Brain Oral Session: Brain Imaging

1002. Bolus-tracking arterial spin labeling: a new marker for aging and age related neurological diseases

M. Kelly, C. Blau, R. Bechara, M. Lynch and C. Kerskens

Trinity College Institute of Neuroscience, Trinity College Dublin, Dublin, Ireland

Objectives: Arterial spin labeling (ASL) can be used to provide a quantitative assessment of cerebral perfusion. The purpose of this study is to develop a quantitative ASL technique, bolus-tracking arterial spin labeling, which quantifies both the mean and capillary transit times (MTT and CTT). It is expected that these parameters will vary under varying physiological and pathophysiological conditions. In order to test this hypothesis, a study was carried out in groups of rats of different ages.

Methods: A Fokker-Planck equation of motion that describes the distribution of labeled water in the brain was derived. ¹ A bolus-tracking ASL sequence was designed to provide concentration-time curves that represent the passage of a bolus of labeled arterial water through the ROI. ROIs were selected in the cerebral cortex, hippocampus and whole brain. The solution to Fokker-Planck equation was fitted to the ASL concentration-time curves. The MTT and CTT were calculated from the first and second moments of the resultant curve respectively.^2,3 Male Wistar rats of varying age (young: 3 to 5 months, middle aged: 12 to 14 months, aged: 22 to 24 months) were used in the study.

Results: The theoretical model was found to be in excellent agreement with the experimental data for all datasets (chi-square ≈0.0001 for 20 iterations). The mean MTT value for the young, middle-aged and aged groups were 1.79s±0.14s, 1.75±0.12s and 2.2s±0.26s respectively (error = 2 std. dev.). The mean CTT value for these groups was 1.39s±0.2s, 1.32±0.21s and 1.93s±0.27s respectively. A significant difference in both the MTT and CTT between the young and middle aged groups and the aged group was identified (one-way ANOVA, P<0.01), as shown in the Figure.

Figure

Significant difference (*one-way ANOVA).

Conclusions: We have developed a new ASL protocol that is capable of consistently and non-invasively measuring both the MTT and CTT. The technique is applied here to an aging study and a marked difference in both transit times for varying age has been demonstrated. Alterations in vascular dynamics due to aging, such as increased vessel tortuosity, atherosclerosis and reduced vascular reactivity are known to affect cerebral perfusion. ⁴ Consequently, a comparison between healthy aged subjects and subjects with various neurological diseases that affect resting CBF using this technique would be of particular interest.

288. Probing perfusion water transport by Asl-Qbold technique

X. He, M.E. Raichle and D.A. Yablonskiy

Mallinckrodt Institute of Radiology, Washington University School of Medicine, St Louis, Missouri, USA

Objectives: Characterization of the distribution channels of pulsed arterial spin labeled (PASL) water among brain tissue compartments can provide important information on the dynamics of the transendothelial and transcytolemmal water transport and help in deciphering of PET and MRI perfusion experiments. Here we demonstrate that the recently proposed technique—quantitative BOLD (qBOLD), ¹ that takes into consideration intrinsic MR relaxation properties of intravascular, intracellular and interstitial tissue compartments, allows effectively addressing these questions.

Methods: This study was approved by Washington University's IRB. All images were acquired on a Siemens 3T Trio MRI scanner. A hybrid of the FAIR ² with QUIPPS ³ and GESSE (Gradient Echo Sampling of Spin Echo)^1,4 sequences was employed to acquire T2* attenuation curves from ASL-labeled water signal with inversion/labeling times TI between 1.2 and 2.2 secs. The MR ASL signal was modeled as originated from capillaries, interstitial space and intra-cellular space. The signals from interstitial space and intracellular space were characterized as exponentials with different T2* time constants. ¹ The signal from partially deoxygenated blood in capillaries was modeled as a ‘powder-type’, ⁵ accounting for distribution of frequencies resulted from different capillary orientations with respect to the external B0 field.

Results: Figure 1A shows a representative set of ASL water images acquired with TI of 1400 ms and corresponding to different gradient echo times TE. Notice the T2* decay of the ASL signal. Averaged across nine studies, T2* for interstitial and intracellular labeled water were 109±24 and 50.7±7.7 ms, respectively. T2* for capillary water was 71±40 ms, consistent with its blood oxygenation level of 75%.

Figure

The time evolution of water fractions in different compartments is shown in Figure 1B. The capillary fraction slightly increases with TI increasing. The dominant signal portion comes from the labeled water in the interstitial (extravascular-extracellular) space, averaged around 70% and declining with TI increasing while the signal fraction from intracellular space is consistently lower than 10%.

