ABSTRACT Purpose 7T pseudo‐continuous arterial spin labeling (PCASL) can benefit from combining 3D‐EPI with high‐performance background suppression (BS) pulses. However, the perfusion signal can be compromised by increased B 1 and B 0 inhomogeneity, and high SAR demands limiting the labeling duration (LD), labeling efficiency and BS inversion efficiency. Methods 7T PCASL SNR efficiency (SNReff) was quantified using a range of LD (500–4000 ms) with optimized BS inversion pulses and 3D‐EPI acquisitions. PCASL sensitivity to B 0 was also investigated by measuring through‐plane field variations along the feeding arteries and acquiring perfusion data across a range of PCASL mean gradients ( G mean , −0.8 to +0.8 mT/m) with in‐plane B 0 corrections. Results Optimized WURST BS pulses achieved ∼0.93 inversion efficiency. PCASL SNReff was maximized when employing a LD of ∼3 s (B 1pcasl = 1 T, duty cycle = 60%), providing ∼25% higher SNReff than 1 s LD typically used at 7T. Arteries with strong negative though‐plane B 0 gradients () at the labeling plane generated low perfusion signal in the corresponding arterial territories when using a G mean of 0.4 mT/m. In these cases, the perfusion signal increased ∼128% by using a higher G mean of 0.8 mT/m. Conclusion 7T PCASL SNReff can be maximized with LD of ∼3 s when using 3D‐EPI and optimized WURST BS pulses. A G mean of 0.8 mT/m improved PCASL robustness to B 0 field gradients along the feeding arteries without the need for subject specific corrections. Use of higher G mean also offers benefits in cases with multiple polarities and non‐linear field variations.
Saïb et al. (Fri,) studied this question.