0135

Fat Suppression in UTE Imaging of Short T2 Tissues Using a Novel Soft-hard Composite RF Pulse
Ya-Jun Ma1, Saeed Jerban1, Hyungseok jang1, Eric Y Chang1,2, and Jiang Du1

1Radiology, University of California, San Diego, San Diego, CA, United States, 2VA San Diego Healthcare System, San Diego, CA, United States

Synopsis

Fat suppression is very important for both high contrast morphological imaging and accurate quantitative MR imaging. However, the conventional fat suppression methods such as chemical shift-based fat saturation are not well-suited for short T2 imaging due to the large direct saturation of short T2 tissues with broad spectra. The purpose of this study was to design a novel fat suppression pulse for ultrashort echo time (UTE) imaging of short T2 tissues with well-preserved short T2 signals using a soft-hard composite pulse.

Introduction

Ultrashort echo time (UTE) imaging with TE less than 100µs has been widely used in imaging short T2 tissue, such as the calcified cartilage, tendons, ligaments, menisci, and bone1,2. However, UTE imaging of the short T2 tissues always suffers from fat contamination due to partial volume effect and off-resonance fat artifacts induced by non-Cartesian UTE acquisition2. Therefore, fat suppression is very important for both high contrast morphological imaging and accurate quantitative UTE imaging of short T2s. In this study, we designed a novel fat suppression pulse for short T2 imaging with well-preserved short T2 signals using a soft-hard composite pulse.

Methods

The newly designed fat suppression pulse consists of one soft pulse and one hard pulse (Figure 1A). The soft pulse, centered on fat frequency with a negative flip angle and designed with a narrow bandwidth, is used to flip only fat magnetization, and is then followed by a short hard pulse with a positive flip angle which flips all magnetizations in the opposite direction. Since fat magnetization experiences both tipping down and tipping back in same flip angles, most of fat magnetizations are back to the longitudinal state. Thus, most of fat signals are subsequently not received by following UTE acquisitions. In addition, the soft pulse has been designed with a narrow bandwidth, so the soft pulse excitation has little effect on the water magnetizations. This makes it possible for the water magnetizations to be effectively excited by the hard pulse. Compared with commonly used fat saturation (FatSat) module with a flip angle (FA) of no less than 90° (Figure 1B), the FA of the soft pulse is typically much lower (i.e. less than 30°). Therefore, both direct and indirect saturation (magnetization transfer effect) of water signals for the proposed soft-hard pulse are much lower than for the FatSat module.

Bloch simulation was performed to compare the excited magnetizations resulting from a single hard pulse, from the proposed soft-hard composite pulse, and from the conventional FatSat for both short and long T2 tissues with T2s of 20,5,2,1,0.5 and 0.3ms, respectively. Pulse parameters are as follows: 1) the soft-hard pulse: duration of the soft pulse, bandwidth, and center frequency are 4.4ms, 500Hz, and -440Hz, duration of the hard pulse is 50µs; 2) the conventional FatSat module: duration of the FatSat pulse, bandwidth, center frequency, and flip angle are 8ms, 500Hz, -440Hz, and 90°. The excitation FAs are all 10° for simulation.

Then, in-vivo knee and tibial imaging were performed to compare both fat suppression and water saturation for the two fat suppression methods listed above. To count for fat suppression and water saturation, a signal suppression ratio (unit in percentage) was defined by the division of the subtracted image between non-fat suppression image and fat suppression image by the non-fat suppression image. A 3D UTE-Cones sequence was used for data acquisition3.

Sequence parameters for UTE-Cones imaging of the knee joint of a healthy volunteer (34 years old male) are shown as follows: FOV=15×15×9.6cm3,matrix=256×256×32,TE=0.032µs,FA=5°,TR=20ms, and scan time=3min36s. In vivo tibial imaging sequence parameters are shown as follows: FOV=12×12×16cm3,matrix=192×192×32,TE=0.032µs,FA=5°,TR=20ms, and scan time=2min45s. Dual-echo tibial data (TE=0/2.2,0.3/4.4,0.6/6.6,0.9/8.8,1.4/11ms) were also acquired for bi-component T2* quantification. Informed consent was obtained from all subjects in accordance with guidelines of the institutional review board.

