FDA Backs Risk-Based Approach for Lab TestsFDA Backs Risk-Based Approach for Lab Tests

Laboratory-developed tests (LDTs) have received considerable public attention lately. In May, FDA objected to San Diego–based Pathway Genomics' attempt to market an over-the-counter genetic test collection kit at Walgreens drug stores because it is outside the boundaries of the regulatory safe harbor for such tests, which are regulated by CMS. The agency has been bothered by broader claims made by test marketers. CDRH Office of In Vitro Diagnostics director Alberto Gutierrez said in June: “Well, the claims [made by the companies] have changed constantly. The original claims from three years ago were very, very vague. For example, the claims they’re making now for the different drugs and how they’re metabolized—those weren’t being made previously.”

Another concern the agency has with LDTs is that the tests have evolved into much more complex devices than FDA anticipated and they are worried about labs introducing unreasonable risks to patient health through uncontrolled design and manufacturing, unsupported claims, as well as unreported malfunctions and device failures. LDTs have evolved in many ways, Mansfield said, and have become a lot more like commercially distributed IVDs in terms of business models used and their ability to test multiple analytes from a single specimen. She said the agency has concern that the status of being an LDT test is “self applied” and that there is no formal regulatory definition. As a result, many labs offer tests developed by others as LDTs, and thus are technically not covered by enforcement discretion. “So we see LDTs being more and more used as loopholes in many cases as a way to go to market quickly without independent premarket oversight,” she said.

Mansfield told the meeting that the agency has been seeing a lot of preliminary information in scientific literature that is being packaged for use as medically actionable information, “and we think in many cases this is premature and there is not enough data to support the claims. We know formalized control of design of LDTs is often lacking... And of course software is uncontrolled, and software design and validation principles are critical to having good software,” she added.

Resource management will be an issue if CDRH pursues a risk-based strategy for regulatory oversight because it will require additional activity from agency personnel, according to Mansfield. “So we have considered a re-visit of currently regulated tests to assess potential for down-classifications,” she said. “So there may be tests out there now that we are actively doing premarket review for and we may look at them and say this premarket review doesn't add much value and let's not do that and instead let's do some higher risk tests... We will use and build our resources according to the need that we see, and we are able to track that very easily by how many submissions are coming in the door, how long reviews are taking, and so on... We could also look at pilots for third-party accreditation of other bodies than FDA that might perform premarket review of lower risk devices and for performing inspections.”

Mansfield said it would be important to develop a listing of who offers what tests “because we don't know what the universe of LDTs is now, and they are not registered and listed with us. NIH has a voluntary genetic test registry and we may be able to coordinate with them and maybe we need to expand beyond genetic tests. At any rate, we will probably need to expand our registration and listing in order to encompass all tests out there.”

The classification of high-risk tests will be important, Mansfield said, and likely will comprise tests where a false result could lead to incorreect and harmful clinical management, an unnecessary invasive procedure, or failure to follow up a serious disease. Examples include companion diagnostics, tests for cancer diagnosis, tests that direct or very strongly influence management of serious disease, and tests for serious or fatal communicable diseases. The underlying factor for deciding which are high-risk tests would likely be the claims made for them.

Moderate-risk tests would tend to have the potential for non-serious injury or injury that is medically manageable, according to Mansfield. They may have relatively easy-to-detect false results and may be adjunctive tests – tests that are used as one part of the totality of information for patient clinical management. Examples include genetic tests where the phenotype is already known and now is being confirmed genetically, tests where there are multiple findings used to direct clinical management but where each finding has specific weight, and tests used to monitor already-detected and -diagnosed disease.

Former FDA senior advisor to the commissioner Mary Pendergast told the meeting that as a “bedrock principle” all tests that are similar in risk should be subject to the same level of FDA oversight regardless of who sells or conducts the tests. Pendergast, who represents testing companies but said her comments were her own, supported arguments made in an earlier Genentech petition calling for agency regulation of LDTs.

 “In my opinion, it is unlikely that physicians or consumers pay attention to who provides a test or what silo of FDA regulation the test falls into,” Pendergast told the meeting. “FDA should conduct research into physician and consumer understanding rather than making decisions based on what the agency thinks it knows, which may be wrong or based on paternalistic assumptions. FDA official have been reported as saying that DTC genetic tests will be regulated as high-risk because consumers will not understand the information they are receiving and they may do something that someone at FDA thinks is irrational. These fears are out of date and paternalistic. While FDA has been wringing its hands over what consumers may or may not do, scientists are studying the issue and these must be considered before any action is taken.”

FDA's actions to determine what a person may or may not know could violate the First Amendment, Pendergast warned. “Truthful and non-misleading information is good. Knowledge is good. Preliminary information is good as long as it is properly described.... The FDA said the same thing 34 years ago when FDA tried to regulate the information pharmacists could give to consumers. The Supreme Court said there is of course an alternative to this highly paternalistic approach. That alternative is to assume this information is not in and of itself harmful, that people will perceive their best interest if only they are well enough informed, and that the best means to that end is to open up the channels of communication. The question is will FDA leave the channels of communication open. The First Amendment requires that the FDA impose no greater burden on speech than is required to stop false and misleading speech.”

Genetic Alliance’s Sharon Terry said her group recommends a classification framework based on relative risk of the information provided by a diagnostic device while considering the context of the use of the test. “The standard should be flexible and dynamic... supported by evidence that has been deemed competent and reliable to make the intended claim, and that also the level of evidence is consistent with what experts in the relative field consider to be sufficient at the time the test is being developed,” she said. “The current system is ill-suited to enable an efficient approval or clearance of advanced diagnostic tests with meaningful claims that reflect how the tests will be used in patient management.”

Agendia CEO Bernhard Sixt said his company supports a tiered risk-based approach. The company currently markets what it describes as the only FDA-cleared breast cancer recurrence test, MammaPrint. “As the nation's leading authority for patient safety,” he said grandiosely, “Agendia believes that only the FDA can regulate the LDT category in a uniform and unbiased manner to ensure that all claims are validated, data is accurate and post-market surveillance is in place... Moreover, Agendia believes that holding the LDT category to the rigors of FDA scrutiny will inspire greater confidence among investors to enter this emerging market, foster a level playing field among molecular diagnostics companies and encourage the best technology to come to the fore in personalized medicine.”

As it stands now, the proposed regulatory approach is not written in stone and no final decisions are being made until all stakeholders provide their input into the discussion, Mansfield said. “We are aware that there is a lot of anxiety over duplication or conflict with CLIA and we intend to avoid duplication, and if we can detect a conflict we will work it out. We are considering using CLIA or other deemed inspectors for our inspection processes... and certainly our goal is to avoid disruption of ongoing critical testing.” More stakeholder updates are planned, she said, and the agency is likely to issue guidance documents and conduct workshops as it moves forward with its regulatory strategy.