点击显示 收起
the University of Western Ontario (B.D.M., T.-Y.L.), London, ON
Robarts Research Institute (B.D.M., T.-Y.L.) London, ON
Lawson Health Research Institute (B.D.M., T.-Y.L.), London, ON
London Health Sciences Centre (D.H.L., I.B.G., V.B., D.P., V.H., R.C.), London, ON
Sunnybrook Health Sciences Centre (A.J.F., D.J.S., S.E.B., R.I.A., S.S.), Toronto, ON
Foothills Medical Centre (S.B.C., A.M.D.), Calgary, AB
Ottawa Health Research Institute (M.J.H., M.G.), Ottawa, ON.
Abstract
Background and Purpose— We investigated whether computed tomography (CT) perfusion–derived cerebral blood flow (CBF) and cerebral blood volume (CBV) could be used to differentiate between penumbra and infarcted gray matter in a limited, exploratory sample of acute stroke patients.
Methods— Thirty patients underwent a noncontrast CT (NCCT), CT angiography (CTA), and CT perfusion (CTP) scan within 7 hours of stroke onset, NCCT and CTA at 24 hours, and NCCT at 5 to 7 days. Twenty-five patients met the criteria for inclusion and were subsequently divided into 2 groups: those with recanalization at 24 hours (n=16) and those without (n=9). Penumbra was operationally defined as tissue with an admission CBF <25 mL · 100 g–1 · min–1 that was not infarcted on the 5- to 7-day NCCT. Logistic regression was applied to differentiate between infarct and penumbra data points.
Results— For recanalized patients, CBF was significantly lower (P<0.05) for infarct (13.3±3.75 mL · 100 g–1 · min–1) than penumbra (25.0±3.82 mL · 100 g–1 · min–1). CBV in the penumbra (2.15±0.43 mL · 100 g–1) was significantly higher than contralateral (1.78±0.30 mL · 100 g–1) and infarcted tissue (1.12±0.37 mL · 100 g–1). Logistic regression using an interaction term (CBFxCBV) resulted in sensitivity, specificity, and accuracy of 97.0%, 97.2%, and 97.1%, respectively. The interaction term resulted in a significantly better (P<0.05) fit than CBF or CBV alone, suggesting that the CBV threshold for infarction varies with CBF. For patients without recanalization, CBF and CBV for infarcted regions were 15.1±5.67 mL · 100 g–1 · min–1 and 1.17±0.41 mL · 100 g–1, respectively.
Conclusions— We have shown in a limited sample of patients that CBF and CBV obtained from CTP can be sensitive and specific for infarction and should be investigated further in a prospective trial to assess their utility for differentiating between infarct and penumbra.
Key Words: blood volume cerebral blood flow cerebral infarction computed tomography penumbra
Introduction
The concept of ischemic penumbra was introduced by Astrup et al as a region of hypoperfused, electrically silent and functionally impaired but viable tissue.1 Since its introduction, penumbra has become the focus of intense imaging research to differentiate it from infarction. Accurate identification of this "tissue at risk" could be used to identify patients who would benefit most from treatment. Currently, tissue plasminogen activator (tPA) is the only approved drug for acute stroke patients in North America that showed significant benefit in clinical trials when administered intravenously <3 hours after stroke.2,3 It has been suggested that thrombolysis may be beneficial up to 6 hours after stroke if given intra-arterially.4 Because of the narrow time window and increased probability of hemorrhage, tPA is limited to 10% of stroke patients in most academic stroke centers and 2.4% of patients in the general population.5 Imaging techniques may help select patients beyond the current 3-hour time window for intravenous thrombolysis because it has been shown in a meta-analysis that the odds of a favorable outcome with tPA does not decrease to 1.0 until 360 minutes after onset.6
Positron emission tomography (PET) is able to differentiate between normal, penumbral, and infarcted tissue in the acute stage of stroke.7,8 However, because of logistic and practical limitations of PET, its clinical use in acute stroke is limited. MRI has also contributed significantly to stroke imaging. Although early studies suggested that various techniques such as perfusion- and diffusion-weighted imaging could differentiate between penumbra and infarct,9 more recent studies have shown that this interpretation of diffusion-weighted imaging and perfusion-weighted imaging may be an oversimplification.10,11 MRI sequences and analysis tools are constantly being developed and improved, and MRI will remain an important part of stroke imaging. These techniques are mainly limited by time in hyperacute stages when magnetic safety checklists and safety of monitoring equipment/personnel is a limiting factor.
