Fulfilling a promise it made to industry this spring, the US Food and Drug Administration this week released for public comment its much-anticipated draft guidance on the submission of certain pharmacogenomics data.
The draft, which the FDA’s Center for Drug Evaluation and Research originally said would be ready for public comment over the summer, is designed to generally instruct industry about when and how pharmacogenomics data should be submitted with certain new drug applications, investigational new drug applications, and biological license applications.
The draft also attempts to define — again generally — the criteria for how and when to submit pharmacogenomics data when drug labels will suggest genetic testing, and how to handle data supporting so-called companion diagnostic products used with therapeutics, much like what Roche and Genentech have created with the strand-alone diagnostic component of their breast cancer drug Herceptin.
Broadly, these tenets may encourage drug makers to ramp up their use of pharmacogenomics technology, such as gene-expression platforms and SNP-genotyping instruments. (The draft said proteomics data, and thus proteomics technologies, do not yet fall under this pharmacogenomics rubric.)
More narrowly, the draft, which can be seen here, has an important short-term objective that addresses a sore point among many drug makers: It specifies ways in which biopharmas may submit pharmacogenomics data voluntarily, and promises, to an extent, that the agency will not use them as critical deciding factors in the drug-approval process.
The goal of this component, described by the FDA as a way to stay up to speed with new pharmacogenomics technologies and applications, will be seen as a way to nurture trust within biopharmas and address the concern that these data could inadvertently be used to jeopardize candidate compounds. In the draft, the agency said it would create a new steering committee to collect these data and help the FDA write additional policy.
“I think the FDA has been very thoughtful, and it’s trying to be inclusive,” said Kevin Rakin, CEO of Genaissance Pharmaceuticals. “I think [the draft] will move the whole field [of pharmacogenomics] forward. It’s a real watershed event in bringing genetic variation into mainstream use.”
Yet the question that remains, at least until the FDA closes its public comment period for the draft on Feb. 2, 2004, is whether the document in the eyes of drug makers is comprehensive enough to begin ramping up genotyping and gene-expression studies.
The draft, released on Monday, comes on the heels of a similar document that seeks to set guidelines for submitting microarray data as components in the drug-approval process. That draft was to have appeared in August [see 6/13/03 SNPtech Reporter]. The new document also follows CDER’s promise to issue a draft guidance next summer for submitting microarray-based toxicogenomics data [see 8/21/03 SNPtech Reporter].
Specifically, the new draft describes a variety of general submission approaches. For example, the agency said that for cases in which drug makers use pharmacogenomic results “in decision making” in animal trials or during clinical development in human trials, “full information on the test system should be submitted to the IND.” In contrast, the agency wrote, “results from earlier feasibility studies done under the same IND to establish the potential usefulness of the pharmacogenomic test” — such as from samples taken during a dose-response study — “should not normally be submitted unless they provide support for the use of the test in clinical decision making.”
The agency wrote that if a pharmacogenomic test “shows promise” for enhancing the dose selection, safety, or effectiveness of a drug, a drug maker “may wish to fully integrate pharmacogenomic data into the drug development program. This could occur if the pharmacogenomic data “are intended to be included in the drug label in an informational manner” — for example, in cases where a CYP450 microarray is used to show a potential for dose adjustments. In these cases, the FDA wrote, the pharmacogenomic test result “may or may not be considered a valid biomarker, and an FDA-approved or widely used commercial pharmacogenomic test may not be available.”
It also suggests that a drug’s “dose selection, safety, or efficacy as described in its label” will be “contingent upon the performance of a pharmacogenomic test or tests.” Specifically, the FDA said that in later phases of drug development, patients “will be tested for drug metabolism genotype and dosed according to the test results.” They “will be selected for efficacy trial entry based on genotype … or gene expression profile” and will be excluded from trials based on these profiles.
“In all of these cases,” the agency wrote, “the FDA recommends co-development of the pharmacogenomic tests and the drug and submission of complete information on the test. ...” The agency said it plans to issue additional guidance about how to co-develop pharmacogenomic tests and drugs, à la Herceptin, “in the near future.” However, the agency recommends that if a new pharmacogenomic test will be used “in therapeutic decision making,” biopharmas should consider obtaining premarket review by CDRH, which oversees in vitro diagnostics, “in conjunction with” the drug-development program.
The draft “strikes a middle ground, but it says to industry where the information is reliable, and if it intends to rely on it [when] it is intended to submit it,” said Gregory Glover, a partner in the Washington, DC office of Boston-based Ropes & Gray LLP who deals with regulatory issues. “It’s a good first step. At the end of the day, the scientific reliability and reproducibility of [pharmacogeomics data] is going to make the FDA much more comfortable with it.”
