Ultrafast spin-echo sequences such as Turbo Spin Echo (TSE) and Half Fourier Acquisition Single Shot Turbo Spin Echo (HASTE) have found wide-spread application. However, these sequences have high RF power deposition which particularly limits their use at high resolutions and high field strengths. Implementing fixed SAR Power Independent of Number of Slices (PINS)-refocusing pulses¹ has been shown to reduce the power deposition to an acceptable level in the closely related Turbo Spin Echo (TSE) sequence^2,3. However, when PINS pulses are used for both excitation and refocusing, slice orientations are limited by the periodic and infinite nature of the PINS slice profile². The use of a multiband (MB) pulse⁴ for excitation and a PINS pulse for refocusing retains the benefits of PINS for reduced power deposition while enabling arbitrary user-defined slice orientations. Multiband excitation pulses have the added benefit of accelerating the acquisition due to the simultaneous excitation of several slices. Blipped CAIPIRINHA, which shifts the simultaneously acquired slices according to their z position, was included in the sequence to improve the quality of MB reconstruction⁵, and TRAPS⁶ was implemented to allow further reductions in power deposition and effective TE.

Standard MB excitation pulses and PINS refocusing pulses were incorporated into the standard Siemens HASTE sequence. In order to reconstruct the images on the scanner, an integrated FLASH reference scan was incorporated into the beginning of the sequence. Images were acquired on a Siemens 3T Prisma system from a healthy volunteer after obtaining written informed consent.

A MB4-HASTE scan was acquired with 3.1 mm resolution in 5 seconds with a 10 s reference scan at 100% SAR (10% SAR per slice): 40 axial slices, 20% slice gap, TR 660ms, TE 61ms, 3.1x3.1x3.0mm resolution, 64x64 matrix, partial Fourier 5/8, flip angle 162 degrees, TRAPS, multiband factor of 4, CAIPI 2. The standard HASTE sequence with the same parameters required 28s and ran at 63% SAR.

A second MB4-HASTE scan was acquired with 1.1 mm resolution in 120s with a 31 s reference scan at 55% SAR ( 3% SAR per slice): 72 axial slices, 20% slice gap, TR 3290, TE 74ms, 1.1 mm isotropic resolution, 192x192 matrix, partial Fourier 5/8, flip angle 120 degrees, TRAPS, multiband factor of 4, CAIPI 2. The standard HASTE sequence with the same parameters required 3 min 58s and ran at 24% SAR.

A MB6-HASTE full-brain scan was acquired with a multiband factor of 6 and 1mm isotropic resolution in 85 seconds with 100% SAR (20% SAR per slice): 120 axial slices, 30% slice gap, TR 4320 ms, TE 99ms, 1.1x1.1x0.8 mm resolution, 192x192 matrix, partial Fourier 5/8, flip angle 120 degrees, TRAPS, multiband factor 6, CAIPI 3. The standard HASTE sequence with the same parameters required 8min 40 s and ran at 25% SAR.

MB4-PINS HASTE shows good image quality which is comparable to a standard HASTE acquisition in both a faster (5s acquisition time) version and a higher resolution (1.1mm isotropic spatial resolution) version (figure 1). MB6-PINS also shows good image quality for a full-brain 1.1mm isotropic resolution protocol (figure 2). These images demonstrate that HASTE scans can be reconstructed using FLASH reference data without significant artifacts or leakage between individual slices. The slightly different contrast observed in the MB4-PINS HASTE compared to the standard HASTE at 3 T seen in figure 1 may be due to differing levels of magnetization transfer². Higher MB factors should be attainable with further protocol optimization, allowing better separation of aliased slices and therefore closer slice spacing. Additionally, the “faster” and “higher resolution” protocols shown here can be further optimized for the desired application by adjusting parameters such as MB factor, CAIPIRINHA shift, GRAPPA factor, ETL and resolution. This would enable the MB-PINS HASTE sequence to be used, on the one hand, for anatomical T2-imaging, and on the other hand, potentially for fMRI experiments.