Perfusion computed tomography (CT) is a relatively new technique that allows rapid qualitative and quantitative evaluation of cerebral perfusion by generating maps of cerebral blood flow (CBF), cerebral blood volume (CBV), and mean transit time (MTT). The technique is based on the central volume principle (CBF = CBV/MTT) and requires the use of commercially available software employing complex deconvolution algorithms to produce the perfusion maps. Some controversies exist regarding this technique, including which artery to use as input vessel, the accuracy of quantitative results, and the reproducibility of results. Despite these controversies, perfusion CT has been found to be useful for noninvasive diagnosis of cerebral ischemia and infarction and for evaluation of vasospasm after subarachnoid hemorrhage. Perfusion CT has also been used for assessment of cerebrovascular reserve by using acetazolamide challenge in patients with intracranial vascular stenoses who are potential candidates for bypass surgery or neuroendovascular treatment, for the evaluation of patients undergoing temporary balloon occlusion to assess collateral flow and cerebrovascular reserve, and for the assessment of microvascular permeability in patients with intracranial neoplasms. This article is a review of the technique, clinical applications, and controversies surrounding perfusion CT.
The advent of thrombolytic therapy for acute nonhemorrhagic stroke has intensified the need for a rapid readily available technique to help identify and quantify the presence and extent of a perfusion deficit (1–3). Magnetic resonance (MR) perfusion, xenon computed tomography (CT), positron emission tomography (PET), and single photon emission computed tomography (SPECT) have all been used to evaluate cerebral perfusion but are hampered by limited availability, cost, and/or patient tolerance (2). Perfusion CT was introduced as a means to rapidly and easily evaluate cerebral perfusion in patients presenting with acute stroke symptoms, most of whom would already undergo unenhanced head CT to exclude acute hemorrhage (2,4). Perfusion CT can be performed quickly with any standard spiral CT scanner, and the perfusion maps can be generated in a short time at a workstation equipped with the appropriate software. At our institution, the technique has been extended for evaluation of cerebrovascular reserve with acetazolamide challenge in patients with stenotic lesions who are potential candidates for bypass surgery or neuroendovascular treatment, for evaluation of collateral flow and cerebrovascular reserve in patients undergoing temporary balloon occlusion (5), and for evaluation of patients with possible vasospasm after subarachnoid hemorrhage (SAH). This technique can also be applied in patients with neoplasms to measure the permeability surface product area (PS) (6–8).
TECHNIQUE
The theory behind this technique is the central volume principle, which relates cerebral blood flow (CBF), cerebral blood volume (CBV), and mean transit time (MTT) in the following equation: CBF = CBV/MTT. Perfusion studies are obtained by monitoring the first pass of an iodinated contrast agent bolus through the cerebral vasculature. There is a linear relationship between contrast agent concentration and attenuation, with the contrast agent causing a transient increase in attenuation proportional to the amount of contrast agent in a given region. Contrast agent time-concentration curves are generated in an arterial region of interest (ROI), a venous ROI, and in each pixel. Deconvolution of arterial and tissue enhancement curves, a complex mathematic process, gives the MTT. CBV is calculated as the area under the curve in a parenchymal pixel divided by the area under the curve in an arterial pixel. The central volume equation can then be solved for CBF (2,4,9).
Perfusion CT scans are obtained at our institution by using a multi–detector row scanner (Lightspeed; GE Medical Systems, Milwaukee, Wis). After unenhanced CT of the whole brain, four adjacent 5-mm-thick sections are selected starting at the level of the basal ganglia. At this level, all three supratentorial vascular territories are visualized. Fifty milliliters of a nonionic contrast agent (300 mg of iodine per milliliter) is injected at a rate of 4 mL/sec. At 5 seconds after initiation of the injection, a cine (continuous) scan is initiated with the following technique: 80 kVp, 190–200 mA, 4 × 5-mm sections, 1-second per rotation for a duration of 50 seconds. If 370 mg of iodine per milliliter is used, a volume of 40 mL is given at a rate of 4 mL/sec with a scan duration of 45 seconds. The injection rates with the deconvolution technique are much lower than those of other CT perfusion techniques, such as the maximal-slope model, and thus are more practical and tolerable for patients (9,10). These lower injection rates do not represent a disadvantage because the deconvolution analysis compares the arterial input time-attenuation curve with that of the tissue to control for bolus dispersion (10). Perfusion maps of good quantitative quality have been obtained in humans with injection rates as low as 1.5 mL/sec (4,10).
