FDA's Evolving Framework for Oversight of Next-Generation Sequencing

On September 9, the US Food and Drug Administration (FDA) announced plans to host two public workshops on "next-generation sequencing" (NGS) that may provide greater clarity regarding the regulatory framework for personalized medicine and the use of individuals' genetic information to predict disease and develop customized treatments.  FDA is holding the workshops on November 12-13 to obtain feedback and information from stakeholders on (1) approaches to developing performance and evidentiary standards for NGS and (2) the use of curated third-party databases of human genetic information to support product approvals and other Agency functions.  FDA will accept written comments on these issues until November 25.¹

I.  Clinical and Regulatory Importance of NGS

DNA sequencing is a process of determining the order in which the four critical building blocks of human DNA appear in a single DNA molecule.  The order of these four chemicals in a DNA molecule determines the instructions or the genetic code for a particular cell or group of cells within the body.  This genetic code forms the human genome, which contains all of the genetic information needed to create an human being.  The ability to identify, manipulate and characterize genes offers enormous potential to predict, detect, or alter genes to treat or prevent disease.  NGS is a general term used to describe new and emerging technologies, including software and algorithms that detect and sequence millions of fragments of DNA from a single sample at the same time.² This technology allows researchers and scientists to sequence a human genome very quickly. 

NGS represents a quantum leap in technology compared to the methods used in the Human Genome Project in the 1990's.  The cost of sequencing a human genome has also fallen dramatically -- from US$300 million in 2001 to about US$1,000 in 2014, according to the Personalized Medicine Coalition.³ The increasing availability of NGS for use in clinical practice has been of interest to regulators and both private and public payers.  The greatest challenge is that these products do not fit neatly into the

traditional regulatory framework for medical devices or drugs, which are defined and regulated based on clinical evidence of safety and effectiveness for a particular intended use or indication for use.  As discussed in detail below, NGS challenges the "intended use" standard for medical devices, as the tests can have an infinite number of intended uses, and can thus be used to identify patients who have various conditions or biomarkers, and may be candidates for any number of treatments.     

II.  The Evolving Regulatory Framework for NGS

The regulatory framework for NGS is complex and is evolving in real time. The FDA regulates NGS test systems as medical devices, and CMS oversees clinical laboratories' compliance with the Clinical Laboratory Improvement Amendments of 1988 (CLIA).  The FDA has proposed a new framework for laboratory developed tests (LDTs) that would end the agency's policy of enforcement discretion, and apply traditional medical device clearance and approval requirements under a risk-based framework.  FDA's Draft LDT guidance (previously described in Arnold & Porter's Advisory), if finalized, will affect laboratories that perform NGS and other genetic testing.  Many clinical laboratories that are conducting NGS testing are accredited by organizations such as the College of American Pathologists and the American College of Medical Genetics, which have established quality standards for NGS.  But those standards are voluntary, and existing CLIA regulations do not address many aspects of NGS testing. Below we describe the evolving regulatory framework for NGS.     

Since November 2013, FDA has reviewed and assessed the clinical validity of laboratory test systems for NGS using its medical device authorities.  But NGS poses several issues for FDA that have pointed to the need for a more flexible regulatory system that can be used to determine analytical and clinical performance.  Because NGS can identify an unlimited number of mutations, it is impossible to demonstrate clinical effectiveness and performance on every possible one.  And because NGS can identify very rare mutations that occur in a small number of patients in the population, it may not be feasible to run clinical trials that study the clinical impact of each individual mutation on patients' disease status. 

In November 2013, FDA announced the approval through the de novo pathway of the Illumina MiSeqDx^® instrument platform and a reagent kit.  The two devices make up the first FDA-regulated test system that allows laboratories to develop and validate sequencing of any part of a patient's genome.  At the same time, FDA approved two assays for cystic fibrosis that are performed on the Illumina MiSeqDx platform.  In this system, the patient's DNA sequence is compared to a reference sequence of the cystic fibrosis gene and to a curated database of information on the genetic variants of patients with cystic fibrosis.  FDA used the database to assess the validity of the 139 genetic variants involved in the assay, rather than requiring Illumina to generate data to support each variant individually.

