Oncogenicity Variant Interpreter (OncoVI) Supports Harmonized Somatic Variant Interpretation in Precision Oncology

Accurate and reproducible interpretation of somatic variants is fundamental for therapy decision-making in patients with cancer. To harmonize and automate oncogenicity classification, Oncogenicity Variant Interpreter (OncoVI), an open-source, Python-based implementation of the Clinical Genome Resource/Cancer Genomics Consortium/Variant Interpretation for Cancer Consortium oncogenicity guidelines, was developed. For each of the guideline criteria, the textual descriptions were interpreted, and publicly available resources were identified to be used as reference. Starting from the genomic coordinates of a variant, OncoVI automatically performs functional annotation, collects relevant evidence from the integrated resources, evaluates each criterion, and provides a final oncogenicity classification. OncoVI achieved an accuracy of 80% on a gold standard set of 93 somatic variants provided by the guidelines, with a sensitivity of 88% for oncogenic/likely oncogenic variants. When applied to a real-world set of 7802 variants from 557 participants previously evaluated by the Molecular Tumor Board (MTB) Erlangen, OncoVI showed 79% concordance with the prior MTB assessment of variant impact on protein function. In addition, expert reassessment of 135 MTB variants, conducted in accordance with the oncogenicity guidelines, further confirmed both the validity of OncoVI implementation and the appropriateness of the identified resources. Taken together, OncoVI provides significant support for the harmonized and reproducible oncogenicity classification of somatic variants across institutions.