Quantitative Cerebral Blood Flow Determinations in Acute Ischemic Stroke
Relationship to Computed Tomography and Angiography
Background and Purpose The advent of new modalities to treat acute ischemic stroke presents the need for accurate, early diagnosis. In acute ischemic stroke, CT scans are frequently normal or reveal only subtle hypodense changes. This study explored the utility and increased sensitivity of xenon-enhanced CT (XeCT) in the diagnosis of acute cerebral ischemia and investigated the relationship between cerebral blood flow (CBF) measurements and early CT and angiographic findings in acute stroke.
Methods The CT scans, XeCT scans, and angiograms of 20 patients who presented within 6 hours of acute anterior circulation ischemic strokes were analyzed.
Results CT scans were abnormal in 11 (55%) of 20 patients. XeCT scans were abnormal in all 20 (100%) patients, showing regions of interest with CBF <20 (mL/100 g per minute) in the symptomatic middle cerebral artery (MCA) territories. The mean CBF in the symptomatic MCA territories was significantly lower than that of the asymptomatic MCA territories (P<.0005). In patients with basal ganglia hypodensities, the mean symptomatic MCA territory CBF was significantly lower than that of patients who did not exhibit these early CT findings (P<.05). The mean symptomatic MCA territory CBF in patients with angiographic M1 occlusions was significantly lower than that of patients whose infarcts were caused by MCA branch occlusions (P<.01).
Conclusions These results show that XeCT is more sensitive than CT in detecting acute strokes and that CBF measurements correlate with early CT and angiographic findings. XeCT may allow for the hyperacute identification of subsets of patients with acute ischemic events who are less likely to benefit and more likely to derive complications from aggressive stroke therapy.
As the neuroprotective and thrombolytic options for the treatment of ischemic stroke continue to expand, the need to rapidly and accurately diagnose cerebral ischemia has become paramount.1 2 3 4 5 6 7 8 An early and complete understanding of the pathophysiological processes of acute stroke in individual patients will likely assist in predicting outcome and guiding treatment. Early CT changes, for example, have been associated with poorer clinical outcomes and with higher complication rates of thrombolytic therapy.9 10 11 Early CT scans in acute stroke, however, are frequently normal.12 The purpose of this study was to determine whether CBF measurements with XeCT are more sensitive than CT in the early detection of acute stroke and how CBF relates to early CT and angiographic findings. A subsequent question will be whether knowledge of the CBF and volume of ischemic brain tissue at risk in the acute setting may be useful in selecting patients who will benefit from treatment from those with an increased risk of developing hemorrhagic complications. To these ends, this is the first study to systematically examine early quantitative CBF measurements in relation to CT and angiographic findings in patients with acute stroke symptoms.
Subjects and Methods
Twenty patients who underwent XeCT CBF studies within 6 hours of the onset of an anterior circulation ischemic stroke at the University of Pittsburgh Medical Center were studied. Patients ranged in age from 26 to 84 years (mean age, 66 years); 9 were male (45%). Fourteen patients presented with right-sided hemipareses or hemiplegias and aphasias (70%); 6 (30%) had left-sided motor deficits. All patients were admitted with the clinical diagnosis of “acute MCA territory stroke”; this group of patients, therefore, is a subset of all patients who presented with acute stroke symptoms by virtue of the time window of <6 hours since onset and the specific anterior circulation symptomatology. Patients with lacunar symptoms, events related to the vertebrobasilar system, visual complaints, or resolving deficits were not included.
All patients underwent CT scanning within 6 hours of the onset of symptoms. The mean interval between the onset of symptoms and performance of CT scans was 230 minutes (3 hours, 50 minutes).
Blinded Interpretation of CT Scans
All CT scans were independently evaluated by two fellowship-trained neuroradiologists who were informed only that the CT scans were those of patients who were clinically suspected to have a stroke. They were blinded to the following: (1) the specific stroke symptoms and the involved side, (2) the CBF data, and (3) the angiographic findings. CT images were evaluated for hypodensity (cortical or basal ganglia), mass effect (sulcal or ventricular effacement), and hyperdensity in the MCA. Findings were recorded as present or absent. Differences were adjudicated by consensus between the two neuroradiologists.
