many controls as possible per experimental sample.
“I don’t know what else to say other than to encourage [applicants] to apply this wonderful technology so that we can learn together on how the data will ultimately be able to provide much greater insight into the mechanisms by which these agents have their biological effects,” Sistare said.