On January 13, 2017, the US Food and Drug Administration (FDA) posted a “discussion paper” in which the agency outlines a substantially revised “possible approach” to the oversight of laboratory-developed tests (LDTs). The discussion paper explicitly states that it is not a final version of the July 2014 draft guidance, and that it does not represent the agency’s “formal position.” Rather, the document represents the latest iteration of the agency’s thinking on LDTs, which the agency posted to “spur further dialogue.
Historically, the US Food and Drug Administration (FDA) has exercised enforcement discretion with respect to most laboratory-developed tests (LDTs), and not required the laboratories that furnish LDTs to comply with the agency’s requirements for medical devices (e.g., registration and listing, premarket review, medical device reporting). In July 2014, the FDA published draft guidance in which it outlined its intent to end enforcement discretion and impose a risk-based regulatory framework for most LDTs. The agency received substantial stakeholder feedback on the draft guidance and ultimately decided not to publish a final version of the guidance prior to the end of the Obama administration.
On January 13, 2017, however, the FDA posted a “discussion paper” in which the agency outlines a substantially revised “possible approach” to the oversight of LDTs. The discussion paper explicitly states that it is not a final version of the July 2014 draft guidance, and that it does not represent the agency’s “formal position.” Rather, the document represents the latest iteration of the agency’s thinking on LDTs, which the agency posted to “spur further dialogue.”
Key principles articulated in the discussion paper include the following
Scope of oversight
Currently marketed LDTs would be “grandfathered”—i.e., not expected to comply with most or all FDA regulatory requirements (e.g., premarket review, Quality System Regulation (QSR), registration and listing)—“unless necessary to protect the public health.” The agency would focus its oversight on “new and significantly modified” high and moderate risk LDTs.
Several categories of new and significantly modified LDTs would also be exempt from agency’s premarket review, QSR, and registration and listing requirements, unless necessary to protect the public health. Under the discussion paper, (1) low risk LDTs; (2) LDTs for rare diseases; (3) traditional LDTs (i.e., LDTs that use components that are legally marketed for clinical use, whose output is the result of manual interpretation by a qualified laboratory professional, and do not use automated instrumentation or software for intermediate or final interpretation); (4) LDTs intended solely for public health surveillance; LDTs used in CLIA-certified, high-complexity histocompatibility labs; and (5) LDTs intended solely for forensic use would not be expected to comply with such requirements. LDTs for “unmet needs” would, however, be subject to premarket submission within 90 days after offering the LDT and certain additional regulatory requirements.
Notwithstanding the above, the FDA expressly retains its ability to enforce its regulatory requirements to the extent required to protect patients from harmful tests. FDA would take enforcement action against LDT developers if an LDT is not analytically or clinically valid; if there is an absence of data to support the LDT’s analytical or clinical validity; if the LDT’s developer is engaged in deceptive promotion; or if there is a reasonable probability that the LDT will cause death or serious adverse health consequences.
Timeline for Implementation
Premarket review of new and significantly modified LDTs would be phased in over four years. The FDA would focus first on the tests for which the consequences of a false result are known to have the highest risk to the patient, with an additional two years to meet QSR requirements. The phase-in would occur as follows:
Year one—serious adverse event and malfunction reporting for LDTs (except traditional LDTs, LDTs solely for public health surveillance, histocompatibility LDTs and LDTs solely for forensic use)
Year two—premarket review for new/modified LDTs with the same intended use as an in vitro diagnostic device (IVD) approved under a premarket approval (PMA) application
Year three—premarket review for new/modified LDTs with the same intended use as a Class II IVD subject to the premarket notification (510(k)) requirement
Year four—premarket review for new/modified LDTs that do not fall into the above categories
Tests introduced between the effective date of the framework and its phase-in date could continue to be offered for clinical use during the period of premarket review.
With respect to analytical validity, FDA anticipates that laboratories that conduct “appropriate” evaluations would not need to collect data beyond that required to establish performance characteristics of an LDT under the Clinical Laboratory Improvement Amendments of 1988 (CLIA). The agency also intends to consider how third-party proficiency testing programs, accepted reference standards and/or certification programs may be leveraged.
With respect to clinical validity, FDA anticipates that labs may be able to establish clinical validity using literature, well-curated databases and other appropriate sources.The agency also notes that once an LDT’s clinical validity has been established, laboratories with subsequent tests could generally, in accordance with applicable regulations, leverage prior evidence when factors such as indications for use, technology and standardization are the same.
These approaches are consistent with the FDA’s proposed approach for analytical and clinical validity for next generation sequencing (NGS) molecular diagnostics.
FDA would review prospective change protocols that outline specific types of anticipated changes, the procedures that will be followed to implement them and the criteria that will be met prior to implementation. Modifications could be made in accordance with the change protocol without the need for a new submission. Premarket review of modifications to an already marketed test would be limited to those modifications that significantly change performance specifications or intended use of the test and are not made in accordance with the test’s approved change protocols.
