点击显示 收起
the Department of Clinical Neurosciences (J.V.G., P.S.J., F.I.A., C.J.P., E.A.W., J.-C.B.), University of Cambridge, UK
the Stroke Unit (E.A.W.), Addenbrooke’s Hospital, Cambridge, UK
the Wolfson Brain Imaging Centre (T.D.F., O.B., F.I.A., T.A.C., J.H.G., J.-C.B.), University of Cambridge Clinical School, UK
the Department of Medicine (C.J.P.), University of Cambridge Clinical School, UK
the Department of Radiology (J.H.G.), University of Cambridge Clinical School, UK.
Abstract
Background and Purpose— MR is widely used to depict still ischemic but viable tissue in acute stroke. However, the relationship between the apparent diffusion coefficient (ADC) and energy failure from reduced oxygen supply are unknown in man.
Methods— Acute carotid-territory stroke patients were studied prospectively with both diffusion tensor–imaging and back-to-back steady-state 15O-PET. Substantial numbers of voxels with oxygen extraction fraction >0.70 (ie, significant ongoing hypoxia) were identified in 3 patients (imaged at 7, 16 and 21 hours after stroke onset). In this voxel population, the quantitative relationships between the ADC and cerebral metabolic rate of oxygen (CMRO2), and ADC and cerebral blood flow (CBF), were assessed.
Results— The ADC remained essentially unchanged until CBF reached values 20 mls/100g per min, beyond which it declined linearly. In contrast, except when severely reduced, the ADC was a poorer predictor of CMRO2. For both CBF and CMRO2, however, the relationship with ADC became steeper with longer times since onset, ie, the same ADC reflected lower perfusion and CMRO2 with elapsed time.
Conclusions— Despite the small sample and late times from stroke onset, the findings indicate that the degree of restricted water diffusion reliably reflects the severity of oxygen deprivation below the penumbral threshold but is less strongly related to metabolic disruption, which may explain why the ADC does not reliably predict tissue outcome. However, the same degree of diffusion restriction may correspond to greater severity of tissue disruption with elapsing time, which has relevance for stroke therapy. Time elapsed since stroke onset should be taken into account when interpreting ADC declines and in voxel-based infarct prediction models.
Key Words: cerebral blood flow cerebral ischemia diffusion magnetic resonance imaging PET stroke
Introduction
Reduced oxygen supply, the hallmark of cerebral ischemia, translates with positron-emission tomography (PET) as increased oxygen extraction fraction (OEF), with 3 operationally-defined high OEF compartments being usually present at any one time after acute ischemic stroke, namely "benign" oligemia, penumbra and core.1–3 MR-based diffusion-weighted (DWI) and perfusion-weighted imaging are increasingly used in the clinical setting to identify the at-risk tissue in acute stroke and guide therapeutic decisions. The DWI lesion provides early evidence of cellular injury, whereas the significantly hypoperfused but DWI-intact tissue, so-called ‘mismatch’, points to the at-risk tissue4 and corresponds well to the high OEF area.5 The DWI lesion is attributable to a reduction of the apparent diffusion coefficient (ADC), related to a water shift from the extra- to the intracellular space caused by energy failure from reduced oxygen supply.4 However, the relationship between ADC changes and energy failure in human stroke is largely unknown. Establishing this relationship for the tissue with high OEF would be important as the level of oxygen consumption in ischemic stroke is a good marker of subsequent infarction.1,2,6–8
Experimental and human studies9–12 have shown that the ADC remains almost unchanged with decreasing cerebral blood flow (CBF) until a certain level of perfusion close to the penumbra threshold is reached, beyond which it declines steeply. However, the corresponding relationship between the ADC and oxidative metabolism has not been demonstrated. The potential normalization of even markedly decreased ADC with reperfusion in both the rat13 and man14–16 highlights the well-reported poor value of the ADC to predict final infarction.14–17 In turn, because the cerebral metabolic rate of oxygen (CMRO2) is a good marker of irreversible damage, it is suggested that the ADC may be less tightly related to the severity of oxidative metabolism impairment than it is to perfusion.18,19 One group only20 has directly studied the ADC-CMRO2 relationship using PET during MCA occlusion in the anesthetized pig. They reported a linear relationship and concluded that the ADC is an indicator of tissue oxygen metabolism. However, extrapolation of this study to human stroke is uncertain. Two previous clinical articles have reported DWI-CMRO2 comparisons in acute stroke,7,21 but neither systematically assessed the ADC-CMRO2 relationship.