The 1-second images are reformatted at 0.5-second intervals, and the 5-mm sections are reformatted into two 10-mm-thick sections. The scans are obtained at 5 mm rather than 10 mm to lessen beam-hardening artifacts in the brain. The reformatted 10-mm-thick sections provide a better signal-to-noise ratio than do the 5-mm-thick sections and improved (0.5-second cine interval) temporal resolution. The overall effective dose required for dynamic CT (2.0–3.4 mSv) is only slightly higher than that required for routine head CT (1.5–2.5 mSv). This dose equivalent is less than the dose equivalent obtained with PET or SPECT and is comparable to that of a single-level xenon CT examination (4,11–13).
CT perfusion data are analyzed at an imaging workstation (Advantage Windows; GE Medical Systems) equipped with commercially available software (CT Perfusion; GE Medical Systems). Post–image-collection processing involves placement of freehand drawn ROIs in an input artery and an input vein, for which contrast-enhancement curves are generated (Fig 1). The ACA or middle cerebral artery (MCA) can be selected as the input artery. A large venous structure such as the torcular herophili is chosen as the input vein. Since the input artery is usually smaller than the input vein, the venous ROI serves to correct for volume averaging in the arterial ROI. The venous ROI may also affect the signal-to-noise ratio, with the peak enhancement value of the venous ROI correlating significantly with the signal-to-noise ratio of CBF and CBV maps (14). The software then generates color-coded CBF, CBV, and MTT maps (Fig 2).
Figure 1a.(a) Transverse CT image obtained during contrast agent administration and (b) time-concentration curves in a healthy adult volunteer. (a) A branch of the anterior cerebral artery (ACA) is chosen as the input artery (white arrow) and the torcular herophili is chosen as the input vein (black arrow). (b) Time-concentration curves are generated for the input artery (green line) and vein (red line).
Figure 1b.(a) Transverse CT image obtained during contrast agent administration and (b) time-concentration curves in a healthy adult volunteer. (a) A branch of the anterior cerebral artery (ACA) is chosen as the input artery (white arrow) and the torcular herophili is chosen as the input vein (black arrow). (b) Time-concentration curves are generated for the input artery (green line) and vein (red line).
Figure 2a. Transverse CT perfusion maps in a healthy adult volunteer show normal perfusion. Various color ramps, selected according to user preference, are used to display the (a) CBF, (b) CBV, and (c) MTT maps.
Figure 2b. Transverse CT perfusion maps in a healthy adult volunteer show normal perfusion. Various color ramps, selected according to user preference, are used to display the (a) CBF, (b) CBV, and (c) MTT maps.
Figure 2c. Transverse CT perfusion maps in a healthy adult volunteer show normal perfusion. Various color ramps, selected according to user preference, are used to display the (a) CBF, (b) CBV, and (c) MTT maps.
ROIs can be placed in the brain parenchyma to yield quantitative data. We typically place six circular ROIs along the periphery of each hemisphere (Fig 3). This method was chosen, with input from our neurosurgery colleagues, to approximate the ROIs used with xenon CT. A similar technique has been used for perfusion-weighted MR and SPECT imaging of the brain (15).
Figure 3a. Transverse (a) CT image and (b) CBF map in a healthy adult volunteer, with circular ROIs (in purple) in place. Identical ROIs appear on CBV and MTT maps (not shown).
Figure 3b. Transverse (a) CT image and (b) CBF map in a healthy adult volunteer, with circular ROIs (in purple) in place. Identical ROIs appear on CBV and MTT maps (not shown).
CLINICAL INDICATIONS
Acute Stroke
Conventional unenhanced head CT is still the primary imaging modality used to evaluate patients presenting with stroke symptoms, to help exclude intracranial hemorrhage and detect signs of brain ischemia (2). Detection of early ischemic changes, such as parenchymal hypoattenuation and/or sulcal effacement, is important for stratifying patients who may benefit from thrombolytic therapy (2,3). Thrombolytic therapy has been reported to be most beneficial in patients with cytotoxic edema involving 33% or less of the MCA territory. It is also important to identify patients who will not benefit from thrombolytic therapy: namely, those with conventional CT findings of cytotoxic edema involving greater than 33% of the MCA territory. The findings of cerebral ischemia or infarction, however, can be subtle or absent. It has been suggested that evaluation of brain perfusion may help in the selection of patients for thrombolytic therapy by allowing identification of those patients with potentially salvageable tissue that is at risk for infarction (the ischemic penumbra) from those with extensive infarct (11). Other techniques to measure cerebral perfusion, such as xenon CT, SPECT, PET, and perfusion-weighted MR imaging, are limited by patient tolerance, availability, and lack of quantitative results (2).
