Improved quantification of brain perfusion using FAIR with active suppression of superior tagging (FAIR ASST)

Abstract

Purpose: To address two problems for perfusion studies in the middle or inferior brain regions: (1) to reduce venous artifacts due to the intrinsic superior labeling of FAIR; (2) to alleviate the discrepancy of the existence of both superior and inferior boluses, but with only the inferior bolus having a temporally defined bolus width with Q2TIPs or QUIPSS.Materials and Methods: Superior tagging suppression methods for FAIR with different combinations of pre-and postinversion superior saturation pulses were evaluated and compared with FAIR with Q2TIPS for producing perfusion maps of superior, middle, and inferior brain regions.Results: One preinversion plus two postinversion superior saturation radio frequency pulses effectively suppressed the superior tagging of FAIR and sufficiently eliminated venous artifacts without negative effects, avoiding the overestimations of cerebral blood flow that can occur in FAIR.Conclusion: FAIR ASST improves FAIR with Q2TIPS and provides more reliable and accurate blood flow estimations for perfusion studies of middle and lower brain regions. FAIR ASST confers the advantages of asymmetric PASL techniques, such as PICORE, in which only the inferiorly labeled blood is used for perfusion quantification, to the symmetric PASL technique FAIR, while preserving the robustness of FAIR against MT effects. THERE ARE TWO major types of pulsed arterial spin labeling (PASL) techniques: symmetric (1-4) and asymmetric (5-7). Symmetric PASL techniques are mainly based on the flow-sensitive alternating inversion recovery (FAIR) technique (3,(8)(9)(10)(11); the symmetric application of inversion pulses relative to the center of the imaging section makes symmetric PASL techniques very robust against magnetization transfer (MT) effects. Cerebral blood flow (CBF) quantification can be achieved by using FAIR with QUIPSS (12) or Q2TIPS (13) and a single blood compartment model (14,15). In FAIR with Q2TIPS, temporal bolus width is defined by performing inferior saturations to null the tail of the labeled blood bolus.However, when performing perfusion studies in regions other than the superior part of the brain, the superior inversion band of FAIR can also label blood, which results in confounds or discrepancies between the actual boluses and the fact that only the inferior bolus is considered by Q2TIPS or QUIPSS in temporal bolus width definition for perfusion quantification using the single-subtraction blood flow quantification method (12,13). First, if there is a contribution of superiorly labeled arterial blood to tissue perfusion, the corresponding temporal bolus width should also be defined by performing superior saturations. A second confound occurs if the superiorly labeled blood contains venous blood; superior labeled venous blood can generate adverse frequency or phase errors and create severe artifacts in specific regions, such as the cerebellum, making blood flow measurements there unreliable. These difficulties also exist when using traditional FAIR-based techniques for perfusion studies of...

Quantifying Cerebellum Grey Matter and White Matter Perfusion Using Pulsed Arterial Spin Labeling

“…The cerebellum GM CBF values of 63.6 ± 5.0 mL/100 g/min estimated using a PCASL multiple inversion time experiment with 3 × 3 × 7 mm 3 voxels [14] and of 58–62 mL/100 g/min estimated using PCASL with 3.44 × 3.44 × 5 mm 3 voxels [21] agree well with the cerebellum GM CBF value of 56.7 ± 5.0 mL/100 g/min for 6 subjects using FAIR with Q2TIPS with (3.5 mm) 3 voxels [17] and the cerebellum GM CBF values of 55–65 mL/100 g/min obtained in the inferior saturation and postlabeling delay experiments of this study. PET [34], SPECT [35–37], and other ASL studies, one with 7 × 3 × 3 mm 3 voxels [20] and others with smaller voxels [16, 18, 19], have reported cerebellum CBF values of 30–48 mL/100 g/min, similar to those obtained with FAIR ASST in this study.…”

“…PET [34], SPECT [35–37], and other ASL studies, one with 7 × 3 × 3 mm 3 voxels [20] and others with smaller voxels [16, 18, 19], have reported cerebellum CBF values of 30–48 mL/100 g/min, similar to those obtained with FAIR ASST in this study. The superior saturation of FAIR ASST reliably reduces contributions of inflow of superiorly labeled blood that artifactually enhance CBF estimates [15, 17] in FAIR, and thus it is not surprising that the cerebellar GM CBF of 43.8 ± 5.1 mL/100 g/min obtained using FAIR ASST with 2.5 × 2.5 × 3.5 mm 3 voxels in this study is lower than those obtained using traditional FAIR [17]. GM CBF measurements by FAIR ASST are comparable to, although insignificantly and slightly lower than, the literature values from PCASL [14, 20–22]; these slight differences of CBF measurements between FAIR ASST and PCASL can possibly be attributed to the use of different ASL techniques as well as CBF quantification models [22].…”

“…GM CBF measurements by FAIR ASST are comparable to, although insignificantly and slightly lower than, the literature values from PCASL [14, 20–22]; these slight differences of CBF measurements between FAIR ASST and PCASL can possibly be attributed to the use of different ASL techniques as well as CBF quantification models [22]. The cerebellar WM CBF of 27.6 ± 4.5 mL/100 g/min from this FAIR ASST study is also lower than the cerebellar WM CBF of 36.7 ± 2.7 mL/100 g/min from FAIR with Q2TIPS [17]. It should be noted that our data indicate volume averaging of GM and WM in some voxels (Figure 8(c)); if this were corrected [38], our average GM CBF values would be somewhat greater and our average WM CBF values would be slightly smaller.…”

Measuring renal tissue relaxation times at 7 T

“…To achieve uniform inversion across and within an imaging slice, the inversion was performed using a slice-selective adiabatic hyperbolic secant RF pulse with an inversion slab that was twice the imaging slice thickness. At 3T, the inversion was achieved with a 15.36 ms hyperbolic secant pulse with 3 kHz bandwidth (36); at 7T an HS4 pulse with 20 ms duration and 1 kHz bandwidth was used (37). T 2 mapping was performed by inserting a Carr-Purcell-Meiboom-Gill (CPMG) style refocusing pulse train (38,39) after the excitation and prior to ss-FSE readout, and varying number of echoes in steps of two (CMPG-prep ss-FSE).…”

Modeling magnetization transfer effects of Q2TIPS bolus saturation in multi-TI pulsed arterial spin labeling

“…For better slice-profile delineation of the bolus saturation pulse and more effective bolus suppression, several thin-profile slices are usually used repetitively on the distal edge of the bolus (Q2TIPS) instead of a single wide saturation pulse (12). Alternative saturation schemes include saturation of regions both superior and inferior of the imaged plane (13) or applying saturation pulses of varying thickness (14). These bolus saturation pulses are applied during, both, control and labeling phase.…”