Here, we aimed to study directly in humans and at the voxel level, the ADC-CMRO2 relationship, as compared with the ADC-CBF relationship, in the hypoxic tissue. Based on the available literature, we hypothesize that in ongoing hypoxic tissue the ADC is closely related to CBF below the penumbra threshold but is less clearly predictive of CMRO2. However, because it has been reported that for a given CBF value the ADC is lower with longer time since stroke onset,11 our secondary aim was to explore whether the ADC becomes more predictive of metabolic decline with increasing time since stroke onset. A (time X severity) dependence of the ADC20,22 would explain at least in part its inconsistent predictive value for subsequent infarction. We applied back-to-back PET to measure the CMRO2 and high-field MR to implement full diffusion tensor (DTI) allowing calculation of mean diffusivity (<D>), a more robust index of water diffusion.23
Materials and Methods
Subjects
Patients with first-ever carotid territory stroke within 21 hours of onset were prospectively enrolled. Main exclusion criteria were previous stroke, lacunar syndrome, hemorrhagic infarct on CT, anticoagulation or thrombolysis (as arterial cannulation was required for the PET study), inability to cooperate or organ failure. Patients were scored with the National Institutes of Health Stroke Scale (NIHSS). The Local Ethics Committee approved the study and informed consent was obtained.
MR
DTI was performed on a 3-T magnet (Med-spec s300, Bruker). The MR technique, parameters and calculation of <D> are detailed elsewhere.24 Briefly, data sets were calculated from 12 noncollinear gradient directions and 5 b values (range: 300 to 1570 seconds/mm2), which effectively characterizes fast water diffusion, avoids contamination by perfusion and reduces CSF artifacts. Except in patient 1 (see Figure 1) full-brain coverage was obtained (field of view 19x19, matrix size 100x100, 27 slices, thickness 5 mm). A T2-weighted FSE sequence was also acquired.
PET
A fully quantitative steady-state 15O PET study using H215O, 15O2 and C15O was performed according to published procedures25,26 on a GE Advance scanner with axial field of view=15.3 cm (General Electrics). Image reconstruction included 68Ge correction for attenuation, scatter, randoms and dead time. Parametric maps of CBF, CMRO2 and OEF were calculated using standard models including correction for cerebral blood volume27,28 using the software PETAN.29 In all patients transcranial Doppler was performed before and between MR and PET scanning.
Image Processing
Using PETAN, the PET parametric maps were coregistered to the T2 sequence using SPM99 (www.fil.ion.bpmf.ac.uk/spm). The average image from the b0 DTI preparation sequence was coregistered to the T2 using the normalized mutual information algorithm in vtk-CISG software package (www.image-registration.com). The matrix transformation file was applied to the b945 images and <D> maps, thus coregistering all the images. Using ANALYZE 5.0, all image modalities were set to PET voxel size (2.34x2.34x4.25 mm) for further analysis.
Data Analysis
To assess the <D>-CMRO2 (and <D>-CBF) relationship in ischemic tissue on a voxel-by-voxel basis, we used high OEF as a marker of ischemia3,21 and selected all voxels with OEF >70%30,31 in the affected hemisphere, using Analyze. The High OEF voxel set was then ‘flipped’ onto the unaffected hemisphere, and the spatial coordinates of all the selected voxels were transferred to the coregistered <D> (thresholded at 1200x10–6 mm2s–1 to eliminate CSF), CBF and CMRO2 maps. We then plotted <D> against CBF and CMRO2 for the High OEF and mirror voxel populations for each patient, and compared the relationships between the 2 hemispheres. To allow formal comparison, linear regressions were computed across the entire voxel data sets, and differences in correlation strength between the affected and nonaffected hemispheres were assessed using standard methods,32 with P<0.05 being significant.
