3D Spectroscopy Evaluation of GBM Treatment Volumes

MR spectroscopic imaging (MRSI) can detect the region of metabolic active tumor via Choline (Cho) and N-acetylaspartate (NAA). Cho is one of the most important up-regulated metabolites in tumors reflecting increased cellular membrane turnover and proliferation; NAA is a marker of neuronal density and viability that is decreased in tumor tissue. MRSI has the ability to provide whole brain coverage at a high spatial resolution and identify complex biochemical patterns associated with tumor histopathology and viability. Here we present a comparison of volumetric MRSI generated Metabolic Tumor Volume (MTV) and standard RT volumes for GBM. MRSI data were acquired at 3T MRI on 64 x 64 x 32 grid, ∼1.0 mL effective voxel size. Maps of Cho and NAA were obtained by automated spectral fitting, which were then transferred to MIM treatment planning station in DICOM format and fused with simulation MRIs. RT volumes are defined as follows: GTV (tumor lesion); CTV (edema); PTV60 = GTV+2.5 cm; PTV46 = CTV+2 cm. Cho and NAA concentrations of normal appearing tissue (NAT) are estimated outside of CTV+5 cm, excluding cerebellum, brain stem and ventricles. Maps of aberrant high Cho (MTVCho) and low NAA (MTVNAA) are constructed using Tukey range test for statistical difference with NAT. We investigated the relationships between the RT volumes and MTVs. MRSI data obtained from eighteen patients with biopsy confirmed GBM prior to any treatment were retrospectively analyzed. RT and MRSI volumes are: (mean [cc] ± SD): GTV: 68 ± 44; CTV: 133 ± 64; PTV60: 368 ± 112; PTV46: 401 ± 125: MTVCho: 66 ± 74; and MTVNAA: 85 ± 47. Significant fractions of MTVCho are outside of GTV (66%) and CTV (48%) (see Table). MTVCho is >60% outside of GTV in 13 (72%) patients while the corresponding number for MTVNAA is 3 (17%). MTVNAA is significantly correlated with CTV (r = 0.67, p = 0.006). To evaluate MTV prospectively we are actively accruing patients on an IRB approved protocol (n = 5). Although in 4 patients the tumor was partially or completely resected, the coverage and quality of MRSI is adequate. MTVNAA is closely overlapping with CTV and does not provide additional metabolic information to MRI with respect to RT volumes. In contrast, significant fractions of MTVCho extend outside of GTV and CTV. Our study indicates that Cho is a marker of tumor infiltration and precursors of lesion formation, which we plan to test further using an ongoing prospective collection of MRSI. The relatively small areas of increased Cho lend themselves to aggressive RT schemes.Poster Viewing Abstract 2134; TableMetabolic volumes outside of treatment volumes (mean % ± SD)GTVCTVPTV60PTV46MTVCho66 ± 2748 ± 2518 ± 1714 ± 15MTVNAA40 ± 2719 ± 94 ± 73 ± 6 Open table in a new tab