Evaluation of acute stroke is one of the main indications for perfusion CT, with the goal of distinguishing infarcted tissue from the penumbra (2,4,9). The latter tissue may be salvageable with the administration of thrombolytic agents, while irreversibly damaged tissue will not benefit from reperfusion and may be at increased risk of hemorrhage after thrombolytic therapy (2,3). It is hypothesized that tissue at risk of infarction will have decreased CBF, normal or elevated CBV secondary to activation of cerebral autoregulatory mechanisms, and elevated MTT, while infarcted tissue will have decreased CBF and CBV with elevated MTT (4,9) (Figs 4, 5).
Figure 4a. Acute infarction in a 76-year-old woman who was unresponsive. (a) Initial transverse head CT image shows subtle hypoattenuation (arrows) in left insula, temporal lobe, and frontal lobe. (b) Transverse CBF map shows decreased CBF (arrows) in left ACA and anterior MCA territories, with quantitative values as low as 11 mL/100 g/min. (c) Transverse CBV map shows decreased CBV (arrows). (d) Transverse MTT map shows slightly elevated MTT (arrows). Changes are consistent with infarction. (e) Follow-up transverse CT image shows large infarct (arrows) in left ACA and MCA territories.
Figure 4b. Acute infarction in a 76-year-old woman who was unresponsive. (a) Initial transverse head CT image shows subtle hypoattenuation (arrows) in left insula, temporal lobe, and frontal lobe. (b) Transverse CBF map shows decreased CBF (arrows) in left ACA and anterior MCA territories, with quantitative values as low as 11 mL/100 g/min. (c) Transverse CBV map shows decreased CBV (arrows). (d) Transverse MTT map shows slightly elevated MTT (arrows). Changes are consistent with infarction. (e) Follow-up transverse CT image shows large infarct (arrows) in left ACA and MCA territories.
Figure 4c. Acute infarction in a 76-year-old woman who was unresponsive. (a) Initial transverse head CT image shows subtle hypoattenuation (arrows) in left insula, temporal lobe, and frontal lobe. (b) Transverse CBF map shows decreased CBF (arrows) in left ACA and anterior MCA territories, with quantitative values as low as 11 mL/100 g/min. (c) Transverse CBV map shows decreased CBV (arrows). (d) Transverse MTT map shows slightly elevated MTT (arrows). Changes are consistent with infarction. (e) Follow-up transverse CT image shows large infarct (arrows) in left ACA and MCA territories.
Figure 4d. Acute infarction in a 76-year-old woman who was unresponsive. (a) Initial transverse head CT image shows subtle hypoattenuation (arrows) in left insula, temporal lobe, and frontal lobe. (b) Transverse CBF map shows decreased CBF (arrows) in left ACA and anterior MCA territories, with quantitative values as low as 11 mL/100 g/min. (c) Transverse CBV map shows decreased CBV (arrows). (d) Transverse MTT map shows slightly elevated MTT (arrows). Changes are consistent with infarction. (e) Follow-up transverse CT image shows large infarct (arrows) in left ACA and MCA territories.
Figure 4e. Acute infarction in a 76-year-old woman who was unresponsive. (a) Initial transverse head CT image shows subtle hypoattenuation (arrows) in left insula, temporal lobe, and frontal lobe. (b) Transverse CBF map shows decreased CBF (arrows) in left ACA and anterior MCA territories, with quantitative values as low as 11 mL/100 g/min. (c) Transverse CBV map shows decreased CBV (arrows). (d) Transverse MTT map shows slightly elevated MTT (arrows). Changes are consistent with infarction. (e) Follow-up transverse CT image shows large infarct (arrows) in left ACA and MCA territories.
Figure 5a. Ischemia without infarction in a 47-year-old woman with severe right internal carotid artery stenosis and episodes of left arm numbness. (a) Transverse perfusion CT map shows decreased CBF (arrows) in right ACA and MCA territories, with quantitative values as low as 21 mL/100 g/min. (b) CBV and (c) MTT are both elevated (arrows) on transverse perfusion maps. Changes are consistent with reversible ischemia. CBV is elevated due to cerebral autoregulatory mechanisms, with vasodilatation in response to decreased perfusion. (d) Follow-up transverse CT image obtained 8 months later shows no evidence of cortical infarction.
Figure 5b. Ischemia without infarction in a 47-year-old woman with severe right internal carotid artery stenosis and episodes of left arm numbness. (a) Transverse perfusion CT map shows decreased CBF (arrows) in right ACA and MCA territories, with quantitative values as low as 21 mL/100 g/min. (b) CBV and (c) MTT are both elevated (arrows) on transverse perfusion maps. Changes are consistent with reversible ischemia. CBV is elevated due to cerebral autoregulatory mechanisms, with vasodilatation in response to decreased perfusion. (d) Follow-up transverse CT image obtained 8 months later shows no evidence of cortical infarction.
