Multi-shot Diffusion-Weighted PROPELLER MRI of the Abdomen

J. Deng, A. Stemmer, F. H. Miller, T. K. Rhee, R. Salem, D. Li, R. A. Omary, A. C. Larson Department of Radiology, Northwestern University, Chicago, IL, United States, Department of Biomedical Engineering, Northwestern University, Chicago, IL, United States, Siemens Medical Solutions, Erlangen, Germany Introduction: Diffusion-weighted imaging (DWI) techniques use water mobility as an exogenous probe for non-invasive interrogation of microstructural tissue properties. Singleshot DW-EPI techniques are routinely used for neuroimaging applications due to relative insensitivity to bulk motion artifacts. However, these single-shot techniques can suffer significant image distortion, chemical shift artifacts, and reduced spatial resolution particularly when extending the imaging field-of-view (FOV) as necessary for abdominal imaging applications. These limitations have significantly complicated routine clinical DWI of the visceral organs. The recently introduced DWPROPELLER strategy offers the potential to overcome these limitations [1,2]. The PROPELLER sequence uses a multi-shot acquisition strategy while permitting segmental phase correction to reduce bulk motion artifacts. In this study, we evaluated the feasibility of using the multi-shot DW-PROPELLER sequence for diffusionweighted imaging of the abdomen. We tested the hypothesis that DW-PROPELLER provides accurate quantitative diffusion measurements while improving qualitative image sharpness, distortion, and artifact levels compared to single-shot DW-EPI. Methods: The PROPELLER technique uses a multi-shot turbo-spin echo (TSE) acquisition strategy with each segment of data acquired as a single rectilinear blade along a ‘propeller-shaped’ k-space trajectory. From each k-space blade, a low-resolution image is reconstructed permitting phase correction of motion artifacts. Following data correction, k-space blade segments are combined using k-space regridding for high resolution image reconstruction. Our implemented pulse sequence was based upon the BLADE pulse sequence (Siemens Medical Solutions implementation of PROPELLER TSE). Motion-probing gradients separated by a slice-selective 180° refocusing pulse provided the requisite diffusion-weighing. Phantom experiments were performed to test the accuracy of DW-PROPELLER quantitative diffusion measurements (three cylindrical vials consisting of distilled water, acetone and ethanol at room temperature). For abdominal imaging studies, DW-PROPELLER image series were compared to corresponding single-shot spin-echo EPI (SS-SE-EPI) diffusion-weighted image series. All imaging experiments were performed using the Magnetom Sonata 1.5T clinical MR scanner (Siemens Medical Solutions, Erlangen, Germany) with highperformance gradient subsystem (40 mT/m maximum amplitude; 200 mT/m/ms maximum slew rate) and flexible six-channel phase-array abdominal imaging coil. Eight healthy volunteers were recruited for this study according to IRB protocol. Volunteer studies (N=8) were performed at a single axial slice position including both liver and pancreatic tissues with slice position chosen based upon T2-weighted TSE scout scans. Both DW-PROPELLER and SS-SE-EPI diffusion-weighted image series were acquired for each volunteer study with the following imaging parameters. DW-PROPELLER: TR/TE = 2000/105ms, 5mm slice-thickness, 400 Hz/pixel BW, 380x380 mm FOV, 128x128 matrix (3.0x3.0x5.0 mm), ETL = 17, 44 segments, free-breathing with respiratory bellows triggering. SS-SE-EPI: TR/TE = 2000/82ms, 5mm slice thickness, 1.5 kHz/pixel BW, 380x260 mm FOV, 70x128 matrix (3.7x3.0x5.0 mm), 6/8 partial Fourier, EPI factor = 70. Non-selective fat saturation and regional superior and inferior pre-saturation slabs were used for both sequences. At a fixed diffusion time, motion probing gradient amplitudes were varied to obtain diffusion weighting of b~0, 10,100 and 502 s/mm. For both sequences, separate image series were acquired with diffusion weighting applied along one of three orthogonal axis defined as (-0.5x, y, z), (x, -0.5y, z) and (x, y, -0.5z). Using DW images with b~10, 100 and 502 s/mm, apparent diffusion coefficient (ADC) maps were separately reconstructed for each direction of diffusion weighting and then averaged to obtain the isotropic diffusion coefficient Dtrace. In each volunteer additional series of DW-PROPELLER images, 192x192 matrix (2.0x2.0x5.0 mm), were acquired to demonstrate the feasibility of improving spatial resolution. Both qualitative and quantitative evaluations were performed to compare DW-SS-SE-EPI and DW-PROPELLER sequences. Isotropic diffusion-weighted images (b~0 and 502 s/mm) and reconstructed ADC images were randomized and qualitatively scored separately by a clinician blinded to the particular acquisition strategy. The ADC images were scored on a two-point scale: 1, organs appeared homogeneous; 2, organs partially or completely disappeared or appeared heterogeneous. Diffusion-weighted images were scored based on the artifacts level, image sharpness and image distortion with 1 as the best quality and 4 as the worst. For quantitative comparisons, mean ADC values of liver and pancreatic tissues in each volunteer were measured using corresponding ROI within DW-SS-SE-EPI and DWPROPELLER ADC maps. A matched pair t-test was used (α=0.05) to test for statistical differences between both qualitative scores and quantitative measurements. Results: Representative diffusion-weighted images (b~0 and 502 s/mm) with reconstructed ADC maps acquired using DW-SS-SE-EPI and DWPROPELLER sequences are shown in Fig. 1 along with an additional 192x192 matrix image series demonstrating the feasibility of improving spatial resolution with DW-PROPELLER. Overall, no image distortion or motion artifacts were observed in the DW-PROPELLER images which provided improved anatomic detail. DW-SS-SE-EPI images were commonly distorted and provided inferior spatial resolution. Phantom Studies: Respective ADC values of water, acetone and ethanol as measured by the DW-PROPELLER sequence were 2.3x10mm/s, 4.9x10 mm/s and 1.3x10mm/s, consistent with those reported previously (2.252.51x10mm/s, 4.5-4.8 x10mm/s and 1.1-1.2 x10mm/s) [3-5]. Qualitative Comparison: Sharpness, distortion, and ADC organ homogeneity scores were significantly improved for DW-PROPELLER images in each category; artifact level scores were improved at b=0 s/mm but not statistically different at b=502 s/mm. Quantitative Comparison: The ADC map of each organ obtained using the DW-PROPELLER sequence was more homogeneous than the ADC map obtained using SS-SE-EPI. Mean Dtrace of liver and pancreatic tissues measured using the DW-PROPELLER sequence were (1.37±0.19)x10 mm/s and (2.06±0.23)x10mm/s respectively compared to (1.17±0.14)x10mm/s and (1.82±0.23)x10mm/s as measured using the DW-SS-SE-EPI sequence (mean±SD, no significant difference, p>0.05). Conclusions: The DW-PROPELLER sequence is a promising technique for multi-shot diffusion-weighted imaging of abdominal organs. DW-PROPELLER improved image sharpness and reduced distortion while providing accurate isotropic water diffusion measurements. Future pre-clinical studies will evaluate the use of DW-PROPELLER techniques for abdominal oncologic imaging applications (lesion detection, characterization, and therapy assessment). [1] Pipe Mag Res Med 2002 47:42-52 [2] Forbes Radiol 2002 225:551-5 [3] James J Mag Res 1973 11:58-61 [4] Cantor J Mag Res 1977 28:157-62 [5] Tofts Mag Res Med 2000 43:368-74.