Comparison of Several 8-element Surface Coil ... · mean= 1.15 Gmax = 1.5 Gmean = 1.6 Gmax = 2.1...
1
The coils configurations tested are shown in Fig 1. The 8-element surface coil arrays (Nova Medical Inc, Wakefield, MA, USA) were preamp decoupled with ultra-low impedance pre-amps (Nova Medical) [4,5] to minimize inductive coil-to-coil coupling while maintaining sample noise loading. Additional isolation was achieved by capacitative nulling of the inductive coupling between adjacent elements. The 4-element array was the standard GE cardiac surface coil array product. The optimum phased array combined SNR (B1-weighted) [5] and SENSE g-factor were computed for various coil geometries and imaging plane orientations based on simulated field maps. Magnetic fields (B1-maps) for each coil were calculated using the Biot-Savart law. The sample noise correlations between coils which are used in calculating the optimum B1- weighted combining and SENSE g-factor [1] were calculated from the magnetic vector potential (A-field). All experiments were conducted using a GE Signa CVi 1.5T MR imaging system which was modified to incorporate a high performance 8-channel digital receiver. Imaging parameters are listed in Table 1. Image reconstruction was performed off-line using MATLAB (The Mathworks, Natick, MA, USA). Several surface coils arrangements were used on a normal healthy volunteer. Care was taken to maintain the same localization for each coil configuration. METHODS METHODS Peter Kellman 1 , J. Andrew Derbyshire 1 , H. Douglas Morris 1 , Patrick J. Ledden 2 , Elliot R. McVeigh 1 1 Laboratory of Cardiac Energetics, National Heart Lung and Blood Institute, Bethesda Maryland 20892 USA 2 Nova Medical Inc, Wakefield, Massachusetts 01880 USA REFERENCES REFERENCES 1. Pruessmann, K.P., et.al. Magn. Reson. Med., 42, 952–962, 1999. 2. Weiger, M., et.al., Magn. Reson. Med., 45, 495-504, 2001. 3. Pruessmann, K.P., et.al. J. Cardiovascular Magn. Reson., 3, 1-9, 2001. 4. Roemer, P.B., et.al., Magn. Reson. Med., 16, 192-225, 1990. 5. Ledden, P.J., et.al., Proc 9th Annual Meeting ISMRM, 1117, 2001. CONCLUSIONS CONCLUSIONS INTRODUCTION INTRODUCTION RESULTS RESULTS Comparison of Several 8-element Surface Coil Configurations for Cardiac Imaging using SENSE • Overall array performance - all arrays tested had same SNR in central region (within 15%) - reasonable agreement with simulation - good pre-amp decoupling achieved • Cardiac SENSE performance - 8 coils provided reasonable quality up to R=3 acceleration + increased robustness to differences in coil positioning or slice orientation + linear array performed slightly better for SAX slice geometries tested + further testing of 2-d array planned - 4 coils degraded rapidly for R>2 due to ill-conditioning + high sensitivity to B1-map errors reduced artifact suppression Figure 5. Axial images of cylindrical phantom with 8-element Nova 1x4 linear arrays illustrating low degree of inductive coupling. R=1 24 beats R=2 12 R=3 9 R=4 6 Table 1 Cine Imaging Parameters scanner: custom modified* GE Signa 1.5T pulse sequence: ECG Gated, Segmented SSFP (FISP) acquisition matrix: 256 freq x 144 phase encodes FOV: 340 x 255 mm 2 Bandwidth: ±125 kHz spatial resolution: 1.3 x 1.8 mm 2 # cardiac phases: 34 TR: 4.1 ms # lines per segment: 6 breath-hold duration: 24 heart beats 12 (R=2 SENSE) 9 (R=3 SENSE) 6 (R=4 SENSE) * using GE LX CV/i custom modified with 8 digital receivers Morris, HD, Derbyshire, JA, Kellman, P., Chesnick, AS, Guttman, MA, McVeigh, ER. A Wide-Bandwidth Multi-channel Digital Receiver and Real-Time Reconstruction Engine for use with a Clinical MR Scanner. ISMRM 2002. Figure 3. Simulated g-factors for rates R= 2,3, & 4 SENSE with doubly oblique short axis cardiac imaging geometry. 