Patients

We studied 60 consecutive right-handed patients with unilateral atherosclerotic occlusion or stenosis of the ICA or MCA (table 1). They were referred to our positron emission tomography (PET) unit during 16 months for evaluation of the haemodynamic effects of ICA or MCA disease as a part of a comprehensive clinical evaluation to determine the necessity of vascular reconstruction surgery (carotid endarterectomy, carotid stenting, or superficial temporal artery to MCA anastomosis). Inclusion criteria were as follows: (1) occlusion or stenosis of the ICA (>60% diameter reduction according to the NASCET criteria) or MCA (>50% diameter reduction) as documented by conventional or magnetic resonance angiography (MRA);18 19 (2) ability to independently carry out daily life activities (score on modified Rankin scale <3); and (3) for patients with symptoms, history of transient ischaemic attack (TIA) or completed stroke involving the relevant ICA or MCA territory. TIA was defined as the development of focal symptoms of presumed ischaemic cerebrovascular origin lasting <24 h. The exclusion criteria were (1) cortical infarction detectable on routine MRI images (T1-weighted, T2-weighted or fluid-attenuated inversion recovery (FLAIR) images); (2) history of TIA or stroke in regions other than the relevant ICA or MCA territory; (3) history of vascular reconstruction surgery; (4) contralateral ICA or MCA stenosis (>50%); (5) unilateral arterial disease with extensive white-matter lesions in both hemispheres; (6) history of consumption of BZR agonists or alcohol abuse; or (7) presence of potential sources of cardiogenic embolism, including recent myocardial infarction (<3 weeks previously), known atrial fibrillation, mitral stenosis, presence of a prosthetic mitral valve, dilated cardiomyopathy, sick sinus syndrome or acute bacterial endocarditis.

In the 36 patients with symptoms, the interval between the last ischaemic event and the PET evaluation ranged from 0.5 to 41 (mean (SD), 8 (10)) months for the 14 TIA patients and from 0.7 to 92 (17 (25)) months for the 22 patients with stroke. In the 24 asymptomatic patients, arterial disease was suspected because of the findings of MRA or echoangiography performed for dizziness, vertigo or loss of consciousness, or as part of screening for cerebral arterial disease concomitant with coronary arterial disease.

To obtain a normal control database for BZR imaging, we studied 10 healthy control subjects (aged 57 (7) years, including seven men), who had no previous history of medical or psychiatric disorders and no history of consumption of BZR agonists or alcohol abuse. The ethics committee of our centre approved the study protocol, and written informed consent was obtained from both patients and control individuals.

Wisconsin Card Sorting Test

The WCST is one of the most frequently used neuropsychological tests and is the most sensitive to frontal-lobe dysfunction.17 In this study, we used the Keio–Fukuoka–Shimane version of the WCST, specifically, the Keio version of the newly modified WCST that can be performed on a personal computer (PC).20 21 In this test, four stimulus cards were shown in the upper row of the PC test screen, and 48 response cards were shown in order in the lower row. By trial and error, the participants determined which stimulus card matched each response card, based on one of the three categories of colour, form and number selected by the computer. Simultaneously, the patients were requested to state which category they selected to evaluate impaired verbal regulation. The computer responded to the answer by voice with ‘correct’ or ‘wrong.’ Once the subject learnt to sort by one rule, after six consecutive correct responses, the initial sorting principle was changed without warning, shifting to a new category. The test continued until all 48 cards were used. The calculated indices included categories achieved, total errors, perseverative errors as defined by Nelson, non-perseverative errors and difficulty maintaining sets. Perseverative errors designated exact repetitions of the immediately preceding incorrect responses, whereas non-perseverative errors included all errors other than perseverative errors. Difficulty maintaining sets designated incorrect responses after two or more consecutive correct responses. For the categories-achieved index, a value of 3 or more for subjects aged ≥65 years old and a value of 4 or more for subjects aged ≤64 years old was determined to be normal.21

Positron emission tomography measurements

We performed PET scans using a whole-body Advance scanner (General Electric Medical System, Milwaukee, Wisconsin), which permits simultaneous acquisition of 35 image slices with interslice spacing of 4.25 mm.22 A transmission scan was performed using ⁶⁸Ge/⁶⁸Ga for attenuation correction in each subject before the tracer administration. In reconstruction of PET data using filtered back projection, images were blurred to 6.0 mm full width at half maximum in the transaxial direction using a Hanning filter.

First, a series of ¹⁵O-gas studies was performed.22 C¹⁵O₂ and ¹⁵O₂ were continuously inhaled through a mask. The one scan time was 5 min. Bolus inhalation of C¹⁵O with scanning for 3 min was used to measure the cerebral blood volume (CBV). Arterial blood was sampled two or three times during each procedure of the ¹⁵O-gas study. After the ¹⁵O-gas study, a ¹¹C-flumazenil (FMZ) study was performed.7 23 ¹¹C-FMZ was synthesised by ¹¹C-methylation of demethylated-FMZ (Hoffmann-La Roche, Basel, Switzerland). Specific activities at the time of injection ranged from 16.2 to 20.2 (mean 17.8) GBq/μmol. After slow intravenous injection of 370–555 MBq of ¹¹C-FMZ into the right antecubital vein using an automatic injector over 60 s, a 50 min dynamic PET scan was started at the time of tracer administration with frame durations of 10 s × 6, 15 s × 8, 30 s × 4, 60 s × 5, 5 min × 4 and 10 min × 2.