Despite advances in various imaging modalities, computed tomography (CT) remains the most used imaging modality in acute stroke. CT scanners are readily accessed around the clock in urban and community hospitals, are relatively inexpensive compared with other imaging techniques, and studies can be performed rapidly. Recent studies have shown that contrast-enhanced CT can add to sensitivity, specificity, and accuracy of routinely performed noncontrast CT scan (NCCT)12–14 for stroke diagnosis and may be able to delineate infarcted and penumbral tissue in acute stroke.13,15,16
In this study, we attempted to determine whether cerebral blood flow (CBF) and cerebral blood volume (CBV) maps obtained by dynamic contrast-enhanced CT (CT Perfusion 3; General Electric Healthcare) at admission could be used to accurately differentiate between penumbra and infarcted brain tissue. Patients were divided into 2 groups: those who experienced recanalization at 24 hours, and those without recanalization at 24 hours. In patients without recanalization, the CBF threshold for infarction was established to define tissue that would progress to infarction without reperfusion. Logistic regression was applied to data points from patients in the recanalized group to define the thresholds for separating infarcted from penumbra. Only patients with recanalization were used in the regression analysis to minimize the effects of tissue progression from penumbra to infarction.
Materials and Methods
Patients
Thirty patients who presented to the emergency department of 4 Canadian academic centers within 7 hours of acute stroke with signs and symptoms of middle cerebral artery occlusion were prospectively recruited. Exclusion criteria were: evidence of brain stem infarcts, intracranial hemorrhage, previous stroke with residual deficit, minor stroke symptoms with National Institutes of Health Stroke Scale score <4, clinically significant hypoglycemia, impaired renal function/known allergy to contrast media, pregnancy, and <18 years of age. Twenty-five patients contributed data to this study, and their information is summarized in the Table. Five patients were excluded after enrollment for the following reasons: 2 showed no ischemia on admission and no infarct on follow-up NCCT, and 1 experienced a large hemorrhage and edema causing damage secondary to ischemia; admission CT perfusion (CTP) scan was performed at the wrong levels in 1 patient; and 1 had excessive movement, and CTP maps could not be calculated. The 25 patients who contributed data were divided into 2 groups based on the 24-hour CT angiography (CTA) data: those who showed partial or complete recanalization within 24 hours (recanalized group: 16 patients, 6 no tPA, 2 intra-arterial tPA, 4 intravenous tPA, and 4 patients received combination intravenous and intra-arterial tPA), and those who remained occluded at 24 hours (occluded group: 9 patients, 5 no tPA; 3 intravenous tPA; and 1 patient received combination intravenous and intra-arterial tPA). Results from this study did not have any effect on treatment decision, and patients were enrolled regardless of whether they received thrombolytic therapy or not. Ethics approval was obtained from participating institutions, and informed consent was obtained from all patients or family members before enrollment in the study.
Summary of Patient Information
Imaging
All patients underwent NCCT, CTA, and CTP on admission, again 24 hours after admission, and follow-up NCCT at 5 to 7 days after stroke. NCCT helical scans were performed from the skull base to the vertex with the following imaging parameters: 120 kVp, 340 mA, 4x5-mm collimation, 1 second/rotation, and table speed of 15 mm/rotation. Each CTP consisted of injection of 0.5 mL/kg (maximum 50 mL) iodinated contrast agent (Iohexol; 300 mg/mL) at a rate of 2 to 4 mL/second followed by a 45-second cine scan at 80 kVp, 190 mA. Scanning was delayed by 3 to 5 seconds after the start of the injection. CTP studies covered a 20-mm slab of the brain from basal ganglia to lateral ventricles with either 4 5-mm or 2–10 mm sections. CTA was performed as follows: 0.7 mL/kg contrast (maximum 90 mL), 5- to 10-second delay, 120 kVp, 270 mA, 1 second/rotation, 1.25-mm thick slices, and table speed 3.75 mm/rotation. CTA covered from the carotid bifurcation through to vertex. Recanalization at 24 hours was classified as complete, partial, or absent by an experienced neuroradiologist (D.H.L.).