To assess the effects of time since stroke onset, the <D>-CMRO2 (as well as the <D>-CBF) relationship was compared among patients (scanned at different times), both visually and using the slopes derived from the regression analysis. In addition, we assessed whether time influenced the <D> values for the Core, Penumbra and Oligemia subsets of the High OEF voxel population, defined by voxels with CBF <8.4, 8.4 to 20 and >20 mls/100 g per minute, respectively.3,6,31 To this end, 1-way ANOVAs taking time from stroke onset (in hours) as the independent variable and <D> (normalized to the mean of the mirror voxels to account for potentially different gray/white mix) as the dependent variable were performed separately for the 3 voxel subsets. So as to verify the findings, the probabilistic voxel-based CMRO2 threshold for the Core of 39 μmol/100 mL per minute was also applied, as this threshold yields a not entirely overlapping voxel population than with the above CBF threshold.6
Results
Clinical Data
Of a series of 6 consecutive patients who completed the imaging protocol, 3 (2 males; age 53 to 84 years; imaged 7, 16 and 21 hours after stroke onset) exhibited substantial numbers of High OEF voxels (1144, 468 and 976, respectively) and form the basis of this report. Their main clinical data are shown in Table 1. They all had extensive clinical deficits (NIHSS range: 10 to 16). In all, transcranial Doppler showed occluded MCA before MR scanning, unchanged between MR and PET. Patient 3 had a distal carotid occlusion with the stroke affecting both the MCA and posterior cerebral artery territories.
Imaging Findings
Figure 1 shows for each patient the coregistered <D>, High OEF voxels, CMRO2, CBF and T2-weighted maps for a subset of illustrative planes.
Table 2 shows the mean (and 1SD) for <D> and the PET variables for the High OEF and Mirror voxel populations in each patient. In addition to the increased OEF (expected), the <D>, CBF and CMRO2 were all significantly reduced in the High OEF as compared with Mirror voxel population for each patient.
The <D>-CBF and <D>-CMRO2 relationships within the High OEF and Mirror voxel populations for each patient are shown in Figure 2. In the Mirror voxel population, <D> was as expected essentially stable across the voxels, with the spread of CBF and CMRO2 values reflecting the fact that the data sets comprised a mix of gray and white matter voxels. Strikingly different relationships characterized the High OEF voxels not only relative to the Mirror population but also between CBF and CMRO2; these observations were consistent across patients, although with some differences. Regarding CBF, the relationship could be described as follows: (1) it was displaced leftward relative to Mirror voxels; (2) <D> started to substantially decline only below CBF values 20 mls/100 g per minute; and (3) the steepest slope was found in Patient 3, studied the latest. Regarding CMRO2, the relationship could be described as follows: (1) it was displaced not only leftward but also downward, ie, <D> values tended to be reduced across the whole dataset; (2) although they also showed a monotonic decline, the slope was much shallower than with CBF; (3) again the slope was steepest in Patient 3; and (4) in each patient, a marked decline in <D> occurred with lowest CMRO2 values. Thus, although severely reduced ADC consistently corresponded to markedly low CMRO2, moderately reduced <D> could correspond to unchanged as well as substantially reduced CMRO2 but was more likely to represent more reduced CMRO2 at later time points.
Analysis of the linear regressions confirmed these observations (Figure 3). In the unaffected hemisphere, all <D>-CBF and <D>-CMRO2 correlations were weak (explaining <10% of the variance32), whereas in the affected hemisphere they were all meaningful; significant between-hemisphere differences were found for both sets of relationships in all patients apart from a trend for the CBF relationship of Patient 1 (P<0.10). Slopes of the regression lines were very weak in the unaffected hemisphere for both relationships. In the affected hemisphere, they were steeper for the <D>-CBF than for the <D>-CMRO2 relationship, and the slopes became gradually steeper with increasing time since stroke onset.
Table 3 shows the normalized CMRO2 and <D> values for the Oligemia, Penumbra and Core voxel subsets. For each subset, there was a highly significant decrease in normalized <D> with increasing time from ictus (P<0.001, ANOVA), most pronounced for the Core. Similar results were obtained when using the CMRO2 instead of the CBF Core threshold (P<0.001, ANOVA; data not shown).
