P3 Artificial intelligence (AI)-driven exome re-analysis and reinterpretation of undiagnosed developmental delay cases

Abstract

<h3>Introduction</h3> Developmental delay (DD) and intellectual disability (ID) affect 1–3% of children. These cases are complex, requiring lengthy analysis due to the number of genes associated with these phenotypes. Congenica’s proprietary AI tool ranks and classifies variants by phenotypic relevance and likely pathogenicity, enabling panel-agnostic analysis for complex presentations. Successful application of Congenica-AI was demonstrated in the re-analysis of undiagnosed epilepsy cases (Trump et al., 2021). <h3>Objectives</h3> To evaluate an AI-first approach in the re-analysis of undiagnosed DD/ID families. <h3>Methods</h3> AI-driven re-analysis of exome data from 113 consented, undiagnosed DD/ID families was performed 2-5 years after initial interpretation. The top 20 Congenica-AI variants were reviewed to identify potential new diagnoses. <h3>Results and Conclusion</h3> AI-driven re-analysis identified 10 new diagnoses, increasing diagnostic yield from 25% to 30.2%, with an average analysis time of 15 minutes per case. Importantly, six new diagnostic variants were outside of the Exomiser top 10, and five were in genes not known to be associated with the patient phenotype at the original time of testing, highlighting the significant value of Congenica’s AI gene-agnostic approach to quickly reinterpret older cases. These results demonstrate the significant role that Congenica-AI can play in making routine re-analysis efficient and feasible.

Ähnliche Arbeiten