Introduction

Time-of-flight (TOF) angiography is considered the clinical gold-standard of acquiring non-contrast enhanced angiographic data. The technique is acquired as static angiogram and therefore information about dynamic blood flow is not available or can only be assumed secondary¹. Arterial Spin Labeling (ASL) on the other hand can be used for time-resolved imaging of blood-flow, but lacks the spatial resolution of TOF angiography². More information about anatomical and flow properties of a given vasculature could be derived by combining both, the high resolution of TOF and the time-resolved information from ASL. Recently such a technique was presented³. However, this approach increased the scan time by several minutes. Therefore, the aim of this study is to present an approach that should be keeping the scan time penalty of acquiring these two imaging sequences as low as possible.

Materials and Methods

Five healthy volunteers underwent MR scanning under the general protocol for sequence development, approved by the local ethics committee. Imaging was performed on a Philips 3T Achieva (Philips, Best, The Netherlands) scanner using a standard 32 channel SENSE Head coil. TOF parameters were: 3D T1-FFE readout with 20° Flip Angle, 0.45*0.45*0.8mm³ voxel size and 135 slices with a scan time of 5 minutes. ASL parameters were: Thick-slab 2D T1-TFE readout with 10° Flip Angle, 1*1mm2 in-plane voxel size, slice thickness: 120mm. 6 tilted thick-slabs were acquired with angular increments of 30° and 6 acquisition time-points with a temporal resolution of 150ms after labeling (Figure 1). Scan time per projection: 8 seconds. To map the time-resolved information from ASL to TOF, at first the ASL data from the different acquisitions was combined to a 3D image using the filtered back projection approach. This was performed for 6, 4 (45° angular increments), and 2 thick-slab acquisitions (90° angular increments) to check whether less data would provide sufficient information as well. Subsequently, the voxel size of the ASL data was adapted according to the spatial resolution of the TOF scan.4. The dynamic ASL data is then used as mask on the TOF data, i.e. each time-frame of the now 3D ASL is mapped on the TOF image. Finally, this results in TOF images showing different states of blood filling (Figure 2). All image post-processing was performed using Matlab R2016b (The Mathworks, Natick, MA).

Results and Discussion

In all scans, the images could be successfully obtained and post-processed. Using the temporal information from ASL on the highly spatial resolved TOF images, it is possible to obtain a high-resolution time-resolved angiogram of the cerebral vasculature. Unlike the recently proposed technique, the scan time penalty is reduced to maximum 48 seconds compared to 5 minutes as it was before using an isotropic ASL dataset3. The final images appear well separated in their arterial filling. The proposed technique is not limited to non-selective imaging, but selective ASL angiography can be used as well. A source of uncertainty are the border zones between the individual filling states of the ASL data. The temporal resolution is 150ms and no cardiac trigger was used, therefore different states of filling are averaged per image acquisition and could draw a less accurate image of the vasculature, which should be explicitly investigated in future studies. The filtered back projection compiled from only 4 ASL acquisitions shows acceptable results yet lacks the delineation of smaller arteries when compared to the filtered back projection using the full ASL dataset (Figure 2). The projection approach using only 2 ASL acquisitions failed to create acceptable images (not shown).

Conclusion

By using filtered back projection data compiled from thick-slab ASL acquisitions, it is possible to map time resolution information to a TOF angiogram with only little scan time penalty (less than one minute). Having the time-resolved information about blood flow available and combined in an image with high spatial resolution, these images may add important information in the diagnosis of cerebrovascular diseases.

Acknowledgements

No acknowledgement found.