Conclusions: In this study, the ASL labeled water has been used to probe the dynamics of the transendothelial and transcytolemmal water transport in normal human brain. Our results indicate that the extravascular labeled water in PASL experiment is mainly confined within the interstitial space implying a very slow dynamic transcytolemmal water transport (estimated apparent lifetime of intracellular water on the order of tens of seconds), consistent with PET tracer kinetic studies.^6,7

150. A comparison of the evolution of the ischaemic penumbra in the SHRSP & WKY rat using diffusion and perfusion MRI

C. McCabe¹, W. Holmes¹, L. Gallagher¹, W. Gsell², D. Graham³, A.F. Dominiczak³ and I.M. Macrae¹

¹Glasgow Experimental MRI Centre, Division of Clinical Neurosciences, University of Glasgow, Glasgow; ²Biological Imaging Centre, MRC Clinical Sciences Centre, Imperial College London, London; ³BHF Glasgow Cardiovascular Research Centre, Division of Cardiovascular & Medical Sciences, University of Glasgow, Glasgow, UK

Objectives: The stroke prone spontaneously hypertensive rat (SHRSP) is a highly pertinent model of human stroke with increased sensitivity to focal ischaemia. ¹ The mechanisms underlying increased stroke sensitivity remain poorly understood. However, impaired collateral flow to penumbral tissue is thought to contribute. Our objective was to compare the size and fate of penumbral tissue in SHRSP and WKY.

Methods: Male SHRSP and WKY rats (12 to 16 weeks old, n = 8) were anaesthetized (1% to 2% isoflurane in 70:30 N₂O:O₂), ventilated and the middle cerebral artery permanently occluded by the intraluminal filament technique. Diffusion-weighted (DWI) and perfusion-weighted (PWI) MRI was performed on a Bruker Biospec 7T/30 cm system with diffusion/perfusion mismatch used to define penumbral tissue.

PWI was run on a single slice within MCA territory using an arterial spin labeling (ASL) sequence. DWI revealed ischaemic damage. DWI and PWI were run every hour from 1 to 6 h post MCAO. Apparent diffusion coefficient (ADC) and relative CBF (rCBF) maps were generated using Image J and thresholded using previously published values^2,3 to calculate mismatch area.

Results: By 1 h post-stroke the ADC derived lesion volume was significantly larger in SHRSP compared to WKY (275±75.2 versus 171±29.2 mm³; P<0.05 unpaired t-test) and increased in both strains from 1 to 6 hrs (275±75.2 to 330±57 mm³, SHRSP; 171±29.2 to 253±43 mm³, WKY; P<0.05 paired t-test between 1h and 6h). In contrast, perfusion deficit area was similar between strains at 1hr (43.6±9.7 versus 39.9±6 versus mm²) resulting in 59% less penumbra in SHRSP than WKY (Figure 1).

Figure 1

Temporal evolution of the mismatch area between the ADC and rCBF area in WKY and SHRSP rats following permanent MCAO. Data presented as mean±S.D., *indicates P<0.05 unpaired t-test (n = 8 per group).

The perfusion deficit remained unchanged in WKY (40.2 mm² at 6hrs) but increased in SHRSP (to 48 mm², P<0.05, paired t-test). In WKY 56% of penumbra was incorporated into the ADC lesion by 6 hrs. In SHRSP, the increase in both ADC lesion and perfusion deficit over time resulted in no overall change in the very small mismatch area.

Conclusions: The present study demonstrates 1) SHRSP have significantly more ischaemic damage and less penumbral tissue than WKY within 1 h of stroke onset; 2) penumbral tissue is recruited into the ADC abnormality over time in both strains; 3) The expanding perfusion deficit in SHRSP predicts greater tissue at risk of infarction. These results could have important implications for the management of stroke patients with pre-existing risk factors such as hypertension and suggest ischaemic damage could progress at a faster rate and over a longer time in the presence of hypertension.