each CTP, a time density curve (TDC), which displays the change in Hounsfield Units for a specified region over the duration of the scan, was obtained from a contralateral artery and from the superior sagittal or transverse sinus. CT Perfusion 3 software, an approved and commercially available product (General Electric HealthCare), was used to calculate quantitative parametric maps of CBF and CBV by deconvolution of tissue enhancement curves and arterial TDC in 2x2 pixel blocks. CBF and CBV derived from deconvolution-based CT imaging has been shown to be quantitative and accurate when compared with PET or Xenon CT imaging.17,18 Partial volume averaging of the arterial input curve was corrected using the venous TDC.19 In addition to CBF and CBV maps, CT Perfusion software was used to create a perfusion-weighted map by averaging cine images over the duration of the first contrast passage through the brain.
Image Analysis
All NCCT, CBF, CBV, and perfusion-weighted images were imported into custom software (IDL v5.6; RSI Inc.) for analysis. Delayed NCCT images were registered to baseline images to adjust for movement between scans. The perfusion-weighted map was used to segment gray and white matter based on thresholds of Hounsfield Units. The resulting gray matter mask was applied to the admission CBF and CBV maps to obtain average gray matter values for each of 3 tissue types: contralateral, penumbra, and infarct. CBF and CBV thresholds were applied to all tissue types to minimize the contribution of vascular pixels because it has been shown that CTP values correlate well with Xenon CT and PET when large vessels are excluded.17,18 Pixels with CBF >100 mL · 100 g–1 · min–1 or CBV >8 mL · 100 g–1 were excluded and not used in calculating average CBF and CBV values for regions of interest (ROIs).
For patients in the occluded group, ischemia was defined as CBF <25 mL · 100 g–1 · min–1, which is consistent with previously published thresholds for infarction in patients without recanalization.20 The area of ischemic gray matter on admission CBF map and final infarct size on the corresponding 5- to 7-day NCCT images were calculated and compared to determine whether this CBF threshold was appropriate for defining tissue that would progress to infarction. Because of the loss of differentiation between gray and white matter in the final infarct, a "mirrored" template of the contralateral gray matter was applied to estimate the area of infarcted gray matter.
For patients in the recanalized group, analysis was performed in the following sequence by observers blinded to the clinical data of the patients: (1) outline infarct ROI on registered 5- to 7-day NCCT (performed by an experienced neuroradiologist; D.H.L.); (2) superimpose infarct ROI on admission CTP images; (3) outline ischemic tissue (<25 mL · 100 g–1 · min–1), if present, on admission CBF map; and (4) outline contralateral hemisphere. The gray matter mask obtained from the perfusion-weighted images was applied, and regions were labeled as infarct, ischemic, and contralateral. The penumbra region was operationally defined in this study as the difference between the infarct and ischemic regions. In this study, infarct regions were defined 5 to 7 days after stroke on an NCCT image, which could lead to an underestimation of the final infarct size because ischemic/penumbral tissue may still be evolving to infarction at this time point.
Statistical Analysis
Mean and SD of National Institutes of Health Stroke Scale score, onset to imaging time, and age were calculated for all patients and are given in the Table. For patients in the occluded group, a paired t test was performed to determine whether ischemic (CBF <25 mL · 100 g–1 · min–1) area on admission was significantly different from final infarct size for the corresponding 5- to 7-day NCCT. For patients in the recanalized group, 1-way ANOVA was performed with Tukey post hoc test for significant differences in CBF and CBV values between penumbra, infarct, and contralateral regions. Significant differences were defined as P<0.05 for all comparisons. Logistic regression analysis was performed using all data points obtained from the recanalized group. CBF, CBV, and an interaction term (including CBF and CBV) were included as potential predictors in the regression model. Sensitivity and specificity for infarction were calculated from the resulting regression model. All statistical analyses were performed using SPSS 13 for Windows (SPSS Inc).