Figure 5c. Ischemia without infarction in a 47-year-old woman with severe right internal carotid artery stenosis and episodes of left arm numbness. (a) Transverse perfusion CT map shows decreased CBF (arrows) in right ACA and MCA territories, with quantitative values as low as 21 mL/100 g/min. (b) CBV and (c) MTT are both elevated (arrows) on transverse perfusion maps. Changes are consistent with reversible ischemia. CBV is elevated due to cerebral autoregulatory mechanisms, with vasodilatation in response to decreased perfusion. (d) Follow-up transverse CT image obtained 8 months later shows no evidence of cortical infarction.
Figure 5d. Ischemia without infarction in a 47-year-old woman with severe right internal carotid artery stenosis and episodes of left arm numbness. (a) Transverse perfusion CT map shows decreased CBF (arrows) in right ACA and MCA territories, with quantitative values as low as 21 mL/100 g/min. (b) CBV and (c) MTT are both elevated (arrows) on transverse perfusion maps. Changes are consistent with reversible ischemia. CBV is elevated due to cerebral autoregulatory mechanisms, with vasodilatation in response to decreased perfusion. (d) Follow-up transverse CT image obtained 8 months later shows no evidence of cortical infarction.
In a recent retrospective study, Eastwood et al (16) demonstrated a statistically significant difference in CBF, CBV, and MTT in the symptomatic hemisphere, compared with those values in the contralateral hemisphere, in patients presenting with acute MCA stroke. The following thresholds for ischemia were chosen: CBF of 0–10 mL/100 g/min, CBV of 0–1.5 mL/100 g, and MTT of greater than 6 seconds. They found that the extent of regional abnormalities on the perfusion maps was greatest with MTT, followed by CBF and CBV. Thus, MTT maps may be the most sensitive indicators of stroke, with changes in CBF and CBV being more specific for distinguishing ischemia from infarction. In the same study, 25% of the patients with acute stroke had no abnormality on the CBV map, which may make this a less accurate indicator in the discrimination between the penumbra and infarcted tissue. These results correlated with those of an experimental rabbit model of ischemia in which perfusion CT findings and postmortem tissue specimens were compared (17). In that study, MTT maps were the most sensitive, particularly with regard to detection of early stages of minor ischemia, while CBF maps showed the best correlation between critical ischemia and postmortem evaluation results. CBV maps were of only moderate diagnostic value. Thus, it may be best to evaluate the CBF and MTT maps for abnormalities first and, if abnormalities are present, to use the CBV map to try to elucidate the underlying pathophysiology (ischemia vs infarct), keeping in mind that CBV may be normal even in cases of infarction.
However, in a prospective study in which patients with acute ischemic stroke were evaluated, Wintermark et al (11) reported that they were able to distinguish infarct from penumbra by defining ischemic tissue (infarct plus penumbra) as cerebral pixels with a greater than 34% decrease in CBF relative to clinically normal areas in the cerebral hemispheres. Within this ischemic area, a CBV threshold of 2.5 mL/100 g was selected, with higher and lower values representing penumbra and infarct, respectively. Using these criteria, Wintermark et al showed a significant correlation between penumbra size on initial perfusion CT images and clinical improvement as measured with the National Institutes of Health Stroke Scale in patients with arterial recanalization (either spontaneous or due to thrombolytic therapy). They also reported a close correlation between infarct size on perfusion CT images obtained at admission and infarct size on delayed diffusion-weighted MR images in patients undergoing recanalization, likely reflecting recovery of the penumbra. In addition, in patients without recanalization there was a close correlation between the combined infarct and penumbra on the initial perfusion CT images and the final infarct size on delayed diffusion-weighted MR images, due to evolution of the penumbra toward infarction.
Quantitatively, reversible paralysis has been demonstrated to occur in monkeys at CBF values of less than 20–23 mL/100 g/min, with a transition to irreversible infarction at a CBF of less than 10–12 mL/100 g/min (4,9,18,19). This transition from ischemia to infarction depends not only on CBF values but also on the duration of the diminution in blood flow, with the infarction threshold increasing over time to 17–18 mL/100 g/min (18–20). Increasing severity and duration of ischemia lead to increasingly severe histologic damage (19). It has been suggested that the upper threshold reflects failure of normal electric activity and the lower threshold reflects energy and ion pump failures, with the potentially salvageable penumbra being between these two levels (20,21). Cessation of electric activity in baboons and cats has been demonstrated at CBF values of 18 mL/100 g/min and 15 mL/100 g/min, respectively, with ion pump failure occurring in the baboon at a CBF threshold of 10 mL/100 g/min (20). In humans undergoing carotid endarterectomy, flattening of the electroencephalogram occurs at CBF values of less than 16–18 mL/100 g/min (20,22).
