MRI offers a number of well-documented advantages over CT, including superior soft-tissue contrast and the lack of ionizing radiation. The need to characterize patient respiratory motion is an important aspect of radiation therapy treatment planning (RTP), where MRI has the potential to exceed the performance of 4D CT, the current standard of care. Possible advantages include target identification, depiction of organs-at-risk, the ability to image for longer periods of time and/or more frequently, and ultimately, better outcomes for patients.¹ One approach to 4D MRI (vs. 4D CT), is to generate a time-resolved 3D volume via a multi-slice, multi-phase MR acquisition, where the images are retrospectively sorted by respiratory phase.² Here, we compare two dynamic, multi-slice pulse sequences, each with navigator-based retrospective respiratory compensation.A cylindrical navigator pulse was added to both single-shot balanced-SSFP and single-shot fast spin echo (aka SSFSE, or HASTE, SSH-TSE, etc.) sequences capable of multi-phase imaging. Additionally, the SSFSE pulse sequence incorporated variable refocusing flip angle (vrf), useful in this context for reduced repetition times (due to lower SAR), and therefore higher frame rates. Phantom and volunteer scanning was performed on a wide-bore 3.0T MRI scanner (MR750w, GE Healthcare, Waukesha, WI). Following informed consent, volunteer images were acquired for both single-shot sequences. Imaging was performed in the coronal plane for 20, 5-mm slices, 20 phases per slice, FOV = 40 x 36 cm, 2x parallel imaging acceleration (ARC, GE Healthcare, Waukesha, WI). Additional imaging parameters for the single-shot bSSFP acquisition were TR/TE = 3.7/1.65 ms, frames per second = 1.57, 35° flip angle, 224 x 256 matrix, ±125 kHz bandwidth, and 1.0 NEX. Additional imaging parameters for the vrfSSFSE acquisition were TR/TE = 642/80 ms, frames per second = 1.56, 256 x 224 matrix, ±125 kHz bandwidth, 0.64 effective NEX, and refocusing flip angle targets³ of 130, 90, 100 and 45 degrees, respectively. Total (arbitrary) acquisition time for both sequences was approximately 4 minutes and 15 seconds for 400 total images. The navigator profiles were used retrospectively to measure the position of the liver dome between imaging shots, which was then used to sort the images retrospectively, based on slice location and respiratory phase.Figures 1 and 2 are composite images, showing all acquired phases for a single slice location, retrospectively sorted by respiratory phase, for both the single-shot bSSFP and vrfSSFSE in vivo acquisitions. The bSSFP images (Figure 1) demonstrate high signal-to-noise, bright-fluid T2/T1-weighted contrast, robustness to motion, and expected off-resonance effects, including banding and fat-water cancelation. The vrfSSFSE images (Figure 2) demonstrate T2-weighted contrast, robustness to off-resonance, a modest (intrinsic) black-blood effect due to the flip angle modulation, and lower relative SNR. As intended, in both cases, the respiratory phase is successfully represented by a monotonic displacement of the diaphragm from end inspiration to end expiration. Pitfalls of both sequences are illustrated in Figure 3, showing typical off-resonance effects encountered with bSSFP and shading due to cardiac motion encountered with vrfSSFSE.⁴In spite of their differences, both single-shot acquisitions demonstrate comparable effectiveness in their ability to thoroughly sample the respiratory cycle and to satisfy the basic requirements for respiratory-guided 4D MR without the need for respiratory bellows or associated reliability concerns. Balanced SSFP is already commonly used in multi-phase imaging, generally prized because of its steady state properties. In this case, the reduced SAR of the vrfSSFSE sequence, however, allows it to match the relatively high frame rate of the bSSFP acquisition (of approximately 1.5 fps).⁵ Additional strengths of the vrfSSFSE acquisition are its familiar T2-weighted contrast, and robustness to off-resonance effects, which may translate to a relative advantage in terms of target and at-risk organ delineation over large fields of view. Relative to CT, both sequences exhibit superior soft-tissue contrast, and the total (arbitrary) acquisition time of approximately 4 minutes, represents an unacceptable exposure time for CT. This long scan time may translate into a valuable advantage for MR, via a more thorough characterization of patient respiratory motion over time than CT can safely achieve. Finally, image artifacts common to both sequences remain a concern, though they may be mitigated to some extent with well-established techniques such as phase cycling, cardiac gating, and/or retrospective rejection of corrupted images.We believe these single-shot sequences offer a straightforward approach to generating 4D MRI data for MR-guided RTP. In spite of their differences, we believe both sequences offer sufficient soft-tissue contrast, with each exhibiting its own set of unique characteristics. Future work will include the evaluation of these techniques in a clinical radiation therapy treatment planning setting.