1 1.1 1.2 1 1.5 2 2.5 3 1 2 3 4 5 R=2 G mean = 1.15 G max = 1.5 G mean = 1.6 G max = 2.1 R=3 R=4 Figure 2. Simulated g-factors for rate R=2 SENSE with doubly oblique short axis cardiac imaging geometry. Figure 4. Short axis cardiac images with comparing SENSE, rates R=1 (full k-space), 2, 3, & 4 for various surface coil arrays. preamplifier decoupled with adjacent elements capacitatively decoupled 2-element overlapped linear array 4-element linear array 4-element 2d array 21 cm (8.3”) 21 cm (8.3”) 21 cm (8.3”) 21 cm (8.3”) 19 cm (7.5”) 21 cm (8.3”) 2 cm (0.75”) 8-element configurations 4-element configuration Figure 3. Block diagram showing the phased-array phase-sensitive SENSE accelerated reconstruction of IR image using a separate reference image acquired after magnetization recovery. The performance of parallel imaging using sensitivity encoding (SENSE) [1] for accelerated acquisition is highly dependent on the number and geometry of surface coils. The SNR loss due to the noise amplification of the SENSE method is characterized by the g- factor. Arrays optimized for cardiac SENSE application using 6-elements have been previously described [2,3]. We present experimental cardiac imaging results comparing SENSE at various rates acquired with several different 8-element surface coil configurations, as well as a 4-coil cardiac array. SENSE was used to reduce the breath-hold time for acquiring cardiac cine images 1 1.05 1.1 1.15 1.2 1.25 imaging plane coils cross-section of body ellipsoid g-factor map approx. breath-hold duration:
Transcript of Comparison of Several 8-element Surface Coil ... · mean= 1.15 Gmax = 1.5 Gmean = 1.6 Gmax = 2.1...
The coils configurations tested are shown in Fig 1. The 8-element surface coil arrays (Nova Medical Inc, Wakefield, MA, USA) were preamp decoupled with ultra-low impedance pre-amps (Nova Medical) [4,5] to minimize inductive coil-to-coil coupling while maintaining sample noise loading. Additional isolation was achieved by capacitative nulling of the inductive coupling between adjacent elements. The 4-element array was the standard GE cardiac surface coil array product.
The optimum phased array combined SNR (B1-weighted) [5] and SENSE g-factor were computed for various coil geometries and imaging plane orientations based on simulated field maps. Magnetic fields (B1-maps) for each coil were calculated using the Biot-Savart law. The sample noise correlations between coils which are used in calculating the optimum B1-weighted combining and SENSE g-factor [1] were calculated from the magnetic vector potential (A-field).
All experiments were conducted using a GE Signa CVi 1.5T MR imaging system which was modified to incorporate a high performance 8-channel digital receiver. Imaging parameters are listed in Table 1. Image reconstruction was performed off-line using MATLAB (The Mathworks, Natick, MA, USA). Several surface coils arrangements were used on a normal healthy volunteer. Care was taken to maintain the same localization for each coil configuration.
METHODSMETHODS
Peter Kellman1, J. Andrew Derbyshire1, H. Douglas Morris1, Patrick J. Ledden2, Elliot R. McVeigh1
1Laboratory of Cardiac Energetics, National Heart Lung and Blood Institute, Bethesda Maryland 20892 USA2Nova Medical Inc, Wakefield, Massachusetts 01880 USA
REFERENCESREFERENCES
1. Pruessmann, K.P., et.al. Magn. Reson. Med., 42, 952–962, 1999.2. Weiger, M., et.al., Magn. Reson. Med., 45, 495-504, 2001.3. Pruessmann, K.P., et.al. J. Cardiovascular Magn. Reson., 3, 1-9, 2001.4. Roemer, P.B., et.al., Magn. Reson. Med., 16, 192-225, 1990.5. Ledden, P.J., et.al., Proc 9th Annual Meeting ISMRM, 1117, 2001.