By using PET to measure CBF in patients with MCA ischemia or infarction, the minimum averaged gray and white matter CBF to maintain cerebral function was 19 mL/100 g/min or greater, while the minimum CBF to preserve tissue viability was 15 mL/100 g/min. There was, however, substantial overlap in CBF values between infarcted areas and viable tissue (22). Touho and Karasawa (23), using xenon CT in patients with anterior circulation ischemia, showed that CBF greater than 19 mL/100 g/min did not result in cortical infarction even if recanalization did not occur. Below this level, progressively lower CBF values resulted in cortical infarction at progressively shorter time intervals, with CBF values less than 9 mL/100 g/min invariably resulting in infarction. In another study with xenon CT (24), mixed cortical gray and white matter regions with CBF less than 6 mL/100 g/min were destined for infarction.
Evaluation and clinical application of such threshold determinations is complicated by being carried out in different species under variable conditions and with differing methods of flow measurement (25). Despite these difficulties, it is evident that the penumbra lies within a narrow range of perfusion and is dependent on small changes in perfusion pressure (26). Factors other than depth and duration of ischemia may also play a role in the viability of the penumbra, including selective vulnerability of specific neuronal populations, antecedent glucose stores, and physiologic conditions during resuscitation (22,27).
Areas of hypoattenuation on unenhanced head CT scans in the study by Eastwood et al (16) had a mean CBF value of 13.1 mL/110 g/min ± 8.4 (SD), which correlates well with previously reported values for xenon CT and PET results in patients with MCA stroke. In a small series comparing quantitative CBF values derived from xenon CT and perfusion CT data (28), there was good correlation between these two methods. Quantitative results can vary depending on the choice of input artery and the placement of ROIs (28,29).
Further studies are necessary to clarify these results and to compare perfusion CT with other methods of measuring cerebral perfusion (16,28). However, these early results show that perfusion CT may enable prediction of which patients will benefit from thrombolysis on the basis of the size of the penumbra and determination of the final infarct size in patients with and patients without recanalization.
Cerebrovascular Reserve
In patients with known chronic cerebral ischemia related to underlying stenotic lesions, it is necessary to distinguish tissue in need of increased CBF (tissue under hemodynamic stress) from tissue with decreased CBF due to decreased metabolic demand (30). Hemodynamic stress can be evaluated by using a tolerance test such as acetazolamide (Diamox; Wyeth, Marietta, Pa) administration in conjunction with quantitative measurement of CBF. Although the exact mechanism of action is uncertain, acetazolamide normally causes vasodilatation of cerebral arterioles and an increase in CBF (30). Patients with hemodynamic stress, however, are already maximally vasodilated due to the utilization of cerebral autoregulatory mechanisms in response to decreased perfusion pressure and cannot respond further to acetazolamide. These patients are considered to be at increased risk of stroke and may benefit from interventions to increase CBF (30–32).
Xenon CT, PET, SPECT, transcranial Doppler sonography, and perfusion-weighted MR imaging have all been used to evaluate cerebrovascular reserve with the acetazolamide test (33–35). With xenon CT to measure CBF, an increase of 20%–40% over baseline is normal, an increase of less than 5% over baseline is indicative of relative hemodynamic insufficiency, and a decrease of 5% or greater from baseline (steal phenomenon) indicates tissue at higher risk of stroke (35,36). Perfusion CT in conjunction with acetazolamide challenge may help to identify such patients (29). Acetazolamide is generally well tolerated. The most common side effects are circumoral numbness, paresthesias, and headache. A case of reversible ischemia has been reported with the use of acetazolamide (32,37).
At our institution, patients undergo a routine perfusion CT examination. Subsequently, 1,000 mg of acetazolamide is given intravenously, followed 20 minutes later by another perfusion CT examination (Fig 6). Whether the same quantitative values can be used for perfusion CT that have been used for xenon CT is yet to be determined in prospective studies comparing the two techniques. The quantitative results potentially available with perfusion CT may be an advantage over qualitative techniques such as SPECT and perfusion-weighted MR imaging (31). The ability to measure CBV and MTT may also be an added advantage of perfusion CT. This technique can also be used to evaluate the efficacy of revascularization procedures (32) (Fig 7).
Figure 6a. Known right common carotid artery occlusion in a 52-year-old man with 3-week history of amaurosis fugax, left-sided weakness and numbness, and left-sided facial droop. (a) Transverse perfusion CT image obtained before acetazolamide challenge shows decreased CBF (arrows) in right ACA and MCA territories. CBV and MTT maps (not shown) revealed elevation of both CBV and MTT. Changes are compatible with reversible ischemia. (b) After acetazolamide injection, there is steal phenomenon in right ACA and MCA territories. Transverse CBF map shows CBF decreased (arrows) from baseline values, indicating hemodynamic stress and tissue at high risk of infarction.
