1189P - Cancer genome interpreter: A data-driven tool for tumor mutation interpretation

Background

Implementing precision oncology in clinical practice faces the challenge of accurately interpreting genomic alterations in tumors, ultimately improving patient outcomes. The rapid development of next-generation sequencing (NGS) and the introduction of large gene panels (including whole exome and whole genome sequencing) into clinical settings underscore the need for advanced systems capable of interpreting increasingly vast genomic data. An additional hurdle in interpreting tumor mutations is the high prevalence of variants of unknown significance (VUS) identified in cancer patients.

Methods

To address these challenges in cancer variant interpretation, we developed the Cancer Genome Interpreter (CGI) tool. Currently, through the EU-funded CGI-Clinics project, we are further optimizing CGI for clinical implementation in collaboration with ten hospitals across five European countries and considering the patient's perspective.

Results

CGI identifies the oncogenic mutations from a tumor and highlights biomarkers of drug response associated with these mutations. A key singularity of the CGI tool is its integration of data-driven approaches in combination with expert-curated databases to comprehensively classify mutations, identifying the oncogenic ones. Specifically, CGI integrates IntOGen, a computational framework for pinpointing cancer driver genes, and BoostDM, a machine learning-based method for identifying cancer driver mutations. These computational methods are tumor- and gene-specific, providing an oncogenicity classification to each mutation, including VUS, while explaining the mutational features used for that annotation. These and other data-driven approaches constitute CGI’s automatic learning platform, which leverages mutation data from thousands of cancer genomes. Continuous access to an increasing number of tumor genomes improves CGI’s interpretation capabilities for current and future cancer patients.

Conclusions

The CGI tool enhances the interpretation of tumor variants from cancer patients by employing data-driven computational methodologies to interpret both known mutations and VUS. CGI is currently undergoing clinical adaptation for its implementation as a clinical decision-support tool.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

Barcelona Biomedical Genomics Group (IRB Barcelona).

Funding

European Union's Horizon Europe programme under grant agreement 101057509 (CGI-Clinics project).

Disclosure

All authors have declared no conflicts of interest.