486. MR-derived oxygen metabolic index (OMI) predicts gray matter infarction better than ADC during the hyper-acute phase of ischemic stroke

A.L. Ford¹, H. An², K.D. Vo³, A.M. Nassief¹, C.P. Derdeyn³, W.J. Powers⁴, W. Lin² and J.-M. Lee¹

¹Neurology, Washington University School of Medicine, Saint Louis, Missouri; ²Radiology, University of North Carolina, Chapel Hill, North Carolina; ³Radiology, Washington University in St Louis, St Louis, Missouri; ⁴Neurology, University of North Carolina, Chapel Hill, North Carolina, USA

Objectives: Distinguishing ‘reversibly’ from ‘irreversibly’ injured brain tissue during acute brain ischemia could potentially extend the current therapeutic window for interventions by individualizing treatment decisions. Prior PET studies have suggested that cerebral metabolic rate of oxygen utilization (CMRO₂) is a good predictor of tissue viability in acute ischemic stroke. More recent studies have suggested that MR-measured diffusion- and perfusion-weighted sequences may be able to determine tissue viability. We have developed a novel MR-imaging approach to measure oxygen metabolic index (OMI), a parameter closely related to CMRO₂. In a cohort of ischemic stroke patients, we directly compared the ability of OMI and ADC, measured within 3.5 h of symptom onset, to predict brain tissue destined to die or remain alive at 1 month.

Methods: Eight ischemic stroke patients were imaged within 2.7±0.8 h of stroke symptom onset. Six patients received intravenous tPA prior to the scan while two patients did not. Dynamic susceptibility contrast method was used to measure cerebral blood flow (CBF), and a previously validated asymmetric spin echo sequence was used to provide a measure of oxygen extraction fraction (OEF). OMI was derived from the product of CBF and OEF. All measurements were normalized to the contralateral unaffected hemisphere. Co-registration and tissue segmentation were performed to align timepoints and separate gray from white matter, respectively. Histograms were constructed to demonstrate the relative frequency of OMI (or ADC) values for tissue destined to die or survive on 1 month FLAIR imaging. ROC analyses were performed to quantify the ability of the parameter to distinguish alive and dead tissue. Areas under the ROC curves (AUC) of the two parameters were compared using paired t-tests.

Results: Representative 3-h OMI, 3-h ADC, and 1 month FLAIR maps are shown for a patient with acute right middle cerebral artery infarction (see Figure). Histograms show relative frequency of 3-h OMI (or ADC) values for ‘dead’ voxels at 1month (red curve) and ‘alive’ voxels at 1 month (green curve). The accuracy of OMI and ADC for separation of dead and alive tissue was quantified using ROC analysis. For gray matter, mean AUC for OMI was 0.78±0.10 compared to mean AUC for ADC, 0.72±0.09 (P = 0.01). For white matter, mean AUC for OMI was 0.73±0.09 compared to ADC, 0.73±0.09 (P = 0.5). For whole brain tissue, mean AUC for OMI was 0.75±0.09 compared to ADC, 0.73±0.08 (P = 0.12).

Conclusions: Hyper-acute measures of OMI were a better predictor of gray matter death than ADC and an equally accurate predictor of white matter death. Thus, MR-OMI shows promise as a potential MR imaging approach to delineate the core of an evolving infarction.

622. RNA interference against aquaporin-4 decreases the apparent diffusion coefficient in the normal brain

J. Badaut^1,2,3, S. Ashwal¹, A. Adami¹, B. Tone¹, R. Recker¹, B. Ternon³ and A. Obenaus^1,4,5

¹Department of Pediatrics, Loma Linda University School of Medicine, Loma Linda, California, USA; ²Clinical and Fundamental Neurosciences, University of Geneva, Geneva; ³Neurosurgery, Lausanne University Hospital, Lausanne, Switzerland; ⁴Radiology; ⁵Radiation Medicine, Loma Linda University School of Medicine, Loma Linda, California, USA

Aim: Diffusion weighted magnetic resonance imaging (MRI) is now widely used in human brain diagnosis. ¹ To date molecular mechanisms underlying changes in Apparent Diffusion Coefficient (ADC) signals remain poorly understood. AQP4, localized to astrocytes, is one of the most highly expressed cerebral AQPs. ² AQP4 is involved in water movement within the cell membrane of cultured astrocytes. ³ We hypothesize that AQP4 contributes to water diffusion and underlying ADC values in normal brain.

Methods: We used an RNA interference (RNAi) protocol in vivo, to acutely knockdown expression of AQP4 in rat brain and to determine whether this was associated with changes in brain ADC values using MRI protocols as previously described. ⁴ RNAi was performed using specific small interference RNA (siRNA) against AQP4 (siAQP4) and a non-targeted-siRNA (siGLO) as a control. The specificity and efficiency of the siAQP4 were first tested in vitro in astrocyte and hippocampal slice cultures. In vivo, siRNAs were injected into the rat cortex 3d prior to MRI acquisition and AQP4 was assessed by western blot (n = 4) and immunohistochemistry (n = 6). Histology was performed on adjacent slices.

