Musculoskeletal MRI Technical Considerations Garry E. Gold, M.D. Professor of Radiology, Bioengineering and Orthopaedic Surgery Stanford University
Outline Joint Structure Image Contrast Protocols: 3.0T
Take-Home Points RF Coil is key - for any MSK imaging Consider 512 freq. matrix Use IR, but selectively Sequences and Imaging Planes Complementary
Outline Joint Structure Image Contrast Protocols: 3.0T
Knee Tissues Bone: cortical, cancellous Marrow: fatty (yellow) hematopoietic (red) Muscles Tendons and ligaments Cartilage: articular vs. fibrocartilage Joint fluid
Outline Joint Structure Image Contrast Protocols: 3.0T
Image Contrast Concepts Pulse Sequences Fat Suppression
Image Contrast Concepts Pulse Sequences Fat Suppression
Contrast - Concepts Sequences determine image contrast based on proton density and relaxation times
Contrast Mechanisms Proton Density T2 Contrast T1 Contrast Fat Suppression Mixed or T2/T1 contrast
T 2 Contrast Short Echo-Time (PD) Long Echo-Time (T2) Signal Joint Fluid Cartilage/Muscle Time Long TR avoids T 1 contrast
Proton Density Fluid: higher signal relative to cartilage
PD PD + Fat Sat
T 2 Fluid: very high signal relative to cartilage Dx?
T 1 Contrast Short TR Long TR Signal Signal Cartilage/Muscle Time Short TE avoids T 2 contrast Joint Fluid Time
T 1 Spin Echo Fluid: low signal relative to cartilage
Image Contrast Concepts Pulse Sequences Fat Suppression
Image Contrast Pulse Sequences Spin Echo Fast/Turbo Spin Echo Inversion Recovery (fast) Gradient Echo Special Purpose: 3D-FSE, T2-mapping
Spin Echo T 1 weighting Short TR, short TE 1.5T: TR 600-800 ms 3.0T: TR 800-1000 ms Longer TR: More slices and better SNR
Fast/Turbo Spin Echo TE ~ 54-70 sufficient for T 2 W Need less slice gap then SE Blurring on PD weighted images Longer TE, shorter ETL decreases blurring Increasing receiver bandwidth decreases blurring Never need conventional PD, T2 T1 FSE possible with short ETL, high BW
Image Space K-Space
PD FSE/TSE - T2 Blurring PD FSE GRE
FSE/TSE - Echo Train Length To avoid significant blurring artifact PD Images: ETL 4-6 T2W Images: ETL 8-12 * depends on receiver bandwidth
Special Sequences 3D FSE (SPACE, CUBE, VISTA) Fast spin echo PD and T2 contrast 3D imaging, little blurring T2 Mapping Cartilage collagen breakdown Muscle after exercise
2D vs 3D Slab Imaging 2D Multislice 3D Slab Shorter scan times Efficient when TR is long and all slices can be interleaved Averaging helps SNR Continuous coverage Better for image reformat
3D-FSE Knee at 3.0T 0.6 0.6 0.6 mm (256 256 200) in 8 minutes Coronal Acquisition Sagittal Reformat Axial Reformat
3D-FSE: Coronal Source 3.0T isotropic resolution of 0.6 mm
3D-FSE: Sagittal Reformat 3.0T isotropic resolution of 0.6 mm
T 2 Mapping Increased T 2 relaxation time is a marker for decreased collagen organization in cartilage Signal TE = 20 TE = 40 TE = 60 Color T2 Map TE = 80 Echo Time (ms) Dardzinski BJ, et al. Radiology, 205: 546-550, 1997.