Figure 6b. Known right common carotid artery occlusion in a 52-year-old man with 3-week history of amaurosis fugax, left-sided weakness and numbness, and left-sided facial droop. (a) Transverse perfusion CT image obtained before acetazolamide challenge shows decreased CBF (arrows) in right ACA and MCA territories. CBV and MTT maps (not shown) revealed elevation of both CBV and MTT. Changes are compatible with reversible ischemia. (b) After acetazolamide injection, there is steal phenomenon in right ACA and MCA territories. Transverse CBF map shows CBF decreased (arrows) from baseline values, indicating hemodynamic stress and tissue at high risk of infarction.
Figure 7a. Transverse perfusion CT maps in a 65-year-old man with progressive worsening of left hand and foot numbness and clumsiness. Angiogram (not shown) revealed severe stenosis of precavernous right internal carotid artery. (a) Decreased CBF (arrows) is shown in right ACA and MCA territories before acetazolamide challenge. CBV was normal to elevated and MTT was elevated (not shown). (b) After acetazolamide injection, CBF map reveals decrease (arrows) in CBF values from baseline in right ACA and MCA territories, consistent with steal phenomenon. (c) After angioplasty and stent placement, repeat pre-acetazolamide challenge perfusion CT shows more symmetric CBF between the two hemispheres. CBV and MTT were also more symmetric (not shown). (d) After acetazolamide injection, CBF map now shows augmentation of flow in both hemispheres, with flow increased over that of baseline.
Figure 7b. Transverse perfusion CT maps in a 65-year-old man with progressive worsening of left hand and foot numbness and clumsiness. Angiogram (not shown) revealed severe stenosis of precavernous right internal carotid artery. (a) Decreased CBF (arrows) is shown in right ACA and MCA territories before acetazolamide challenge. CBV was normal to elevated and MTT was elevated (not shown). (b) After acetazolamide injection, CBF map reveals decrease (arrows) in CBF values from baseline in right ACA and MCA territories, consistent with steal phenomenon. (c) After angioplasty and stent placement, repeat pre-acetazolamide challenge perfusion CT shows more symmetric CBF between the two hemispheres. CBV and MTT were also more symmetric (not shown). (d) After acetazolamide injection, CBF map now shows augmentation of flow in both hemispheres, with flow increased over that of baseline.
Figure 7c. Transverse perfusion CT maps in a 65-year-old man with progressive worsening of left hand and foot numbness and clumsiness. Angiogram (not shown) revealed severe stenosis of precavernous right internal carotid artery. (a) Decreased CBF (arrows) is shown in right ACA and MCA territories before acetazolamide challenge. CBV was normal to elevated and MTT was elevated (not shown). (b) After acetazolamide injection, CBF map reveals decrease (arrows) in CBF values from baseline in right ACA and MCA territories, consistent with steal phenomenon. (c) After angioplasty and stent placement, repeat pre-acetazolamide challenge perfusion CT shows more symmetric CBF between the two hemispheres. CBV and MTT were also more symmetric (not shown). (d) After acetazolamide injection, CBF map now shows augmentation of flow in both hemispheres, with flow increased over that of baseline.
Figure 7d. Transverse perfusion CT maps in a 65-year-old man with progressive worsening of left hand and foot numbness and clumsiness. Angiogram (not shown) revealed severe stenosis of precavernous right internal carotid artery. (a) Decreased CBF (arrows) is shown in right ACA and MCA territories before acetazolamide challenge. CBV was normal to elevated and MTT was elevated (not shown). (b) After acetazolamide injection, CBF map reveals decrease (arrows) in CBF values from baseline in right ACA and MCA territories, consistent with steal phenomenon. (c) After angioplasty and stent placement, repeat pre-acetazolamide challenge perfusion CT shows more symmetric CBF between the two hemispheres. CBV and MTT were also more symmetric (not shown). (d) After acetazolamide injection, CBF map now shows augmentation of flow in both hemispheres, with flow increased over that of baseline.
