Standards for Personalized Medicine

Abstract

In a recent hearing of the U.S. House of Representatives Science and Technology Committee, I had the opportunity to testify about genetic testing and other innovative technologies as the Committee considered the issue: How our National Institute of Standards and Technology (NIST) can more effectively influence innovation in life sciences. I told the Committee that I believe that the success of the biomedical research enterprise and America as a leader in innovation depends on NIST providing standards upon which to build personalized medicine, including high quality precision molecular diagnostics. Currently, each developer or service laboratory offering genetic tests maintains its own standards, leading to inconsistencies between providers, laboratories, protocols, and test results. Americans believe that they are receiving healthcare in the form of diagnostic tests that are high quality, accurate, valid, useful, and consistent. No doubt in most instances they are, but since there are limited ways to compare one laboratory’s protocol to another, or one test to another, there is no certainty in this supposition. Most measurements are relative, internal to one laboratory, one state (in the case of newborn screening), or one technology company. Every manufacturer applies relevant measurement technology with their own standard references and controls and as a result makes comparisons of truth and accuracy extremely difficult to assess, if not impossible in some circumstances. In some instances, these technologies, genetic testing among them, are quickly migrating into complex clinical circumstances and having a dramatic impact on decisionmaking. These applied technologies have far and away outpaced the current state and federal oversight regimes and regulatory schemas. We need new measurement and information standards up to the task of 21st century genomic medicine. The current standard for laboratory performance in the United States is the Clinical Laboratory Improvement Amendment (CLIA). CLIA is structured in such a way that laboratories can create their own internal standards and quality systems. At the present time, almost every single standard for every single test is unique to the test provider. This has created an untenable morass of controversy and confusion. Only a small number of standard analytes exist upon which the 2700 genetic tests currently listed in GeneTests (www.ncbi.nlm.nih.gov=sites=GeneTests=?db1⁄4GeneTests) are based. Because of the variability of methods across the labs performing these tests in the United States and beyond, the actual number of unique tests might be closer to hundreds of thousands. Further, the current online catalog of tests is voluntary, with a vast number of tests being offered to patients not being on the list. Over the last few years, many organizations and advisory bodies have called for a mandatory national test registry. Without such a system we are only guessing at who is doing genetic testing, what tests are being used, and how genetic testing is being delivered to patients today. Organizations such as NIST could work with industry, professional associations, and medical specialty societies to help create a benchmark for accuracy in measurement science that would bring biomedical research and technologies a level of evidence they sorely need to move, with confidence, into the era of personalized medicine. In all cases, handling, storage, and preparation methods all have influence on the accuracy of a laboratory measurement. It is very difficult to control for all these dynamically complex variables in a measurement science. In reality, NIST has been organizationally paralyzed in part because of the large number of variables, technologies, and all manner of testing approaches. It is bewildering to figure out where to start, and the agency seems to be overwhelmed by the challenge. If the biomedical testing universe is too big for everything to be tackled at once, then NIST should begin by producing methods and consensus standards as a first stage for instance. The human genome, in all of its complexity, is a fixed repository. It is a recipe for biological processes that has 23,000 functioning genes and probably 100,000 alternate transcripts that could be mapped today and catalogued. Many of these genes and their alterations will be linked to various diseases and delivered as genetic tests within the next few years. NIST could collaborate in a much more effective way with the U.S. Food and Drug Administration (FDA) in the submissions they receive and integrate standards more frequently into the regulatory regime. Certainly at first we would be demanding more of a near perfection standard from these new technologies than what was cleared under the FDA’s predicate standard, but one hopes science moves us all forward. These technologies, in genetics, genomics, laboratory science and imaging, are migrating into health care via multiple avenues. It is critical that patients know that these GENETIC TESTING AND MOLECULAR BIOMARKERS Volume 14, Number 2, 2010 a Mary Ann Liebert, Inc. Pp. 155–156 DOI: 10.1089=gtmb.2010.1509