| ObjectiveAt3.0T higher-fields magnetic resonance (MR), to prospectively and intraindividually compare abdomen diffusion-weighted imaging (DWI) obtained with different b values between dual-source parallel radiofrequency (RF) transmission and conventional single-source RF transmission in terms of image quality and measured apparent diffusion coefficient (ADC).Materials and MethodsThe written informed consent was obtained from each participating volunteer and a total of eighteen healthy male volunteers (mean age,27.3years; age range,24-33years) with no prior history or findings related to abdomen disease were enrolled in our study. From July,2011to August, eighteen healthy male volunteers sequentially underwent DWI of the abdomen at a commercially available clinical3.0-T MR imaging system (Achieva3.0T TX; Philips Healthcare) equipped with a dual-source parallel RF transmission system. The system used in our study allows one to perform both dual-source and conventional single-source excitation with the quadrature body coil and thus enables an intraindividual comparative study design without repositioning.Two engineers (Philips Healthcare) assisted in the design of the sequence and imaging protocol. Before DWI was started, Bi mapping was performed by using dual-source RF transmission and conventional single-source RF transmission, respectively. This was done for quantitative assessment of RF fields (B1) homogeneity on B1map. Acquisition parameters were as follows:repetition time/echo time (TR/TE) =517/3.3ms, field of view (FOV)=326×278mm, matrix size=64×58, voxel size (mm)=5×5×6, dual flip angle=60°and120°, bandwidth=4293.7Hz, a single section and a section thickness of6mm. The acquisition time was10s during a single breath hold.The DW imaging protocol included a series of different b values:0/100,0/500,0/800and0/100/800s/mm2. Each b value was performed twice:one by using dual-source parallel RF transmission (with two independent RF sources), another by using conventional single-source RF transmission (with a single RF source). Respiratory triggering technique was used and the other sequence parameters were as follows:TR/TE=2400/40ms, FOV (mm)=326×278, matrix size=108×89, voxel size (mm)=3×3×6, water-fat shift (bandwidth)=9.9pixel/43.8Hz, halfscan=0.698, Echo train length (EPI factor)=39, number of acquired signals=4, sense factor=2.5, RF excitation (degrees)=90°, section thickness/intersection spacing (mm)=6/1, number of sections=24, spectral presaturation with inversion recovery (SPIR) for fat suppression.ADC map of each DWI series was calculated on a workstation with standard software (Diffusion Calculation, Philips Medical Systems). In addition, higher b value100/800included in0/100/800DW images was selected for ADC map calculation to separate perfusion and true diffusion effects. Therefore, a total of ten ADC map calculations were made for each volunteer (five different b values, dual-source parallel RF transmission and single-source RF transmission). Data were analyzed as the following two steps:(1) Image AnalysisSubjective Image Quality. Two radiologists (10and4years experience in body radiology, respectively) independently evaluated both the dual-source parallel RF transmission DWI images and the conventional single-source RF transmission DWI images in a randomized order during two separate reading sessions. The two sessions were separated1week. The observers were fully blinded to RF transmission and b value. Images obtained with the b-values of0,100,500, and800s/mm2were evaluated on a4-point scale (1=poor image quality, major standing-wave artifacts;2=fair image quality, moderate standing-wave artifacts;3=good image quality, mild standing-wave artifacts;4=excellent image quality, no standing-wave artifacts). Signal to Noise Ratio:Signal to noise ratio (SNR) was calculated on both the dual-source parallel RF transmission DWI images and the conventional single-source RF transmission DWI images at b values of0,100,500, and800s/mm2for posterior right hepatic lobe, lateral left hepatic lobe, spleen and pancreatic tail. Parallel imaging prohibits the use of a conventional method of measuring noise in which the region-of-interest (ROI) is placed in the air. Thus, as recommended in the literature, we used Standard Deviation (SD) of normal organ signal intensity as an estimate of local noise. In order to avoiding the interference from tissues with different properties and various artifacts, large blood vessels and biliary ducts should be avoided and regions with good homogeneity should be chosen when place ROI. Using copy and paste function to ensure the ROI concordance between dual-source and single-source RF transmission images.Quantitative Assessment of B1Homogeneity:The actual flip angles excitated by90°RF pulse were measured and the coefficient of variation (CV, equal to standard deviation divided by mean) of the actual flip angles was calculated on dual-source parallel RF transmission Bi map and the conventional single-source RF transmission B1map for the posterior right hepatic lobe, lateral left hepatic lobe, pancreatic tail and spleen. The consistency of the actual flip angles distribution (uniformity of local signal intensity distribution) was compared between the dual-source parallel RF transmission and the conventional single-source RF transmission to assess the homogeneity of B1field.(2) ADC Measurements of Abdominal OrgansROI measurements were obtained in the posterior right hepatic lobe, lateral left hepatic lobe, pancreatic tail and spleen. For each anatomic location studied,3nonoverlapping elliptic ROIs were placed on the images. ADC measurements were obtained from both the dual-source parallel RF transmission images and the conventional single-source RF transmission images at each b value (0/100,0/500,0/800,0/100/800,100/800s/mm2). Therefore, a total of120data points per volunteer were collected:3ROIs per anatomic location,4anatomic locations,5different b values, dual-source parallel RF transmission and the conventional single-source RF transmission.All statistical analyses were performed with SPSS, version17.0(SPSS, Chicago, Ill). A series of paired Wilcoxon tests were used to compare the subjective image quality, SNR of each anatomic location, the actual flip angles and CV of each anatomic location on B1map between dual-source parallel RF transmission images and the conventional single-source RF transmission images. In order to assess interobserver agreement, we calculated the kappa statistic for two observers. Kappa values of0.4or less were considered to indicate positive but poor agreement, while those of0.41-0.75and greater than0.75indicated good and excellent agreements, respectively. At each b value, the