XeCT Cerebral Blood Flow Testing
All patients underwent XeCT scans within 6 hours from the onset of symptoms. In all cases, XeCT scans were obtained immediately after CT scans. Quantitative CBF studies were performed on standard CT scanners to which an independent system for xenon delivery and CBF calculation were added (Xe/CT System, Diversified Diagnostic Products). While patients inhaled a 33% Xe/67% O2 mixture (XeScan, Praxair Pharmaceutical Gases) for 4.5 minutes, CT images were obtained at three levels through the brain. Setup and computer calculation time requires approximately 10 additional minutes, making the total time required to obtain CBF images 15 minutes.
Interpretation of XeCT Scans
XeCT CBF data were analyzed by means of a computerized data analysis program that calculates the mean CBF within a series of 2-cm circular ROIs distributed throughout the cortical and subcortical areas (Fig 1⇓). Three axial CT scan slices were studied for most (>90%) patients, yielding between 55 and 65 ROIs per patient (Fig 2⇓). A total of approximately 1200 ROIs were analyzed in this manner. ROIs within areas that corresponded to artifact on CT scans were excluded from the analysis. ROIs with mean CBF <8 mL/100 g per minute were identified and tabulated separately for additional analysis.
Fifteen of 20 (75%) patients underwent digital subtraction angiography within 6 hours of onset of symptoms. In all 15 patients, selective injection of the symptomatic common carotid artery was performed by a transfemoral approach. In some patients, selective injection of the internal carotid artery was also performed.
CT Scan Findings
CT scans were normal in 9 (45%) of 20 patients. In the 11 (55%) patients with abnormal CT scans, 4 (36%) showed hypodensities in the basal ganglia; 11 (100%) showed cortical hypodensities; 4 (36%) showed cortical mass effect; and 4 (36%) showed hyperdensities (thromboses) in the MCA. All abnormalities were found on the symptomatic side; no abnormalities were detected on the asymptomatic side.
CBF in Symptomatic Versus Asymptomatic MCA Territories
XeCT scans were abnormal in all patients (100%), showing ROIs with CBF <20 mL/100 g per minute within the symptomatic MCA territories in all cases. The mean CBF in the symptomatic MCA territories for the group was 15±8.1 mL/100 g per minute compared with 34±9.2 mL/100 g per minute for the asymptomatic MCA territories; this difference was statistically significant (P<.0005, t test) (Fig 3⇓). CBF in the symptomatic MCA territory ranged between 4 and 37 mL/100 g per minute among individual patients (Fig 4⇓).
The mean percentage of ROIs with CBF <8 mL/100 g per minute in the symptomatic MCA territories was 34±26.5 compared with 0.3±1.1 in the asymptomatic MCA territories; this difference was statistically significant (P<.0005, t test). The percentage of ROIs with mean CBF <8 mL/100 g per minute ranged between 0 and 88 among individual patients.
Angiograms were performed in 15 (75%) of 20 patients in this study. Of the patients who underwent cerebral angiography, M1 occlusions were found in 11 (73%) of 15 patients, while distal branch occlusions were found in 3 patients (20%); 1 patient (7%) had a normal angiogram. Four patients with M1 occlusions also had ipsilateral ICA occlusions; this group was not analyzed separately and is referred to simply as patients with M1 occlusions.
Relationship Between CBF and Early CT Changes
The mean CBF of the symptomatic MCA territories of the group of patients who had basal ganglia hypodensities (7 mL/100 g per minute) was significantly lower than that of patients who did not have these findings (17 mL/100 g per minute) (P<.05, t test). There were no significant differences in the CBF between groups in the asymptomatic control hemispheres (31 versus 35 mL/100 g per minute, P=.15, t test) (Table⇓, Fig 5⇓).
Similarly, the mean percentage of ROIs with mean CBF <8 mL/100 g per minute in the symptomatic MCA territories of patients who had basal ganglia hypodensities (62%) was significantly greater than that of those who did not have these findings (27%) (P<.05, t test). There were no significant differences between groups in the percentage of ROIs with CBF <8 10 mL/100 g per minute in the asymptomatic control hemispheres.
In patients with a mean symptomatic MCA territory CBF <10 mL/100 g per minute, 57% had basal ganglia hypodensities. No patient with a mean CBF >10 mL/100 g per minute in the symptomatic MCA territories had a basal ganglia hypodensity. These associations were statistically significant (P<.01, Fisher’s exact test).
There were no significant differences among the mean CBFs of the symptomatic MCA territories of patients with cortical hypodensities, cortical mass effect, or M1 hyperdensity signs on their CT scans compared with patients who did not have these findings (Table⇑). Similarly, there was no significant differences between the mean % of ROIs with mean CBF <8 cc/100 g per minute in patients with these findings compared to those without.
