Blood flow measurements using MRI and Arterial Spin Labeling: a comparison with radioactive microspheres

Aim:

Arterial Spin Labeling (ASL) is a Magnetic Resonance Imaging (MRI) technique for quantitative blood flow measurements. While the principal validity of the technique has been shown, e.g. for human brain investigations, its practical applicability and accuracy depends very sensitively on the specific experimental setting. The purpose of this work was the evaluation of ASL for perfusion measurements in the rat brain by a comparison with microsphere derived regional perfusion information using dedicated small animal PET and SPECT systems.

Methods:

MRI measurements were performed first, immediately followed by the microsphere measurements. Before measurements, catheters were implanted through the right carotid artery into the left ventricle of the heart for administration of radio-labeled microspheres (20). The in-vivo distribution of radio-labeled microspheres was evaluated by PET (microPET P4, Siemens) using Cu-64 and Ga-68 microspheres. For SPECT (NanoSPECT, Bioscan) measurements Tc-99m microspheres were used. MRI perfusion measurements were performed in a 7T small animal system (BioSpec 70/30, BRUKER) with the vendor provided ASL protocol using a FAIR (flow-sensitive alternating inversion recovery) sequence with an adiabatic hyperbolic secant inversion pulse (length-bandwidth product: 80). The global and selective T1 images were used for calculation of perfusion values.

Results:

For normal rat brain (without catheter) we measured perfusion values using FAIR-ASL ranging from 1.2 to 1.4 ml/min/g in the caudate putamen. The implantation of the catheter created differences in the perfusion between the right and left hemisphere of the brain (due to the partial blocking of the right carotid artery), which are apparent from the left/right differences in the microsphere distribution. These differences are visible in the ASL-derived perfusion as well, ranging from 25 - 60%. ASL-derived perfusion exhibits substantial inter- and intra-individual variability, the cause of which is currently under investigation.

Conclusions:
Quantitative perfusion measurements in the rat brain using ASL seem possible but are very susceptible to minor deviations from the optimal setup (e.g. concerning shimming of the magnetic field and motion artifacts). Overall regional contrast is on average concordant with regional distribution of microspheres. In order to be useful for routine application in small animal imaging, ASL data acquisition and data evaluation needs to be further optimized. A final calibration via a quantitative comparison with radio-labeled microspheres seems mandatory.