Results
Occluded Group
For patients without recanalization (average onset to imaging time 256 minutes), total ischemic area on admission CBF map was not significantly different from final infarct area, suggesting that a CBF threshold of 25 mL · 100 g–1 · min–1 was appropriate for defining tissue that would progress to infarction if recanalization did not occur (penumbra). Average admission CBF and CBV values for regions that were infarcted on the 5- to 7-day NCCT were 15.1±5.67 mL · 100 g–1 · min and 1.17±0.41 mL · 100 g–1, respectively.
Recanalized Group
Sixteen patients in the recanalized group (average onset to imaging time 133 minutes) contributed a total of 36 penumbra and 33 infarct regions (Figure 1). Penumbral tissue had a significantly lower CBF (25.0±3.82 mL · 100 g–1 · min–1) than contralateral brain tissue (37.3±5.01 mL · 100 g–1 · min–1), and infarcted tissue had a significantly lower blood flow (13.3±3.75 mL · 100 g–1 · min–1) than penumbral tissue (Figure 2A). CBV values in the penumbra (2.15±0.43 mL · 100 g–1) were significantly higher than contralateral brain tissue (1.78±0.30 mL · 100 g–1) and were significantly lower in the infarct (1.12±0.37 mL · 100 g–1) compared with contralateral and penumbral tissue (Figure 2B).
Logistic regression analysis identified the interaction term between CBF and CBV as best predictor for differentiating between penumbra and infarct data points, significantly better than CBF or CBV thresholds alone (P<0.05). A nonlinear line (Figure 1) can be derived from the logistic regression model that provides the maximum separation between penumbra and infarct data points in the CBV versus CBF plot, defined by the equation CBFxCBV=31.3, where the units of CBF are mL · min–1 · (100 g)–1 and that of CBV are mL · (100 g)–1. The model classified any pair of CBV and CBF values above this line as penumbra and points below the line as infarct. Sensitivity for infarction (97.0%), specificity for infarction (97.2%), and overall accuracy (97.1%) are obtained by comparing the predicted group membership (from the regression model) to the actual group membership (defined on the 5- to 7-day NCCT).
Discussion
In this study, penumbral regions were characterized by a mismatch between CBF and CBV, whereas infarcted areas showed a matched decrease in both parameters.13,14 In the penumbra, CBF was reduced, and CBV was maintained or elevated above contralateral values. Infarcted tissue showed a matched decrease in both CBF and CBV, which was unique to infarcted regions (Figure 3). Some patients in the occluded group showed a matched decrease (low CBF, low CBV) in some portions of ischemic tissue and mismatch (low CBF, normal/elevated CBV) in other portions. Lack ofnormal/elevated CBV) in other portions. Lack of recanalization resulted in infarct of entire ischemic (CBF <25 mL · min–1 · 100 g–1) area (Figure 3), suggesting that our CBF threshold for defining tissue that would progress to infarction was appropriate for this study. This model differs from previous work because it incorporates an interaction between blood flow and blood volume that resulted in a greater sensitivity and specificity for infarction than either CBF or CBV thresholds alone. The improved sensitivity and specificity of this model is a result of the matched decrease in CBF and CBV for infarcted tissue. Because infarct CBF is slightly lower and CBV much lower than penumbra tissue, the separation between the 2 groups of data points is maximized when the product of both values is used (Figure 2C). This method therefore minimizes the overlap between penumbra and infarct values when compared with using either a CBF or CBV threshold alone (Figure 2A and 2B). Although CBFxCBV values improve differentiation between penumbra and infarct, the CBF and CBV independently provide important information regarding the status of the tissue, particularly compared with contralateral values. This is illustrated by the fact that the CBF map alone was used to define tissue at risk of infarction and comparison of CBFxCBV shows considerable overlap between penumbra and contralateral values (Figure 2C).