CONCLUSIONSCONCLUSIONS
INTRODUCTIONINTRODUCTION
RESULTSRESULTS
Comparison of Several 8-element Surface Coil Configurations for Cardiac Imaging using SENSE
• Overall array performance- all arrays tested had same SNR in central region (within 15%)- reasonable agreement with simulation- good pre-amp decoupling achieved
• Cardiac SENSE performance- 8 coils provided reasonable quality up to R=3 acceleration
+ increased robustness to differences in coil positioning or slice orientation+ linear array performed slightly better for SAX slice geometries tested+ further testing of 2-d array planned
- 4 coils degraded rapidly for R>2 due to ill-conditioning+ high sensitivity to B1-map errors reduced artifact suppression
Figure 5. Axial images of cylindrical phantom with 8-element Nova 1x4 linear arrays illustrating low degree of inductive coupling.
R=124 beats
R=212
R=39
R=46Table 1 Cine Imaging Parameters
scanner: custom modified* GE Signa 1.5Tpulse sequence: ECG Gated, Segmented SSFP (FISP)acquisition matrix: 256 freq x 144 phase encodesFOV: 340 x 255 mm2
Bandwidth: ±125 kHzspatial resolution: 1.3 x 1.8 mm2
# cardiac phases: 34TR: 4.1 ms# lines per segment: 6breath-hold duration: 24 heart beats
12 (R=2 SENSE)9 (R=3 SENSE)6 (R=4 SENSE)
* using GE LX CV/i custom modified with 8 digital receiversMorris, HD, Derbyshire, JA, Kellman, P., Chesnick, AS, Guttman, MA, McVeigh, ER. A Wide-Bandwidth Multi-channel Digital Receiver and Real-Time Reconstruction Engine for use with a Clinical MR Scanner. ISMRM 2002.
Figure 3. Simulated g-factors for rates R= 2,3, & 4 SENSE with doubly oblique short axis cardiac imaging geometry.
1 1.1 1.2 1 1.5 2 2.5 3 1 2 3 4 5
R=2 Gmean= 1.15Gmax = 1.5
Gmean = 1.6Gmax = 2.1
R=3 R=4
Figure 2. Simulated g-factors for rate R=2 SENSE with doubly oblique short axis cardiac imaging geometry.
Figure 4. Short axis cardiac images with comparing SENSE, rates R=1 (full k-space), 2, 3, & 4 for various surface coil arrays.
preamplifier decoupled withadjacent elements capacitatively decoupled
2-element overlapped linear array4-element linear array 4-element 2d array
21 cm(8.3”)
21 cm (8.3”)
21 cm(8.3”)
21 cm (8.3”)
19 cm(7.5”)
21 cm (8.3”)
2 cm (0.75”)
8-element configurations 4-element configuration
Figure 3. Block diagram showing the phased-array phase-sensitive SENSE accelerated reconstruction of IR image using a separate reference image acquired after magnetization recovery.
The performance of parallel imaging using sensitivity encoding (SENSE) [1] for accelerated acquisition is highly dependent on the number and geometry of surface coils. The SNR loss due to the noise amplification of the SENSE method is characterized by the g-factor. Arrays optimized for cardiac SENSE application using 6-elements have been previously described [2,3]. We present experimental cardiac imaging results comparing SENSE at various rates acquired with several different 8-element surface coil configurations, as well as a 4-coil cardiac array. SENSE was used to reduce the breath-hold time for acquiring cardiac cine images
1 1.05 1.1 1.15 1.2 1.25
imaging plane
coils
cross-section of body ellipsoid
g-factor map
approx. breath-hold duration:
