FDA and the Regula/on of Next Genera/on Sequencing

Transcription

1 FDA and the Regula/on of Next Genera/on Sequencing David Litwack, Ph.D. Personalized Medicine Staff Office of In Vitro and Radiological Health, FDA

2 In Vitro Diagnos/cs in the Age of Precision Medicine Precision Medicine Low/medium resolu.on technology High resolu.on technology ( omics ) Detect a finite number of analytes (usually one) One test one disease Clinical evidence from randomized controlled trials research separate from prac.ce Undefined (millions?) One test many diseases Clinical evidence from learning health systems- merging of research and prac.ce

3 FDA s Risk Based Approach to Regula/on Premarket submission 3 classifica.on levels: Class I: common, low- risk devices 510k (usually exempt) Class II: more complex, moderate risk 510k Class III: most complex, high risk - PMA Clinical trial using inves@ga@onal device 3 categories: IDE exempt Nonsignificant risk Significant risk Risk is based on consequences of a false result in the context of clinical care Decision is based on risk- benefit analysis (safety and effec.veness) Risk is based on the consequences of a false result in the context of a clinical protocol Decision is based on risk (safety only)

4 Elements of FDA Premarket Review Analy.cal performance Correctly detects analyte Accuracy, precision, limits of detec.on/measurement Clinical performance Correctly iden.fies disease/condi.on Clinical sensi.vity, clinical specificity, predic.ve values Labeling

5 Special issues with NGS Don t usually predefine what the test will detect Even a single gene test could detect previously unobserved varia.on Unprecedented ability to detect rare variants OYen lack specific intended use Can simultaneously diagnose mul.ple condi.ons Incidental findings Physicians and pa.ents have trouble understanding results Rapidly evolving technology High throughput allows discovery to outpace understanding

6 Clearance of NGS PlaHorm and Assays Intended Use Performance Clinical Performance MiSeqDx PlaNorm Universal Kit 1.0 targeted sequencing of human genomic DNA. Clinical and cell line samples Well- standardized panel with known variants Performance demonstrated on a representa.ve set of variants use with the MiSeqDx instrument. See above NA NA Cys@c Fibrosis Clinical Sequencing Assay re- sequences the protein coding regions and intron/exon boundaries of the CFTR gene Valida.on of both specific variants and CFTR normal sequence Well- established associa.on of CFTR and CF; expert interpreta.on Cys@c Fibrosis 139 Variant Assay simultaneously detect 139 clinically relevant cys.c fibrosis disease- causing muta.ons and variants of the CFTR gene Specific valida.on of 139 variants Use of the CFTR2 database (JHU) for evidence

7 Analy/cal evalua/on strategies (plahorm) Accuracy - Select and validate adequate subset of gene.c markers à inference that planorm as a whole analy.cally valid Enrich with analy.cally challenging markers, classes of varia.ons Include clinically relevant markers Homopolymeric regions, indels, repeats, CNVs, redundant sequences, samples across the genome / chromosomes, etc.

8 CFTR2 Database ( Datatype Name/ Associated Nomenclature Associated Clinical Captured Provides a standardized muta.on name and muta.on by amino acid and nucleo.de number (rela.ve to the CFTR gene) Provides the following relevant clinical characteris.cs: Average sweat chloride value at.me of diagnosis Range of FEV1 percent predicted value based on age group Percentage of pa.ents with posi.ve Pseudomonas aeruginosa culture Percentage of pancrea.c insufficient individuals Func@onal Tes@ng Notes the results in vitro laboratory tests performed for applicable muta.ons. Specifically, assesses protein processing and matura.on, CFTR dependent chloride current, and gene splicing. Literature Review Notes research previously completed on this par.cular muta.on. Annota@on History Provides a history of changes and.mestamps of any revisions to the annota.on.

9 Illumina Special Controls Specify the informa.on that should be in the label Include: Read depth necessary for claimed sensi.vity Data demonstra.ng performance characteris.cs of instrument

10 Registered and Listed Sequencers High Throughput DNA Sequence Analyzer: A high throughput sequencing technology performing targeted DNA sequencing of amplicons from a defined gene.c region or a subset of genes in human genomic DNA from a clinical sample. Class 2 exempt no premarket submission needed Illumina MiSeqDx Life Technologies Ion PGM Dx Vela Sentosa SQ301

11 Myriad BRACAnalysis CDx Companion diagnos.c for treatment with olaparib an aid in iden.fying ovarian cancer pa.ents with deleterious or suspected deleterious germline BRCA variants eligible for treatment with Lynparza (olaparib). Test is a combina.on of PCR and Sanger sequencing Defined variant classifica.on process conducted by in- house commijee

12 Precision Medicine Ini/a/ve FDA will develop a new approach for evalua.ng Next Genera.on Sequencing tests A new approach will facilitate the genera.on of knowledge while ensuring that the tests are accurate and reliable. NIH research FDA patient regulation clinical NGS

