Genetic Algorithm for Predicting Primary Imatinib Resistance in Locally Advanced or Metastatic Gastrointestinal Stromal Tumors

Gastrointestinal stromal tumors (GISTs) are the most common mesenchymal neoplasms, primarily driven by KIT or PDGFRA mutations. While tyrosine kinase inhibitors (TKIs), such as imatinib, have transformed treatment, resistance remains a significant challenge. Alternative TKIs offer options but with increased toxicity and inconsistent efficacy. This study investigates artificial intelligence-driven radiomics for predicting primary TKI resistance in GIST patients. Sixteen patients were classified as imatinib-sensitive or -resistant based on CT scans taken at baseline and one year post-treatment. Expert radiologists manually segmented CT scans and extracted the volume of interest, composed of the principal mass and metastasis. A total of 93 radiomic features were extracted from each axial slice from the axial view. A genetic algorithm was implemented for feature selection and model hyperparameter optimization. The Support Vector Classifier demonstrated the highest accuracy (95%), underscoring the potential of AI-driven radiomics to guide personalized GIST treatment.