Fat suppression using water excitation for improved spatial resolution compared with 2-point Dixon at 3.0 T for DCE breast MRI
Courtney K Morrison1, Leah C Henze Bancroft1, Kang Wang2, James H Holmes2, Frank R Korosec1,3, and Roberta M Strigel1,3

1Medical Physics, University of Wisconsin-Madison, Madison, WI, United States, 2Global MR Applications and Workflow, GE Healthcare, Madison, WI, United States, 3Radiology, University of Wisconsin-Madison, Madison, WI, United States

Synopsis

Fat suppression can be achieved in a variety of ways, including using water-only excitation or using a Dixon method. Each of these methods exhibits strengths and limitations. In this work, we evaluated the characteristics of water excitation compared to a 2-point Dixon fat suppression method used in high spatiotemporal resolution DCE breast MRI.

Introduction

Several techniques have been proposed to improve the temporal resolution of dynamic contrast-enhance (DCE) breast MRI while maintaining the clinically-necessary high spatial resolution.1-4 One such technique is DIfferential Subsampling with Cartesian Ordering (DISCO) which most commonly uses a 2-point Dixon method for removal of fat signal.5 However, the echo spacing at 3.0 T required for the Dixon method limits the number of readout points that can be acquired in each echo and thus limits the achievable spatial resolution. Fat suppression can also be achieved using a spatial-spectral radiofrequency pulse that excites only water. This technique allows for higher spatial resolution but is more affected by B0 and B1 inhomogeneities. Given the potential for trade-offs between these methods, in this work we investigated the characteristics of these two types of fat suppression when used with DISCO at 3.0 T for DCE breast MRI.

Methods

DISCO uses a pseudorandom k-space sampling scheme that fully samples the center of k-space but undersamples the periphery of k-space in each time frame. Images are then reconstructed by view sharing peripheral k-space data from nearby frames. This sampling scheme was incorporated into a spoiled gradient echo sequence. Fat suppression was achieved in one of two ways: with a spatial-spectral radiofrequency pulse (water excitation) or with 2-point Dixon fat-water separation (Dixon).6 Twenty-five patients undergoing clinical breast MRI were recruited for this study. Patients were scanned on a 3.0 T scanner (Discovery MR 750w, GE Healthcare, Waukesha, WI) using a 16-channel breast coil (Sentinelle, Invivo, Gainsville, FL). Clinical DCE MRI was acquired after injection of 0.1 mmol/kg of gadobenate dimeglumine (MultiHance, Bracco Inc., Milan, Italy). A single, fully-sampled, mask phase of both DISCO water excitation and DISCO-Dixon were acquired immediately following the clinical DCE MRI. The order of the water excitation and Dixon sequences was randomized. DISCO-Dixon was acquired with the highest in-plane resolution achievable for the selected imaging parameters (1.0 x 1.0 mm2, limited by echo spacing) while water excitation was acquired at a clinically-desired in-plane resolution (0.8 x 0.8 mm2). Other imaging parameters were: FOV = 32 x 32 cm2, 124 slices with thickness of 1.6 mm, flip angle = 10 degrees, ARC parallel imaging factor of 3 x 1, bandwidth = ±83.33 kHz for water excitation and ±166.67 kHz for DISCO-Dixon. The acquisition time for the fully-sampled mask was 1:17 for water excitation and 1:19 for DISCO-Dixon. Both sets of DISCO images were scored by three radiologists with 1, 6, and 22 years of experience for diagnostic image quality (1-non-diagnostic, 3-adequate, 5-excellent diagnostic quality) in the following categories: depiction of anatomy, degree of fat suppression, uniformity of fat suppression, water signal shading, and overall diagnostic confidence. Radiologists also ranked their preference for the images based on overall image quality and the quality of fat suppression.