Temporary Balloon Occlusion
Temporary balloon occlusion is performed in patients in whom arterial sacrifice or prolonged temporary occlusion is considered as part of the surgical or endovascular therapy (38,39). Temporary balloon occlusion in conjunction with a quantitative analysis of CBF can help identify patients who will not tolerate permanent occlusion despite clinically passing the temporary balloon occlusion test (38,39). Patients who have clinically passed the temporary balloon occlusion test but who have CBF of less than 30 mL/100 g/min as measured with xenon CT had a 56% incidence of temporary or permanent neurologic deficit, while those with a CBF of greater than 30 mL/100 g/min had a stroke incidence of 7% (40). The use of an absolute CBF value of less than 30 mL/100 g/min as a criterion for the success or failure of temporary balloon occlusion was corroborated in an outcome study in which intracarotid injection of xenon 133 was used to measure CBF (41). The use of acetazolamide challenge during temporary balloon occlusion may help identify patients who are already maximally vasodilated in response to the balloon occlusion alone and cannot vasodilate further in response to the acetazolamide. In such cases, CBF may be maintained during the occlusion by exhausting autoregulatory reserves (42). The goal is to identify patients who clinically pass temporary balloon occlusion but have low CBF or abnormal response to acetazolamide who may not tolerate sacrifice or prolonged occlusion. Xenon CT, transcranial Doppler sonography, SPECT, PET, and perfusion-weighted MR imaging have all been used to evaluate cerebral perfusion during temporary balloon occlusion (38,43).
At our institution, patients undergo angiography and balloon occlusion, during which time they are clinically evaluated for 30 minutes. Patients who pass the clinical portion of the examination are brought to the CT suite with the balloon in place. A perfusion CT scan is obtained with the balloon inflated and again with the balloon deflated. The balloon is reinflated, 1,000 mg of acetazolamide is injected intravenously, and a final perfusion CT scan is obtained.
As with assessment of cerebrovascular reserve, it is unclear whether the same quantitative parameters can be used for perfusion CT as have been used for xenon CT. Initial results from our institution indicate that asymmetry of flow between the two hemispheres and response to acetazolamide challenge may be a better indicator than a CBF value of less than 30 mL/100 g/min (5) (Figs 8, 9). One patient with flow of 17–23 mL/100 g/min in watershed areas ipsilateral to balloon occlusion but with normal response to acetazolamide underwent successful permanent occlusion without stoke (5). However, further investigation is warranted.
Figure 8a. Images in a 45-year-old man with recent onset of left-sided headache and retro-orbital pain. CT, MR, and angiographic images (not shown) indicated giant left supraclinoid carotid aneurysm. Temporary balloon occlusion was performed to determine if patient could tolerate embolization of the aneurysm and sacrifice of the left internal carotid artery. Perfusion CT was performed as part of the evaluation. (a) Transverse CT image obtained during perfusion CT shows large enhancing aneurysm (arrows) in left frontal region. (b) Transverse perfusion CT map with balloon inflated in left internal carotid artery shows relatively symmetric CBF throughout both hemispheres except in area corresponding to the aneurysm. (c) Transverse CBF map with balloon deflated again demonstrates symmetric CBF except in area of aneurysm (arrows). (d) With balloon inflated after acetazolamide injection, transverse CBF map shows normal augmentation of flow in both hemispheres except in area corresponding to aneurysm. Patient did well after treatment of aneurysm, with no neurologic deficit.
Figure 8b. Images in a 45-year-old man with recent onset of left-sided headache and retro-orbital pain. CT, MR, and angiographic images (not shown) indicated giant left supraclinoid carotid aneurysm. Temporary balloon occlusion was performed to determine if patient could tolerate embolization of the aneurysm and sacrifice of the left internal carotid artery. Perfusion CT was performed as part of the evaluation. (a) Transverse CT image obtained during perfusion CT shows large enhancing aneurysm (arrows) in left frontal region. (b) Transverse perfusion CT map with balloon inflated in left internal carotid artery shows relatively symmetric CBF throughout both hemispheres except in area corresponding to the aneurysm. (c) Transverse CBF map with balloon deflated again demonstrates symmetric CBF except in area of aneurysm (arrows). (d) With balloon inflated after acetazolamide injection, transverse CBF map shows normal augmentation of flow in both hemispheres except in area corresponding to aneurysm. Patient did well after treatment of aneurysm, with no neurologic deficit.
Figure 8c. Images in a 45-year-old man with recent onset of left-sided headache and retro-orbital pain. CT, MR, and angiographic images (not shown) indicated giant left supraclinoid carotid aneurysm. Temporary balloon occlusion was performed to determine if patient could tolerate embolization of the aneurysm and sacrifice of the left internal carotid artery. Perfusion CT was performed as part of the evaluation. (a) Transverse CT image obtained during perfusion CT shows large enhancing aneurysm (arrows) in left frontal region. (b) Transverse perfusion CT map with balloon inflated in left internal carotid artery shows relatively symmetric CBF throughout both hemispheres except in area corresponding to the aneurysm. (c) Transverse CBF map with balloon deflated again demonstrates symmetric CBF except in area of aneurysm (arrows). (d) With balloon inflated after acetazolamide injection, transverse CBF map shows normal augmentation of flow in both hemispheres except in area corresponding to aneurysm. Patient did well after treatment of aneurysm, with no neurologic deficit.
