By Diane Allingham-Hawkins, PhD, FCCMG, FACMG, Senior Director, Genetic Test Evaluation Program and Technical Editing
On February 3, 2014, I had the privilege of attending an Institute of Medicine (IOM) roundtable titled Assessing Genomic Sequencing Information for Health Care Decision-Making. This meeting, held in Washington, DC, brought together a wide range of stakeholders with the primary goal of identifying key challenges related to the use of high throughput sequencing technology for clinical purposes and suggesting pragmatic approaches to address those challenges.
First, some context: The advent of high throughput next-generation sequencing (NGS) platforms has enabled laboratories to generate vast amounts of sequencing data in a relatively short period of time. These platforms can sequence many genes or even whole exomes (all the coding regions of every gene) or whole genomes (the entire sequence of all chromosomes, including coding and non-coding regions) quickly and at relatively low cost. This has led to the development of multigene panels, which include up to 100 or more genes that are reportedly associated with a given condition. In addition, a number of providers now offer whole exome sequencing (WES) on a clinical basis. With few exceptions, there is little or no published evidence that these multigene panels or WES have any impact on patient treatment decisions or outcomes. Development of these tests has been technology driven, rather than clinically driven.
In addition, NGS is limited in that it is less accurate than conventional Sanger sequencing methods, with gaps in the sequenced region(s) and the inability to detect all kinds of genetic variants (i.e., NGS will not detect large deletions or duplications). Furthermore, NGS detects a large number of variants of unknown clinical significance. The various NGS platforms use complex computer algorithms to “filter out” variants that have little or no clinical relevance; unfortunately, the algorithms are different between platforms, which means that different platforms could deliver different results.
The IOM meeting was broken into 3 sessions. The first session focused on how evidence is gathered and evaluated. All of the speakers in this session agreed that an objective method is needed to evaluate which information obtained by NGS is medically actionable and that findings that are not considered medically actionable should not be reported. There were a variety of perspectives on how best to achieve this, however. Dr. Madhuri Hedge from Emory University School of Medicine summed up the issues well when she stated that NGS is not a standalone test and needs to be improved technically in order to be used on a clinical basis. In order to define the “medical exome,” Dr. Hedge indicated that we need to establish evidence of gene: disease associations, improve the technical performance of NGS, standardize analysis methods, and define medically actionable genetic variants.
The second session focused on decision making around genomic sequencing data and included speakers addressing coverage determination, professional guideline development, and patient care. Reimbursement for multigene panels and WES varies widely from insurer to insurer and the reasons for coverage or noncoverage are not always clear. The use of evidence in guideline development is inconsistent and the resulting recommendations are often based on consensus and expert opinion rather than a strong evidence base. Finally, in the area of patient care, there are a number of challenges with using genomic sequencing data, including, but not limited to, patient understanding of the risks, benefits, and limitations of such testing; unclear results with many variants of unknown clinical significance; a lack of transparency in the analysis methods used; and ensuring that patients give informed consent for testing. One of the more controversial issues in WES currently is the return of so-called “incidental findings,” results that are not directly related to the reason for testing but which may have medical implications for the patient and their family members. Recently, the American College of Medical Genetics and Genomics (ACMG) recommended that a minimum list of 56 incidental findings should be returned to patients following WES, even if the patient does not wish to receive this information. These recommendations have been criticized as paternalistic and have also raised concerns regarding resources to ensure adequate follow-up. For a discussion of these issues, please see my previous blog post here.
The final session was a wrap-up session with many of the speakers from earlier in the day providing their input regarding next steps. Sadly, I was unable to stay for this session due to the timing of my flight.
For me, the dialogue at the meeting was heartening in that the serious issues with NGS technology are being considered and, hopefully, addressed. The significant technical issues and the lack of evidence regarding clinical utility really mean that this is research rather than clinical testing at this time. I felt that our D2 ratings for these types of tests were supported by the discussions. I am hopeful that some meaningful change and resolution will emerge from this meeting and other similar meetings. In the meantime, let us leave this technology where it belongs: in the research lab.