Relationship Between Angiographic Findings and CBF
When only patients who underwent angiography were considered, of patients with angiographically documented M1 occlusions, mean CBF in the symptomatic MCA territory was 12 mL/100 g per minute compared with 30 mL/100 g per minute in patients who did not have M1 occlusions. This difference was statistically significant (P<.01, t test) (Fig 6⇓). This statistically significant difference did not exist in the asymptomatic MCA territory, in which the CBFs were 35 and 39 mL/100 g per minute in patients with and without M1 occlusions, respectively (P=.5, t test) (Fig 6⇓).
When only patients who underwent angiography were considered, the mean percentage of ROIs with CBF <8 mL/100 g per minute in the symptomatic MCA territories of patients who had angiographically documented M1 occlusions was 40% compared with 6% in those who did not have M1 occlusions. This difference was very statistically significant (P<.005, t test). This statistical difference did not exist in the asymptomatic MCA territory, in which no patients had a ROI with mean CBF <8 mL/100 g per minute.
Of patients who had angiographically documented M1 occlusions, 91% had a mean CBF in the symptomatic MCA territory <20 mL/100 g per minute. No patient without an M1 occlusion at angiography had CBF in the symptomatic MCA territory <20 mL/100 g per minute. All patients (100%) with CBFs in the symptomatic MCA territory <20 mL/100 g per minute had M1 occlusions. These associations were very statistically significant (P<.005, Fisher’s exact test).
Relationship Between Angiographic Findings and CT Changes
There were no statistically significant associations between cortical or basal ganglia hypodensities, cortical mass effect, or M1 hyperdensity and the finding of an M1 occlusion at angiography.
CBF Measurements in the Diagnosis of Acute Cerebral Ischemia
Although decreased CBF is the initial event in the pathophysiological process of acute ischemic stroke, this is the first study to report quantitative CBF measurements using XeCT in human subjects within 6 hours of the onset of symptoms. An important result of this study is that CBF measurements by the XeCT method were 100% sensitive and 100% specific for determining the ischemic etiology of the acute neurological deficits seen in this population of patients, despite the fact that CT scans were abnormal in only 55% of patients. All patients had regions of brain with decreased perfusion in the critical ischemic range (<20 mL/100 g per minute) to explain their neurological deficits. Furthermore, although the mean CBF for the entire MCA territory was >20 mL/100 g per minute in some patients with MCA branch occlusions, there was a statistically significant quantitative difference between the mean CBFs of the affected MCA territories (15±8.1 mL/100 g per minute) compared with the unaffected side (34±9.2 mL/100 g per minute) for the study population.
The finding that the mean CBF of the symptomatic MCA territory was sometimes normal despite harboring focal areas of ischemia has important implications. Studies that measure CBF by averaging values over a large area of brain tissue or a vascular territory will sometimes miss small regions of ischemia buried within areas of normal or hyperemic flow. Studies using 133Xe, because of their lack of resolution combined with this “look-through” effect, historically created a degree of pessimism about the utility of all CBF measurements in the management of acute cerebral ischemia.13
The finding of CBF in the asymptomatic MCA territory of 34±9.2 in this study also has important implications. Studies that determine whether ischemia is present in a nonquantitative manner by comparing the CBF of the affected side to the contralateral side by assuming that the unaffected side has a “normal” CBF of 50 mL/100 g per minute will sometimes underestimate regions of ischemia when actual flows in the unaffected side are much lower, as they were in many patients in this study. Single photon emission tomography studies,14 15 16 therefore, lack the anatomical resolution and numeric quantification of CBF to provide rich physiologic information in the early management of cerebral ischemia, furthering the erroneous notion that CBF measurements are of little clinical utility in the acute management of stroke. Similarly, although positron emission tomography studies have provided early prognostic CBF information in cerebral ischemia in small groups of patients, they are expensive, technically demanding studies that lack the resolution and direct anatomic correlation to be well suited for routine clinical use in acute stroke management.17
XeCT Technique and Selection of ROIs
The clinical XeCT technique has been described in detail previously.18 19 It is important to note that XeCT is extremely well suited for use in the diagnosis and management of acute cerebral ischemia. Nearly all patients who are undergoing an urgent stroke evaluation will require a CT scan to rule out hemorrhage as the cause of the sudden focal deficit. The addition of a XeCT scan to the conventional CT scan requires only an additional 4.5 minutes of scanning while the patient inhales stable Xe, a mild, short-acting anesthetic that has no significant effect on CBF calculations.20 After an additional 10 minutes of computer calculation work, a color CBF map that correlates with the CT scan is generated (15 minutes from initiation of study to data output). ROIs can be automatically selected by the computer program or can be customized to measure particular areas of interest. In this study, standard ROIs were generated for all patients in the study so that comparisons could be easily made. For this study, approximately 20 circular (2 cm diameter) ROIs were generated for three standard CT-slice levels. Mean CBF for each vascular territory could then be calculated.
