Health care systems join NIH in studying significance of mutations
Health plans face a daunting task in keeping up with the increased demand for genetic testing. The National Institutes for Health (NIH) says that 16,000 tests for 4,200 medical conditions and 2,800 genes are in its Gene Test Registry.
Genetic testing is expanding quickly, driven by the availability of tests, guideline recommendations for genetic testing, sensationalistic marketing of tests that identify scary heritable diseases, and incredibly powerful assay technologies that have reduced costs considerably.
“We have increased the number of policies relating to genetic testing to close to 100 policies,” says Stanley Harris, MD, medical director of Horizon Blue Cross Blue Shield of New Jersey. “The fact that we have so many policies related to genetic testing reflects our attempt to provide guidance on the medical management of these tests based on evidence published in the peer-reviewed literature.”
One underlying reason for the increased number of policies is that “Many tests do not have evidence that supports a positive coverage decision because clinical utility – how the test will improve the management of the patient or improve the patient’s health outcome – has not been adequately demonstrated,” says Harris.
In May, Horizon implemented a new medical policy for panel testing of cancers that provides insight regarding the requirements for coverage – analytic validity, clinical validity, and clinical utility – of genetic tests. Horizon determined that panel testing is still investigational because these tests do not demonstrate validity and utility.
Regarding clinical validity, Horizon’s policies noted there were no published studies of the analytic validity of panel tests. To demonstrate validity, a panel test would have to show, consistently and reproducibly, sensitivity in detecting variants in the genes it targets. But it can be hard to show that samples used in a study are highly likely to include all of the variants frequently enough that the test’s ability to detect them can be assessed.
Horizon says “the clinical validity of the panels as a whole cannot be determined because of the many different mutations in the large number of potential cancers that [they] can be used for. Clinical validity would need to be reported for each specific mutation for a particular type of cancer.”
To determine clinical utility, Horizon says, “Controlled trials are required in which a strategy of cancer mutation testing followed by targeted treatment based on mutation analysis is compared to standard treatment without mutation testing.”
In genetics, clinical utility depends on the quality of the clinical interpretation. The cost of machine-based sequencing of an entire genome is moving toward the magic $1,000 mark, but experts jest that the power of the technologies to detect an increasing number of variants is leading to the $1 million interpretation. The NIH recognizes this problem and is funding work to improve the clinical interpretation of genetic tests.
The standards used to evaluate panel tests apply equally to tests of single genes, and very few genetic tests have passed muster.
The NIH has gone one step further to express concern about other limitations of genetic testing. It says “systematic and comprehensive profiling of genetic variations is severely underdeveloped in many patient-oriented research or clinical settings. As a result, information on few genomic variants is used in clinical practice.”
Genetic testing routinely identifies novel variants with unknown significance as well as a large number of insignificant variations. Thus, curation – the process of determining a variant’s pathogenicity and looking in genetic databases to see the extent to which those variations have been found previously – is a necessity, requiring a database with a large number of confirming cases and very detailed information about the characteristics of each case.
The NIH has initiated two projects, ClinVar and ClinGen, that may improve the reporting of test results and help to demonstrate the clinical utility of genetic testing.
The NIH says there are about 2,000 separate databases on specific genes and diseases, and many labs have their own protocols for assessing whether the genomic variants they have identified are clinically relevant. It describes a need to agree on the evidence required to decide whether the effects of a variant are medically relevant and to make that evidence available to the public as well as the research and clinical communities.
ClinVar is a project to establish an authoritative reference database with standardized information on genetic variations and also to standardize reporting on genetic variations. It is a public and freely accessible database of genotype (specific genetic alterations), phenotype (observable physical or biochemical traits), clinical interpretation, and supporting evidence.
A major goal is to enable the ongoing evolution and development of knowledge regarding variations and associated phenotypes. ClinVar accepts reports from research and clinical labs and it aggregates information to transparently reflect both consensus and conflicting assertions of clinical significance.
The ClinGen project complements the ClinVar database project. It includes researchers at the Geisinger Health System and Boston’s Partners Healthcare, who will design and implement a framework for evaluating which variants play a role in disease and are relevant to patient care.
The investigators will develop and curate authoritative information on what is expected to be millions of genomic variants relevant to human disease and the thousands that are useful for clinical practice.
“One laboratory reported that it was to able save $76,000 in a single month by reviewing orders for genetic tests and correcting orders for the wrong or unnecessary tests,” according to Michael Watson, PhD, of the American College of Medical Genetics and Genomics.
ClinGen will help to give guidance on how to take the vast amount of data being gathered and put it to effective use in clinical practice, says Michael Watson, PhD, executive director of the American College of Medical Genetics and Genomics (ACMG), a ClinGen grantee.
“The starting point is to create clinical domain work groups in cancer, cardiovascular disease, and metabolic disorders, and the focus is on evaluating variants for their clinical relevance,” says Watson.
Christa Martin, PhD, is a ClinGen investigator who is also director of the autism and developmental medicine institute at Geisinger. “The specialized work groups have been brought together for one purpose: to improve clinical interpretation and bring genomics into everyday medicine,” she says.
The work groups will also set standards for curation and study ways to get this information into electronic health records.
ClinGen is seen as a tool that will help professional organizations develop clinical practice guidelines.
Separately, ACMG is working to improve the quality of genetic testing by focusing on standards for reporting sequence variations. “We are coming to the end of a public process to develop a practice guideline for interpreting sequence variation, not at a gene level but at a genome sequencing level where many issues come into play,” says Watson.
The new guideline deals with the data elements, procedures, and standards for reporting the significance of sequencing variations. “The guideline identifies variant characteristics that are strong indicators of pathogenicity,” Watson says. “For example, when a variant is de novo in a family, it is highly likely that it is pathogenic, particularly if it is in a gene that is associated with a particular disease. There are many types of variant data, and each data element differs in terms of strength of indicators of pathogenicity. The guideline sets data standards for test reports and for conclusions, saying a variant is pathogenic, probably pathogenic, a variant of unknown significance, probably benign, or benign.”
ClinVar and ACMG practice guidelines deal with the quality of test results, but that is only part of the picture. Experts point out that health plans need to pay attention to the full range of genetic testing activities, including support for clinicians in ordering tests, understanding test results, and subsequent patient management decisions.
“One laboratory reported that it was able to save $76,000 in a single month by reviewing orders for genetic tests and correcting orders for the wrong or unnecessary tests,” says Watson. “Doctors have ordered the RET gene test, thinking it is for Rett syndrome, but they are entirely different.”
Amber Trivedi, the senior vice president of InformedDNA, a genetic counseling company, says, “There is a real need for support of clinicians who do not specialize in genetics. Even with simple tests, we see a lot of difficulty with understanding the meaning of test reports. Even for tests that are very well known and have excellent reports, we see providers and patients who do not necessarily know what the results mean and whether or not they should act on it. A BRCA test patient may be told, ‘You’re negative, you’re good,’ but based on family history alone, the patient may meet American Cancer Society guidelines for receiving annual breast MRI.”
