1. Automated variant re-evaluation is labor-balanced and gives clinically relevant results: Hereditary cardiac disease as a use case.
- Author
-
Grosen A, Lautrup CK, Bahsen E, Jensen HK, and Lildballe DL
- Abstract
Background: Genetic findings influence clinical care of patients suspected of hereditary cardiac diseases. As additional knowledge arises over time, the classification of genetic variants may change. The labor cost associated with systematic manual reevaluation for reported variants is substantial. We applied an automated variant classifier for reevaluation of previous reported variants to assess how such tools may assist in manual reevaluation., Methods: Historically (2010-2022), patients (N = 2987) suspected of inherited cardiomyopathies or ion-channel disorders were screened for genetic variants in at least one of up to 114 genes. We had reported 1455 unique variants, of which 742 were among the 14 most relevant genes. In the 14-gene-group, we compared our reported classification to that of an autoclassifier and manually reevaluated variant classification of all variants. Among the remaining genes (N = 100), only variants where the autoclassifier predicted change of clinical impact, such as variant of uncertain significance to likely pathogenic or oppositely, were manually reevaluated., Results: We identified 9% (66/742) of variants with clinical impact in the 14-gene-group. Of these, 91% could have been identified solely evaluating the 120 variants where the autoclassifier had predicted a change of clinical impact. In the 100 remaining genes, a change of clinical impact was identified in 3% (22/713) after manual reevaluation., Conclusion: Using an autoclassifier reduces the workload to identify variants likely to have a change in variant class with clinical impact. Hence, we recommend using such tools to identify the variants most relevant to manually reevaluate to improve patient care., (Copyright © 2024 The Authors. Published by Elsevier Masson SAS.. All rights reserved.)
- Published
- 2024
- Full Text
- View/download PDF