Two different measures of cerebral ischemia were used in this study. Regional ischemia was felt to be present when the CBF was <20 mL/100 g per minute in a vascular territory. Previous controlled studies in healthy volunteers have shown that the normal CBF in these mixed gray and white matter ROIs is 51±10 mL/100 g per minute.18 Mixed cortical flow values less that 20 mL/100 g per minute are, therefore, clearly abnormal. In studies of vasospasm after aneurysmal subarachnoid hemorrhage, infarction was not likely to occur in patients with similar mixed cortical CBF values greater than 18 mL/100 g per minute. Mixed cortical CBF values less than 15 mL/100 g per minute, on the other hand, were predictive of infarction in this population.21 22
In addition to CBF measurements averaged over the entire MCA territory, an assessment of the volume and severity of the ischemic insult was attempted by quantifying the percentage of ROIs with CBF <8 mL/100 g per minute. This level was chosen because flows in this range were felt to represent irreversible tissue injury. The percentage of ROIs with CBF <8 mL/100 g per minute was a marker of ischemia that paralleled the overall mean CBF in the MCA territory in this study. It is believed that the percentage of ROIs with CBF <8 mL/100 g per minute provides an indirect measurement of the volume of the most severely (perhaps irreversibly23 ) affected ischemic tissue and will therefore be an important predictor of the ultimate infarction. Indeed, the percentage of ROIs with CBF <8 mL/100 g per minute was even more strongly associated with the presence of an M1 occlusion at angiography than the mean CBF of the MCA territory.
Relationship Between Early CT Findings and CBF
Our study has shown, for the first time, a statistical relationship between early CT changes and quantitative CBF measurements in acute stroke. Patients with basal ganglia hypodensities had significant, associated decreases in CBF in the affected MCA territories compared with patients with normal CT scans. An early basal ganglia hypodensity was significantly associated with a mean CBF in the affected MCA territory of <10 mL/100 g per minute. The M1 hyperdensity sign was not associated with CBF in this study or to other CT changes.
Acute CT changes have been reported within 6 hours of the onset of cerebral ischemia.9 Basal ganglia hypodensities,24 cortical hypodensities and mass effect,25 and the M1 hyperdensity sign26 27 are the most commonly reported early CT findings. Our finding of a 55% detection rate for finding at least one of these early findings within 6 hours is consistent with the literature, although Tomura et al24 reported early CT changes in 23 of 25 patients within 6 hours of onset of MCA or carotid distribution infarctions.
Newer diffusion-weighted and perfusion imaging MRI techniques have also been able to identify changes in the early phases of ischemic stroke.28 29 30 Sorensen et al31 found abnormal diffusion-weighted MRI scans in 9 of 11 patients with acute stroke within 10 hours; all of these patients had normal CT scans.
That early basal ganglia CT changes were associated with lowered CBFs in this study is consistent with the known physiology of CT changes in acute ischemia.32 33 Hypodensity, the first sign of infarction on CT, is likely due to edema formation secondary to blood-brain barrier breakdown.34 The likelihood of development of a cerebral infarction depends on both the degree and duration of hypoperfusion.23 Therefore, patients with marginal CBFs between 15 and 20 mL/100 g per minute) may require more time for infarction to occur; CT evidence of infarction was less likely to have been present in <6 hours in these patients. Patients with critically lowered CBF (<10 cc/100 g per minute), on the other hand, were more likely to develop early CT evidence of infarction. That CT changes in the basal ganglia (rather than cortical hypodensities) were associated with significantly lower CBFs in the MCA territory likely relates to more proximal occlusions in the M1 segment that interrupt lenticulostriate supply to the basal ganglia, a region with minimal collateral blood supply compared with the cortex. Critically low CBF in the basal ganglia is likely to occur very early on in a proximal M1 occlusion, leading to rapid CT changes. Patients with more distally located occlusions will have preserved blood supply to the basal ganglia and presumably greater overall MCA territory CBF.