The College of American Pathologists (CAP) has developed proficiency testing for NGS that includes standards for wet bench and bioinformatics processes for a set of test samples, including sample preparation, library generation, DNA amplification, sequencing, and data analysis, and compares the laboratory's results to dozens of reference variants.  CAP also has developed a Molecular Pathology Checklist that includes a section on NGS and is updated annually.⁴ Importantly, laboratories that are accredited by the CAP and other accreditors pay fees to access the proficiency testing and checklist standards.  The standards are voluntary, and some laboratories that are marketing NGS tests, for claims that they can drive choices of drug therapy, are not accredited.

III.  The Developing Regulatory Approach for NGS

In his State of the Union address in January 2015, President Obama announced the launch of a Precision Medicine Initiative (PMI), which includes a request for US$215 million investment in the President's 2016 Budget. The PMI "will pioneer a new model of patient-powered research that promises to accelerate biomedical discoveries and provide clinicians with new tools, knowledge, and therapies to select which treatments will work best for which patients."⁵  It will "leverage advances in genomics, emerging methods for managing large data sets, and health information technology to accelerate biomedical discoveries, all while protecting patient privacy."  NGS technology is a key component of the PMI framework, as NGS "is capable of detecting the billions of bases in the human genome, and in so doing identify the approximately 3 million genetic variants an individual may have."⁶

In recent testimony to the US Senate, Dr. Jeffrey Shuren, Director of the FDA Center for Devices and Radiological Health, stated that "a critical mass of genomic data…has been accumulated by researchers and clinicians" and that "new regulatory approaches will be needed to enable the Agency to provide appropriate oversight, in a way that is more suitable to the complexity and data-richness of this new technology."  FDA has stated that NGS is "broad and indication-blind testing" which potentially conflicts with the FDA's mandate to review devices based on their intended use.  Other technologies capable of detecting genetic variation are generally designed to capture predefined data points that are known in advance of testing, and therefore are more suited to regulation under traditional approaches, according to FDA.

In December 2014, FDA released a white paper, Optimizing FDA's regulatory oversight of next generation sequencing diagnostic tests - preliminary discussion paper, and in February 2015 held a public meeting to obtain input from stakeholders on this new approach. The white paper describes the scope of FDA's regulatory approach as:

FDA would assess analytical performance using quality-based standards for NGS test performance that would be created in collaboration with experts in genomics.  The FDA believes that such a standard would assure that an NGS test meets an acceptable level of performance, and that the results generated are reliable and accurate.  FDA asked stakeholders to comment on: (1) value of a standards-based approach to regulatory review of NGS tests (2) content of standards to be developed that will assure that conformity to the standard will assure test accuracy and reliability, (3) who should develop such standards, and (4) appropriate mechanisms to ensure compliance.

FDA proposed to assess clinical performance by using high-quality curated genetic databases that provide information on genetic variants and their association with disease.  NIH has created the ClinVar database, which pulls together information about the association of specific mutations with diseases from many researchers and clinical laboratories.  NIH also runs a program called ClinGen that compares the findings from different laboratories and published peer-reviewed studies in a systematic way.  That information is synthesized so that the ClinVar database includes a complete picture of the known medical information on each variant.  FDA requested comments on how to leverage genetic databases to evaluate clinical performance, and whether and how to communicate information about less-well-understood variants. 

On February 19, 2015, the FDA announced that it was approving 23andMe's direct-to-consumer genetic test for Bloom Syndrome Carrier Status.  The FDA approval also classified carrier screening tests as Class II and stated the agency's intent to exempt future submissions for carrier screening tests from FDA premarket review. FDA also announced plans to issue a notice announcing its intent to exempt the tests and receive comments from stakeholders for 30 days.  FDA's press release states that this move "creates the least burdensome regulatory path for autosomal recessive carrier screening tests with similar uses to enter the market." 

As conditions of the approval, FDA also required:

• Product labeling, stating that "the test is intended only for postnatal carrier screening in adults of reproductive age, and the results should be used in conjunction with other available laboratory and clinical information for any medical purposes."
• Consumer information,similar to that FDA has required for other direct-to-consumer tests, on how to find a medical professional for genetic counseling. 

