Local MR Perfusion in Acute Stroke

Local Perfusion MRI in Acute Stroke

Purpose: Demonstration of “local” perfusion, i.e. all blood flow within a region whether supplied antegrade or delayed and dispersed through the collateral network, containing information regarding collateral supply, infarct growth rate, and flow augmentation treatment response in acute stroke.

Methods and Materials: Two methods of local perfusion (i.e. “local” quantitative cerebral blood flow, qCBF in ml/100g/min) are included: 1) dynamic susceptibility contrast (DSC) MRI with a local arterial input function for delay and dispersion correction , 2) intravoxel incoherent motion (IVIM) which is by nature independent of blood flow path or delay. Standard perfusion is included with DSC using a traditional global-AIF. These are compared to standard AIF-dependent DSC perfusion. Experiments in a model of middle cerebral artery occlusion were performed and native collateral circulation was quantified via x-ray DSA 30 minutes post-occlusion. Some also underwent flow-augmentation aimed to boost collateral supply to the affected territory. DSC MRI was acquired 2.5hrs post-occlusion and post-processed first with a standard global-AIF, and second with a Local-AIF. A subset with IVIM MRI was acquired and post-processed to returned local-qCBF from fD*. Infarct volume was measured with standard DWI at 2.5hrs and 4hrs post-occlusion.

Results: (1) The degree of collateralization at x-ray correlated more strongly with DCS local-qCBF in the ischemic penumbra (R² = 0.81) than with standard DSC qCBF (R² = 0.05); inclusion of delayed and dispersed flow for a “local” DSC perfusion demonstrated ability to capture collateral flow in the affected territory. (2) IVIM local-qCBF in the penumbra correlated with DSC local-qCBF (R²=0.69); the local nature of IVIM, as it is intravoxel rather than tracer-based, may by its very nature automatically include all perfusing blood, including collateral flow, in the affected territory. (3) Correlation was weaker between IVIM and standard DSC (R²=0.13); standard DSC may have false hypoperfusion when it does not account for a robust collateral supply, while IVIM inherently includes it. (4) Both DSC and IVIM local-qCBF predicted final 4hr infarct volume (R²=0.78, R²=0.61 respectively) while standard DSC was a weaker predictor (R²=0.12). The volume of hypoperfused lesion from DSC and IVIM local-qCBF both predicted final infarct volume with good sensitivity and correlation (slope=2.08, R²=0.67, slope=2.50, R²=0.68 respectively). As collateral flow can delay or slow infarction, inclusion of collateral flow using “local” perfusion imaging may improve prediction of infarction.

Conclusions: In acute stroke, local perfusion with Local-AIF DSC or IVIM can assess tissue status and collateral supply in acute stroke more accurately than conventional Global-AIF DSC. These findings support using local perfusion that includes delayed and dispersed retrograde flow when imaging the affected territory of acute stroke. Further, as IVIM is non-contrast, it can allow longitudinal and repeated measurement of local-qCBF may include collateral circulation, simultaneous perfusion and diffusion images, and improve prediction of infarct growth.