Metabolic spatial heterogeneities in brain tumours biopsies by NMR microscopy

Introduction HR-MAS metabolic profiles in biopsies has been useful for classifying and subtyping brain tumours [1,2]. However, HR-MAS can exclusively provide the average metabolic profile of the biopsy studied for very heterogeneous lesions as brain tumours. Metabolic spatial distribution by in vivo MRS imaging (MRSI) has been proven very helpful for improving the diagnosis, prognosis and treatment selection [3,4]. MRSI can cover the whole tumour and as well the contralateral region, but the resolution is limited by in vivo clinical MR equipments. MRSI and SV techniques can be applied by NMR microscopy and can be combined for obtaining detailed spatial and average biochemical information on intact biological tissues. Purpose The aim of this communication is to obtain the metabolic profiles of human glioma tumour biopsies, in particular GBM, by using MRSI and SV spectra from NMR microscopy. In addition, an exhaustive comparison between the biochemical information obtained by both techniques and the correlation with pathology results will be carried out. Material and Methods Small pieces of human high grade glioma tumours (GBM), coming from surgical resection (15 20 mg) were introduced in a ZrO2 rotor provided with an insert to achieve an inner ‘cylindrical’ cavity of 50 microlitres, D2O was used to complete the volume. The rotor was introduced in a regular 5mm HR NMR tube and afterwards in a Micro5 imaging probehead with rf insert of 5 mm. A 600 MHz standard-bore magnet was used for NMR microscopy. Orthogonal FLASH (gradient echo) imaging, and RARE (spin echo, fast imaging technique) images were used to allocate the SV (PRESS sequence) and the MRSI volume of interest in selected regions of the biopsy and the D2O solvent. VAPOR was used for water suppression. The temperature of acquisition was 273.7K.The biopsies were afterwards analyzed by conventional histology. Automated 1 and 2 shims (FASTMAP) strategy was used to achieve field homogenization in the volume including the ROI of MRSI sequence and also in the SVs located at different rotor positions and in the same plane than the MRSI. PRESS sequence was used either for acquiring SV or MRSI , with TE/TR of 12/2100 ms. The dimensions were 1mm for SV, for MRSI the FOV was 10 x 5 mm with a thickness of 1mm and for the ROI inside of the rotor was 6 x 2.5 mm. Results Metabolic profiles obtained in each SV (average spectrum) and MRSI (spatial metabolite distribution) spectra were well resolved and with good S/N, as shown in Figures 1 and 2. An adequate correlation between SV average spectrum and the addition of 9 MRSI spectra for almost identical location was obtained for the three different regions on the biopsy samples, as in Figure 2. Noteworthy, a significant amount of metabolites were detected in the solvent region almost immediately the biopsy was in contact with the D2O, middle location in Figure 2. A good correlation between metabolic profiles and histology has been found for the brain tumour biopsies studied, the larger the necrotic content the larger the lipids amount, as for the biopsy in Figure 1 (ca. 80% necrotic composition). However, for the biopsy with high cellularity content, Figure 3 with 90% tumour, the different Cho compounds show high intensity signals and low lipids concentration. Different metabolic profiles have been observed in distinct biopsy voxels from different tissue regions, as in Figures 1 and 3. The average spectrum from the whole sample, tissue and solvent, were similar to the HR-MAS spectrum for a similar biopsy piece for the different GBM studied. Discussion/Conclusion Similar metabolic profiles in tissue or in the surrounding solution can be obtained by using SV and MRSI techniques. High content in lipids (in mostly necrotic tissue) and Choline derivatives (in mostly active tumour tissue) were detected together with absence of NAA. MV profiles in solvent region were almost constant, but spectra in tissue showed metabolic profiles heterogeneity. Metabolic profiles were very similar to the profiles achieved by HR-MAS in another piece from the same biopsies, and also coherent with the histological post-HR-MAS study. References [1] Martinez-Bisbal et al. NMR Biomed. 2004 Jun;17(4):191-205; [2] Monleón et al. J Proteome Res. 2008 Jul;7(7):2882-8. Epub 2008 May 29; [3] Celda et al. Adv Exp Med Biol. 2006;587:285-302; [4] García-Gómez at al. MAGMA 2008, in press. Acknowledgements Ministerio de Educación y Ciencia del Gobierno de España (SAF2007-6547) and European project: eTUMOUR (contract no. FP6-2002-LIFESCIHEALTH 503094).