Purpose: We sought to retrospectively investigate differences in correlation of absolute metabolites concentrations and metabolite ratios with histopathologic parameters of stereotactic biopsies from the border zone of gliomas.
Materials and methods: Proton magnetic resonance spectroscopic imaging (H-MRSI) with high spatial resolution was performed in 10 glioma patients at 1.5 T. Magnetic resonance spectroscopic imaging (MRSI) data were coregistered to a 3D MR data set used for stereotactic procedures. Metabolite concentrations of choline-containing compounds (Cho), creatine (Cr), and total N-acetyl-aspartate (tNAA) in addition to metabolite ratios of Cho/Cr, tNAA/Cr, and Cho/tNAA were calculated for voxel positions at biopsy loci with low tumor cell infiltration. Metabolite values were correlated with histopathologic findings expressed as a relative (% tumor infiltration, %TI) and an absolute parameter (tumor cell number, TCN).
Results: We found a strong negative linear correlation for tNAA with %TI (R = -0.773, P < 0.001) and TCN (R = -0.769, P < 0.001) but no correlation for Cho (P > 0.05). On the other hand, the metabolite ratio of Cho/Cr showed a moderate positive linear correlation with %TI (R = 0.523, P = 0.012) and TCN (R = 0.486, P = 0.019), but no correlation was obtained for tNAA/Cr (P > 0.05). Differences in correlation between tNAA and Cho as well as tNAA/Cr and Cho/Cr were significant for both %TI (P = 0.012 and P = 0.024) and TCN (P = 0.016 and P = 0.040) using a t test.
Conclusions: We conclude that absolute values of tNAA are more significant than Cho in the detection of low tumor cell infiltration. On the basis of absolute metabolite values, neuronal damage seems to be more distinct than cell membrane proliferation in the border zone of gliomas. Furthermore, the calculation of metabolite ratios versus Cr for the border zone may yield to misleading results because Cr is not constant in this area.