• J F Schneider, A Viola, S Confort-Gouny, K Ayunts, Y Le Fur, P Viout, M Bennathan, F Chapon, D Figarella-Branger, P Cozzone, and N Girard.
• Centre de résonance magnétique biologique et médicale, UMR-CNRS 6612, faculté de médecine, université de la Méditerranée, Marseille, France. jacques.schneider@ukbb.ch
• J Neuroradiology. 2007 Mar 1; 34 (1): 49-58.

AbstractThe correct assessment of the four most frequent infratentorial brain tumors in children (medulloblastoma, ependymoma, pilocytic astrocytoma and infiltrating glioma) has always been problematic. They are known to often resemble one another on conventional magnetic resonance (MR) imaging. We tested the hypothesis whether the combined strength of diffusion-weighted imaging (DWI) and proton MR spectroscopy (MRS) could help differentiate these tumors. Seventeen children with untreated posterior fossa tumors were investigated between January 2005 and January 2006 with conventional MR imaging and combined DWI and MR spectroscopy using a single-voxel technique at short and long echo time (TE) of 30 ms and 135 ms respectively. Apparent diffusion coefficient (ADC) values were retrieved after regions of interest were manually positioned within non necrotic tumor core. Water signal was quantified and metabolite signals were compared and analyzed using linear discriminant analysis. When a combination of ADC values and normalized metabolites was used, all tumors could be discriminated against one other. This could only be achieved when metabolites were normalized using water as an internal standard. They could not be discriminated when using metabolite ratios or ADC values alone, nor could they be differentiated using creatine (Cr) as an internal reference even in combination with ADC values. In conclusion, linear discriminant analysis and multiparametric combination of DWI and MRS, although not replacing histology, fully discriminates the four most frequent posterior fossa tumors in children, but metabolites have to be normalized using water and not Cr signal as an internal reference.

16 keywords...

hide…