3.0T Knee Cartilage MRI Protocol PD-FSE 3D-FSE-Cube 3D MERGE Coil and cartilage thickness help determine MRI parameters T2 Map
Image Contrast Concepts Pulse Sequences Fat Suppression
Fat Suppression Why fat suppression is important: Contrast Dynamic range Chemical shift Methods: Frequency selective chemsat Inversion recovery Fat/Water separation or IDEAL
Fat Suppression for Contrast PD FSE Fat-Sat PD FSE Coronal Wrist Coronal Wrist Radial cyst was otherwise iso-intense with fat
Fat Suppression for Dynamic Range SPGR: TR 50, TE 7, 40 Flip, 2.5 mm
Chemical Shift Artifact Bone-marrow can obscure cartilage due to chemical shift BW + 16 BW + 16 64 khz
Fat Sat: Sensitive to B 0 Inhomogeneity water fat ω
Failure of Fat Saturation
Inversion Recovery (STIR) TI 150-160 ms (1.5 T) 100 ms (0.5 T) Better for inhomogeneous field Air-tissue interfaces Metal artifact decreased Lower SNR than FSE Medium TE, ~ 50 works well
Fat sat IR Fat sat IR
IDEAL Fat/Water Separation Source Images F W F W Least Squares Estimation Water F W Fat 3x Scan Time = 5 minutes SNR efficiency is preserved Post-processing can correct large field inhomogeneities (Reeder 2003, 2004)
IDEAL FSE Water Fat + Water Combined Reeder, Pelc et al., 2003 Fat
Take-Home Point IDEAL results in the best image quality Use IR, but selectively (if you have the choice) consider shorter TE (~35-50 ms)
Outline Joint Structure Image Contrast Protocols: 3.0T
Protocol Principles: 3.0T Reduce slice thickness Keep TR long on PD, T2 (4000-5000 ms) Frequency matrix 512 Multi-channel, parallel capable coils Increase bandwidth on non-fat suppressed images Short echo train on T1w FSE (2-3) Balance echoes around TE for FSE
High Resolution Knee Protocol Use: High quality knee imaging Goal: Keep imaging time to about 30-45 min while having outstanding quality Possible to scan a knee in 45 min with table turn around 8 channel knee coil
High Resolution Knee Protocol Axial PD FSE TR/TE = 5000/20 Fat saturation 2.5 mm slices 416 x 320 2 averages 14 cm FOV 26 slices ETL = 8 +/- 32 khz BW
High Resolution Knee Protocol Coronal T1 FSE TR/TE = 1000/15 No Fat saturation 2.5 mm slices 512 x 320 2 averages 14 cm FOV 18 slices ETL = 3 +/- 41 khz BW
High Resolution Knee Protocol Coronal T2 FSE TR/TE = 5000/60 Fat saturation 2.5 mm slices 416 x 320 2 averages 14 cm FOV 22 slices ETL = 8 +/- 32 khz BW
High Resolution Knee Protocol Sagittal PD FSE TR/TE = 5000/15 No fat saturation 2.5 mm slices 512 x 320 2 averages 16 cm FOV 30 slices ETL = 6 +/- 41 khz BW
High Resolution Knee Protocol Sagittal T2 FSE TR/TE = 5000/54 Fat saturation 2.5 mm slices 384 x 320 2 averages 16 cm FOV 30 slices ETL = 8 +/- 32 khz BW Flow Comp, S Sat
High Resolution Knee Protocol Coronal 3D FSE TR/TE = 1500/35 Fat saturation 0.6 mm slices 320 x 288 0.5 averages 17 cm FOV 200 slices ETL = 35 +/- 50 khz BW Reformat at 2 mm slices
High Resolution Scan Time Comparison Sequence 1.5T (slice, mm) 3.0T (slice, mm) Axial PD 3:10 (4) 6:00 (2.5) Cor T1 5:10 (4) 3:30 (2.5) Cor T2 3:10 (4) 6:00 (2.5) Sag PD 4:16 (3) 5:00 (2.5) Sag T2 4:48 (3.5) 5:00 (2.5) Cor 3D FSE x 5:00 (0.6) Total 20:34 30:30
Rapid Knee Protocol Use: Routine knee imaging Goal: Keep imaging time to a minimum while having acceptable quality Possible to scan a knee in 15 min with table turn around
Rapid Knee Protocol Axial PD FSE TR/TE = 5000/35 Fat saturation 4.0 mm slices 320x224, 1 nex 14 cm FOV 26 slices ETL = 8 32 khz BW ARC 1.8
Rapid Knee Protocol Coronal T1 FSE TR/TE = 1000/20 No Fat saturation 4 mm slices 384x224, 1 nex 16 cm FOV 18 slices ETL = 4 32 khz BW ARC 1.8
Rapid Knee Protocol Coronal T2 FSE TR/TE = 4000/54 Fat saturation 4 mm slices 320x224, 1 nex 16 cm FOV 22 slices ETL = 8 32 khz BW ARC 1.8
Rapid Knee Protocol Sagittal PD FSE TR/TE = 5000/35 No fat saturation 3 mm slices 384x224, 1 nex 14 cm FOV 30 slices ETL = 8 32 khz BW ARC - none
Rapid Knee Protocol Sagittal T2 FSE TR/TE = 6400/60 No fat saturation 3 mm slices 320x224, 1 nex 14 cm FOV 30 slices ETL = 10 32 khz BW Flow Comp, SI Sat
Scan Time Comparison Sequence 1.5T (nex) 3.0T (nex) Axial PD 3:10 (2) 1:25 (1) Cor T1 5:10 (2) 1:43 (1) Cor T2 3:10 (2) 2:24 (1) Sag PD 4:16 (2) 2:30 (1) Sag T2 4:48 (3) 2:40 (1) Total 20:34 10:42
Summary Joint Structure Image Contrast Protocols: 3.0T
Thank You