T2* Perfusion Imaging at 7.0 Tesla using 3D PRESTO: Initial results

Purpose: 7.0 Tesla ultra high field systems are currently introduced for clinical imaging, especially focusing on high resolution structural imaging and fMRI applications. Single-shot EPI at 7.0 Tesla will suffer from severe distortions, while multi-shot techniques can often not establish the required temporal resolution. Furthermore, a shorter TE will be required at this high field strength which is often not possible with a single-shot approach. We investigated the use of a multi-shot technique with echo-shifting (PRESTO) to overcome some of the distortions and to prevent low temporal resolution. The aim of this study was to test the feasibility of contrast enhanced T2*- Perfusion at 7.0T.

Methods: The T2*-perfusion imaging study was performed in one male thirty-one year old volunteer using a 3D PRESTO sequence (TR 12.35ms/TE 18.5ms, effective TR 1s., Slice thickness 3.5mm, 64×64 Matrix, FOV250mm, rFOV 70%, 20 section with no gap covering the supratentorial brain tissue, 150 dynamic scans, total scan time 146s.). The imaging was conducted at a 7.0T system (Achieva 7.0T, Philips Medical Systems, Cleveland, OH, USA). A contrast medium dose of 0.05 mmol/Kg bodyweight (5ml), Gadovist®, Schering, Berlin, Germany was injected via a 14G intravenous catheter after the 5th dynamic acquisition. The data were transferred to an off-line workstation (Medx 3.4.3, Medical Numerics, Inc., Sterling, Virginia, USA). Dedicated perfusion analysis software was used to calculate parametric maps for MTT (Mean Transit Time), CBV (cerebral blood volume) and the CBF (cerebral blood flow). The maximal T2* (Smax induced signal drop at the bolus peak was measured. The source images of the perfusion data were inspected visually for image distortions. The parametric maps were judged for image quality, and the number of non fitted voxels was rated. The calculated time-intensity-curves from the dynamic data as visually inspected and the sharpness of the bolus peaks were evaluated. Quantification of the perfusion data was done using the Medx package. AIF was estimated in the large brain feeding vessels-CBF was calculated after AIF deconvolution. A region of interest analysis was performed in several regions of the white and gray matter for MTT, CBV, CBF, and Smax.

Results: Only minor image distortions were seen in the source data. Clear and sharp bolus peaks could be visualized for the gray and white matter without a plateau shaped saturation at bolus peak time. The mean MTT was 11.9s for the gray matter and 9.2s for the white matter. Gray matter CBV was 34.4, and 8.1 for the white matter respectively. A mean CBF of 54.3ml per minute per 100g was measured in the white matter and a CBF of 165.4ml per minute per 100g was measured in the gray matter. The mean measured Smax for the gray matter was 54.7% and 21% for the white matter.

Conclusion: 3D perfusion imaging (PRESTO) is feasible at 7.0T even without parallel imaging techniques. Further significant improvements in image quality and resolution can be expected when parallel imaging techniques and multi-channel coils are available for 7.0 T