iPAT - Integrated Parallel Acquisition Technique

PAT – Parallel Imaging Techniques Employs reconstruction algorithms and arrays of coils   Each coil independently and simultaneously images a given volume   Uses spatial information inherent in local array or matrix coils Vendors Image-based algorithm k-space-based algorithm Siemens mSENSE GRAPPA GE ASSET* ARC* Phillips SENSE* None* Hitachi RAPID* (k-RAPID)* Toshiba SPEEDER* None* * Competitor Information Disclaimer: Competition product descriptions, comparison and specifications contained in this document are based on interpretation of available data at the time this material was being prepared and may require independent verification.  Specifications have been obtained from competition brochures, websites and other independent published sources Coil Compatibility with iPAT Minimum of 2 Coil Elements Coils Siemens Integrated Panoramic Array (IPA) and Matrix coils 3rd Party coils with multiple coil elements Coil elements must be aligned in phase-encoding direction    Note: check with coil manufacturer to determine if the coil is compatible with parallel imaging Coil Auto-Calibration Prerequisite for calculation is exact knowledge of coil sensitivity maps Coil profiles determined by calibration No calibration prior to the initial scan required Siemens PAT mode reconstruction algorithms mSENSE – modified SENSitivity Encoding Image based GRAPPA – GeneRalized Autocalibrating Partially Parallel Acquisition k-space – Full FoV calculated without aliasing mSENSE – image-based method      GRAPPA – k-space-based method Aliasing original FOV > object Aliasing original FOV > object mSENSE – Pre-aliasing artifact GRAPPA – No Pre-aliasing Artifact PAT Mode = mSENSE Accel. Factor PE = 2 FoV Phase = 66.7% Auto-Calibration PAT Mode = GRAPPA Accel. Factor PE = 2 FoV Phase = 66.7% Auto-Calibration mSENSE Accel. Factor PE = 2 Phase FoV 50% Auto-Calibration GRAPPA Accel. Factor PE = 2 Phase FoV 50% Auto-Calibration Reduced measurement time   Shorter breath-holds   Increase resolution   Single Shot EPI and HASTE Decrease distortion artifacts Decrease blurring iPAT – 2D & 3D Sequences PAT Mode GRAPPA mSENSE Accel. factor PE Ref lines PE iPAT – 2D Sequences PAT Mode GRAPPA mSENSE Accel. factor PE Maximum number corresponds to number of receiver channels used Increasing value reduces signal-to-noise Ref lines PE iPAT – 2D Sequences PAT Mode GRAPPA mSENSE Accel. factor PE Maximum based on number of Coils aligned PE direction Increasing value reduces signal-to-noise Typical values used: 2, 3 & 4 Ref lines PE Value automatically adjusts based on sequence and protocol parameters iPAT – 3D Sequences PAT Mode GRAPPA mSENSE Accel. factor 3D Slice-selection direction Maximum number corresponds to the number of receiver channels used Ref lines 3D iPAT – 3D Sequences PAT Mode GRAPPA mSENSE Accel. factor 3D Slice-selection direction Maximum number corresponds to the number of receiver channels used Ref lines 3D Value automatically adjusts based on sequence and protocol parameters iPAT – 2D & 3D Sequences Reference scan mode Method reference lines are measured Options Integrated Separate Self-calibrated tPat Intelligent consultant for optimal Accel. factor PE Tim Assistant turns Red when Siemens coils are not used Incorrect coil elements Incorrect phase-encoding direction Number of coil elements ≥ Accel. factor PE Coils not aligned in the phase-encoding direction Acceleration factor entered is greater than recommended Acceleration factor 3D is greater than recommended for iPAT2 3rd Party Coils utilized Array or Matrix coils required for PAT Minimum of 2 coil elements   Array or Matrix coil elements aligned in