OtherBASIC SCIENCE INVESTIGATIONS

Experimental and Clinical Evaluation of Iterative Reconstruction (OSEM) in Dynamic PET: Quantitative Characteristics and Effects on Kinetic Modeling

Ronald Boellaard, Arthur van Lingen and Adriaan A. Lammertsma
Journal of Nuclear Medicine May 2001, 42 (5) 808-817;

Article Figures & Data

Figures

  • FIGURE 1.
    FIGURE 1.

    Image noise (COV) as function of emission scan NEC (NECEx) for Tx of 2, 5, and 30 min for FBP reconstruction (A) and OSEM 2 × 12 reconstruction (B). Results obtained with other transmission scan durations have been omitted for clarity.

  • FIGURE 2.
    FIGURE 2.

    Image noise (COV) as function of emission scan NEC (NECEx) obtained with 2 × 12, 3 × 24, and 12 × 24 iterations of OSEM.

  • FIGURE 3.
    FIGURE 3.

    (A) Image noise (COV) as function of emission scan NEC (NECEx) obtained with 5-cm-diameter and 20-cm-diameter phantoms. (B) Ratio of image noise within insert region (COVinsert) of NEMA phantom between OSEM 2 × 12 and FBP as function of image contrast, defined as ratio of activity concentration (AC) between insert and background (BG) regions.

  • FIGURE 4.
    FIGURE 4.

    Bias as function of image contrast, defined as ratio of activity concentration (AC) between insert and background regions (BG). Note that image contrast <1 corresponds to cold spots in hot background.

  • FIGURE 5.
    FIGURE 5.

    (A) Line spread functions for FBP, OSEM 1 × 16, OSEM 2 × 12, and OSEM 3 × 24. (B) Activity profiles across edge of cold insert and warm background of NEMA phantom for FBP, OSEM 1 × 16, OSEM 2 × 12, and OSEM 3 × 24.

  • FIGURE 6.
    FIGURE 6.

    Percentage difference of lung tumor MRglu values (μmol/mL/min) between OSEM and FBP data as function of MRglu values derived from FBP data using standard two-tissue compartment model with blood volume (Vb) parameter (A) and Patlak analysis (B). Data were obtained for nine lung tumors.

  • FIGURE 7.
    FIGURE 7.

    (A) Example of time–activity curve from ROI on aorta ascendens for last 35 min of FDG scan. (B) Bias of left ventricular (LV) activity concentration (AC) as function of image contrast for OSEM 2 × 12 with and without smoothing and OSEM 4 × 16 with and without smoothing.

Tables

  • TABLE 1

    Values for C1, C2, and Ratio C2/C1 for Various Reconstruction Methods

    ReconstructionC1C2C2/C1
    FBP307860.28
    OSEM 2 × 12553550.10
    OSEM 3 × 242,4601850.075
    OSEM 12 × 243.1 × 1054.6 × 1030.015

    • Values calculated with Equation 1.

  • TABLE 2

    Values of C1 and PL of Equation 4 for Various OSEM Iterations

    ReconstructionC1PL
    OSEM 2 × 125530.50
    OSEM 3 × 242,4600.43
    OSEM 12 × 243.1 × 1050.35
  • TABLE 3

    Ratio of ROI Values (OSEM/FBP) for Interval 45–60 Minutes of FDG Scan

    TissueROI2 × 124 × 16G5
    2 × 124 × 16
    MyocardiumShort axis, template ROI1.101.081.081.05
    Axial slices, isocontour ROI1.131.211.011.02
    Tumor2.5- to 5-cm diameter1.14—0.99—
    BrainGray matter0.981.040.940.99

    • G5 = with smoothing.

    • Data are given for OSEM 2 × 12, 4 × 16 with and without 5-mm FWHM Gaussian smoothing of reconstructed image. Uncertainty of these ratios is 2% (1 SD) or less.

  • TABLE 4

    Ratio of MRglu (OSEM/FBP) for Various OSEM Reconstruction Techniques

    TissueROI2 × 124 × 16G5
    2 × 124 × 16
    MyocardiumShort axis, template ROI1.111.091.081.05
    Axial slices, isocontour ROI1.091.151.021.02
    Tumor2.5- to 5-cm diameter1.14—1.01—
    BrainGray matter0.971.020.940.98

    • G5 = with smoothing.

    • Data are given for OSEM 2 × 12, 4 × 16 with and without 5-mm FWHM Gaussian smoothing of reconstructed image. Uncertainty of these ratios is 3% (1 SD) or less for brain and myocardium. For tumors, larger variation of this ratio was observed, as shown in Figure 6.

Back to top

In this issue

Print
Download PDF
Article Alerts
Email Article
Citation Tools
Bookmark this article

Jump to section

Related Articles

  • No related articles found.

Cited By...

More in this TOC Section

Similar Articles