Results and Discussion

Simulation results (Figure 2) demonstrated that the excitation induced by the proposed soft-hard pulse approaches that of the standard single pulse when T2 is getting shorter (i.e., the side lobes in the two sides of fat spectrum are getting smaller). Even though there are larger side lobes for longer T2 tissues, they have negligible effect on the water peak due to the narrower water spectrum. Therefore, the proposed soft-hard composite pulse has very little effect on the water signal excitation. The conventional FatSat module has better fat suppression than the soft-hard pulse. However, there are much stronger direct saturations on water using this method, especially for shorter T2 tissues.

In vivo knee and tibial UTE imaging (Figures 3-4) demonstrated excellent fat suppression for both the proposed soft-hard excitation and the product FatSat module. However, the short T2 signals for patellar tendon, PCL, menisci, and bone were much better preserved in soft-hard excitation images compared with images with FatSat. In addition, the long T2 muscle signals are also highly saturated by the FatSat module, which is probably due to the magnetization transfer effect.

As shown in Figure 5, much better bi-component fitting was obtained from the images with soft-hard excitation as compared to the non-fat suppression images.

Conclusion

The proposed soft-hard composite pulse can effectively suppress fat signal and preserve signals for both short and long T2 tissues.

Acknowledgements

The authors acknowledge grant support from GE Healthcare, NIH (1R21AR073496, R01AR068987, R01AR062581), and the VA Clinical Science and Rehabilitation R&D Awards (I01CX001388 and I01RX002604).

References

1. Du J, Carl M, Bydder M, Takahashi A, Chung CB, Bydder GM. Qualitative and quantitative ultrashort echo time (UTE) imaging of cortical bone. J Magn Reson 2010;207(2):304-11.

2. Robson MD, Gatehouse PD, Bydder M, Bydder GM. Magnetic resonance: an introduction to ultrashort TE (UTE) imaging. J Comput Assist Tomogr 2003;27:825-846.

3. Gurney PT, Hargreaves BA, Nishimura DG. Design and analysis of a practical 3D cones trajectory. Magn. Reson. Med. 2006;55:575–582.

Figures

Figure 1 The proposed soft-hard composite water excitation pulse (A) and conventional FatSat module (B). A soft RF pulse centered on fat frequency with a negative flip angle was used to flip the fat magnetization only, followed by a short hard pulse with a positive flip angle to flip all the magnetizations in the opposite direction (A). Since the fat magnetizations experienced both tipping down and tipping back, most of the fat magnetizations were not excited. On the other hand, the water magnetizations were effectively excited by the hard pulse. The commonly used FatSat module was shown in (B) for comparison.

Figure 2 Bloch simulations for the excitations of a single hard pulse (blue curves), the proposed soft-hard composite pulse (red curves), and the conventional FatSat module (yellow curves). Both transversal (A-F) and longitudinal (G-L) magnetizations were calculated for tissues with different T2s of 20, 5, 2, 1, 0.5, and 0.3ms. Much lower saturation effect was observed for the proposed soft-hard composite pulse compared with the FatSat module.

Figure 3 In vivo knee imaging with excitations of a single hard pulse (A-C), the proposed soft-hard water excitation pulse (D-F), and the conventional FatSat module (G-I). Fat signals were well suppressed by both the proposed soft-hard composite pulse and the FatSat pulse. However, the short T2 signals (blue arrows in D-F) were much better preserved in the water excitation images (D-F) compared with images from FatSat (G-I). Furthermore, muscle, ligament, and tendon signals were much better preserved, as can be seen in the signal suppression ratio images of water excitation (J-L) compared with the ratio from FatSat (M-O).

Figure 4 In vivo tibial imaging with excitations of a single hard pulse (A-C), the proposed soft-hard water excitation pulse (D-F), and the conventional FatSat module (G-I). Fat signals were well suppressed by both the proposed soft-hard composite pulse and the FatSat pulse. However, the short T2 bone signal was much better preserved in the water excitation images (D-F) compared with images from FatSat (G-I). Muscle and bone signals were much better preserved as can be seen in the signal suppression ratio images of water excitation (J-L) compared with the ratio of FatSat (M-O).

Figure 5 Quantitative T2* measurement for in vivo tibial bone imaging with the proposed soft-hard composite excitation (A and C) and a single hard pulse excitation without fat suppression (B and D). Selected tibial images with a TE of 0.032 ms are shown in A and B. The marrow fat was well suppressed with the proposed water excitation pulse (A) compared with the non-fat suppression image (B). Much better bi-component fitting was obtained from the soft-hard water excitation images compared with the non-fat suppression images.

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)
0135