Posted 27 September 2016
By Zachary Brennan
Diagnostic companies Illumina, Roche, AstraZeneca and others are offering a peek inside how they view next-generation sequencing (NGS) diagnostic devices, with the release of 20 comments on recent US Food and Drug Administration (FDA) draft guidance.
"Next generation sequencing has the capability to replace previous methods with a single approach to accomplish what might have required several different tests in the past," FDA explains. "In contrast to human sequencing diagnostics, infectious disease sequencing diagnostics generally require rapid and actionable results, sometimes within hours, as delayed or incorrect initial diagnoses can result in fatalities."
The draft guidance from May proposes to regulate the tests as complete systems that encompass all processes from specimen collection through obtaining clinical results. It also tells companies that in order to validate NGS-based tests for infectious diseases, they should use an "alternative comparator" sourced from public databases of microbial genetic sequences.
FDA says it has already developed one such database, called FDA-ARGOS (FDA dAtabase for Regulatory Grade microbial Sequences) to provide sponsors with "a set of validated regulatory-grade microbial genomic sequence entries."
The DNA sequencing powerhouse Illumina commended FDA for its ”one system” approach for the evaluation of infectious disease NGS diagnostic devices based on the use of testing data generated in real world use, ie. clinical laboratories.
However, Illumina takes issue with the process described in the clinical evaluation section of the draft as it appears borrowed from the referenced FDA document, “Highly Multiplexed Microbiological/Medical Countermeasure In Vitro Nucleic Acid Based Diagnostic Devices – Guidance for Industry and Food and Drug Administration Staff.”
The draft also calls for a process composed of two separate parts: The first is a prospective negative percent agreement study of 1500 patients enrolled with the signs and symptoms of the infections of interest and the second is a positive percent agreement study of well-defined archived samples known to be positive for the infections or markers of interest.
“While this seems a good approach toward addressing difficult to find positive samples, we would note that some targeted and agnostic sequencing panels may be directed at relatively common pathogens or disease markers. In such a case a more conventional cross sectional multi-site study of a population selected to reflect that of the test’s intended use would constitute a satisfactory approach to establishing test performance,” Illumina says.
The company also notes that the numbers described in the current document (1500 negatives with 100 of each comparative method for each detected organism or marker, 50 positives) “seem somewhat arbitrary. We note FDA itself appears to suggest an alternative approach to establishing numbers for studies of both negative and positive percent agreement based on statistical requirements to meet agreed on preset performance requirements. We think there is great merit in this approach.”
Illumina is also calling for more of an explanation for how the ARGOS database will be developed and used as currently, the complexity “of the database is limited, especially when it comes to phenotype-genotype correlations. It would be helpful to increase the number of isolates and to develop a learning algorithm over time. It would also be useful to specify if every new isolate needs to be deposited and receive a regulatory grade or if only reference genomes up to a certain homology need to be of regulatory grade. It is not entirely clear if the database is being considered a part of the device itself or being used simply as a comparative measure to help establish device performance. The scope of use should be described in a more clear and transparent manner.”
Meanwhile, Roche Diagnostics is requesting that FDA consider the applicability of its proposed approach in the draft guidance to apply to all IVDs.
“There are many IVDs, such as multiplex PCR-based assays or mass spectrometry-based tests, that share significant similarities with NGS-based tests and could benefiot from the same novel, innovative regulatory approaches,” Roche says. “For example, the use of curated databases for establishing clinical validity should also be applicable to multiplex PCR-based tests, and a representative testing approach could be applied to a mass-spectrometry-based test with a more general, agnostic intended use.”
And unlike Illumina, Roche says that FDA’s plan to regulate infectious disease NGS diagnostic devices as “systems” is “diametrically opposed to FDA’s expressed policies and precedent,” Eric Kolodziej, global head of quality and regulatory affairs, wrote. He explained that numerous manufacturers “specialize in a particular discipline associated with NGS so the evaluation of such IVDs as one system would mean FDA is “limiting the innovative tools that such specialized manufacturers could bring to the market to benefit public health.”
Likewise, AstraZeneca questioned FDA on whether there might be opportunities to consider “modular” approvals to facilitate what it calls a “mix and match” approach.
“For example could the specimen collection ‘module’ be approved and applied to other relevant processes without duplicating the review? Could this be applied to other ‘modules’ such as ‘sample preparation’, ‘Sequencing chemistry/data collection’ etc.? If so, how could the approval of such modules be shared so as to avoid unnecessary duplication of experimental work and FDA review activity?” AstraZeneca questioned.