Artifact reduction in 3D radial imaging with out-of-volume saturation pulses
Jacob Macdonald1, Oliver Wieben1,2, Scott K Nagle2, and Kevin M Johnson1

1Medical Physics, University of Wisconsin - Madison, Madison, WI, United States, 2Radiology, University of Wisconsin - Madison, Madison, WI, United States

Synopsis

Streaking artifacts in radial acquisitions from undersampling or data inconsistencies can reduce SNR and make it difficult to discern features in low signal areas. Anatomy that is outside of the imaging volume of interest but within the excitation volume can contribute to these artifacts. We used out-of-volume spatial saturation pulses to suppress these streaking artifacts with minimal scan time penalties. In-vivo acquisitions with spatial saturation showed equal or superior quality in all cases. They should be implemented whenever the additional SAR can be tolerated.

Purpose

3D radial acquisition methods are capable of acquiring images in which artifacts from undersampling and data inconsistencies manifest as relatively benign, incoherent streaking artifacts in high contrast imaging scenes1. However, in cases in which quantification is desired or when the tissue of interest has naturally low signal, such as the lungs, streaking artifacts can lead to decreased apparent SNR and errors in derived parametric maps. As Figure 1 shows, a source of streaking artifacts can be anatomy outside of the imaging volume of interest, in this case the subject's arms. Artifacts are accentuated when this anatomy undergoes motion. The purpose of this pilot study is to investigate the potential for outer volume suppression to improve image quality for body imaging with 3D radial sampling.

Methods

Four healthy volunteers were imaged on a 3.0T scanner (Discovery MR750, GE Healthcare, Waukesha, WI) with a 3D ultrashort echo time (UTE) sequence2 (TR/TE=2.9/0.1ms, FA=4o, res=1.25mm isotropic, #projections=40,000, axial slab excitation). For out-of-volume saturation, two successive quadratic phase 90o saturation pulses (TR=12.6ms; BW=12.5kHz; PW=10ms) were played to the right and left of the field of view of interest (the lungs). The pulses were played every 80 TRs (257 ms), the minimum time allowable while observing SAR limits. A 32-channel, full torso coil was used for the free-breathing acquisition with respiratory gating using respiratory bellows and a gating efficiency of 50%. Each subject was scanned with and without out-of-volume saturation bands. Scan times were 3:54 (min:sec) without saturation and 4:07 with saturation. These two scans were repeated with the volunteer instructed to move their upper arms throughout the duration of the scan to simulate motion common with alert pediatric subjects. This resulted in 4 scans for each volunteer. An experienced cardiothoracic radiologist ranked the images by overall image quality for each subject from best to worst (1=best, 4=worst).

Results

Figure 2 shows the effect of out-of-volume suppression on the visibility of streaking artifacts. Figure 3 shows a table of the rankings assigned to the images for each subject by the radiologist. Figure 4 highlights a region in the lung where the effect of motion and the spatial saturation pulses was most apparent. In the coronal views of another subject (Figure 5), use of out-of-volume saturation resulted in more feature detail for the cases with movement. These cases corroborated the results of the image rankings.

Discussion

Saturation pulses were effective at reducing streaking artifacts, especially near the anterior surface of the arm. Radiologist scoring suggested that saturation did not have a substantial effect in the images without arm motion, but produced better images when motion was present. The scan time penalty of implementing these pulses was minimal, as they only increased scan time by 13 seconds (5%) in this case. While saturation proved effective at nulling the tissue and muscle, the short T1 of fat allowed for regrowth between successive saturation pulses and had considerable signal (Figure 2). This suggests either the need for fat saturation or more frequent saturation, both of which are limited by the increase in SAR. The design of SAR minimized suppression could substantially increase the efficacy of saturation, which is roughly ~60% for the 257ms spacing between pulses.

Conclusion

We have demonstrated the feasibility of using out-of-volume saturation pulses with a 3D radial UTE acquisition to reduce streaking artifacts and improve image detail in low contrast and low SNR areas, particularly when patient motion is present. These saturation pulses are worth implementing in 3D radial acquisitions if the extra SAR can be tolerated, given that they only increased scan time by 13 seconds (5%) in this case and appeared to improve image quality and detail.

Acknowledgements

We gratefully acknowledge GE Healthcare for their research support.

References

1. Block WF, Barger AV, Mistretta CA. Vastly Undersampled Isotropic Projection Imaging. Proceedings of the International Society of Magnetic Resonance in Medicine. 2000; 8: 161.

2. Johnson KM, Fain SB, Shiebler ML, Nagle S. Optimized 3D Echo Time Pulmonary MRI. Magnetic Resonance in Medicine. 2013; 70(5): 1241-1250.

Figures

Figure 1: Coronal abdominal slice of a subject displaying prominent streaking artifacts (red arrow) originating at the arms and continuing into the body.

Figure 2: Images of subject 2 moving their arms with (A) no saturation pulses and (B) saturation every 80 TRs. The logarithm was taken to enhance the appearance of streaking artifacts (red arrow). The blue arrow shows the suppression of most tissue in the arms while fat remains somewhat brighter.

Figure 3: Rankings of each acquisition technique for each subject by a trained radiologist. A ‘1’ indicates the best image quality, while a ‘4’ is the worst image quality.

Figure 4: Axial slices of the lung in subject 2 under all four imaging conditions. The red box identifies an area where effect of the saturation bands on the SNR of the lung structure is especially noticeable.

Figure 5: Coronal slices of the lung in subject 3 under all four imaging conditions. The saturation pulses appear to provide improved image quality for imaging with movement.



Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)
1831