Discussion
This is the first study to combine fully quantitative PET and mean diffusivity calculated from a DTI sequence in acute stroke to examine the relationships between the ADC and oxygen consumption in hypoxic tissue, ie, the target for reperfusion therapy. Although only 3 patients contributed to this study, they all had acute MCA stroke and exhibited high OEF at scanning time, so can be considered representative. The other 3 patients who completed this demanding imaging protocol exhibited spontaneous reperfusion.33
Two sets of findings, both consistent with our working hypotheses, emerged. Firstly, <D> remained essentially unchanged with declining perfusion until CBF reached critical values, below which it linearly declined, whereas it linearly declined with CMRO2 but the relationship was shallow, ie, the ADC was a poorer predictor of CMRO2 over a wide range of values except when severely reduced, where it consistently predicted very low oxygen consumption. Secondly, the relationship between <D> and CMRO2 (and <D> and CBF) was more pronounced with time since stroke onset, ie, the same degree of water diffusion restriction corresponded to more reduced CMRO2, and hence greater tissue disruption, with elapsing time.
Our findings regarding the <D>-CBF relationship are entirely consistent with previous studies, all using MR to assess perfusion. Although experimental10,20,34,35 and clinical studies12,36 have shown the general relationship, only few have done so in a quantitative and detailed way. In gerbils, the ADC declined only when CBF fell beyond 15 to 20 mL/100 g per minute.9 In stroke patients, Lin et al,11 using a voxel-based approach, reported an abrupt ADC drop below a mean CBF threshold of 21 mL/100 g per minute (range: 15 to 24), below which it declined continuously. We now confirm these earlier results using gold-standard PET. Altogether, a reduced ADC reflects tissue that is hypoxic below the threshold that defines the penumbra–ie, at immediate risk of, or already affected by, irreversible damage.
One group only20 has directly studied the ADC-CMRO2 relationship. Using PET in a pig model, these authors reported a linear relationship between relative ADC and CMRO2 during MCA occlusion, and concluded that the ADC is a reliable indicator of tissue oxygen metabolism. Our results showing a linear relationship between absolute CMRO2 and <D> would be consistent with this earlier report. However, in the present study the observed CMRO2 gradients were shallower, such that moderate ADC reductions were weakly predictive of concomitant CMRO2, whereas severe ADC reductions were predictive of marked CMRO2 reductions. Close inspection of Sakoh et al’s data (Figure 420) indicates that in fact the behavior of the ADC-CMRO2 relationship is largely consistent between the 2 studies, as their data also show linearity in the moderate levels of ischemia until relative CMRO2 50% is reached (corresponding to a relative ADC 75%), whereas below this level the CMRO2 appeared to decline precipitously in the face of small changes in ADC, to reach baseline for ADC 65%. Our data (Figure 2) also suggest a steeper ADC-CMRO2 relationship for markedly reduced CMRO2 values in each patient, and the analysis of the data according to the 3 hypoxic tissue compartments also suggests marked ADC declines in the Core (see Table 3). Apart from the species, additional differences between ours and this earlier study20 include (1) the anesthesia used, which reduces oxidative metabolism37; (2) the complex cerebrovascular anatomy of the pig38; and (3) the use of large region-of-interest based on CBF thresholds rather than on high OEF as here, potentially merging tissues with different physiological states.
Our results regarding the ADC-CMRO2 relationship are also consistent with a recent PET-MR comparison study in human stroke.7 Although these authors did not systematically study this relationship, they report relative preservation of the ADC in the penumbra despite substantial decrements in relative CMRO2 (absolute values were not measured) within 6 hours of onset; however, marked ADC declines were found for relative CMRO2 values below 0.45 which corresponded to the infarction threshold. Like these authors, we find that the ADC can be well preserved in hypoxic tissue unless there is markedly reduced CMRO2 and/or irreversible damage (see Figure 2 and Table 3). Over and above this earlier study,7 we show here that the ADC declines become more reliably associated with marked reductions in oxygen metabolism with times since onset >6 hours. Also, we show these relationships at the voxel level rather than with region-of-interest, and with absolute rather than relative CMRO2.
