Exploratory evaluation of two-dimensional and three-dimensional methods of FDG PET quantification in pediatric anaplastic astrocytoma: a report from the Pediatric Brain Tumor Consortium (PBTC).

Abstract

The rationale of this study was to investigate the feasibility of three-dimensional (3D) methods to analyze (18)F-fluoro-deoxy-glucose (FDG) uptake in children with anaplastic astrocytoma (AA) in a multi-institutional trial, to compare 3D and two-dimensional (2D) methods and explore data associations with progression-free survival (PFS). 3D tumor volumes from pretreatment MR images (fluid attenuation inversion recovery and postgadolinium) of children with recurrent AA on a phase I trial of imatinib mesylate were co-registered to FDG positron emission tomography (PET) images. PET data were normalized. Four metrics were defined: the maximum ratio (maximum pixel value within the 3D tumor volume, normalized), the total ratio (cumulative pixel values within the tumor volume, normalized) and tumor mean ratio (total pixel value divided by volume, normalized). 2D analysis methods were compared. Cox proportional hazards models were used to estimate the association between these methods and PFS. Strongest correlations between 2D and 3D methods were with analyses using postcontrast T1 images for volume of interest (VOI). The analyses suggest 3D maximum tumor and mean tumor ratios, whether normalized by gray matter or white matter, were associated with PFS. This study of a series of pretreatment AA patients suggests that 3D PET methods using VOIs based on postcontrast T1 correlate with 2D methods and are related to PFS. These methods yield an estimate of metabolically active tumor burden and may add prognostic information after tumor grade is determined. Future rigorous multi-institutional protocols with larger numbers of patients will be required for validation.