wagshul y học

Pulse Sequences Mark Wagshul, PhD Director, MR Research Center Department of Radiology Stony Brook University MRI Pulse Sequences  Spin echo and gradient echo sequences basic methods of MRI contrast   3D techniques Preparation techniques secondary contrast methods   Fast imaging Gating and k-space segmentation Important basic concepts   Images are not acquired in real space, but in inverse-space  k-space Image contrast is obtained by manipulating magnetic or biological properties of spins (e.g relaxation, precession frequency, magnetic susceptibility, diffusion)   Image contrast is (almost) always obtained at the expense of signal size “Static” fields (e.g B0,G) are along z, RF fields are to z Basic Spin Echo Sequence TE/2 RF pulses slice select Position of BOTH spin-echo and gradient echo phase encode readout TE The step encoding process z = z =  / Gz BW/ Gz  y) =  Gyy ef encoded PRIOR to x)readout =  Gxxt encoded DURING readout Double-echo spin echo - PD/T2 Basic Gradient Echo Sequence RF Pulse Slice Select Phase Encode e-iGx x eiGxxt Readout Major diferences between spin-echo and gradient echo     SE – 90 degree pulse, uses all Mz per pulse GE – variable pulse angle, partial use of Mz, allows for faster TR’s SE – refocuses T2*, allows longer TE’s GE – T2* weighting, generally requires short TE’s SE – usually used for T2 weighted images GE – usually used for T1 weighted images, or for speed SE - T1 weighting adjusted with TR GE - T1 weighting adjusted with TR AND flip angle 3D imaging Reduced gradient for “slab” selection secondary phase encode:  (z) =  Gzzton b a l S e m u l Vo Bo v  Fast Imaging  SE methods     GE methods     Fast spin echo (FSE) Echo planar imaging (EPI) Steady state imaging (PSIF, CE-FAST) Spoiled GE (SPGR) Echo planar imaging (EPI), Spiral Steady state imaging (True-FISP, FIESTA, Balanced FFE) Parallel imaging (SENSE, SMASH) Fast Spin Echo      Collect multiple k-space lines per 90 degree pulse New parameters – interecho spacing and echo train length (ETL), will determine overall slice TR and attainable TE’s Typical ETL – 8, but can collect the entire image in a single shot (e.g 128 PE’s) TE determined by echo position of center of kspace For large ETL  increasing T2 signal loss in later echoes lead to image blurring (loss of edge of kspace) ETL Fast Spin Echo Inter echo spaci ng Phase unwrap k1 k2 k3 k4 FSE k-space coverage }k4 }k3 }k2 } } k1 k2 }k3 }k4 Optimum sampling will place more central portions of k-space close to the k=0 echo Echo Planar Imaging (EPI)     Multi-echo gradient echo readout Multi-shot – uses standard gradients and multiple acquisition windows Single shot – uses sinusoidal readout gradient and “blipped” phase encode single acquisition window  requires “rebinning” to reconstruct Overall T2* weighting across the echo train  blurring and distortion Phase encode Readout EPI k-space coverage Alternate strategy – spiral, much more efficient k-space coverage Gx Gy EPI Applications  Functional MRI – BOLD imaging   Perfusion   T2* weighting, detects changes in blood oxygenation T2* weighting, detects rapid passage of paramagnetic contrast agent Diffusion  Diffusion-weighted, detects differences in water diffusion, e.g due to stroke, EPI used for rapid coverage of whole brain with minimum motion SPGR (Spoiled GE)   Short-TR, low flip angle GE Inter-echo coherent affects are spoiled using     RF phase cycling Inter-echo gradient pulses Ideally, gradient spoiling uses varying (random) gradient pulses, in practice often just use fixed large gradient pulse with phase unwrapping Gradient spoiling condition – G*x*ton > 2 True FISP (Free Induction Steady state Precession)     GE w/o spoiling, magnetization remains in transverse plane More efficient because magnetization is “reused” Contrast ~ T1/T2, excellent for blood/myocardium contrast BUT, very sensitive to field inhomogeneity Rephasing period = TR, so require B < 2TR, usually need TR < ms  Main application to date – near real-time cardiac FLASH True FISP Parallel Imaging     Uses spatial information available from an array of RF coils to fill in portions of k-space Maximum reduction factor = number of coils SENSE – obtain smaller FOV images/coil and utilize sensitivity maps of coils to correct aliasing SMASH – use spatial harmonic parameters of coils to fill in parts of k-space Gating and k-space segmentation   Image acquisition is gated to some physiological signal, e.g ECG or respiration BUT, extremely inefficient if only one k-space line covered per period  collect multiple segments of k-space per period k1 k2 k3 k4 k5 k6