Our results suggest that the degree of water diffusion restriction in the penumbra is a good reflection of the severity of oxygen deprivation from reduced blood supply, but is much less dependent on the actual level of current metabolic disruption. It is well established that in hypoxic conditions the oxygen metabolism starts to become disrupted only beyond a certain point of reduced oxygen supply (ie, for CBF <20 mls), whereas it is maintained with less reduced oxygen supply (ie, CBF between 20 and 50 mls).39,40 Within the penumbral range (ie, CBF between 8 to 20 mls/100 mls per minute), the OEF reaches maximal values and the CMRO2 is initially relatively preserved but declines markedly in the lower penumbral range and with elapsing time.3,7 This may explain why even moderately severe ADC reductions do not consistently predict irreversible tissue damage, and can be permanently reversed.14–16 That the ADC can be substantially restricted in hypoxic tissue despite only mild reductions in CMRO2 is intriguing and may reflect the fact that in the initial stages of brain hypoxia (ie, oligemia and moderate penumbra) anerobic glycolysis is triggered, resulting in intracellular lactic acidosis that causes a change in water diffusion.41 Conversely, in the more severe penumbra range the CMRO2 decreases rapidly but the ADC remains almost stable probably because of the relative preservation of ATP function maintaining membrane homeostasis.7 Finally, the CMRO2 declines precipitously in the core because of failure of ATP production, leading to cell depolarization and marked ADC declines.
With elapsing time since stroke onset the <D>-CMRO2 relationship became steeper, ie, <D> more reliably predicted metabolic impairment. There is no previous comparable data. There was a similar effect on the <D>-CBF relationship, entirely consistent with Lin et al.11 To tease out further this effect, we determined the <D> values for the core, penumbra and oligemia subsets of the High OEF voxels. Acknowledging the small sample, the results suggest that <D> significantly declines with elapsing time, particularly so for the core. These findings are consistent with previous animal work22,42,43 which led to the hypothesis that the ADC threshold of ischemic tissue viability may be time-dependent.20 For instance, Rohl22 demonstrated in the pig that the ADC significantly decreased over time in the caudate-putamen, taken to represent the core. In humans, it is well established that across patients the ADC within the DWI lesion declines with elapsing time.44,45 Consistent with our findings, a recent study16 found that the ADC of tissue rescued by reperfusion increased with elapsing time, suggesting the viability threshold of the ADC is time-dependent. Such a dependence on time of the relationship between the ADC and tissue damage provides a plausible explanation for the discrepancies among studies attempting to identify an ADC threshold for infarction.14–18
Although the number of patients studied was small, limiting generalized conclusions, the data suggest that interpretation of ADC declines after stroke should take into account not only the degree of ischemia but also time elapsed since onset, which has clear clinical implications. Accordingly, duration might be worth adding to ADC and/or perfusion data in infarct prediction models.46,47
A limitation of our study is the late imaging times from stroke onset, beyond most current therapeutic time windows. However, the study provides valuable insight into some of the fundamentals of the water diffusion/energy failure relationship in acute human stroke. This may have relevance in the design of future therapy trials using imaging-based patient selection, currently being extended beyond the classic 3- to 6-hour time window.48
Acknowledgments
The authors thank Diana J. Day, RN, for help with patient recruitment and imaging studies, and Dr John C. Clark, Profs DK Menon and JD Pickard and the radiographers at the WBIC for their help in this project.
Sources of Funding
This research was supported by Medical Research Council (UK) grant number G0001219 to J.C.B. The Stroke Association supports E.A.W. J.V.G. was in receipt of a Raymond and Beverley Sackler Fellowship.
Disclosures
None.
References
Baron JC. Mapping the ischaemic penumbra with PET: Implications for acute stroke management. Cerebrovascular Disease. 1999; 9: 193–201.
Heiss WD. Ischemic penumbra: Evidence from functional imaging in man. J Cereb Blood Flow Metab. 2000; 20: 1276–1293.
