NIMG-22. COMPARISON OF NON-INVASIVE, ML-GENERATED WHOLE-BRAIN SPATIAL RADIOPATHOMIC MAPS DERIVED FROM ANATOMICAL AND DIFFUSION-WEIGHTED MRI ON PATIENTS WITH NEWLY DIAGNOSED-GLIOMA

• DOI: 10.1093/neuonc/noad179.0718
• Publication Date: 2023-11-11T23:09:51Z
• Authors: Oluwaseun Adegbite; Nate Tran; Annette Molinaro; Joanna Phillips; Peter LaViolette; Samuel Bobholz; Jacob Ellison; Yan Li; Tracy Luks; Anny Shai; Devika Nair; Javier Villanueva...; Jennifer Connelly; Mitchel Berger; Shawn Hervey-Jumper; Manish Aghi; Susan Chang; Janine Lupo
• Keywords: Brain tumor

Abstract INTRODUCTION Spatial heterogeneity in the glioma microenvironment is difficult to capture through singular biopsy samples. Identification of malignant tumor regions can help guide diagnosis and treatment planning. This study compares radiopathomic maps from 2 strategies derived anatomical diffusion-weighted MRI: 1) a support vector machine model trained on an MRI dataset with tissue samples known spatial coordinates generate probability KI-67 cellularity, 2) bagged random-forest 40+ person brain bank autopsy coregistered end-of-life cellularity (TPM). We hypothesize that TPMs will correlate predictions, while values among methods correlate. METHODS pathology autopsy- sample-based models were generated 55 patients newly-diagnosed glioblastoma scanned anatomical, diffusion-weighted, metabolic imaging. Mean extracted each map within T2-hyperintense, non-enhancing lesion (NEL), contrast-enhancing (CEL), normal-appearing white matter (NAWM), 5mm diameter spheres centered at location where tissues taken. Spearman correlation analysis was used elucidate strength linear relationship selected between two maps. RESULTS Correlations TPM highest most significant sample locations CEL (ρ=0.306, p=0.002 ρ=0.312, p=0.044 respectively). Cellularity autopsy-based showed no newly diagnosed locations, potentially highlighting influence treatment. CONCLUSION examines current state ML-generated mapping, similarities CEL, but need account for effects subsequent progression when comparing beyond different time points.