PET: An Overview

							PET: An Overview

Psychiatric Neuroimaging Program
Vanderbilt University
Didactic Session
7/24/07
James Wantuck
PET Outline
  Basic Science          Data Analysis
     Particle Physics       Tracer Modeling
     Sources of error       Binding Potential Map/
                              Compartmental Analysis
  Experimental Design       Reference Tissue Method
     Our methods            Format Wars
        Overview            Preproccesing
        Setoperone          Cluster Analysis
     Types of inquiry       Region of Interest
        Chemistry            Analysis
        Metabolism          MNI to Talairach
        Pharmacology
    Positron Emission Tomography




         Radiotracer Technique                  Coincidence
         Pos + Ele = Gamma Ray                  Background Radiation

Image from: http://dels.nas.edu/ilar_n/ilarjournal/42_3/4203_Positron.shtml
   Sources of Noise 1
   Problem - Radioactive Decay                        Solution - Collect More
    as a Poisson Distribution and                       Counts, limit on spatial
    spatial separation                                  resolution




Image from: http://wlap.physics.lsa.umich.edu/umich/phys/satmorn/2003/20030322/real/sld011.htm
    Sources of Noise 2
         Problem – Photon      Solution - Distance &
          Attenuation with       Density accounted
          Distance and Density   for




Images from: http://imd-www.es.hokudai.ac.jp/~gnishi/scatt.html
  Sources of Noise 3
 Problem -Photon Scatter               Solution - Measure Photon E,
                                         but overlap persists




Image from http://www.petnm.unimelb.edu.au/pet/detail/nucphysics.html
Sources of Noise 4

  Problem - Spatial    Solution -
   Resolution - FWHM     Incorporate
                         Structural Info
    Types of measurements
    with PET
        Neuronal Metabolism
            rCGM or rCBF
            Oxygen-15 labeled water, or fluorine-18 labeled
             fluorodeoxyglucose
        Pharmacology
            Most drugs are organic and thus can be labeled with
             Carbon-11
        Brain Chemistry
            Neurotransmitter synthesis, receptors,
             transporters, and release (sensitive down to 10-
             12M!)



Kaufman, M.J. Brain Imaging in Substance Abuse. New Jersey: Human Press Inc., 2001
 Serotonin Overview

 5-HTTLPR is
  promoter region of
  SERT gene
 SS causes less
  SERT, LaLa causes
  more SERT
 Δ in serotonin in
  synapse should alter
  postsynaptic
  receptor, 5-HT2A       http://www.humanillnesses.com/images/hdc_0
                         000_0001_0_img0054.jpg
    Setoperone v. Serotonin




    (18F)-Setoperone is an
     agonist at the 5-HT2A
     receptor with Ka = 1nM
Seeman P: Receptor Tables: Drug Dissociation Constants for
Neuroreceptors and Transporters, vol 2. Toronto, SZ Research, 1993
    Tracer Modeling
     The basis for most Pet studies –
      numerical value
     Time activity curves  Parametric
      Images – 3 methods
          Kinetic – non linear least squares
          Graphical – linear
          Equilibrium – no change over time



Kaufman, M.J. Brain Imaging in Substance Abuse. New Jersey: Human Press Inc., 2001
  Compartmental Analysis
       Reference Tissue Method
                     - A variantof the equilibrium method
                     - Uses a region that is devoid of the object of
                     interest



     Cp – [Plasma]
     Cr – [Reference]
     Cf – [Free]
     Cb – [Bound]



A A Lammertsma, C J Bench, S P Hume, S Osman, K Gunn*, D J Brooks and R S J Frackowiak. 1995. Comparison of
Methods for Analysis of Clinical [11C]Raclopride Studies. Journal of Cerebral Blood Flow & Metabolism (1996) 16, 42–52;
   The Math