Increased CBV in the penumbra is a result of direct autoregulatory responses by the brain to maintain CBF by dilating the precapillary vessels in response to decreased perfusion pressure.21 Reduced CBV in infarcted tissue has been reported previously, but the mechanism for it is not completely understood.13–15,22 One possible explanation for matched decrease in CBF and CBV is a result of failure of autoregulation in response to severe hypoperfusion.21 Other theories that have been proposed for this matched decrease are metabolic mechanisms such as neuronal death, resulting in significantly elevated extracellular potassium concentrations causing vasoconstriction.23 CBV thresholds for infarction identified in this study were lower than those defined by Wintermark et al13 but could be a result of vascular pixel elimination or algorithm differences for calculation of CBV and CBF, possibly because of partial volume corrections. Additionally, the CBV threshold we identified varies with CBF and is not constant across various CBF values. Using logistic regression (Figure 1), sensitivity and specificity for infarction were 97.0% and 97.2%, respectively, comparable to, or higher than, previous studies using MRI or contrast-enhanced CT.11,12,14 Therefore, this model could possibly be applied to ischemic regions in the acute stage of stroke to determine whether they are viable (penumbra) or infarcted.
This study is not without limitations. First, CBF and CBV thresholds were derived from an ROI analysis, and the exact time of recanalization could not be assessed within the first 24 hours, possibly allowing some conversion of penumbra to infarct between the admission CTP and recanalization. A second factor that may have caused an overestimation of the thresholds for infarction is the variability in onset to imaging time, as it is known that the CBF threshold for infarction increases with time.20 Applying our thresholds would therefore result in an overestimation of infarct size for some patients, particularly those at earlier time points. In thrombolysis treatment of acute stroke, with the possibility of hemorrhagic complication, this overestimation would make our method more conservative than optimistic with respect to treatment decision. Additionally, defining infarcted tissue at 5 to 7 days after stroke on an NCCT image could result in errors because the tissue may still be evolving to complete infarction beyond the first week. Limitations of this CT imaging technique are limited anatomical coverage, use of iodinated contrast agent, and exposure to x-ray films. Advances in CT hardware and imaging techniques will lead to increased anatomical coverage or a reduction in radiation dose.24,25 Studies have shown very good tolerance to iodinated contrast by patients,26 and the x-ray dose from a CTP study is about double that of a whole-head NCCT (T.-Y.L., unpublished data, 2005).
Defining the penumbra and infarct using CBF and CBV values from CTP could help in selecting patients for thrombolytic therapy within and possibly outside the current 3- to 6-hour treatment window, where it has been shown that penumbra may persist for >12 hours.27,28 CT imaging is available around the clock in most hospitals, is more rapidly accessible to stroke patients than MRI, and can be used to exclude intracranial hemorrhage, identify occlusion site, and possibly differentiate infarct from penumbra. Our study provides preliminary evidence that CBF and CBV derived from an admission CTP can identify infarct from penumbral tissue with sensitivity and specificity. This technique needs to be tested in a larger, randomized, prospective trial to examine its efficacy and whether it could be used to guide treatment decisions and possibly improve clinical outcome.
Acknowledgments
We would like to thank the stroke and CT teams at Sunnybrook and Women’s College Health Sciences Centre, Ottawa Health Research Institute, Foothills Medical Centre, and London Health Sciences, all of which helped collect and interpret data.
Sources of Funding
We thank the Canadian Stroke Network for supporting this study and General Electric HealthCare for partial support.
Disclosures
T.-Y.L. is a consultant to General Electric HealthCare on the CT Perfusion software. Portions of the software are copyrighted by Lawson Health Research Institute and Robarts Research Institute.
References
Astrup J, Siesjo BK, Symon L. Thresholds in cerebral ischemia—the ischemic penumbra. Stroke. 1981; 12: 723–725.
