Measurement of regional cerebral blood flow using iodine-123-IMP and SPECT with and without partial volume correction

Objectives: N-isopropyl-p-[123I] iodoamphetamine (123I IMP) has been widely used as a cerebral blood flow (CBF) tracer for single photon emission computed tomography (SPECT). The autoradiographic (ARG) method that employs a two-compartment model has been proposed as a method of quantitating CBF using 123I IMP, in which distribution volume (Vd) is assumed to be uniform in the brain and an arterial input function is obtained by the calibration of standard input function with one-point arterial blood sampling. Several investigators reported CBF in normal subjects measured by the 123I IMP ARG method, although there were differences in the normal CBF among various institutions. Many authors have reported differences in regional CBF between young and old subjects, males and females as well as morphological differences between them. Due to a relatively poor spatial resolution, SPECT as well as positron emission tomography (PET) images suffer from partial volume effects (PVEs). Recently, Methods for the correction of PVEs have been reported for PET, and applied to SPECT using the 99mTc-ethylcysteinate dimer. It is expected that apparent differences resulting from regional morphological differences of the brain would be corrected and functional or physiological differences could be detected following PVE correction. Some authors reported that PVE correction could differentiate a pathological CBF decrease due to a neurodegenerative disease from an apparent CBF decrease with normal aging, in the present study, we aimed to determine normal CBF values obtained using 123I-IMP and the ARG method in elderly subjects with and without PVE correction using magnetic resonance (MR) imaging, and to examine the regional differences in effects of PVE correction. Methods: Thirty-three healthy subjects (18 males and 15 females; mean age: 64.8±6.5 years) participated. SPECT images were obtained at 30 min of mid-scan-time after an intravenous infusion of 111 MBq 123I-IMP, lasting for 1 rain, and a three-head SPECT scanner (MultiSPECT III, Siemens, Germany) equipped with a fan beam collimator. Attenuation correction was carried out numerically by assuming the object shape to be an ellipse for each slice and the attenuation coefficient to be uniform (0.08/cm). Correction for scatter photons was not performed. Quantitative CBF images were calculated by the ARG method with one-point arterial blood sampling from the brachial artery, which was performed 10 min after IMP infusion. Vd was fixed at 40 ml/ml. All MR imaging studies were performed using a 0.5 Tesla MR system (Signa, GE, U.S.A.). A three-dimensional volumetric acquisition of a Tl-weighted gradient echo sequence produced a gapless series of thin transverse sections using an SPGR sequence. Image processing was performed using SPM2. The correction of SPECT data for PVEs was performed as follows. Each MR image was segmented into the grey matter (GM), white matter (WM) and cerebrospinal fluid space (CSF) images. Those were then co-registered and resliced to adjust voxel size to corresponding SPECT images, following convolution with a 7 mm FWHM isotropic Gaussian kernel approximating the SPECT spatial resolution. PVE-corrected CBF images were produced using the convoluted grey and white matter MR images. The CBF images before and after the PVE correction were then spatially normalized. ROI analysis was performed using WFU-PickAtlas software. Twelve ROIs were placed in each hemisphere, and averaged for homologous ROIs in both hemispheres. The correction factor (CF), which was defined by dividing CBF after PVE correction by that before PVE correction, was measured for these ROIs, as well as CBF before and after PVE correction. Differences in CBF before and after PVE correction and CF among ROIs and between genders were assessed using the analysis of variance (ANOVA). The effects of age on CBF before and after PVE correction and CR were also examined. Results: CBF for the grey and white matter; and PVE-corrected CBF for the grey matter were 24.1 ±1.8 ml/100 g/min, 18.2±1.4 ml/100 g/min, and 40.9±2.8 ml/100 g/min in males, and 31.6±2.5 ml/100 g/min, 22.9±1.8 ml/100 g/min and 54.0±4.2 ml/100 g/min in females, respectively. Females showed statistically significantly higher CBF than males for both before and after PVE correction, but there was no gender effect on CF. There were satistically significant regional differences among ROIs in CBF before and after PVE correction, and in CE CBF before PVE correction was the lowest in the occipital cortex (22.4±5.5 ml/100 g/min) and the highest in the precuneus (33.0±7.6 ml/100 g/min), whereas after PVE correction, it was the lowest in the inferior occipitotemporal cortex (38.4±8.0 ml/100 g/min) and the highest in the precuneus (57.5±12.0 ml/100 g/min). CF was the lowest m the cerebellum (1.37±0.05) and the highest in the superior frontal gyrus (1.92±0.10) and the occipital cortex (1.92±0.11). There was no statistically significant age effect on CBF before and after PVE correction and on CF. Conclusion: The normal CBF values of this study were lower than those obtained by previous studies, which could be caused by the difference in the spatial resolution and scatter component of the SPECT scanner and reconstruction algorithm used. There was a regional difference in the effect of PVE correction measured using CF. CF tended to be low in the inferior part of the brain and high in the superior part of the brain. This regional difference in CF could be a result the regional differences in the volume of the gray matter contained within each voxel Although PVE-corrected CBF values were higher than those before PVE correction, these values remained lower than those obtained using 133Xe and H215O PET, which could be induced by the limited first-pass extraction of IMP. (authors)