Innovative Prediction of VEGF Levels and Prognosis in Gastric Cancer through PET/CT-Based Radiomics

Abstract

Background Gastric cancer (GC) remains a major challenge in oncology due to its late diagnosis and poor prognosis. Predicting Vascular Endothelial Growth Factor (VEGF) levels and survival outcomes accurately can significantly enhance therapeutic decision-making. This study introduces an innovative approach utilizing [18F] FDG PET/CT radiomics to predict VEGF status and survival outcomes, aiming to improve personalized treatment strategies in GC. Methods We performed a retrospective analysis of gastric cancer patients who underwent [18F] FDG PET/CT scans. Radiomics features were extracted from these scans and subjected to Least Absolute Shrinkage and Selection Operator (LASSO) regression to develop a predictive Radiomics Score (RS). The effectiveness of RS in predicting VEGF status and survival was assessed using ROC curve analysis and Cox regression models, respectively. Results The RS demonstrated excellent predictive capabilities with an Area Under the Curve (AUC) of 0.861 in the training cohort and 0.857 in the validation cohort for VEGF status. It also significantly predicted overall survival, with patients having higher RS experiencing worse outcomes (Hazard Ratio = 5.063, p < 0.05). Conclusion This study successfully develops and validates a radiomics-based model using [18F] FDG PET/CT that predicts both VEGF levels and survival in gastric cancer patients. This model provides a foundation for non-invasive, precision oncology approaches that can significantly impact clinical practice by facilitating targeted treatment plans.