Tissue plasminogen activator for acute ischemic stroke. The National Institute of Neurological Disorders and Stroke rt-PA Stroke Study Group. N Engl J Med. 1995; 333: 1581–1587.
Hacke W, Kaste M, Fieschi C, Toni D, Lesaffre E, von Kummer R, Boysen G, Bluhmki E, Hoxter G, Mahagne MH. Intravenous thrombolysis with recombinant tissue plasminogen activator for acute hemispheric stroke. The European Cooperative Acute Stroke Study (ECASS). J Am Med Assoc. 1995; 274: 1017–1025.
Furlan A, Higashida R, Wechsler L, Gent M, Rowley H, Kase C, Pessin M, Ahuja A, Callahan F, Clark WM, Silver F, Rivera F. Intra-arterial prourokinase for acute ischemic stroke. The PROACT II study: a randomized controlled trial. Prolyse in Acute Cerebral Thromboembolism. J Am Med Assoc. 1999; 282: 2003–2011.
Douglas VC, Tong DC, Gillum LA, Zhao S, Brass LM, Dostal J, Johnston SC. Do the Brain Attack Coalition’s criteria for stroke centers improve care for ischemic stroke Neurology. 2005; 64: 422–427.
Hacke W, Donnan G, Fieschi C, Kaste M, von Kummer R, Broderick JP, Brott T, Frankel M, Grotta JC, Haley EC Jr, Kwiatkowski T, Levine SR, Lewandowski C, Lu M, Lyden P, Marler JR, Patel S, Tilley BC, Albers G, Bluhmki E, Wilhelm M, Hamilton S. Association of outcome with early stroke treatment: pooled analysis of ATLANTIS, ECASS, and NINDS rt-PA stroke trials. Lancet. 2004; 363: 768–774.
Baron JC. Positron emission tomography. Stroke: Pathophysiology, Diagnosis, and Management. 3rd ed. New York, NY: Churchill Livingstone; 1998: 101–119.
Heiss WD, Sobesky J, Smekal U, Kracht LW, Lehnhardt FG, Thiel A, Jacobs AH, Lackner K. Probability of cortical infarction predicted by flumazenil binding and diffusion-weighted imaging signal intensity: a comparative positron emission tomography/magnetic resonance imaging study in early ischemic stroke. Stroke. 2004; 35: 1892–1898.
Barber PA, Darby DG, Desmond PM, Yang Q, Gerraty RP, Jolley D, Donnan GA, Tress BM, Davis SM. Prediction of stroke outcome with echoplanar perfusion- and diffusion-weighted MRI. Neurology. 1998; 51: 418–426.
Kidwell CS, Saver JL, Mattiello J, Starkman S, Vinuela F, Duckwiler G, Gobin YP, Jahan R, Vespa P, Kalafut M, Alger JR. Thrombolytic reversal of acute human cerebral ischemic injury shown by diffusion/perfusion magnetic resonance imaging. Ann Neurol. 2000; 47: 462–469.
Kucinski T, Naumann D, Knab R, Schoder V, Wegener S, Fiehler J, Majumder A, Rother J, Zeumer H. Tissue at risk is overestimated in perfusion-weighted imaging: MR imaging in acute stroke patients without vessel recanalization. AJNR Am J Neuroradiol. 2005; 26: 815–819.
Wintermark M, Fischbein NJ, Smith WS, Ko NU, Quist M, Dillon WP. Accuracy of dynamic perfusion CT with deconvolution in detecting acute hemispheric stroke. AJNR Am J Neuroradiol. 2005; 26: 104–112.
Wintermark M, Reichhart M, Thiran JP, Maeder P, Chalaron M, Schnyder P, Bogousslavsky J, Meuli R. Prognostic accuracy of cerebral blood flow measurement by perfusion computed tomography, at the time of emergency room admission, in acute stroke patients. Ann Neurol. 2002; 51: 417–432.