Consistent with the lack of a relationship between the M1 hyperdensity sign and CBF, the M1 hyperdensity sign has been shown to be an unreliable sign of infarction with a sensitivity of 50%.9 This may be because typical CT slices do not necessarily precisely cut through the M1 segments and is consistent with the findings of others.9
Relationship Between CBF and Angiographic Findings
This study shows, for the first time, that patients with acute anterior circulation ischemic strokes with angiographic M1 occlusions have significantly decreased quantitative CBF measurements in the MCA territory compared with patients without M1 occlusions. A mean CBF <20 mL/100 g per minute in the symptomatic MCA territory was 91% sensitive and 100% specific for an M1 occlusion. CBF measurements with XeCT can therefore reliably predict the anatomic lesion causing an acute neurological deficit in the very early stages of the stroke process, well before CT changes give an indication of the anatomic locus and size of the lesion.
Although one might expect a relationship between CT changes and angiographic findings from a physiological standpoint, there was no relationship between early CT changes and the findings of angiography in this study. Such a relationship has been cited in a previous study.35 Although this may be attributed to the small population size, the fact that CBF measurements were significantly associated with angiographic findings in this same group reinforces the strength of that association.
Implications for the Management of Acute Stroke
The results of this study show that ischemic insults to the brain can be uniformly detected with quantitative CBF measurements several hours before conventional CT scans reveal abnormalities. This knowledge should allow for a rapid prediction of the anatomy of the final infarct. Earlier and more accurate enrollment of patients into neuroprotective and thrombolytic protocols for the acute management of stroke may be possible.
Furthermore, that the degree and extent of decreased CBF correlated with the findings of angiography may allow for more selective use of angiography and may assist in the acute triage of patients with evolving stroke. Patients with patterns of ischemia on CBF testing consistent with distal branch occlusions or smaller infarcts, for example, may be more appropriately treated with intravenous thrombolytic therapy.8 A regional CBF of <20 for the entire MCA territory was strongly associated with an M1 occlusion; these patients may prove to be better candidates for intra-arterial thrombolytic therapy.
The finding that early basal ganglia CT changes correlate with statistically lower CBF values supports the clinical notion that early CT changes are a risk factor for poor outcome9 and hemorrhagic conversion after thrombolytic therapy.10 11 However, early CBF studies may identify subgroups of patients with critically low CBFs who have yet to develop CT changes but who nevertheless carry a similar increased risk of developing hemorrhagic complications of thrombolytic therapy or a decreased likelihood of benefiting from intervention.15 17 XeCT provides a practical means of determining CBF measurements that may ultimately allow earlier and more accurate stratification of the risks and benefits of aggressive management of acute stroke.
Selected Abbreviations and Acronyms
|CBF||=||cerebral blood flow|
|MCA||=||middle cerebral artery|
|ROIs||=||regions of interest|
The authors thank Robert L. Williams, MD, for serving as a blinded reader of the CT scans; Carol Barch, RN, for assisting in the clinical care of these patients; and John May for his management of the XeCT CBF film library.
Reprint requests to Howard Yonas, MD, Department of Neurological Surgery, University of Pittsburgh Medical Center, 200 Lothrop St, Suite B-400, Pittsburgh, PA 15213.
- Received June 12, 1997.
- Revision received July 8, 1997.
- Accepted July 29, 1997.
- Copyright © 1997 by American Heart Association
Brott TG, Haley EC, Levy DE, Barsan W, Broderick J, Sheppard GL, Spilker J, Konagable GL, Masset S, Reed R, Mahler JR. Urgent therapy for stroke, Part I: pilot study of tissue plasminogen activator administered within 90 minutes. Stroke. 1992;23:632-640.
Hacke W, Kaste M, Fieschi C, Toni D, Lesaffre E, von Kummer R, Boysen G, Bluhmki E, Hoxter G, Mahagne MH, Hennerici M, for the ECASS Study Group. Intravenous thrombolysis with recombinant tissue plasminogen activator for acute hemispheric stroke: the European Cooperative Acute Stroke Study (ECASS). JAMA. 1995;274:1017-1025.
Haley EC Jr, Brott TG, Sheppard GL, Barsan W, Broderick P, Marler JR, Kongable GL, Spilker J, Massey S, Hansen CA, Torner JC. Pilot randomized trial of tissue plasminogen activator in acute ischemic stroke. Stroke. 1993;24:1000-1004.