phase encoding direction   Number of coil elements must be ≥ Accel. factor PE   SNR is reduced when iPAT is used and Accel. Factor PE is increased CAIPIRINHA – Controlled Aliasing In Parallel Imaging Results IN Higher Acceleration 3D VIBE Fat Saturation (or) 3D DIXON pulse sequence CAIPIRINHA pattern distributes k-space points more uniformly iPAT2 Reconstruction – True 3D reconstruction Improved image quality Less residual artifacts Lower noise amplification Reordering Shift 3D CAIPIRINHA Mode Free PAT mode - CAIPIRINHA Normal iPAT2 used Reordering Shift 3D = 0 Body Tra Body Cor Breast Total PAT Factor Abdomen Fast Excitation Pulse Short TR/TE Fast Fat Suppression QFS – Quick Fat Sat SPAIR Dixon Breast Breast coil selected Motion insensitive reordering Inline parameter card T1 weighted SPAIR Matrix Coil Modes – combines array or matrix coil elements in left to right direction CP Mode Dual Mode Triple Mode Cluster Group of coil elements L-R direction Ring Group of Clusters (coil elements) in A-P direction CP Matrix Coil Mode GRAPPAx2 Coils Head 1-4 Dual Matrix Coil Mode GRAPPAx2 Coils Head 1-4 Triple Matrix Coil Mode GRAPPAx2 Coils Head 2 & 4 Dual Matrix Coil Mode GRAPPAx2 Phase A>>P Coils NE1, NE2, SP1 Triple Matrix Coil Mode GRAPPAx2 Phase A>>P Coils NE2 & SP1 CP Matrix Coil Mode No PAT Pre-scan Normalize On Phase FoV 75% Coils BO1, 2; SP2-4  Triple Matrix Coil Mode GRAPPAx2 Pre-scan Normalize On Phase FoV 75% Coils BO1,2; SP 2-4  Triple Matrix Coil Mode GRAPPAx2 Pre-scan Normalize Off Phase FoV 75% Coils BO1, 2; SP 2-4 Head Matrix (1.5T) – 12 coil elements A>>P = 3 H>>F = 2 L>>R = 3 Triple mode; 2 Dual mode     Head Matrix (3T) – 12 coil elements A>>P = 3 L>>R = 3 Triple mode; 2 Dual mode     Neck Matrix – 4 coil elements A>>P = 2 H>>F = 2 L>>R = 2 Dual mode Body Matrix – 6 coil elements A>>P = 2 e.g., combined with spine matrix coil H>>F = 2 L>>R = 3 Triple mode; 2 Dual mode     Spine Matrix – 24 coil elements H>>F = 4 L>>R = 3 Triple mode; 2 Dual mode     Body Matrix and Spine Matrix combination A>>P = 2 H>>F = 2 L>>R = 3 Triple mode; 2 Dual mode Head/Neck 20 Head 16 H>>F = 2 coil elements L>>R = 4 coil elements 8 anterior and 8 posterior   Neck 4 H>>F = 1 coil element L>>R = 4 coil elements 2 anterior and 2 posterior   Body 18 H>>F = 3 coil elements L>>R = 6 coil elements Spine 32 H>>F = 8 coil elements L>>R = 4 coil elements     Peripheral Angio 36 H>>F = 6 coil elements L>>R = 6 coil elements     Shoulder Large/Small 16 H>>F = 2 coil elements A>>P = 8 coil elements   Hand/Wrist 16 H>>F = 3 Rows of coil elements 1 Row – Phalanges L>>R = 4 coil elements 2 Rows – Metacarpals L>>R = 4 coil elements 3 Rows – Carpals L>>R = 8 coil elements Foot/Ankle 16 A>>P = 2 Rows of coil elements 1 row – forefoot-midfoot L>>R = 4 coil elements 2 rows – Hindfoot-ankle L>>R = 12 coil elements iPAT Disadvantages Signal loss Artifacts Resulting from inappropriate parameter modifications iPAT Advantages PAT averaging Time savings Single Shot Imaging – Reduced distortions and blurring EPI HASTE Cervical Spine   Standard Reconstruction PAT Averaging Conventional 1 Average TA – 3:54 GRAPPA 2 2 Averages TA – 4:18 Brain – EPI Sequence High Resolution Matrix – 256 x 256      No PAT GRAPPAx2 GRAPPAx3 FLASH 2D No PAT TA 4:19  FLASH 2D GRAPPA 2 TA 2:25  Without iPAT TA: 4 minutes 56 seconds GRAPPA PAT 2 TA: 2 minutes 49 seconds No iPAT Phase A>>P Coils SP 3 & SP 4; No Anterior Coil GRAPPAx2 Phase A>>P Coils SP 3 & SP 4; No Anterior Coil GRAPPAx2 Phase H>>F Coil 4 Channel Large Flex GRAPPAx2 Phase A>>P Coil 4 Channel Large Flex GRAPPAx2 Phase H>>F Averages 1 Matrix 336x448 Coils SP 3-5 GRAPPAx3 Phase H>>F Averages 1 Matrix 336x448 Coils SP 3-5 