13 Why is FDA talking about a new approach now? Previous FDA experience with review of NGS plakorm/tests Sequence and analy.cal performance data has been accumulated in mul.ple centers that makes new approaches possible Current efforts by the community to create technical and performance standards are progressing Efforts at improving database quality and cura.ng evidence have advances Rapid transla.on into clinic

14 Goals for FDA Regulatory Oversight of NGS Efficiency to allow innova.on while protec.ng pa.ents Compress the.me between discovery and clinical use Take into account the state of the technology and current prac.ces An.cipate future developments in technology and prac.ce

15 Desired Features of a Regulatory Framework Accommodate mul.ple test configura.ons on a single instrument. Varia.ons in library, informa.cs, etc. depending on use and what is being detected (e.g., CNVs, SNVs) Users can select the right configura.on if they have proper controls in place. Enable improvements by incorpora.ng the ability to easily modify a test Users can modify if they have proper controls in place Recognize that evidence suppor.ng the clinical relevance of results can come from the community, not from the test developer. Need to take advantage of high quality sources of evidence and community efforts Enable physicians and pa.ents to access and understand genomic test results to advance clinical decision- making FDA controls to ensure analy.cal and clinical performance

16 Analy/cal: Performance Standards as a Regulatory Tool for NGS Premarket Review Analy.cal performance for each test reviewed by FDA Performance Standards FDA recognizes standards that assure analy.cal performance Cer.fica.on of conformance

17 Applying Analy/cal Standards Premarket Review Manufacturers of assays and components Laboratories that purchase kits for use Performance Standards Laboratories that: assemble tests from components develop soyware perform interpreta.on

18 Using Databases to Assess the Clinical Performance of NGS Tests Replicate experience with CFTR2 on a large scale Concept of a regulatory grade curated database that provides evidence on the strength of associa.on between variants and diseases. Under one approach, if accepted by FDA, any test developer (manufacturer or laboratory) could use asser.ons supported by a predefined level of evidence to support clinical claims without any further requirements. Quality concepts Annota.on (pa.ent, diagnos.c, etc.) Versioning Source of tes.ng results Procedures and prac.ces Sustainability Through PMI, FDA will assess and, if necessary, upgrading exis.ng databases to assure sufficient quality for regula.on.

19 Computa/onal solu/ons to analy/cal performance Use vast genomic datasets now in existence to create new tools to assess the analy.cal performance of NGS tests Is soyware possible that will verify the quality of an individual NGS run? Addi.onal concepts include the establishment of systems to validate NGS soyware. As part of its effort in the PMI, FDA will enable the development of this soyware and make it freely available to the genomics community.

20 Possibili/es for FDA Oversight of NGS Tests Using Performance Standards Regular review and acceptance of standards Regular review and acceptance of databases and/or clinical asser.ons about variants Regular accredita.on of inspectors Adverse event repor.ng Premarket review of components and kits

21 Obtaining Public Input on FDA s NGS Discussion Paper Discussion paper posted to FDA website in Dec 2014 Docket for wrijen comments opened; closed Mar FDA workshop on regula.on of NGS held on Feb Some ini.al feedback to FDA: Support for a standards- based approach Different standards will be needed for different intended uses Need to work with external stakeholders

22 Some PMC Comments The preliminary dra3 assumes that FDA will regulate laboratory developed NGS tests as medical devices. The preliminary dra3 assumes that any processes for oversight of NGS diagnoscc tests will be equally applicable to laboratories and IVD manufacturers without the creacon of a differencated pathway for either group Labeling concerns for LDTs RestricCons should allow laboratory clinical consultacon, including for uses outside the intended use IUO NGS test informacon should be used in the context of research to direct pacents to clinical trials We encourage FDA to also remain open to the consideracon of other opcons of flexible clinical assessment of variant informacon where the most appropriate and effeccve regulatory oversight may not involve curated databases. PMC also encourages FDA to consider applying the criteria discussed in the context of NGS oversight to other advanced genecc technologies.

23 Future Work Standards architecture Technical standards for a specific intended use Technical standards for multiple use cases Policy development Comment analysis and formal response Organiza.on of technical efforts Analy.cal standards (e.g., GIAB) Clinical performance (pilot with ClinVar/ClinGen) Informa.cs, including pen- source computa.onal solu.ons for valida.ng NGS test performance Engagement of external stakeholders DraY guidance or white papers Follow on public workshops on specific topics Development of a final policy

24 Some Important Issues to Address Communica.on of evidence Incen.vizing data- sharing Track provenance of diagnos.c test results EHRs Interoperability Linking EHRs to databases Capturing longitudinal outcomes Guidance on panels should only the ac.onable genome be reported? Assuring database quality developing best prac.ces, etc. Development of regulatory policy