Probabilistic map of blood flow distribution in the brain from the internal carotid artery
Introduction
Population-based structural and functional maps of the brain provide effective tools for the analysis and interpretation of complex and individually variable brain data (Toga and Thompson, 2001). Probabilistic maps of brain structures based on magnetic resonance image (MRI) data or cytoarchitectonic data have been previously developed and applied to a variety of research areas in neuroscience and medicine (Amunts and Zilles, 2001, Evans et al., 1993, Mazziotta et al., 1995). These applications include the fully automated segmentation of brain MRI and its application to partial volume correction in positron emission tomography (PET) (Collins et al., 1995, Rousset et al., 1998). The regional uptake of radiotracers in PET and in single photon emission computed tomographic (SPECT) images can be quantified by averaging regional intensities, which are weighted using the probabilistic maps of certain brain regions after they have been spatially normalized into the standardized coordinates (Kang et al., 2001). PET and SPECT images have been quantified using the spatial normalization techniques and probabilistic maps to investigate metabolic and perfusion abnormalities in epilepsy patients, age-related alternations in cerebral glucose metabolism, and functional connectivity in hearing disorders (Kang et al., 2003, Lee et al., 2001, Lee et al., 2002). The interpretation of statistical analyses performed in standard brain space is also supported by user-interactive tools designed to assess probabilistic anatomical localizations (Kim et al., 2002, Lancaster et al., 2000).
In addition to anatomical, functional, and cytoarchitectonical probabilistic maps, population-based probabilistic maps of the cerebrovascular system would also be another important tool for medical and scientific purposes. From the medical aspect, cerebrovascular probabilistic maps appear to be more applicable than any other probabilistic maps since they can be applied to numerous basic and clinical investigations related to cerebrovascular disease. The quantification of regional cerebral perfusion in the internal carotid artery (ICA) territory is especially important because the ICA supplies blood to substantial parts of the brain and is the vessel most frequently involved in major cerebrovascular diseases (Bamford et al., 1991, Johansson et al., 2000, Mead et al., 1998).
Several perfusion imaging techniques have been used to assess the effects of revascularization and regional perfusion status in cases of acute or chronic cerebrovascular occlusive disease. Contrast angiography provides anatomical information about stenosis or new blood vessel formation after bypass surgery; however, sometimes it cannot directly provide information about tissue perfusion. Accordingly, functional images representing cerebral perfusion have been developed, which include perfusion weighted or diffusion-weighted MRI (PWI or DWI), perfusion SPECT, and PET (Baird et al., 1994, Baird et al., 1997, Camargo, 2001). Perfusion SPECT is the best established of these modalities, and its use is increasing (Camargo, 2001, Giubilei et al., 1990, Laloux et al., 1995, Watanabe et al., 1999). Thus, the development of probabilistic maps of the ICA territory based on perfusion images has been encouraged. These probability maps might obviate the need of invasive study to delineate ICA territory in individual studies for the clinical routine examination of patients. Brain PET or SPECT imaging following the direct injection of radiotracer for perfusion assessment into the unilateral ICA via an arterial catheter permits the acquisition of anatomical information about the blood supply from the ICA. In addition, such radiotracers should have high extraction fractions so that it can be substantially extracted after its first pass to prevent it reaching other areas after recirculation. However, such imaging studies with invasive catheterization cannot be conducted in healthy volunteers because of the risks of infection, stroke, and intracerebral hemorrhage associated with catheterizing arterial vessels.
Therefore, in this study, we used retrospective brain SPECT images, which were acquired after injecting Tc-99m-HMPAO into the ICA during an intracarotid amobarbital procedure in epilepsy patients (IAP SPECT or Wada SPECT) (Fig. 1). This procedure was performed preoperatively to predict postoperative memory and language deficit after unilateral temporal lobectomy (Kim et al., 1999, Kim et al., 2000). One of the technical challenges associated with generating a probabilistic map of the ICA territory using IAP SPECT images was the anatomical standardization of these images since a standard IAP SPECT template was unavailable. Moreover, anatomical information for the standardization of brain was insufficient in IAP SPECT because radiotracer reaches limited brain areas in IAP SPECT images. In this paper, we present procedures for the anatomical standardization of IAP SPECT images and resulting products of probabilistic maps of blood flow distribution from the ICA and include the results of preliminary clinical applications of these probabilistic maps.
Section snippets
Subjects
Basal interictal SPECT (standard brain SPECT imaging without intervention) and IAP SPECT, data were collected retrospectively from an existing database consisting of scans performed between July 1995 and April 1996 at Seoul National University Hospital. All images were reexamined by two nuclear medicine physicians. Images with poor image quality due to a small injection dose or severe perfusion defect due to anatomical abnormalities were excluded. If any anatomical abnormality other than
Perfusion and extent probabilistic maps
Fig. 3 shows the SPM Tc-99m-HMPAO SPECT template and the spatially normalized basal and IAP SPECT images of the same subject. Spatial normalizations of the basal and IAP SPECT images were successful in all subjects who were included in this study.
Fig. 4 shows the extent probabilistic maps for the vascular territory of left and right ICAs, which are superimposed on a standard TI MRI template. Blood supplied from the ICA reached to the wide brain areas including frontal, inferior parietal, and
Discussion
The anatomical standardization of IAP SPECT images is the most critical procedure involved in the generation of the ICA probabilistic maps. To perform anatomical standardization, IAP SPECT images were coregistered to the basal SPECT and transformed into a standard space using the spatial normalization parameters obtained from basal SPECT images. Two intensity-based automatic algorithms (Friston et al., 1995, Maes et al., 1997) for the image registration did not provide acceptable results: After
Acknowledgments
This work was supported in part by the Korean Ministry of Science and Technology and in part by BK21 Human Life Sciences.
References (46)
- et al.
Classification and natural history of clinically identifiable subtypes of cerebral infarction
Lancet
(1991) - et al.
Quantification of F-18 FDG PET images in temporal lobe epilepsy patients using probabilistic brain atlas
NeuroImage
(2001) - et al.
Developmental hemispheric asymmetry of interregional metabolic correlation of the auditory cortex in deaf subjects
NeuroImage
(2003) - et al.
Microvascular anatomy of the hippocampal formation
Surg. Neurol.
(1992) - et al.
A probabilistic atlas of the human brain: theory and rationale for its development
NeuroImage
(1995) - et al.
Advances in cytoarchitectonic mapping of the human cerebral cortex
Neuroimaging Clin. N. Am.
(2001) - et al.
Reperfusion after thrombolytic therapy in ischemic stroke measured by single-photon emission computed tomography
Stroke
(1994) - et al.
Sensitivity and specificity of 99mTc-HMPAO SPECT cerebral perfusion measurements during the first 48 hours for the localization of cerebral infarction
Stroke
(1997) Hierarchical chamfer matching: a parametric edge matching algorithm
IEEE Trans. Pattern Anal. Mach. Intell.
(1998)Brain SPECT in neurology and psychiatry
J. Nucl. Med.
(2001)