FDA would focus its QS assessment on three specific QS requirements—design controls, acceptance activities, and procedures for implementing corrective and preventive actions (CAPA). Otherwise, FDA would rely on a laboratory’s certification to CLIA QS requirements. Initial inspections would be educational in nature.
Adverse Event Reporting
Laboratories would be required to report serious adverse events for all tests except traditional LDTs, LDTs intended solely for public health surveillance, certain histocompatibility LDTs and LDTs intended solely for forensic use. In the future, it may be possible to decrease or eliminate certain reporting requirements by leveraging real-world data.
The policies set forth in the discussion paper represent a substantial shift from—and would generally result in less extensive oversight of LDTs as compared to—the draft guidance. However, substantial uncertainty remains with respect to several aspects of the discussion paper framework, including the following
How would FDA decide whether a test is an LDT? In the draft guidance, the FDA states that a test is only an LDT insofar as it is designed, manufactured and used within a single clinical laboratory. Certain stakeholders criticized this definition as overly narrow, and noted that many tests currently offered as LDTs would not fit this definition (e.g., because the test incorporates key components that were manufactured by a third party). The discussion paper does not specifically address whether the agency intends to retain this narrow definition, or if the agency may be willing to adopt a broader definition of the term to bring additional tests within the scope of continued enforcement discretion.
How would FDA decide whether agency oversight of a previously marketed (and therefore, potentially grandfathered) LDT is “necessary to protect the public health”? The agency’s willingness to consider any “grandfathering” of previously marketed LDTs represents a significant shift in the agency’s position. That being said, the agency does not provide any guidance on how it will decide whether agency oversight of a previously marketed LDT is “necessary to protect the public health.” For example, would the agency only take action against LDT developers if it has evidence of actual patient harm (e.g., from adverse event reports)? Or could the agency take the position that oversight of any test with certain high-risk intended use(s) is necessary to protect the public health, regardless of evidence of actual harm?
In practice, what would an LDT developer need to do to avail itself of the “unmet needs” pathway? Under the discussion paper, laboratories would have up to 90 days after first offering an LDT for an unmet need to send a premarket submission to the FDA or an accredited third-party reviewer, and could continue to offer the test during the review of such submission. It is unclear, however, whether the agency intends to allow retain this flexibility in all circumstances. For example, what if the developer submits a 510(k), but the agency believes that a PMA is required? Or what if the developer’s submission relies almost entirely on retrospective registry data, while the agency believes that prospective data are required?
How would the phase-in period apply to LDTs that do not have the same intended use as any approved or cleared IVD? As currently written, the framework would require LDTs that have the same intended use as an IVD approved under a PMA to submit a PMA two years after the finalization of the proposal, and for LDTs that have the same intended use as an IVD cleared under a 510(k) to submit a 510(k) three years post-finalization. As such, LDTs that do not have the same intended use as an approved or cleared IVD do not appear to be subject to the premarket submission until four years post-finalization. The extent to which the agency intends to follow this timeline is unclear, however, particularly insofar as LDTs for unmet needs are often precisely the tests about which FDA has expressed significant regulatory concerns and these require premarket submission within 90 days after first offering.
What guidance would be available for laboratories required to comply with QSR requirements? Under the discussion paper, laboratories that develop LDTs would only be subject to a subset of FDA’s QSR requirements (i.e., design controls, acceptance activities, and corrective and preventive actions). That being said, clinical laboratories have not historically been subject to such requirements, and as such, may find it difficult to come into compliance with such requirements (particularly design controls). It is unclear, however, the extent to which the agency will provide laboratories with additional guidance to facilitate their compliance with such requirements.
How stringent will FDA’s requirements be with respect to prospective change protocols? Historically, the agency has taken the position that many changes to an FDA-approved or cleared kit would require a submission to the agency; as such, allowing laboratories to make modifications to LDTs without requiring a new submission if the change is made under a pre-approved protocol would represent a substantial change in the agency’s position. That being said, insofar as a laboratory would be required to comply with the requirements set forth in a pre-approved protocol, this revised position will only meaningfully enhance patient access to modified tests if the requirements of the protocol are reasonably achievable for test developers. The agency does not, however, describe or provide examples of what it would require in such a protocol.
How will FDA address the practical issues associated with regulating LDTs as medical devices? Unlike an IVD test kit, an LDT is not a physical item that is transferred in interstate commerce. Therefore, what is the regulated “device” within the laboratory test operation? To whom would LDT labeling be addressed—the performing laboratory or the ordering physician? And perhaps most importantly, how will FDA find the resources necessary to timely review the additional submissions that would come in under the framework?