The guidance “will force a strategic thinking about what … pharmacogenomics’ role should be” in pharmas, said Genaissance’s Rakin. “I think we’ll see more investment from that in this area.” He said he’s noticed over the last six months “a lot of increase in potential partnering discussions” at Genaissance. He said some Genaissance officials will attend a two-day meeting in Washington, DC, beginning on Nov. 13 at which the FDA and industry will discuss the agency’s pharmacogenomics plans. The meeting is sponsored by the Drug Information Association.
The Biotechnology Industry Organization, which counts as members more than 1,200 drug and diagnostics companies, was more reserved. “I think [pharmacogenomics data] are potentially valuable,” said Gillian Woollett, BIO’s vice president for science and regulatory affairs. “The question is, ‘How do you turn it into practical value?” Woollett said the draft is “a healthy sign that the agency is trying to anticipate” the development of a technology. “This is just a small step on an iterative process. I think [the draft] is tremendous food for thought.” Woollett said she will also attend the Nov. 13 conference.
Are My Data Safe?
As many drug makers will say, one major stumbling block preventing them from sending the FDA certain pharmacogenomics data not directly linked to the review of INDs, NDAs, or BLAs is trust — specifically, the touchy issue of whether the FDA will use voluntarily submitted pharmacogenomics data later on in the drug-approval process.
“It’s been a catch-22 within pharma,” said Rakin. “They’ve been saying, ‘Well, do we get penalized if we generate some of this data? What’s the framework from the FDA?’” And the FDA “has been saying, ‘We need more and more data to keep issuing guidance.’”
Or as Robert Rubin, a professor at MIT’s Program on the Pharmaceutical Industry, told SNPtech Reporter last spring: “Everybody is suspicious of everybody else, and the whole system has been an adversarial one for too long. The industry has some issues.”
To an extent, Janet Woodcock, who oversees CDER — and whose team, along with the FDA’s Center for Biologics Evaluation and Research, composed the new draft guidance — has conceded that industry “is afraid we will over-interpret these data in a way that will be harmful for drug development or it will slow down their drug development.” She also said in the past that “the relationship [between the] FDA [and] the pharmaceutical industry was somewhat adversarial.”
The draft itself acknowledges this, saying that drug makers “have been reluctant to embark on programs of pharmacogenomics testing” during the drug-development process “because of uncertainties in how the data will be used by the FDA” in its review process. “We need to get this issue settled,” Woodcock said at a May pharmacogenomics conference at Yale [see 5/2/03 SNPtech Reporter].
Today, the FDA considers “most pharmacogenomic data” to be exploratory, and does not require them to be submitted as part of an IND, NDA, or BLA. “However,” the draft states, “to be prepared to appropriately evaluate the anticipated future submissions, FDA scientists need to develop an understanding of relevant scientific issues.”
The agency said that some of these issues include “types of genetic loci or gene-expression profiles being explored by the pharmaceutical industry for pharmacogenomic testing; the test systems and techniques being employed; the problems encountered in applying pharmacogenomic tests to drug development; and the ability to transmit, store, and process large amounts of complex pharmacogenomic data streams with retention of fidelity.”
“Therefore,” the draft explained, “the FDA is requesting that sponsors conducting such programs consider providing pharmacogenomic data … voluntarily, when such data are not otherwise required under IND and NDA or BLA regulations.”
The agency has even created a category for this kind of submission — the Voluntary Genomic Data Submission — and stipulated in the draft that these data “are not required to be submitted.” The FDA said it also will create an agency-wide Interdisciplinary Pharmacogenomic Review Group, or IPRG, that will “review VGDSs, to work on ongoing policy development, and to advise review divisions dealing with pharmacogenomic data.”
In a statement released hours after the FDA issued its draft, BIO called this portion of the draft guidance a “mechanism for comfortably sharing exploratory pharmacogenomic information with the FDA — without stalling ongoing drug development.” The biotechnology industry “supports the concept of such voluntary genomic data submissions,” the group wrote.
“The idea of data not being submissible in a drug-review dossier until it’s more validated is key,” agreed Rakin.
Still, some questions may linger. “The remaining doubt will be, ‘Well, if you give something to the agency, can they really ignore it?’” said Ropes & Gray’s Glover. “’Are they really not going to use it in a way that is going to be harmful to us? Are they going to basically do what they’re going to do?’ I think that kind of comfort level is only going to come from time, where people get experience about submitting information that isn’t quite ready for prime time, and seeing what the FDA does with it,” he said.