Results

A representative patient case is shown in Figure 1. In 24 of 25 cases, both DISCO image sets were scored as providing good diagnostic confidence with mean values of 4.0 and 4.5 for water excitation and Dixon, respectively (Figure 2). The mean scores for the depiction of anatomy, degree of fat suppression, uniformity of fat suppression, and water signal shading were 4.2, 3.9, 3.4, and 4.0 for water excitation and 4.4, 4.8, 4.6, and 4.4 for Dixon (Figure 3). All radiologists scored one technically challenging case poorly for both sequences; when compared, this case received a similar poor assessment of the standard clinical images. Reader 1 (6 years experience) preferred water excitation to Dixon in 15/25 cases for overall image quality and in 11/25 cases for fat suppression. Reader 2 (22 years experience) preferred water excitation to Dixon in 8/24 cases and scored one case as equivalent for overall image quality and preferred Dixon for fat suppression in all cases. Reader 3 (1 year experience) preferred Dixon for overall image quality and fat suppression for all 25 cases.

Conclusion and Discussion

DISCO water excitation allowed for higher spatial resolution whereas DISCO-Dixon provided more uniform and a higher degree of fat suppression; however sequence preference varied between readers and both sequences provided diagnostic image quality. This study focused on the assessment of different fat suppression techniques at 3.0 T with DISCO to ensure that the image quality and diagnostic confidence of the current standard-of-care DCE breast MRI method was maintained; however these techniques have other possible advantages that were not included in this evaluation, including higher temporal resolution and fat images generated using the Dixon technique.

Acknowledgements

Acknowledgements: We would like to thank Drs. Frederick Kelcz and JulieAnn Stover for participating in the reader study. We are grateful for support from our institutional departmental R&D Fund, the Radiological Society of North America, NIH (T32CA009206), and GE Healthcare.

References

1. Dougherty L, Isaac G, Rosen MA, et al. High frame-rate simultaneous bilateral breast DCE-MRI. Magn Reson Med 2007;57(1):220-225.

2. Ramsay E, Causer P, Hill K. Plewes D. Adaptive bilateral breast MRI using projection reconstruction time-resolved imaging of contrast kinetics. J Magn Reson Imaging 2006;24(3):617-624.

3. Saranathan M, Rettman DW, Hargreaves BA, Lipson JA, Daniel BL. Variable spatiotemporal resolution three-dimensional Dixon sequence for rapid dynamic contrast-enhanced breast MRI. J Magn Reson Imaging 2014;40(6):1392-9.

4. Hermann KH, Baltzer PA, Dietzel M, et al. Resolving arterial phase and temporal enhancement characteristics in DCE MRM at high spatial resolution with TWIST acquisition. J Magn Reson Imaging 2011;34:973-982.

5. Saranathan M, Rettman DW, Hargreaves BA, Clarke SE, Vasanawala SS. DIifferential Subsampling with Cartesian Ordering (DISCO): a high spatiotemporal resolution Dixon imaging sequence for multiphasic contrast enhanced abdominal imaging. J Magn Reson Imaging 2012;35(60:1484-1492.

6. Ma J. Breath-hold water and fat imaing using a dual-echo two-point Dixon technique with an efficient and robust phase-correction algorithm. Magn Reson Med 2004;52(2):415-419.

Figures

Figure 1. Shown are post-contrast, T1-weighted, axial images of the bilateral breasts from a representative patient acquired using DISCO water excitation (top) and DISCO-Dixon (bottom). The DISCO-Dixon image had a higher degree of fat suppression, but fine details were better depicted in the DISCO water excitation image.

Figure 2. Pooled diagnostic confidence scores are shown for DISCO water excitation and DISCO-Dixon. Both sequences provided good diagnostic confidence, indicated by the median (black lines) and mean (dotted lines with diamonds) scores. There was one outlier (dots) that all three readers scored as “limited diagnostic confidence” for water excitation.

Figure 3. Pooled image quality scores from all three readers are shown for DISCO water excitation and DISCO-Dixon. Median (black lines), mean (dotted lines with diamonds), and outliers (dots) are indicated. The lowest scores are from a technically challenging case with a similarly poor assessment of the clinical MRI.



Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)
2720