Blood Water Dynamics

Transcription

1 Bioengineering 208 Magnetic Resonance Imaging Winter 2007 Lecture 8 Arterial Spin Labeling ASL Basics ASL for fmri Velocity Selective ASL Vessel Encoded ASL Blood Water Dynamics Tissue Water Perfusion: Brain: 0.0s - Heart: 0.0s - Kidney: 0.0s - Lung: 0.0s - Skeletal Muscle: s - E~95% in brain 0-0s 0.5-s s Arteries Capillaries Veins?

2 Arterial Spin Labeling Inflow and T Decay Tag by Magnetic Inversion Acquire image of tissue + tagged blood Control Acquire image of tissue + relaxed blood ASL Signal = Control - Tag! CBF EPISTAR and FAIR EPISTAR FAIR RAW EPISTAR FAIR CONTROL - TAG TI (ms)

3 ASL vs Dynamic Susceptibility Contrast Tracer Kinetics Invasiveness Inherently sensitive to: Can also measure: Problems: Applications: ASL Magnetically modified water (T /2 =s) Delivery None CBF Transit delay, CBV Transit delay in pathology Low SNR fmri, tumors, vascular reserve, neurodegeneration, psychiatry, stroke, CAD? DSC Gd-DTPA Bolus passage Low CBV MTT, CBF AIF, high flow, vascular permeability stroke, tumors, CAD Types of ASL Imaging Plane CASL Flow Dependent Inversion Plane PASL Pulsed Inversion Slab Global Velocity Selective Tag VSASL

4 Continuous vs Pulsed ASL Imaging Plane M z T Decay CASL Flow Dependent Inversion Plane Bolus Width Delay Image CASL vs PASL Larger ASL signal Higher SAR MT effects Single tag plane Tagging Efficiency CASL M z Good Pulses Simulate Efficiency Measure Efficiency PASL

5 Pseudo-continuous ASL n Garcia et al - ISMRM 2005 Pseudo-continuous vs Continuous: Higher gradient during RF pulses -> MT Effects dramatically reduced Easier to implement Slightly higher SAR Transit Delay - PASL #M Task Rest "& & TI (s) TI < "t : #M($%)=0 "t < TI < "t+& : #M($%)=2' 0B (TI-"t) CBF e -TI/TB "t+& < TI : #M($%)=2' 0B (&) CBF e -TI/TB

6 SPATIALLY VARYING DELAY 500 ms ANATOMICAL CALCULATED PERFUSION DELAY MAP 0 ms QUIPSS II TI #$% INVERSION TAG AFTER SATURATION ACQUIRE IMAGES If TI <& and #TI>"t then #M=2' 0B TI f e -TI 2/T B

7 Water Exchange Tissue Water Extraction ~95% Capillary Transit ~s Arteries Capillaries Veins On the T time scale, tag behaves like short lived microspheres TI B=22 T shift changes relaxation rate of tag during TI B=0 T 2 /T 2 * shift changes ASL signal during image acquisition Slice Profile Effects EPISTAR Mz Good Pulses Presaturation Background Suppression FAIR Mz

8 Tagging Geometry Know the anatomy Flow Dispersion Velocity Distribution W = width / V Looks pretty plug-like Different in pathology? 0 V 2V V Activity Rest Plug W Laminar

9 ASL in Different Tissues Brain T close to blood Long T 2 Easy to image Wong et al UCSD Lung No tissue water pool Hopkins, Levin et al UCSD Muscle Heart T lower than blood, exchange more important It moves, it perfuses in fits and starts with cardiac cycle An et al NYU Frank et al UCSD Towards Cardiac ASL at UCSD Brain ASL using balanced SSFP Tag Timing Systole Diastole Systole Diastole Tag Image Tagging Geometry First Results

10 CBF and BOLD Time Series RAW CBF BOLD For Pulsed ASL with Presat: Control - Tag = ASL Signal Control + Tag = Tissue Signal M z t Simultaneous Flow and BOLD fmri Anatomy CBF Change BOLD Change