Koenig M, Kraus M, Theek C, Klotz E, Gehlen W, Heuser L. Quantitative assessment of the ischemic brain by means of perfusion-related parameters derived from perfusion CT. Stroke. 2001; 32: 431–437.
Wintermark M, Reichhart M, Cuisenaire O, Maeder P, Thiran JP, Schnyder P, Bogousslavsky J, Meuli R. Comparison of admission perfusion computed tomography and qualitative diffusion- and perfusion-weighted magnetic resonance imaging in acute stroke patients. Stroke. 2002; 33: 2025–2031.
Eastwood JD, Lev MH, Wintermark M, Fitzek C, Barboriak DP, Delong DM, Lee TY, Azhari T, Herzau M, Chilukuri VR, Provenzale JM. Correlation of early dynamic CT perfusion imaging with whole-brain MR diffusion and perfusion imaging in acute hemispheric stroke. AJNR Am J Neuroradiol. 2003; 24: 1869–1875.
Kudo K, Terae S, Katoh C, Oka M, Shiga T, Tamaki N, Miyasaka K. Quantitative cerebral blood flow measurement with dynamic perfusion CT using the vascular-pixel elimination method: comparison with H2(15)O positron emission tomography. AJNR Am J Neuroradiol. 2003; 24: 419–426.
Wintermark M, Thiran JP, Maeder P, Schnyder P, Meuli R. Simultaneous measurement of regional cerebral blood flow by perfusion CT and stable xenon CT: a validation study. AJNR Am J Neuroradiol. 2001; 22: 905–914.
Eastwood JD, Provenzale JM, Hurwitz LM, Lee TY. Practical injection-rate CT perfusion imaging: deconvolution-derived hemodynamics in a case of stroke. Neuroradiology. 2001; 43: 223–226.
Hossmann KA. Viability thresholds and the penumbra of focal ischemia. Ann Neurol. 1994; 36: 557–565.
Powers WJ, Grubb RL Jr, Raichle ME. Physiological responses to focal cerebral ischemia in humans. Ann Neurol. 1984; 16: 546–552.
Lev MH, Segal AZ, Farkas J, Hossain ST, Putman C, Hunter GJ, Budzik R, Harris GJ, Buonanno FS, Ezzeddine MA, Chang Y, Koroshetz WJ, Gonzalez RG, Schwamm LH. Utility of perfusion-weighted CT imaging in acute middle cerebral artery stroke treated with intra-arterial thrombolysis: prediction of final infarct volume and clinical outcome. Stroke. 2001; 32: 2021–2028.
Kuschinsky W, Wahl M. Local chemical and neurogenic regulation of cerebral vascular resistance. Physiol Rev. 1978; 58: 656–689.
Roberts HC, Roberts TP, Smith WS, Lee TJ, Fischbein NJ, Dillon WP. Multisection dynamic CT perfusion for acute cerebral ischemia: the "toggling-table" technique. AJNR Am J Neuroradiol. 2001; 22: 1077–1080.
Hsieh J, Wei Y, Wang G. Fractional scan algorithms for low-dose perfusion CT. Med Phys. 2004; 31: 1254–1257.
Smith WS, Roberts HC, Chuang NA, Ong KC, Lee TJ, Johnston SC, Dillon WP. Safety and feasibility of a CT protocol for acute stroke: combined CT, CT angiography, and CT perfusion imaging in 53 consecutive patients. AJNR Am J Neuroradiol. 2003; 24: 688–690.
Karonen JO, Liu Y, Vanninen RL, Ostergaard L, Kaarina Partanen PL, Vainio PA, Vanninen EJ, Nuutinen J, Roivainen R, Soimakallio S, Kuikka JT, Aronen HJ. Combined perfusion- and diffusion-weighted MR imaging in acute ischemic stroke during the 1st week: a longitudinal study. Radiology. 2000; 217: 886–894.
Heiss WD, Huber M, Fink GR, Herholz K, Pietrzyk U, Wagner R, Wienhard K. Progressive derangement of periinfarct viable tissue in ischemic stroke. J Cereb Blood Flow Metab. 1992; 12: 193–203.