What question did this study set out to answer?

The aim is to develop a machine learning framework that prioritizes actionable genetic variants for cancer treatment.

April 19, 2026

Abstract LB447: Somatic DiagAI: Automated scoring of drug-variant associations to support clinical decision-making in cancer genomics

Key Points

The aim is to develop a machine learning framework that prioritizes actionable genetic variants for cancer treatment.
Utilized a dataset of 604 patients analyzed via TSO 500 gene panels.
Integrated multiple databases for variant characterization, including JAX-CKB and gnomAD.
Applied expert-curated variant annotations according to Compermed guidelines.
Developed a scoring model that quantifies genetic and pharmacological parameters.
Generated a predictive score ranging from 0 to 100 for drug-variant associations.
Enhanced variant interpretation through systematic integration of biological significance and pharmacological characteristics.
Provided a framework aimed at improving the accuracy of clinical decision-making in oncology.

Abstract

Abstract Large-scale gene panel sequencing has emerged as a critical tool for identifying tumor biomarkers that guide treatment optimization and clinical trial enrollment in contemporary oncology. However, the rapid evolution of genomic knowledge shows significant challenges for pathologists in maintaining current expertise regarding drug-variant associations and accurately prioritizing clinically actionable genetic alterations. SeqOne has developed somatic DiagAI, a novel machine learning framework designed to systematically prioritize actionable variants by computing a quantitative score (0-100) for each potential drug-variant association. The model integrates four critical parameters: (1) the biological significance of genetic variants, (2) pharmacological characteristics of associated therapeutics, (3) concordance between molecular profiles and approved treatment indications, and (4) patient-specific clinical context. The model quantifies the contribution of each component and generates a score that makes predictions interpretable and traceable. The model was trained and validated using a cohort of 604 patients analyzed via TSO 500 gene panels, with documented clinical indications and expert-curated variant annotations. Variant characterization was enhanced through integration of established databases including JAX-CKB and gnomAD, with clinical interpretation performed according to Compermed guidelines. This approach provides a systematic framework for variant prioritization that addresses the growing complexity of precision oncology decision-making. Citation Format: Nicolas Duforet Frebourg, Mitchell Nemcek. Somatic DiagAI: Automated scoring of drug-variant associations to support clinical decision-making in cancer genomics abstract. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts) ; 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86 (8Suppl): Abstract nr LB447.

Bookmark

Abstract LB447: Somatic DiagAI: Automated scoring of drug-variant associations to support clinical decision-making in cancer genomics

Key Points

Abstract

Cite This Study

Also Consider

Also Consider