PO64VOLUME BASED ANALYSIS OF GLIOBLASTOMAS. PROGRESSION IN IMAGING TECHNIQUES

Abstract

INTRODUCTION: Response assessment for Glioblastoma (GBM) is often assessed using the RANO criteria based on linear measurements. However linear measurements may be an inaccurate reflection of volume for tumours with cystic necrotic elements or complex geometric shapes. Automated computer-assisted volumetric analysis may offer a more accurate assessment in such situations. This study explores the differences between linear and volumetric tumour measurements. METHOD: 10 patients with heterogeneously enhancing GBM on 3D post contrast T1 volumetric MRI sequence were selected for this study. Spherical volume of GBM was estimated from 2D measurements (linear) by taking the average of two perpendicular diameters. Volumetric analysis was performed by manually contouring gadolinium enhanced sections of tumour on Agfa PACS. RESULTS: Mean tumour volume was 13.19 cm3 and 10.06 cm3 using linear and volumetric techniques respectively. The nonparametric Wilcoxon signed-rank test (p = 0.092) informs that distributions of 2D Volume and Volumetry are different at the 10% significance level. The one-sided nonparametric sign test (p = 0.05) informs that the 2D Volume is greater than Volumetry for 8 out of 10 observations. CONCLUSION: Linear measurements can over estimate the volume of tumour compared to volumetric analysis. Measuring the change in volume of GBM may provide a more accurate assessment of response than conventional linear measurements used for existing RANO criteria, particularly for necrotic tumours. The benefit of volumetric analysis to assess response in GBM requires prospective validation. Integration of this technique into a prospective study will be discussed.