The Food and Drug Administration has released a 25-page Draft Guidance document addressing FDA oversight of certain applications of Next Generation Sequencing (NGS)-based technology. The Draft Guidance − "Use of Standards in FDA Regulatory Oversight of Next Generation Sequencing (NGS)-based In Vitro Diagnostics (IVDs) Used For Diagnosing Germline Diseases" − aims to “provide recommendations for designing, developing, and validating NGS-based tests for germline diseases that FDA believes are appropriate for use in providing a reasonable assurance of the analytical validity of such tests.”
Although FDA, as well as the National Institutes of Health (NIH), appears to view this guidance as necessary to advance President Barack Obama’s Precision Medicine Initiative, in trying to achieve this end, the Draft Guidance infuses additional regulatory uncertainty about the pathway for NGS-based technologies that could have the opposite effect.
Next-Generation Sequencing refers to the rapid analysis of large segments of DNA in order to determine the order of the nucleic acid base pairs (i.e., the sequence) that make up our genome. Some variations (also called mutations) in the order of base pairs can cause genetic diseases − such as cystic fibrosis − or increased risk of diseases − such as BRCA1- and BRCA2-associated breast cancer. Other variations can affect an individual’s response to certain types of drug therapy. NGS is performed using a sequencing “platform,” or instrument, that “reads” and provides an output of the sequence in a targeted region of the genome. NGS-based testing is widely seen as having the potential to revolutionize healthcare by enabling real-time, affordable access to an individual’s genetic information, which can illuminate both current health and future disease risk.
The standards proposed in the Draft Guidance address areas including test design considerations, test performance characteristics, performance evaluation, variant annotation and filtering, and test reports. The scientific merit and implementation feasibility of these highly technical proposed standards are best evaluated by experts including clinical geneticists and laboratorians, and therefore are beyond the scope of this Alert. But, as explained below, the Draft Guidance leaves several threshold questions unanswered:
- Which tests will be subject to its recommendations?
- To what entities will the recommended standards apply (e.g., instrument developers, clinical laboratories, algorithm developers)?
- Is the scope of the guidance in fact limited to analytical validity considerations, as FDA asserts, or in fact does it encompass the clinical validity of using the test results in a patient treatment context?
- Will FDA require premarket review for NGS-based tests, and, if so, what type of review and approval?
Which tests will be subject to FDA’s recommendations?
The purported scope of the draft guidance is “NGS-based tests with a broad intended use for suspected germline diseases.” The guidance in fact excludes from its scope the vast majority of current uses of NGS-based testing, specifically, stand-alone diagnostic testing, screening, microbial genome testing, risk prediction, cell-free DNA testing, fetal testing, pre-implantation embryo testing, tumor genome sequencing, RNA sequencing, or use as companion diagnostics. The Draft Guidance states that these applications are excluded because such tests “may have other analytical characteristics not addressed by the recommendations presented here.”
Consistent with the language of the Draft Guidance, FDA personnel stated in a July 7 stakeholder call that the initial focus would be on the use of NGS to identify a genetic cause for potentially heritable conditions, such as developmental disability, but that the agency’s scope could broaden in the future. Neither the FDA personnel nor the Draft Guidance provides a rationale for carving out this application or explains why it would present substantially different analytical validation considerations from other uses of NGS.
To whom does the Draft Guidance apply?
Also left unclear is which entities would be subject to FDA’s recommendations. FDA describes NGS-based tests as medical devices subject to the agency’s statutory device authority. To be sure, a few sequencing platforms − e.g., those with a clinical intended use such as diagnosis of cystic fibrosis − have received 510(k) clearance. However, other sequencing platforms do not include a clinical intended use, which arguably is a prerequisite for FDA application of its device authorities.
Furthermore, in practice, clinical NGS-based tests are performed in clinical laboratories, which sequence patient DNA using NGS platforms, analyze the variants using filtering algorithms, and issue a test report to the ordering health care provider to inform his or her patient’s care. Such NGS-based tests are therefore “laboratory developed tests” or LDTs, which currently are not regulated by FDA. FDA issued draft guidance outlining a regulatory framework for LDTs in 2014 (see those documents here and here) and, although the agency’s legal authority to regulate LDTs remains unclear, FDA plans to release final LDT guidance in 2016. Oddly, during the recent stakeholder call, FDA personnel asserted that the NGS Draft Guidance is unrelated to the proposed LDT framework, even though most NGS-based tests meet the agency’s proposed definition of LDTs and therefore foreseeably would be included within the scope of the final LDT guidance.
