Survival prediction in head and neck cancer utilizing vision graph neural network on PET/CT images

Head and neck cancers present significant challenges in treatment due to high rates of locoregional recurrences and distant metastasis, making accurate survival prediction essential for effective patient management. This study introduces a novel approach to survival prediction by integrating computed tomography (CT) and positron emission tomography (PET) imaging data within a Vision-graph neural network (GNN) architecture. The Vision-GNN model excels in capturing complex anatomical structures and spatial relationships by processing medical images as graphs. By combining the detailed anatomical information from CT with the functional insights from PET, our model provides a comprehensive representation of the tumor environment. The model was trained and validated using the "Head-Neck-PET-CT" dataset from the Cancer Imaging Archive (TCIA), achieving an accuracy of 0.8590, and a ROC-AUC score of 0.9088. These results demonstrate the efficacy of our approach in enhancing survival predictions for HNC patients.