TAMPA (GenomeWeb) – At the American College of Medical Genetics and Genomics annual meeting here today, Stanford University genetic counselor Mitchel Pariani touched on discordant classifications that can occur at clinical centers and testing laboratories tasked with attempting to determine the potential pathogenicity of a variant in a disease-related gene.
"Discordant classifications and recommendations for predictive testing on family members is confusing for patients and may be dangerous if relatives are mistakenly dismissed from clinical screening based on a negative test for something one lab says is clinically valid, but another does not," Pariani and his co-author Jazzmin Williams wrote in the abstract for their presentation.
During an oral platform presentation during a whole-exome sequencing-focused session, Pariani — a member of Euan Ashley's lab at the Stanford Center for Inherited Cardiovascular Disease (SCICD) — showed data generated at the Stanford Marfan Center as part of a review of variants identified in FBN1, a gene implicated in cardiomyopathies, arrhythmias, and other conditions.
For their review, Pariani and his team considered test results for 147 patients seen at the center between the spring of 2002 and late summer of 2015 whose samples had been scrutinized for FBN1 variants at one of a dozen different testing laboratories.
To get a sense of the consistency of variant classification interpretation between labs and clinics, the team analyzed FBN1 variants identified in 110 patient samples using a variant classification scheme developed at SCICD as well as a revised classification standards and guidelines that the ACMG published in Genetics in Medicine in 2015 in collaboration with the Association for Molecular Pathology and the College of American Pathologists.
The team then classified each of the variants — nearly 70 percent of which were missense mutations — as likely benign and/or benign, likely pathogenic and/or pathogenic, or as a variant of unknown significance within each classification system.
In addition to comparing variant interpretations between the ACMG and SCICD protocols, the researchers compared results from both classifications with those made by the original testing labs, looking for cases in which a given variant had been placed in discordant categories.
When they looked at variants assessed using the SCICD protocol and those called at the testing labs, for example, they uncovered more than three-dozen instances of discordant variant classifications. Most of those differences appeared to involve variants falling in conserved, functionally important protein-coding domains of FBN1 that were not sufficiently taken into consideration by the SCICD interpretation, Pariani noted.
Five variants were discordantly classified relative to the testing lab results when the team used the ACMG variant interpretation criteria. Again, the majority of those seemed to stem from differences in knowledge about the specific functional regions in FBN1's protein product.
Such findings point to the importance of reviewing variant classifications and revisiting interpretation guidelines, Pariani explained, particularly since FBN1 is thought to be a relatively well-studied gene. However, he noted, these results also highlight the need for ongoing communication between testing labs, clinical labs, and others to ensure that each is up to date on variants that may have been reclassified based on updated protein domain knowledge and similar data.
"Some laboratories reclassify variants without notifying treatment clinics, which indicates the need for genetic counselors to re-review variants and contact the laboratories on their own," Pariani and Williams wrote.