GRAPPAx4 Phase H>>F Averages 1 Matrix 336x448 Coils SP 3-5 GRAPPAx2 Phase H>>F Averages 1 Matrix 336x448 Coils SP 3-5 GRAPPAx2 Phase H>>F Averages 1 Matrix 240x320 Coils SP 3-5 GRAPPAx2 Phase H>>F Averages 2 Matrix 240x320 Coils SP 3-5 iPAT Recommendations Sufficient SNR SNR is reduced when iPAT is used and Accel. Factor PE is increased Array coils required for PAT Minimum of 2 coils (receiver channels) Coil elements aligned in phase-encoding direction Coils must be ≥ than Accel. Factor PE Anatomy fully covered? Tim: Optimal Matrix Mode used? Only with B-line software Protocols labeled with iPAT (e.g., p2) Upon successful completion of this course you will be able to: Summarize Parallel Imaging Techniques and the acronyms used by various vendors Explain how iPAT works and the reasoning behind using iPAT Describe the Coil Compatibility, Calibration, PAT Mode, and parameters used to optimize iPAT Describe the iPAT Fundamentals, PAT Averaging, VIBE CAIPIRINHA technique and recommendations when using iPAT Explain the Matrix Coil Mode for B-Line software and what Maximum Acceleration factors can be used Identify the various Siemens coil designs for D and E-Line software Identify the iPAT disadvantages, advantages and recognize how to reduce iPAT artifacts Measuring with a rectangular FoV   Shorter acquisition times due to less phase-encoding steps   Slightly less SNR due to less averaging effects   Same spatial resolution  Readout Phase measured PE lines (echoes) skipped PE lines (PAT) Coil Calibration Array coil elements Under- sampled k-space coil 2        coil 1 coil 2               coil 1 GRAPPA Recon GRAPPA Recon FFT FFT Calculation of 'artificial' echoes Complete k-space FFT & array comb. Full FOV without aliasing Knowledge of the coil sensitivity profiles Coil Calibration Array coil elements coil 2               coil 1 Under- sampled k-space Knowledge of the coil sensitivity profiles FFT FFT FFT RecFoV with aliasing SENSE Recon Equation System Full FOV without aliasing measured PE lines (echoes) skipped PE lines (PAT)   1 Readout direction in-plane = echo   Only 1 phase-encoding direction in-plane (PE)   1 possible PAT direction  Readout Phase 1 readout direction in-plane = echo Two different phase-encoding directions are used for PAT PATPE – 1 Phase encoding direction in-plane PAT3D – 1 Phase-encoding direction through-plane  (= partition-encoding direction, “3D”) Combine PATPE and PAT3D Apply Accel. factor PE both directions simultaneously Functionality called iPAT2 Measured Phase-Encoding Skipped Phase-Encoding Measured 3D-Encoding Skipped 3D-Encoding   Phase Readout 3D Symphony Non-Tim systems only – Coils turned On in Phase Oversampled area PhOv area   PhOv area (1) Every nth point in phase-encoding direction is acquired (where n = Accel. factor PE)   (2) Every nth point in slice-encoding direction is acquired (where n = Accel. factor 3D)   (3) With CAIPIRINHA, the acquired pattern can be shifted in slice-encoding direction. The relative shift of measured neighboring slice-encoding lines is given by the Reordering Shift 3D. 1 2 3 Coils Coils TA = 11s GRAPPA PAT = 3 in PE GRAPPA PAT = 3 in 3D CAIPIRINHA PAT 3 (1x3_1) CAIPIRINHA PAT 3 (1x3_2) FoV coverage 50 cm with CP Mode and 4 RF channels 4 CP Coil Elements 1 RF Channel per Coil Cluster 3 Coil Elements of a Cluster "act" like 1 CP Element 3 RF Channels per Coil Cluster 3 Coil Elements of a Cluster "act" like 3 CP Elements FoV coverage 50 cm with Triple Mode and 12 RF channels PAT L>>R 3 12 Coil Elements 2 RF Channels per Coil Cluster 3 Coil Elements of a Cluster "act" like 2 CP Elements 8 CP Coil Elements FoV coverage 50 cm with Dual Mode and 8 RF channels PAT L>> R 2