MR-based radiation therapy treatment planning (RTP) may offer a number of benefits over the existing, CT-based standard of care, such as the lack of ionizing radiation, and superior soft-tissue contrast for improved target and at-risk organ delineation. Many challenges remain, however, with motion management identified as one of the top needs for MR-based RTP, especially in the chest and abdomen.¹ Various approaches to characterizing respiratory motion with MRI have been proposed, however this remains an active area of research.^2,3 The purpose of this work was to develop and test a multi-slice multi-phase single-shot fast spin echo (SSFSE, or HASTE, SSH-TSE, etc.) sequence incorporating variable refocusing flip angle (vrfSSFSE) and an interleaved cylindrical navigator, for rapid, T2-weighted imaging with retrospective respiratory compensation.A vrfSSFSE pulse sequence was modified to support ungated multi-slice multi-phase imaging and the acquisition of a single, cylindrical navigator profile between imaging shots (Figure 1). This navigator profile was collected as a retrospective reference only, and not utilized for gating. Phantom and volunteer scanning was performed on a wide-bore 3.0T MRI scanner (MR750w, GE Healthcare, Waukesha, WI). T2-weighted images of a phantom undergoing periodic motion in the S/I direction were acquired with the multi-slice multi-phase vrfSSFSE pulse sequence and the following imaging parameters: coronal plane, 3 slices, 40 phases, TR = 677 ms (1.48 frames per second), TE = 80 ms, FOV = 40 x 36 cm, 5 mm slice, 256 x 224 matrix, ±125 kHz bandwidth, 0.64 effective NEX, 2x acceleration (ARC, GE Healthcare, Waukesha, WI), and refocusing flip angle targets⁴ of 130, 90, 100 and 45 degrees, respectively. The navigator profile was placed in the S/I direction across the edge of the phantom. Following informed consent, T2-weighted volunteer images were acquired for 20 slices and 20 phases with the same parameters listed above, and with slightly modified parameters, as follows: full-NEX, 3x ARC acceleration, TR = 458 ms (2.18 frames per second), TE = 95 ms, and a minimum refocusing flip target of 60°. (Note, the imaging TR includes both the length of the navigator and vrfSSFSE waveforms.) In all cases, the acquisition order was interleaved for both slice and phase to allow maximum recovery time between subsequent acquisitions of the same slice. For in vivo imaging, the cylindrical navigator was placed across the dome of the liver to record its position before each imaging shot. The displacements measured at the navigator TE were interpolated via cubic spline to each respective vrfSSFSE image TE, and then used to retrospectively sort the images by S/I position.Phantom imaging results are summarized in Figure 1, which shows the navigator-derived position data through time (1a), the same position data, retrospectively sorted by image location and position (1b). The retrospectively sorted phantom image data for a single, representative image location is shown in Figure 2, where the green line represents the phantom’s maximum excursion in the S-direction. In vivo results are summarized in Figure 3, showing a composite image of all phases for a single slice acquired with a minimum refocusing flip angle of 90° (3a), and all phases of the same slice acquired with a minimum refocusing flip angle of 60° (3b). Figure 4 includes two examples of typical, intermittent image shading due to cardiac motion for minimum refocusing flip angles of 90° vs. 60°, respectively.These results demonstrate the performance of a dynamic, T2-weighted vrfSSFSE pulse sequence with an integrated cylindrical reference navigator for retrospective respiratory compensation. As previously reported⁵, the reduced SAR associated with vrfSSFSE cuts the minumum TR by an approximate factor of 2, which allows higher frame rates for multi-phase imaging than could otherwise could be safely achieved with SSFSE at 3.0T. The full-Fourier acquisition has been shown to provide improved sharpness and image SNR⁵, though for the relatively low-resolution acquisition tested here, the sole apparent advantage of the more aggressive vrfSSFSE protocol (60° minimum flip) is in the nearly 50% higher frame rate. The shortened TRs available with vrfSSFSE are achieved at the expense of increased motion sensitivity.⁶ As seen in the full-Fourier acquisition, the liver signal is less consistent due to increased susceptibility to cardiac motion. In the context of a redundant, multi-phase acquisition, however, the opportunity to reject such image artifacts is greater than for routine, anatomical imaging.We have demonstrated the feasibility of a rapid, dynamic, T2-weighted vrfSSFSE sequence for navigator-guided retrospective respiratory compensation. We believe this is one approach to motion management in MRI that may prove to be useful in the context of MR-guided radiation therapy treatment planning.