Care in Specialized Centers and Data Sharing Increase Agreement in Hypertrophic Cardiomyopathy Genetic Test Interpretation

Abstract

Background - Clinically impactful differences in the interpretation of genetic test results occur between laboratories and clinicians. To improve the classification of variants, a better understanding of why discrepancies occur and how they can be reduced is needed. Methods and Results - We examined the frequency, causes, and resolution of discordant variant classifications in the Sarcomeric Human Cardiomyopathy Registry (SHaRe), a consortium of international centers with expertise in the clinical management and genetic architecture of hypertrophic cardiomyopathy. Of the 112 variants present in patients at >1 center, 23 had discordant classifications among centers (20.5%; Fleiss κ, 0.54). Discordance was more than twice as frequent among clinical laboratories in ClinVar, a public archive of variant classifications (315/695 variants; 45.2%; Fleiss κ, 0.30; P<0.001). Discordance in SHaRe most frequently occurred because hypertrophic cardiomyopathy centers had access to different privately held data when making their classifications (75.0%). Centers reassessed their classifications based on a comprehensive and current data summary, leading to reclassifications that reduced the discordance rate from 20.5% to 10.7%. Different interpretations of rarity and co-occurrence with pathogenic variants contributed to residual discordance. Conclusions - Discordance in variant classification among hypertrophic cardiomyopathy centers is largely attributable to privately held data. Some discrepancies are caused by differences in expert assessment of conflicting data. Discordance was markedly lower among centers specialized in hypertrophic cardiomyopathy than among clinical laboratories, suggesting that optimal genetic test interpretation occurs in the context of clinical care delivered by specialized centers with both clinical and genetics expertise.