MR diffusion and perfusion parameters: relationship to metabolites in acute ischaemic stroke

Background Magnetic resonance (MR) diffusion and perfusion imaging are used to identify ischaemic penumbra, but there are few comparisons with neuronal loss and ischaemia in vivo. The authors compared N-acetyl aspartate (NAA, found in intact neurons) and lactate (anaerobic metabolism) with diffusion/perfusion parameters.
Methods The authors prospectively recruited patients with acute ischaemic stroke and performed MR diffusion tensor, perfusion (PWI) and proton chemical shift spectroscopic imaging (CSI). We superimposed a 0.5 cm voxel grid on the diffusion-weighted images (DWI) and classified voxels as ‘definitely abnormal,’ ‘possibly abnormal’ or normal on DWI appearance, and ‘mismatch’ for voxels in DWI/PWI mismatch areas. The authors compared metabolite (NAA, lactate), perfusion and apparent diffusion coefficient (ADC) values in each voxel type.
Results NAA differentiated ‘definitely’ from ‘possibly abnormal,’ and ‘possibly abnormal’ from ‘mismatch’ (both comparisons p<0.01) voxels, but not 'mismatch' from 'normal' voxels. Lactate was highest in 'definitely abnormal,' and progressively lower in 'possibly abnormal,' 'mismatch,' than 'normal' voxels (all differences p<0.01). There was no correlation between NAA and ADC or PWI values, but high lactate correlated with low ADC (Spearman r = -0.41, p=0.02) and prolonged mean transit time (Spearman r=0.42, p=0.02). Conclusion ADC and mean transit time indicate the presence of ischaemia (lactate) but not cumulative total neuronal damage (NAA) in acute ischaemic stroke, suggesting that caution is required if using ADC and PWI parameters to differentiate salvageable from non-salvageable tissue. Further refinement of the DWI/PWI concept is required prior to more widespread use.