Furlan M, Marchal G, Viader F, Derlon JM, Baron JC. Spontaneous neurological recovery after stroke and the fate of the ischemic penumbra. Ann Neurol. 1996; 40: 216–226.
Baird AE, Warach S. Magnetic resonance imaging of acute stroke. J Cereb Blood Flow Metab. 1998; 18: 583–609.
Sobesky J, Zaro Weber O, Lehnhardt FG, Hesselmann V, Thiel A, Dohmen C, Jacobs A, Neveling M, Heiss WD. Which time-to-peak threshold best identifies penumbral flow A comparison of perfusion-weighted magnetic resonance imaging and positron emission tomography in acute ischemic stroke. Stroke. 2004; 35: 2843–2847.
Marchal G, Benali K, Iglesias S, Viader F, Derlon JM, Baron JC. Voxel-based mapping of irreversible ischaemic damage with PET in acute stroke. Brain. 1999; 122 (Pt 12): 2387–2400.
Shimosegawa E, Hatazawa J, Ibaraki M, Toyoshima H, Suzuki A. Metabolic penumbra of acute brain infarction: A correlation with infarct growth. Ann Neurol. 2005; 57: 495–504.
Powers WJ, Grubb RL Jr, Darriet D, Raichle ME. Cerebral blood flow and cerebral metabolic rate of oxygen requirements for cerebral function and viability in humans. J Cereb Blood Flow Metab. 1985; 5: 600–608.
Busza AL, Allen KL, King MD, van Bruggen N, Williams SR, Gadian DG. Diffusion-weighted imaging studies of cerebral ischemia in gerbils. Potential relevance to energy failure. Stroke. 1992; 23: 1602–1612.
Pierce AR, Lo EH, Mandeville JB, Gonzalez RG, Rosen BR, Wolf GL. MRI measurements of water diffusion and cerebral perfusion: Their relationship in a rat model of focal cerebral ischemia. J Cereb Blood Flow Metab. 1997; 17: 183–190.
Lin W, Lee JM, Lee YZ, Vo KD, Pilgram T, Hsu CY. Temporal relationship between apparent diffusion coefficient and absolute measurements of cerebral blood flow in acute stroke patients. Stroke. 2003; 34: 64–70.
Hamon M, Marie RM, Clochon P, Coskun O, Constans JM, Viader F, Courtheoux P, Baron JC. Quantitative relationships between ADC and perfusion changes in acute ischemic stroke using combined diffusion-weighted imaging and perfusion MR (DWI/PMR). J Neuroradiol. 2005; 32: 118–124.
Li F, Han SS, Tatlisumak T, Liu KF, Garcia JH, Sotak CH, Fisher M. Reversal of acute apparent diffusion coefficient abnormalities and delayed neuronal death following transient focal cerebral ischemia in rats. Ann Neurol. 1999; 46: 333–342.
Fiehler J, Foth M, Kucinski T, Knab R, von Bezold M, Weiller C, Zeumer H, Rother J. Severe adc decreases do not predict irreversible tissue damage in humans. Stroke. 2002; 33: 79–86.
Kidwell C, Saver J, Mattiello J. Thrombolytic reversal of acute human cerebral ischaemic injury shown by diffussion/perfusion magnetic resonance imaging. Ann Neurol. 2000; 462–469.
Loh PS, Butcher KS, Parsons MW, MacGregor L, Desmond PM, Tress BM, Davis SM. Apparent diffusion coefficient thresholds do not predict the response to acute stroke thrombolysis. Stroke. 2005; 36: 2626–2631.
Warach S. Tissue viability thresholds in acute stroke: The 4-factor model. Stroke. 2001; 32: 2460–2461.
Fiehler J, Fiebach JB, Gass A, Hoehn M, Kucinski T, Neumann-Haefelin T, Schellinger PD, Siebler M, Villringer A, Rother J. Diffusion-weighted imaging in acute stroke–a tool of uncertain value Cerebrovasc Dis. 2002; 14: 187–196.
