Transfer of normal 99mTc-ECD brain SPET databases between different gamma cameras

Medical Signal and Image Processing Department (MEDISIP), Faculty of Applied Sciences, Ghent University, Belgium
European journal of nuclear medicine and molecular imaging (Impact Factor: 5.38). 03/2001; 28(4):435-449. DOI: 10.1007/s002590000461


A stereotactic, normal perfusion database is imperative for optimal clinical brain single-photon emission tomography (SPET). However, interdepartmental use of normal data necessitates accurate transferability of these data sets. The aim of this study was to investigate transfer of three normal perfusion databases obtained in the same large population of healthy volunteers who underwent sequential scanning using multihead gamma cameras with different resolution. Eighty-nine healthy adults (46 females, 43 males; aged 20-81 years) were thoroughly screened by history, biochemistry, physical and full neurological examination, neuropsychological testing and magnetic resonance imaging. After injection of 925 MBq technetium-99m labelled ethyl cysteinate dimer (ECD) under standard conditions, 101 scans were acquired from all subjects (12 repeat studies) on a triple-head Toshiba GCA-9300A (measured average FWHM 8.1 mm). Ninety-one sequential scans were performed on a dual-head Elscint Helix camera (FWHM 9.6 mm) and 22 subjects also underwent imaging on a triple-head Prism 3000 (FWHM 9.6 mm). Images were transferred to the same processing platform and reconstructed by filtered back-projection with the same Butterworth filter (order 8, cut-off 0.9 cycles/cm) and uniform Sorensen attenuation correction (=0.09). After automated rigid intrasubject registration, all subjects were automatically reoriented to a stereotactic template by a nine-parameter affine transformation. The databases were analysed using 35 predefined volumes of interest (VOIs) with normalisation on total VOI counts. For comparison, the high-resolution data were smoothed with a 3D Gaussian kernel to achieve more similar spatial resolution. Hoffman phantom measurements were conducted on all cameras. Partial volume effects after smoothing varied between -6.5% and 10%, depending on VOI size. Between-camera reproducibility was 2.5% and 2.7% for the Toshiba camera versus the Helix and the Prism database, respectively. The highest reduction in between-camera variability was achieved by resolution adjustment in combination with linear washout correction and a Hoffman phantom-based correction. In conclusion, transfer of normal perfusion data between multihead gamma cameras can be accurately achieved, thereby enabling widespread interdepartmental use, which is likely to have a positive impact on the diagnostic capabilities of clinical brain perfusion SPET.

8 Reads
  • [Show abstract] [Hide abstract]
    ABSTRACT: Reliable and high-resolution reference data for regional cerebral blood flow measured with single-photon emission tomography (SPET) are necessary for optimal clinical and research use. Therefore, a large dataset of normal technetium-99m labelled ethylene cysteine dimer (ECD) perfusion SPET in carefully screened healthy volunteers with an age range spanning six decades was created, with correction for non-uniform attenuation and scatter and based on an anatomically standardised analysis. Eighty-nine healthy volunteers, stratified for gender (46 females, 43 males; age 20-81 years), were included. Twelve volunteers underwent repeated 99mTc-ECD SPET after 2.5&#452.3 weeks. An automated whole-brain volume of interest analysis with MANOVA as well as voxelwise analysis using SPM99 was conducted. Average intersubject variability was 4.8% while intrasubject reproducibility was 3.0%. An age-related decline in tracer uptake was found in the anterior cingulate gyrus, bilateral basal ganglia, left prefrontal, left lateral frontal and left superior temporal and insular cortex (all P=0.001-0.02). There was an overall increase in right/left asymmetry with age, which was most pronounced in the frontal and temporal neocortex. The most significant correlations between AI and age decade were found in the prefrontal (R=0.35, P=0.001) and superior temporal neocortex (R=0.43, P<0.001). Women had significantly higher uptake in the right parietal cortex (P<0.001), while men showed higher uptake in the cerebellum and the left anterior temporal and orbitofrontal cortex (all P<0.01). This normative dataset allows age- and gender-specific patient and group assessment of 99mTc-ECD perfusion SPET under a wide variety of clinical circumstances in relation to normal variations and highlights the importance of both age- and gender-specific normal datasets for optimal analysis sensitivity.
    European journal of nuclear medicine and molecular imaging 01/2001; 28(7):873-887. DOI:10.1007/s002590100549 · 5.38 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Radionuclide angiography with technetium-99m ethyl cysteinate dimer (ECD) allows non-invasive estimation of absolute cerebral blood flow (CBF), either by graphical Patlak-Gjedde analysis (PGA) or by spectral analysis (SA). Other methods estimate CBF by means of single-point arterial or venous sampling. The aim of this study was to evaluate radionuclide scanning and single-point venous sampling as potential clinical non- to minimally invasive methods for CBF determination in a large set of carefully screened healthy volunteers over the adult age range. Eighty-three carefully screened healthy volunteers (20-81 years, 43 males, 40 females) underwent planar radionuclide angiography with 925 MBq 99mTc-ECD. After correction for camera dead-time loss, hemispheric CBF was calculated from brain perfusion indices (BPI): BPI(G) for PGA and BPI(S) for SA. Of the volunteers, 49 also underwent venous sampling 6 min post injection, from which the lipophilic octanol extraction fraction and hemispheric brain fractionation index (BFI) were determined. All datasets were correlated and evaluated as a function of age and gender. Intrasubject variability for the BPI measurements was assessed in 11 volunteers by repeat study within 2 weeks of the first acquisition. Graphical and spectral analysis BPIs were strongly correlated (R=0.846, P<0.00001). This correlation coefficient increased to R=0.903 for the 74 cases in which graphical analysis was not hampered by temporal tracer retention in cervicobrachial venous valves. The BFI was weakly correlated to both BPI indices (BPI(G): R=0.34, P=0.02; BPI(S): R=0.31, P=0.04). The right hemisphere showed significant asymmetry for BPI(S) (AI=2.7%+/-4.3%, P<0.001), in correspondence with previous 99mTc-ECD data. BPI(G), BPI(S) and BFI were all inversely related to age, with an increased gradient after the age of 55 years, while there was no significant gender difference. The ratio of BPI(G) to BIP(S), which is a measure of the cerebral extraction fraction for 99mTc-ECD, was not dependent on age. Intersubject variability was 15.5% for both radionuclide scanning-based methods and 18.2% for venous sampling, and in all cases was independent of age. A much lower intrasubject variability was observed for BPI(S) (7.2%) than for BPI(G) (12.6%). This study provides reference values for normal perfusion indices assessed by graphical and spectral analysis. The results also indicate that spectral analysis allows the most reproducible estimate of hemispheric perfusion by means of an operator-independent and objective approach. Whereas accurate calibration of normal BPI(S) values to hemispheric CBF with established methods needs to be performed, non-invasive calculation of regional absolute CBF using 99mTc-ECD is possible by application of a linearisation algorithm.
    European Journal of Nuclear Medicine 07/2001; 28(7):862-72. DOI:10.1007/s002590100559 · 5.38 Impact Factor
Show more