IV.  Stakeholder Views on NGS Regulation

On February 20, 2015, the day after the approval of 23andMe's test for Bloom Syndrome, FDA held a public meeting to receive public comments on the approach outlined in the white paper released in December 2014.  The meeting attracted nearly 1,000 participants. FDA is reviewing the comments from stakeholders in professional organizations and industry who attended the meeting. Key themes that emerged in the comments FDA received include:

Regulation of NGS is an extension of the ongoing debate over FDA oversight of LDTs.  Comments from stakeholders who oppose FDA regulation of LDTs altogether or who desire a minimal role for the FDA in reviewing high-risk LDTs repeated several familiar arguments.  Those include 1) statements that the FDA lacks authority to regulate LDTs;⁷ 2) recommendations to modernize CLIA regulations so that they more closely reflect the state of laboratory practice, particularly with regard to molecular testing;⁸ and 3) comments that the selection of variants of interest, interpretation of test results, and off-label use of a test system are the practice of medicine.⁹ FDA's role in evaluating performance characteristics of instruments, test kits, software, and reagents, and ensuring that test performance is consistent with claims made in marketing, were generally supported.¹⁰ 

Professional organizations will likely have a role in setting standards, but that role is not fully defined.  Stakeholders expressed the view that in general, guidelines for analytic performance should be clear, predictable, and scientific. Stakeholders agreed with FDA's proposal to rely on analytical performance guidelines developed by professional organizations.  Several specific professional organizations were recommended, including the American College of Medical Genetics and Genomics, the Association for Molecular Pathology, the College of American Pathologists, and the Clinical Laboratory Standards Institute.  These organizations individually and collectively develop and maintain clinical guidelines and consensus statements, create reference materials for proficiency tests, administer deemed accreditation programs, and publish new medical research on the significance of clinical variants.  The Association for Molecular Pathology's comment letter on FDA's NGS white paper cites several peer-reviewed publications that describe standards for NGS published from 2012 to the present.¹¹

Limitations of relying on well-curated clinical variant databases.  Stakeholders generally supported the notion of FDA's developing processes and incentives for curation and data sharing, including standardizing formats and nomenclature in reports, as well as conveying the strength of evidence for reported variant-disease associations.  But several commenters stated that publicly available clinical variant databases are of variable quality, are in nascent stages, are sometimes being developed in silos, and thus may not always be effective tools for regulatory oversight. Stakeholders supported the continuing use of databases that are in widespread use by laboratory professionals today, as well as the potential to use datasets created in clinical trials.  Some comments reflected views on FDA's appropriate role vis-à-vis the practice of medicine and the protection of intellectual property.¹² 

Need for regulatory flexibility to promote innovation.  Commenters agreed that NGS oversight should be flexible to keep pace with clinical and scientific advancements.  Manufacturers should be permitted to have open platform instruments and a mechanism to easily and quickly update the test system. Stakeholders noted that the clinical application of a test can change quickly as new information emerges about a genetic variant.  Some stakeholders stated that information from Investigational Use Only NGS tests should be available to be used in research to identify patients for clinical trials.  Others supported the notion of incorporating real time observational data collected through learning healthcare systems. 

IV.  Future Outlook

When taken together, the approach to assessing analytic and clinical validity of NGS described in the white paper and foreshadowed by the approval of the 23andMe test for Bloom Syndrome signal that FDA's regulatory approach to NGS is flexible, and that FDA is using its current authorities to achieve policy objectives.  FDA used the de novo process to review the first NGS test system as well as the Bloom Syndrome test, and by classifying the 23andMe test as a Class II device signals a willingness to approve future carrier screening tests through the 510(K) process. 

FDA is meeting with experts to develop standards and resources that would form the basis for a new oversight proposal for NGS.  On August 5, an FDA Voice blog described a new initiative called precisionFDA, an "informatics community and supporting platform."¹³  FDA plans to launch precisionFDA in a beta version in December 2015.  It will be "a crowd-sourced, cloud-based platform" that will "supply an environment where the community can test, pilot, and validate new approaches…and share and cross-validate their tests or results against crowd-sourced reference material."  To ensure the security of genomic data, it will give researchers "access to secure and independent work areas where, at their discretion, their software code or data can either be kept private, or shared with the owner's choice of collaborators, FDA, or the public."  Publicly available information in precisionFDA will include "a wiki and a set of open source or open access reference genomic data models and analysis tools developed and vetted by standards bodies, such as the National Institute of Standards and Technology."     

FDA has stated publicly that the agency expects to release more specific regulatory proposals in white papers and potentially industry guidance.  The convening of the November 12 and 13 workshops will offer stakeholders an opportunity to comment publicly on the agency's approach.  Further, written comments will be accepted until November 25. The regulatory framework for NGS will continue to evolve as FDA and the medical device industry begin negotiating a user fee agreement to replace the one that expires in 2017.