Thesis

Concordance in classification of hypertrophic cardiomyopathy variants is markedly higher among expert centers than among clinical labs

It has become increasingly clear that there are often marked differences in how variants in Mendelian disease genes are classified by different laboratories. Efforts to improve variant classification strategies will be aided by a thorough understanding of why differences in classifications exist. To gain such insights, we examined discordance in sarcomere variant classifications in SHaRe, a consortium of multiple international centers with expertise in cardiomyopathy genetics. We evaluated the frequency of disagreement in variant classifications between centers submitting to SHaRe and compared it to the frequency of disagreement in classifications among clinical laboratories submitting to ClinVar. The frequency of discordance in ClinVar was two and a half times higher than in SHaRe. We then assessed the severity of discordance as severe (i.e. (likely) benign vs. (likely) pathogenic), moderate (i.e. VUS vs. pathogenic or benign), or modest (i.e. VUS vs. likely benign or likely pathogenic). Sixteen of the 21 discordant variants from SHaRe were further analyzed for severity and reasons of discordance. The majority of discordant classifications in SHaRe were modest (9/16), while about half were moderate (159/314) in ClinVar. To identify the sources of discordance in SHaRe, we compared the data and rationale that each center used in making their classifications. Among the most frequent cause of discordance were differences in privately held data (testing lab –11/16; clinical center –10/16), in lit data used (8/16), and in interpretations of the same data (7/16). The less frequent discordance among cardiomyopathy genetics centers may be due to classification of variants by experts or due to more thorough follow-up of genetic findings including segregation analysis. Taken together, these data suggest that variant classification can be improved by evaluation at specialized centers. Our findings also underscore the importance of sharing of privately held data as this was a frequent cause of discordance.

Relationships

Items