Results: siAQP4 application on primary astrocyte cultures induced a 76% decrease in AQP4 expression after 4 days. In hippocampal slice cultures; we also found a significant decrease in AQP4 expression in astrocytes after siAQP4. In vivo, injection of non-targeted siRNA (siGLO) tagged with CY3 allowed us to show that GFAP positive cells (astrocytes) were positively stained with CY3-siGLO, showing efficient transfection. Western blot and immunohistochemical analysis showed that siAQP4 induced a ∼30% decrease in AQP4 expression without modification of tissue properties or cell death. After siAQP4 treatment, a significant decrease in ADC values (∼50%) were observed without altered of T2 values.

Conclusions: Together these results suggest that AQP4 reduces water diffusion through the astrocytic plasma membrane and decreases ADC values. Our findings demonstrate for the first time that astrocytic AQP4 contributes significantly to brain water diffusion and ADC values in normal brain. These results open new avenues to interpretation of ADC values under normal physiological conditions and in acute and chronic brain injuries.

Statement of financial support: Supported by the Swiss Science Foundation (FN 3100AO-108001 and 31003A-122166), Swissheart Foundation, Novartis Foundation, Department of Pediatrics Research Fund, NASA Cooperative Agreement NCC9–149.

1001. Ultra-fast high resolution computerized tomography angiography of the murine cerebrovascular system in vivo

S.J. Schambach¹, S. Bag¹, C. Isaza¹, C. Groden¹, M.A. Brockmann¹ and L. Schilling²

¹Division of Neuroradiology, ²Division of Neurosurgical Research, Medical Faculty Manheim, University of Heidelberg, Mannheim, Germany

Background: Animal models developed in rats and mice are of major importance in preclinical cerebrovascular research. Points of interest include the investigation of the morphology and function of the arterial, capillary and venous vessels. However, due to their extremely small caliber, in vivo examination of these vessels is extremely difficult. In the present study we present a new method to provide fast 3D in vivo analysis of cerebral murine vessels using volume computed tomography-angiography (vCTA).

Methods: An industrial X-ray inspection system equipped with a multifocus cone beam X-ray source and a 12-bit direct digital flatbed detector was used. The distance between the object, i.e., the animal's head and the detector is variable allowing for a zoom-like alteration of the geometric magnification. Mice were anesthetized with a ketamine/xylazin mixture (10 mg/0.5 mg per 100 g body weight) injected intraperitoneally. A 23G catheter was inserted into a tail vein and used for infusion of iodinated contrast medium (Imeron300®, 350 μl) during each scan period.

A quick scan algorithm was developed consisting of 180° rotation within 40secs and continuous recording of pictures at 30 Hz frequency (fps). During the scan procedure the animal rotates while the X-ray tube and the detector do not move. Images were reconstructed using a filtered back projection algorithm. Image analysis was performed by maximum intensity projection (MIP) and 3D volume reconstruction.

Results: All mice tolerated the infusion of the iodinated contrast agent well. The smallest voxel size of raw data to be achieved with the entire neurocranium depicted was 16 μm. The anatomy of the arteries at the base of the brain making up the circle of Willis was nicely detectable in all animals. The posterior communicating artery connecting the anterior and the posterior part of the circulation was well developed in all C57/black6-mice while it was only small or completely absent in all balb/c mice studied. In these animals the posterior cerebral artery (PCA) was more prominently expressed than in C57/black6-mice. In accord, the mean diameter of the basilar artery (BA) was suspiciously larger in C57/black-6 mice (221±27 μm; mean±S.D.) than in balb/c animals (154±4 μm). In contrast, the diameter of the internal cerebral artery (ICA) did not differ (197.3±27.4 in C57/black6 versus 181.8±38.7 μm in balb/c). Mean vessel diameters could be measured all animals. Dilatation ranging from 57% in the BA to 9.2% in the middle cerebral artery (MCA) was observed during ventilation with a hypoxic/hypercapnic gas mixture (5% CO₂, 12% O₂).

Conclusions: Ultra-fast in vivo vCTA of the murine cerebral vasculature is feasible at a resolution down to 16 μm. The technique allows the assessment of the vascular pattern and of vessel calibre changes in living mice even repeatedly, thus providing a worthwhile new tool to monitor in small laboratory animals different features of the cerebrovasculature in a non-invasive manner.