IMG-143. AI-Guided RANO 2.0 Assessment and Pseudoprogression Incidence in Glioblastoma: A Comparative Analysis of Bidimensional and Volumetric Measurements

Abstract Pseudoprogression (PsP) represents a diagnostic challenge in glioblastoma (GBM) management. This ongoing study evaluates AI-guided bidimensional and volumetric measurements for RANO 2.0 assessment, including T2/FLAIR analysis, and determines PsP incidence rates in a retrospective GBM cohort. Retrospective analysis of GBM patients with post-treatment MRI surveillance is underway. AI-guided measurements with human oversight include: (1) bidimensional target lesion measurements per RANO 2.0 criteria, (2) 3D volumetric tumor segmentation of enhancing components, and (3) AI-guided volumetric T2/FLAIR assessment of non-enhancing disease. Pseudoprogression is being confirmed by clinical stability and imaging resolution based upon confirmation of progression by RANO 2.0. Inter-method agreement is assessed using Bland-Altman analysis and Cohen’s kappa. T2/FLAIR volumetric correlation with pseudoprogression versus true progression patterns is being analyzed. Data collection and analysis are in progress. Enhancing-to-T2/FLAIR volume ratios are being evaluated as predictive biomarkers for pseudoprogression identification. Final T2/FLAIR correlation patterns with pseudoprogression, PsP incidence rates, and detection sensitivity metrics will be reported. Multiparametric AI-guided approaches may significantly improve pseudoprogression identification and clinical decision-making in GBM patients.