Overall survival prediction in glioblastoma patients using structural magnetic resonance imaging (MRI): advanced radiomic features may compensate for lack of advanced MRI modalities

• DOI: 10.1117/1.jmi.7.3.031505
• Publication Date: 2020-06-09T15:26:21Z
• Authors: Spyridon Bakas; Gaurav Shukla; Hamed Akbari; Guray Erus; Aristeidis Sotiras; Saima Rathore; Chiharu Sako; Sung Min Ha; Martin Rozycki; Russell T. Shinohara; Michel Bilello; Christos Davatzikos
• Keywords: 03 medical and health sciences; 0302 clinical medicine

Purpose: Glioblastoma, the most common and aggressive adult brain tumor, is considered noncurative at diagnosis, with 14 to 16 months median survival following treatment. There increasing evidence that noninvasive integrative analysis of radiomic features can predict overall progression-free survival, using advanced multiparametric magnetic resonance imaging (Adv-mpMRI). If successfully applicable, such markers considerably influence patient management. However, patients prior initiation therapy typically undergo only basic structural mpMRI (Bas-mpMRI, i.e., T1, T1-Gd, T2, T2-fluid-attenuated inversion recovery) preoperatively, rather than Adv-mpMRI provides additional vascularization (dynamic susceptibility contrast-MRI) cell-density (diffusion tensor imaging) related information. Approach: We assess a retrospective cohort 101 glioblastoma available from previous study, which has shown an initial feature panel (IFP, intensity, volume, location, growth model parameters) extracted yield accurate stratification. focus on demonstrating equally prediction models be constructed augmented panels (ARFPs, integrating morphology textural descriptors) solely widely Bas-mpMRI, obviating need for Adv-mpMRI. 1612 distinct tumor subregions build multivariate stratified as long-, intermediate-, or short-survivors. Results: The classification accuracy utilizing protocols IFP was 72.77% degraded 60.89% when Bas-mpMRI. ARFP Bas-mpMRI improved 74.26%. Furthermore, Kaplan–Meier demonstrated superior subjects into short-, long-survivor classes Conclusions: This quantitative evaluation indicates in feasible features, compensate lack Our finding holds promise generalization across multiple institutions may not have access better inform clinical decision-making about interventions trials.