11 Estimation of CMRO 2 Changes with Combined CBF and BOLD measures Hypercapnia Contralateral Stefanovic et al 2004 Ipsilateral T. Liu, UCSD

12 ASL with very low task frequencies - Wang et al., MRM 2003 Transit Delays in Diagnostic ASL Problem: Long transit delays (delay >>T ) limit the diagnostic potential of ASL in some disease states Solution: Apply tag pulses that are not spatially selective but velocity selective

13 Velocity Selective ASL Tagging based purely on velocity Tag has no spatial selectivity Transit delay eliminated M z 0 Image Tag 0. 0 Velocity (cm/s) Control Tag READOUT 90 x 80 y -90 x Tag Time 0. 0 Velocity (cm/s) Velocity Profile Effects Control M z Image Tag 0 Tag 0. 0 Velocity (cm/s) CBF * T v

14 Velocity Profile Effects Control M z Image Tag 0 Tag 0. 0 Velocity (cm/s) CBF * T v Velocity Profile Effects Control M z Image Tag 0 Tag 0. 0 Velocity (cm/s) CBF * T v

15 3D Velocity Selective ASL VS Tag/Control 3D Spiral FSE Global Sat. Inversion pulses for background suppression Wong et al, Velocity Selective ASL, MRM, 2006 VS-ASL Pulse Design Requirements: Good velocity selectivity Flat response below cutoff velocity Off resonance insensitive B insensitive Approaches: Norris (( -G -80-G -( 2 -G 2-80-G 2 ( n ) de Rochefort ISMRM 2004 p.96 VS inversion? Hyperechoes (( -G -( 2 80 (-( 2 )-G -(- ( )) B insensitive(bir-4) pulse elements (3T and up) Optimize flip angles, phases, and gradients

16 Velocity Selective ASL Vascular Territory Imaging Hendrikse et al MRM

17 Vascular Territory Imaging - Abstract #667 Guenther Regional Perfusion Imaging - Abstract #669 Zimine et al Vessel Encoded Pseudo-Continuous ASL CYCLE : ) i = ) z CYCLE 2: ) i = ) z + (i%2)* CYCLE 3: ) i = ) z + ) xy CYCLE 4: ) i = ) z + ) xy + (i%2)* ) z = i+g z tz ) xy = (i%2)*(a b) Efficiency (C-T)/2 Control V=5-40cm/s Tag

18 Vessel Encoded Pseudo-Continuous ASL y=ax -> x=a + y Conventional ASL: " y = y % " $ ' A = ( % $ ' # & # & y 2 " x = $ V% ' # S& " V $ % " ' = A + y = 0.5 * ( %" $ ' y % $ '. # S& # & # & y 2 3 Vessel Encoding (Guenther): " $ $ $ $ # y y 2 y 3 y 4 % "( ( % " R% ' $ ' $ ' ' ( ( = $ ' $ L ' ' $ ( ( ' $ B' ' $ ' $ ' & # & # S& -> " R% "( ( % " $ ' $ ' $ $ L ' = 0.25 * $ ( ( ' $ $ B' $ ( ( ' $ $ ' $ ' $ # S& # & # y y 2 y 3 y 4 % ' ', ' ' & Vessel Encoded PCASL Cycle Vessel - - Vessel 2 - -

19 Vessel Encoded PCASL Summary - Vessel Encoded ASL High efficiency Hadamard encoding strategy Pseudo-Continuous Tagging: High SNR Vessels selected at one point in space Applications?

20 Remote Detection - Seeley et al JMR 2004 S(t) " # M t ( r )e ir k (t ) r dr Excite Detect Decode Image Excite and Phase Encode M " # M t ( r )e ir k r dr FLOW Detect Decode Image S " M Application Specific Brain ASL Methods Application Quantitative CBF - Normal to High Flow Flavor of ASL Pseudo-continuous Why? High SNR Simultaneous CBF / BOLD Quantitative - Low Flow Vascular Territory Pulsed Velocity Selective Vessel Encoded Pseudo-CASL Good CBF/BOLD separation Transit Delay Insensitive High SNR efficiency