Development a nomogram integrating deep learning-radiomics, pathomics and Vasari features to predict prognosis of glioblastoma patients

Motivation: The prognosis for glioblastoma remains dismal. An accurate assessment of prognostic stratification is crucial for guiding clinical decisions. Goal(s): Our aim to develop a nomogram that offers precise patient survival predictions. Approach: First, We analysis 194 glioblastoma patients clinical features using Cox regression, extraction radiomics features, and pathomics features based on MRI images and WSIs, respectively. Finally, constructed nomogram to predict OS. Results: The combined model C-index of 0.887 and 0.791 for OS in train and test cohort. The ROC curve of the combined model at predicts 1-, 2-, and 3-year OS produced AUC of 0.871, 0.725, and 0.833, respectively. Impact: The combined model nomogram, created through multimodal data integration of clinical characteristics, Deep learning radiomics signatures, and pathomics features, enhanced the prognostic risk stratification for patients with glioblastoma.