measured ADC values of each anatomic location were compared between dual-source parallel RF transmission and conventional single-source RF transmission by using the within-subject factor repeated measure analysis of variance (ANOVA). P-values less than0.05were considered statistically significant.Results(1) Image analysisAt each of the selected b values, subjective image quality within the abdomen was statistically better at dual-source parallel RF transmission imaging than at conventional single-source RF transmission imaging (P<0.0001for all comparisons). There was a good to excellent agreement between observers for image quality assessment (Kappa values from0.75to1.00). Comparing with the conventional single-source RF transmission, dual-source parallel RF transmission prevent the standing-wave artifacts at b-0DW images (the scores of both the observers were4.00) and significantly reduced the artifacts at b-100,500and800DW images.At each of the selected b-values (0,100,500and800s/mm2), there was a trend toward better SNR at dual-source parallel RF transmission imaging than at conventional single-source RF transmission imaging for all the organs, despite only the lateral left hepatic lobe achieved the significant difference (P<0.001). Among the various anatomic locations, spleen has the highest SNR at dual-source parallel RF transmission imaging, while the lateral left hepatic lobe has the lowest SNR at conventional single-source RF transmission imaging.Quantitative assessment of B1homogeneity showed that the actual flip angles of each anatomic location on dual-source parallel RF transmission Bi map were significantly higher than on the conventional single-source RF transmission B1map. CV of the actual flip angles of each anatomic location on dual-source parallel RF transmission B1map was significantly lower than on the conventional single-source RF transmission B1map, which suggested that the actual flip angle distribution (local signal intensity distribution) of each anatomic location is more consistent on dual-source parallel RF transmission images than on conventional single-source RF transmission images. Additionally, the local signal intensity distribution between different anatomic locations and between different subjects tends to be more consistent on the dual-source parallel RF transmission imaging, comparing with the conventional single-source RF transmission imaging.(2) ADC Measurements of Abdominal OrgansThe main effects analysis showed:there was no significant (P=0.074) main effect of RF transmission on the measured ADC values. It means when considering all b values and all anatomic locations together, the measured ADC values showed no significant difference between the dual-source parallel RF transmission imaging and the conventional single-source RF transmission imaging. The main effects of the other two within-subject factors, b values and anatomic locations, were significant (P<0.0001). It means the measured ADC values between different b values and between different anatomic locations achieved significant difference.The two-way interaction effects analysis showed:RF transmission and b values obtained the significant two-way interaction effects when considering all anatomic locations together. At b value of0/100, measured ADC values on dual-source parallel RF transmission images were significantly lower than on the conventional single-source RF transmission images (P<0.0001). At b value of0/100/800and100/800, measured ADC values on dual-source parallel RF transmission images were significantly higher than on the conventional single-source RF transmission images (P value was0.007,0.031, respectively). At b value of0/500and0/800, measured ADC values between dual-source parallel RF transmission images and conventional single-source RF transmission images showed no significant difference (P value was0.437,0.236, respectively). Additionally, RF transmission and anatomic locations also obtained the significant two-way interaction effects when considering all b values together. Among all of the anatomic locations, only measured ADC values of spleen on dual-source parallel RF transmission images were significantly lower than on conventional single-source RF transmission images (P<0.0001). Measured ADC values of other anatomic locations showed no significant difference between dual-source parallel RF transmission images and conventional single-source RF transmission images. The three-way interaction effects analysis showed:RF transmission, b values and anatomic locations obtained the significant three-way interaction effects. Thus, the simple effect was analyzed and the results were as follow:for lateral left hepatic lobe, the measured ADC values of the dual-source parallel RF transmission images were significantly higher than the conventional single-source RF transmission image at b=0/800(P<0.001),100/800(P=0.001),0/100/800(P=0.001) and lower than the conventional single-source RF transmission image at b=0/100(P<0.001); for spleen, the measured ADC values of the dual-source parallel RF transmission images were significantly lower than the conventional single-source RF transmission image at b=0/100(P<0.001),0/500(P=0.047),0/800(P=0.012); for posterior right hepatic lobe and pancreatic tail, there was no significant difference in the measured ADC values between the dual-source parallel RF transmission images and the conventional single-source RF transmission images at any b value.Combining with the simple effect analysis and the two-way interaction effect analysis between RF transmission and b value, it showed that there was no significant difference on ADC measurements between the parallel and the conventional RF transmission images for all organs at b=0/500, except a slight statistically significant differences for spleen (P=0.047).Conclusion1. Comparing with the conventional single-source RF transmission, the dual-source parallel RF transmission significantly improved the image quality and the homogeneity of the RF field.2. Comparing with the conventional single-source RF transmission, the dual-source parallel RF transmission improved the SNR for all organs at each b-value, of which only the lateral left hepatic lobe achieved the significant difference.3. At either lower or higher b value, the measured ADC values of lateral left hepatic lobe and spleen between the dual-source parallel RF transmission and the conventional single-source RF transmission were significantly different.4. At b value of0/500, the minimum difference on ADC measurements between the dual-source parallel RF transmission images and the conventional single-source RF transmission images was achieved for all abdominal organs. |
PDF Full Text Request