We calculated the cerebral blood flow (CBF), cerebral metabolic rate of oxygen (CMRO₂) and oxygen extraction fraction (OEF) using the steady-state method.24 The CMRO₂ and OEF were corrected on the basis of the CBV.25 The binding potential (BP) of ¹¹C-FMZ was calculated using dynamic data and Logan graphical analysis with reference tissue, in which the pons was used as a reference region.23 26 We selected two tomographic planes corresponding to the pons on the average tissue activity image obtained from the early phase of dynamic PET data with time frames of 1–3 min. We manually placed on these two planes an irregular region of interest (ROI) which included almost all of the pixels that had activity in the range from the maximum pixel value to 50% of the maximum value. This method was validated using the data on normal volunteers in comparison with identification of the pons on a coregistered MRI,23 although it is ideal to set the ROIs on the coregistered MRI images in each patient. These ROIs were transferred to the dynamic PET data for the creation of parametric images of FMZ BP.

Magnetic resonance imaging

We performed MRI using a Signa unit (General Electric, Milwaukee, Wisconsin) operated at a field strength of 1.5 T. The imaging protocol consisted of a T2-weighted spin echo (repetition time/echo time ((TR/TE)=3000/88.8 ms), T1-weighted spin echo (TR/TE=550/11.2 ms) and FLAIR (TR/TE=8002/158 ms, inversion time=2000 ms) imaging series. The slice thickness was 5 mm, and the intersection gap was 2.5 mm.

Cerebral ischaemic lesions were identified as hyperintense lesions on T2-weighted MR images. FLAIR images were used to distinguish infarcts from dilated perivascular spaces. On the slice with the largest area of ischaemic lesions, the largest diameter of a lesion (A) and the largest diameter perpendicular to A (B) were measured, and A×B (mm²) was calculated as an index of the size of the subcortical ischaemic lesion. The total quantity of subcortical ischaemic lesions was calculated by summing the indices for all lesions in the hemisphere affected by arterial disease. A single investigator who was blinded to the clinical status of the patients, including other imaging data, reviewed all scans.

Data analysis

A three-dimensional stereotactic surface projection (3D-SSP) technique was used to analyse FMZ BP, CBF and CMRO₂ parametric data.27 Analysis by 3D-SSP anatomically normalises the individual PET data to the standard brain and compares the regional voxel data between patients and controls. This procedure was performed using the interface software iSSP (version 3.5; Nihon Medi-Physics Corporation, Nishinomiya, Japan). In brief, stereotactic reorientation of each brain into a standard shape is conducted using non-linear warping with the anterior commissure–posterior commissure line as the anatomic line of reference. The data are extracted by projecting the cortical activity onto the brain surface. In the standard stereotactic system, pixels located on the outer and medial surfaces of both hemispheres are predetermined along with 3D vectors perpendicular to the surface at each pixel. For each predetermined surface pixel, the algorithm searches for the highest pixel value in a direction inward along the vector to a 6-pixel depth into the cortex in an individual's anatomically standardised PET image set and assigns the maximum value to the surface pixel. Therefore, each brain was stereotactically transformed into a standard surface image format, which enabled us to compare the resultant cortical projections between the patients and controls. This 3D-SSP surface projection technique minimises the effects of cortical atrophy. For analysis of FMZ BP data, to negate the effect of fluctuations in whole-brain values and to extract changes caused by ICA or MCA disease, the pixel values of an individual's image set were normalised to the mean cerebellar value before the analysis. To quantify the decrease in FMZ BP, pixel-by-pixel Z scores were used, and Z scores ((mean normalised pixel value for controls–normalised pixel value for each patient)/SD for controls) were calculated for each surface pixel. A positive Z score relative to the controls represents a reduced FMZ BP in the patients. We considered Z scores >2 to be abnormal. To quantitatively assess the degree of FMZ BP reduction in each patient, the abnormally decreased BZR index (extent (%) of pixels with Z scores >2 compared with controls × average Z score in those pixels) in the cerebral cortex of the MCA, anterior cerebral artery (ACA) or posterior cerebral artery (PCA) territory was calculated using the stereotactic extraction estimation (SEE) method.28 According to the Talairach Daemon, anatomical information was segmented into the gyrus level. Because vascular territories vary among individuals, we arbitrarily determined the MCA, ACA or PCA territories according to the atlas prepared by Kretschmann and Weinrich.29 The MCA territory included the middle and inferior frontal gyri; precentral gyrus; superior and inferior parietal gyri; angular gyrus; postcentral gyrus; supramarginal gyrus; superior, middle, inferior and transverse temporal gyri; and superior and middle occipital gyri. The ACA territory included the superior and medial frontal gyri, orbital gyrus, rectal gyrus, paracentral lobule, subcallosal gyrus, orbital gyrus, precuneus, and cingulate and anterior cingulate gyri. The PCA territory included the inferior occipital gyrus, cuneus, fusiform gyrus, lingual gyrus, parahippocampal gyrus, posterior cingulated gyrus and uncus. For the analysis of CBF and CMRO₂ data, absolute values for each territory were calculated using the 3DSSP and SEE methods. The regional value of OEF was recalculated using the regional values of CBF, CMRO₂ and the arterial oxygen content.

Statistical analysis

Hemispheric differences in PET parameters in the three arterial territories were tested using a paired t test; significance was established at p<0.05. We compared the values of PET or clinical variables between WCST-abnormal (low categories achieved for their age) and normal groups using the Mann–Whitney U test or Fisher exact test, as appropriate; significance was established at p<0.05. Correlations between the total number of errors in the WCST and the BZR index for the subregions in the MCA or ACA territory of the affected hemisphere were analysed using multiple regression analysis after controlling for the effect of the patient's age; significance was established at p<0.005.