Gass A, Niendorf T, Hirsch JG. Acute and chronic changes of the apparent diffusion coefficient in neurological disorders–biophysical mechanisms and possible underlying histopathology. J Neurol Sci. 2001; 186 Suppl 1: S15–23.
Sakoh M, Ostergaard L, Gjedde A, Rohl L, Vestergaard-Poulsen P, Smith DF, Le Bihan D, Sakaki S, Gyldensted C. Prediction of tissue survival after middle cerebral artery occlusion based on changes in the apparent diffusion of water. J Neurosurg. 2001; 95: 450–458.
Sobesky J, Zaro Weber O, Lehnhardt FG, Hesselmann V, Neveling M, Jacobs A, Heiss WD. Does the mismatch match the penumbra Magnetic resonance imaging and positron emission tomography in early ischemic stroke. Stroke. 2005; 36: 980–985.
Rohl L, Sakoh M, Simonsen CZ, Vestergaard-Poulsen P, Sangill R, Sorensen JC, Bjarkam CR, Gyldensted C, Ostergaard L. Time evolution of cerebral perfusion and apparent diffusion coefficient measured by magnetic resonance imaging in a porcine stroke model. J Magn Reson Imaging. 2002; 15: 123–129.
Le Bihan D, Mangin JF, Poupon C, Clark CA, Pappata S, Molko N, Chabriat H. Diffusion tensor imaging: Concepts and applications. J Magn Reson Imaging. 2001; 13: 534–546.
Green HA, Pena A, Price CJ, Warburton EA, Pickard JD, Carpenter TA, Gillard JH. Increased anisotropy in acute stroke: A possible explanation. Stroke. 2002; 33: 1517–1521.
Coles JP, Fryer TD, Smielewski P, Rice K, Clark JC, Pickard JD, Menon DK. Defining ischemic burden after traumatic brain injury using 15O PET imaging of cerebral physiology. J Cereb Blood Flow Metab. 2004; 24: 191–201.
Coles JP, Fryer TD, Bradley PG, Nortje J, Smielewski P, Rice K, Clark JC, Pickard JD, Menon DK. Intersubject variability and reproducibility of (15) O PET studies. J Cereb Blood Flow Metab. 2005; 26: 48–57.
Frackowiak RS, Lenzi GL, Jones T, Heather JD. Quantitative measurement of regional cerebral blood flow and oxygen metabolism in man using 15o and positron emission tomography: Theory, procedure, and normal values. J Comput Assist Tomogr. 1980; 4: 727–736.
Lammertsma AA, Baron JC, Jones T. Correction for intravascular activity in the oxygen-15 steady-state technique is independent of the regional hematocrit. J Cereb Blood Flow Metab. 1987; 7: 372–374.
Smielewski P, Coles JP, Fryer TD, Minhas PS, Menon DK, Pickard JD. Integrated image analysis solutions for pet datasets in damaged brain. J Clin Monit Comput. 2002; 17: 427–440.
Marchal G, Beaudouin V, Rioux P, de la Sayette V, Le Doze F, Viader F, Derlon JM, Baron JC. Prolonged persistence of substantial volumes of potentially viable brain tissue after stroke: A correlative pet-ct study with voxel-based data analysis. Stroke. 1996; 27: 599–606.
Guadagno JV, Warburton EA, Aigbirhio FI, Smielewski P, Fryer TD, Harding S, Price CJ, Gillard JH, Carpenter TA, Baron JC. Does the acute diffusion-weighted imaging lesion represent penumbra as well as core A combined quantitative PET/MRI voxel-based study. J Cereb Blood Flow Metab. 2004; 24: 1249–1254.
Snedecor GW, Cochran, William G. Statistical methods. Iowa State University Press; 1989.
Guadagno JV, Warburton EA, Jones PS, Fryer TD, Day DJ, Gillard JH, Carpenter TA, Aigbirhio FI, Price CJ, Baron JC. The diffusion-weighted lesion in acute stroke: Heterogeneous patterns of flow/metabolism uncoupling as assessed by quantitative positron emission tomography. Cerebrovasc Dis. 2005; 19: 239–246.