Haley EC Jr, Levy DE, Brott TG, Sheppard GL, Wong MC, Kongable GL, Torner JC, Marler JR. Urgent therapy for stroke, Part II: pilot study of tissue plasminogen activator administered 91-180 minutes from onset. Stroke. 1992;23:641-645.
Mori E, Tabuchi M, Yoshida T, Yamadori A. Intracarotid urokinase with thromboembolic occlusion of the middle cerebral artery. Stroke. 1988;19:802-812.
von Kummer R, Meyding-Lamadé U, Forsting M, Rosin L, Rieke K, Hacke W, Sartor K. Sensitivity and prognostic value of early CT in occlusion of the middle cerebral artery trunk. AJNR Am J Neuroradiol. 1994;15:9-15.
Yokogami K, Nakano S, Ohta H, Goya T, Wakisaka S. Prediction of hemorrhagic complications after thrombolytic therapy for middle cerebral artery occlusion: value of pre- and post-therapeutic computed tomographic findings and angiographic occlusive site. Neurosurgery. 1996;39:1102-1107.
Levy DE, Brott TG, Haley EC, Marler JR, Sheppard GL, Barsan W, Broderick P. Factors related to intracranial hematoma formation in patients receiving tissue-type plasminogen activator for acute ischemic stroke. Stroke. 1994;25:291-297.
Bryan RN, Levy LM, Whitlow WD, Killian JM, Prezios TJ, Rosario JA. Diagnosis of acute cerebral infarction: comparison of CT and MR imaging. AJNR Am J Neuroradiol. 1991;12:611-620.
Halsey JH Jr, Nakai K, Wariyar B. Sensitivity of rCBF to focal lesions. Stroke. 1981;12:631-635.
Brass LM, Walovitch RC, Joseph JL, Léveillé J, Marchand L, Hellman RS, Tikofsky RS, Masdeu JC, Hall KM, Van Heertum RL. The role of single photon emission computed tomography brain imaging with 99mTc-bicisate in the localization and definition of mechanism of ischemic stroke. J Cereb Blood Flow Metab. 1994;14(suppl 1):S91-S98.
Giubilei F, Lenzi GL, DiPiero V, Pozzilli C, Pantano P, Bastianello S, Argentino C, Fieschi C. Predictive value of brain perfusion single-photon emission computed tomography in acute ischemic stroke. Stroke. 1990;21:895-900.
Marchal G, Serrati C, Rioux P, Petit-Taboué MC, Viader F, de la Sayette V, le Doze F, Lochon P, Derlon JM, Orgogozo JM, Baron JC. PET imaging of cerebral perfusion and oxygen consumption in acute ischaemic stroke: relation to outcome. Lancet. 1991;341:925-927.
Gur D, Good WF, Wolfson SK Jr, Yonas H, Shabason L. In vivo mapping of local cerebral blood flow by xenon-enhanced computed tomography. Science. 1982;215:1267-1268.
Good WF, Gur D. Xenon-enhanced CT of the brain: effect of flow activation on desired cerebral blood flow measurements. AJNR Am J Neuroradiol. 1991;12:83-85.
Fukui MB, Johnson DW, Yonas H, Sekhar L, Latchaw RE, Pentheny S. Xe/CT cerebral blood flow evaluation of delayed symptomatic cerebral ischemia after subarachnoid hemorrhage. AJNR Am J Neuroradiol. 1992;13:265-270.
Tomsick T, Brott T, Barsan W, Broderick J, Haley EC, Spilker J. Thrombus localization with emergency cerebral CT. AJNR Am J Neuroradiol. 1992;13:257-263.
Sorensen AG, Buonanno FS, Gonzalez RG, Schwamm LH, Lev MH, Huang-Hellinger FR, Reese TG, Weisskoff RM, Davis TL, Suwanwela N, Can U, Moreira JA, Copen WA, Look RB, Finkelstein SP, Rosen BR, Koroshetz WJ. Hyperacute stroke: evaluation with combined multisection diffusion-weighted and hemodynamically weighted echo-planar MR imaging. Radiology. 1996;199:391-401.
Yonas H, Gur D, Claassen D, Wolfson SK, Moossy J. Stable xenon enhanced computed tomography in the study of clinical and pathologic correlates of focal ischemia in baboons. Stroke. 1988;19:228-238.
Bozzao L, Bastianello S, Fantozzi LM, Angeloni U, Argentino C, Fieschi C. Correlation of angiographic and sequential CT findings in patients with evolving cerebral infarction. AJNR Am J Neuroradiol. 1989;10:1215-1222.