Thus, the implications for clinical laboratories, as well as for third-party bioinformatics tools developers − as opposed to conventional medical device companies − are as yet unclear.
Do the standards extend to the clinical validation and interpretation?
Furthermore, although the standards proposed in the NGS Draft Guidance ostensibly are limited to ensuring the analytical validity of NGS platforms (i.e., whether the platform reliably can identify the correct order of the base pairs), the scope of the guidance would appear to expand FDA authority beyond NGS instruments themselves and to also encompass core clinical laboratory activities. For example, FDA recommends the inclusion of certain information on the test report, citing the requirements of 21 C.F.R. § 809.10. However, this regulation historically has applied only to the labeling of traditional in vitro diagnostic (IVD) devices − such as reagents and instruments − and not to LDTs or test reports issued by clinical laboratories to healthcare providers.
Similarly, the Draft Guidance provides recommendations on the selection and documentation of “filtering algorithms,” which are used “to identify and prioritize candidate causal variants or genes from exome or genome sequencing.” Such algorithms are not necessarily intrinsic to the sequencing platforms, but rather are contained in software independently developed by clinical laboratories or third-party bioinformatics experts, and FDA’s jurisdiction to regulate such filtering algorithms is therefore unclear. Furthermore, the identification and prioritization of causal variants – i.e., those variants that are determined to explain a patient’s clinical symptoms − requires the addition of clinical judgment and therefore relates to clinical, rather than analytical validity.
Will FDA require premarket review?
The Draft Guidance also leaves uncertain whether NGS-based germline tests will be subject to premarket review. At the outset, FDA asserts that, because such tests are a “new type” of device, they would presumptively be class III and therefore would require a full application for premarket approval (PMA) prior to clinical use. Why FDA considers this subset of NGS tests to be fundamentally different, i.e., a “new” intended use, from other NGS-based tests, such as those currently used to diagnose cystic fibrosis or predict the risk of certain cancers, is not explained. Furthermore, the Draft Guidance contemplates that NGS-based germline tests may be suitable candidates for de novo classification and marketing based on a premarket notification (a 510(k)), depending on the specific test, and allows that special controls “could be developed” for these types of tests, which would enable exemption from 510(k) requirements.
So, in short, the regulatory pathway for NGS-based tests for diagnosing germline diseases remains unclear and may run the gamut from PMA to de novo or traditional 510(k) to exemption from premarket review.
Finally, the Draft Guidance appears to make eligibility for exemption from premarket review of analytical validity contingent on a demonstration of clinical validity:
“If FDA were to classify NGS-based tests for germline diseases in class II (i.e. in response to a de novo request), FDA would consider exempting such class II NGS-based tests from premarket notification requirements. In determining whether a 510(k) would be necessary to provide a reasonable assurance of the safety and effectiveness of the test, FDA would rely upon [considerations including] assurance of the clinical validity of the test.”
In other words, a demonstration of clinical validity could be necessary for exemption from prior clearance or approval of analytical validity, an outcome that seems both counterintuitive and contrary to the guidance document’s claimed focus on analytical validity alone. While the FDA, in a companion Draft Guidance, proposes a “recognition process” for public genetic variant databases as a means to establish clinical validity, this proposal also is in its preliminary stages, leaving uncertain whether such databases will in practice be an efficient means to demonstrate clinical validity to the agency’s satisfaction.
Given the many questions left open by the NGS Draft Guidance, all stakeholders – including clinical laboratories, NGS-platform developers, and bioinformatics tools developers – should carefully review its potential implications for their business models. In particular, they should evaluate whether the FDA’s recommendations for test design, development, and validation are scientifically sound and feasible to implement in practice, and the degree to which they may constrain the ability for clinical laboratories and third-party interpretive services providers to communicate relevant health information to healthcare providers. While guidance documents are not legally binding, they reflect FDA’s current interpretation of the agency’s statutory authority.
Comments to the NGS Draft Guidance will be accepted through October 6, 2016, and stakeholders should consider submitting comments requesting that FDA clarify key questions raised by the document and, more broadly, provide a clear statement of the agency’s legal authority to regulate NGS-based testing more generally.