Quast MJ, Huang NC, Hillman GR, Kent TA. The evolution of acute stroke recorded by multimodal magnetic resonance imaging. Magn Reson Imaging. 1993; 11: 465–471.
Perez-Trepichio AD, Xue M, Ng TC, Majors AW, Furlan AJ, Awad IA, Jones SC. Sensitivity of magnetic resonance diffusion-weighted imaging and regional relationship between the apparent diffusion coefficient and cerebral blood flow in rat focal cerebral ischemia. Stroke. 1995; 26: 667–674; discussion 674–665.
Fiehler J, Knab R, Reichenbach JR, Fitzek C, Weiller C, Rother J. Apparent diffusion coefficient decreases and magnetic resonance imaging perfusion parameters are associated in ischemic tissue of acute stroke patients. J Cereb Blood Flow Metab. 2001; 21: 577–584.
Bergsneider M, Hovda DA, Lee SM, Kelly DF, McArthur DL, Vespa PM, Lee JH, Huang SC, Martin NA, Phelps ME, Becker DP. Dissociation of cerebral glucose metabolism and level of consciousness during the period of metabolic depression following human traumatic brain injury. J Neurotrauma. 2000; 17: 389–401.
Burbridge B, Matte G, Remedios A. Complex intracranial arterial anatomy in swine is unsuitable for cerebral infarction projects. Can Assoc Radiol J. 2004; 55: 326–329.
Powers WJ. Cerebral hemodynamics in ischemic cerebrovascular disease. Ann Neurol. 1991; 29: 231–240.
Sette G, Baron JC, Mazoyer B, Levasseur M, Pappata S, Crouzel C. Local brain haemodynamics and oxygen metabolism in cerebrovascular disease. Positron emission tomography. Brain. 1989; 112 (Pt 4): 931–951.
Kohno K, Hoehn-Berlage M, Mies G, Back T, Hossmann KA. Relationship between diffusion-weighted mr images, cerebral blood flow, and energy state in experimental brain infarction. Magn Reson Imaging. 1995; 13: 73–80.
Miyabe M, Mori S, van Zijl PC, Kirsch JR, Eleff SM, Koehler RC, Traystman RJ. Correlation of the average water diffusion constant with cerebral blood flow and ischemic damage after transient middle cerebral artery occlusion in cats. J Cereb Blood Flow Metab. 1996; 16: 881–891.
Hoehn-Berlage M, Eis M, Back T, Kohno K, Yamashita K. Changes of relaxation times (t1, t2) and apparent diffusion coefficient after permanent middle cerebral artery occlusion in the rat: Temporal evolution, regional extent, and comparison with histology. Magn Reson Med. 1995; 34: 824–834.
Schlaug G, Siewert B, Benfield A, Edelman RR, Warach S. Time course of the apparent diffusion coefficient (ADC) abnormality in human stroke. Neurology. 1997; 49: 113–119.
Copen WA, Schwamm LH, Gonzalez RG, Wu O, Harmath CB, Schaefer PW, Koroshetz WJ, Sorensen AG. Ischemic stroke: Effects of etiology and patient age on the time course of the core apparent diffusion coefficient. Radiology. 2001; 221: 27–34.
Wu O, Koroshetz WJ, Ostergaard L, Buonanno FS, Copen WA, Gonzalez RG, Rordorf G, Rosen BR, Schwamm LH, Weisskoff RM, Sorensen AG. Predicting tissue outcome in acute human cerebral ischemia using combined diffusion- and perfusion-weighted mr imaging. Stroke. 2001; 32: 933–942.
Shen Q, Ren H, Fisher M, Duong TQ. Statistical prediction of tissue fate in acute ischemic brain injury. J Cereb Blood Flow Metab. 2005; 25: 1336–1345.
Hacke W, Albers G, Al-Rawi Y, Bogousslavsky J, Davalos A, Eliasziw M, Fischer M, Furlan A, Kaste M, Lees KR, Soehngen M, Warach S. The desmoteplase in acute ischemic stroke trial (DIAS): A phase II MRI-based 9-hour window acute stroke thrombolysis trial with intravenous desmoteplase. Stroke. 2005; 36: 66–73.