   Parametric Map of BP
   BP = Bmax/Kd = k3/k4




                                Cp = Metabolite-corrected plasma concentration (kBq ml-1)
                                     Cr = Concentration in reference tissue (kBq ml-1)
                          Cf = Concentration of free (i.e. not specifically bound) ligand (kBq ml-1)
                                Cb = Concentration of specifically bound ligand (kBq ml-1)
                     K1 = Rate constant for transfer from plasma to free compartment (ml ml-1 min-1)
                         k2 = Rate constant for transfer from free to plasma compartment (min-1)
                          k3 = Rate constant for transfer from free to bound compartment (min-1)
                          k4 = Rate constant for transfer from bound to free compartment (min-1)
                  K1' = Rate constant for transfer from plasma to reference compartment (ml ml-1 min-1)
                      k2' = Rate constant for transfer from reference to plasma compartment (min-1)

A A Lammertsma, C J Bench, S P Hume, S Osman, K Gunn*, D J Brooks and R S J Frackowiak. 1995. Comparison of
Methods for Analysis of Clinical [11C]Raclopride Studies. Journal of Cerebral Blood Flow & Metabolism (1996) 16, 42–52;
Obtaining Binding Potential




        Measure binding in Rref and binding in
         ROI, and subtract out Rref to get BP in
         ROI.
A A Lammertsma, C J Bench, S P Hume, S Osman, K Gunn*, D J Brooks and R S J Frackowiak. 1995. Comparison of
Methods for Analysis of Clinical [11C]Raclopride Studies. Journal of Cerebral Blood Flow & Metabolism (1996) 16, 42–52;
  Reference Tissue Method

        Advantages:
               No arterial cannulation
               Independent of plasma curve and hence the
                tracer , so there is no error in measurement
                of metabolites
        Disadvantages
               Reference tissue must be assumed to be
                void of object of interest and to be modeled
                by only one compartment

A A Lammertsma, C J Bench, S P Hume, S Osman, K Gunn*, D J Brooks and R S J Frackowiak. 1995. Comparison of
Methods for Analysis of Clinical [11C]Raclopride Studies. Journal of Cerebral Blood Flow & Metabolism (1996) 16, 42–52;
Drawing Reference Regions
Drawing Reference Regions
                 Draw in the
                  cerebellum about
                  10 levels
                 Avoid edges,
                  spurious
                  activation, midline
                 Encompass equal
                  gray and white
Preprocessing
  Data must be converted
   into format that SPM can
   read. Use MRIcro
  Convert from raw data
   format into NIFTI files by
   using your known image
   dimensions, voxel size,
   and orientation
   information
  Our data is 256x256x170
   voxels, 1mmx1mmx1mm
   voxel size, and
   2byte/pixel volume
Preprocessing –
Normalization and
Smoothing
  Normalize the data to
   MNI (ICBM-152)
   space using the
   T1.nii SPM template
  Do the structural,
   then BP image
  Then smooth the
   data using 8mm
   Gaussian kernel
ROI Analysis




 Using a completely automated parametric
  approach within SPM
ROI Analysis

  Cluster Analysis
  SPM utilizes t-test to look for difference
   between groups in all voxels
  Subjects grouped by genetics (SS or LL)
  No regard to structure or function
  Once difference found, we must
   determine structures involved
ROI Analysis
  Brodmann’s Areas
   ROI approach
  MarsBar SPM add-
   on looks for
   statistical difference
   in ROIs, inputed by
   user
  Allows some
   structural
   significance to be
   ascertained
      MNI to Talairach?

       Uses the ICBM2TAL
        method described by
        Lancaster et al.
       4x4 Affine Transform
       MTT Algorithm for
        conversion embedded
        into GingerALE, a
        program developed by
        brainmap.org
Lancaster et al. 2007. Bias between MNI and Talaraich coordinates
analyzed using the ICBM-152 brain template. Human Brain Mapping.
 The End

 Thank you
  for listening.
  Please
  enjoy this
  animation of
  5-HT
  pathways

                   http://www.brainchannels.com/evolution/evolutionmedi
                   a/serotonin.gif