Figure 8d. Images in a 45-year-old man with recent onset of left-sided headache and retro-orbital pain. CT, MR, and angiographic images (not shown) indicated giant left supraclinoid carotid aneurysm. Temporary balloon occlusion was performed to determine if patient could tolerate embolization of the aneurysm and sacrifice of the left internal carotid artery. Perfusion CT was performed as part of the evaluation. (a) Transverse CT image obtained during perfusion CT shows large enhancing aneurysm (arrows) in left frontal region. (b) Transverse perfusion CT map with balloon inflated in left internal carotid artery shows relatively symmetric CBF throughout both hemispheres except in area corresponding to the aneurysm. (c) Transverse CBF map with balloon deflated again demonstrates symmetric CBF except in area of aneurysm (arrows). (d) With balloon inflated after acetazolamide injection, transverse CBF map shows normal augmentation of flow in both hemispheres except in area corresponding to aneurysm. Patient did well after treatment of aneurysm, with no neurologic deficit.
Figure 9a. Images in a 33-year-old man with history of medullary thyroid cancer and metastasis to right cavernous sinus with right cavernous sinus syndrome. Perfusion CT was performed in conjunction with temporary balloon occlusion to determine if patient could tolerate right carotid sacrifice as part of aggressive skull base excision of metastases. (a) Transverse perfusion CT map with balloon inflated before acetazolamide challenge shows globally low CBF in both hemispheres (quantitative values, 13-29 mL/100 g/min) but with asymmetrically lower flow in much of right hemisphere. There was corresponding increased CBV and MTT in the right hemisphere (not shown). (b) Transverse CBF map with balloon deflated shows that CBF is more symmetric but still globally low. (c) Transverse CBF map obtained with balloon inflated after acetazolamide injection shows multiple areas of steal. Because of these findings, patient underwent superficial temporal artery-to-MCA bypass before resection of metastases. (d) Transverse diffusion-weighted MR image (b = 1,000 sec/mm2) obtained after surgery shows graft has become occluded and patient has developed right MCA stroke (arrows).
Figure 9b. Images in a 33-year-old man with history of medullary thyroid cancer and metastasis to right cavernous sinus with right cavernous sinus syndrome. Perfusion CT was performed in conjunction with temporary balloon occlusion to determine if patient could tolerate right carotid sacrifice as part of aggressive skull base excision of metastases. (a) Transverse perfusion CT map with balloon inflated before acetazolamide challenge shows globally low CBF in both hemispheres (quantitative values, 13-29 mL/100 g/min) but with asymmetrically lower flow in much of right hemisphere. There was corresponding increased CBV and MTT in the right hemisphere (not shown). (b) Transverse CBF map with balloon deflated shows that CBF is more symmetric but still globally low. (c) Transverse CBF map obtained with balloon inflated after acetazolamide injection shows multiple areas of steal. Because of these findings, patient underwent superficial temporal artery-to-MCA bypass before resection of metastases. (d) Transverse diffusion-weighted MR image (b = 1,000 sec/mm2) obtained after surgery shows graft has become occluded and patient has developed right MCA stroke (arrows).
Figure 9c. Images in a 33-year-old man with history of medullary thyroid cancer and metastasis to right cavernous sinus with right cavernous sinus syndrome. Perfusion CT was performed in conjunction with temporary balloon occlusion to determine if patient could tolerate right carotid sacrifice as part of aggressive skull base excision of metastases. (a) Transverse perfusion CT map with balloon inflated before acetazolamide challenge shows globally low CBF in both hemispheres (quantitative values, 13-29 mL/100 g/min) but with asymmetrically lower flow in much of right hemisphere. There was corresponding increased CBV and MTT in the right hemisphere (not shown). (b) Transverse CBF map with balloon deflated shows that CBF is more symmetric but still globally low. (c) Transverse CBF map obtained with balloon inflated after acetazolamide injection shows multiple areas of steal. Because of these findings, patient underwent superficial temporal artery-to-MCA bypass before resection of metastases. (d) Transverse diffusion-weighted MR image (b = 1,000 sec/mm2) obtained after surgery shows graft has become occluded and patient has developed right MCA stroke (arrows).
Figure 9d. Images in a 33-year-old man with history of medullary thyroid cancer and metastasis to right cavernous sinus with right cavernous sinus syndrome. Perfusion CT was performed in conjunction with temporary balloon occlusion to determine if patient could tolerate right carotid sacrifice as part of aggressive skull base excision of metastases. (a) Transverse perfusion CT map with balloon inflated before acetazolamide challenge shows globally low CBF in both hemispheres (quantitative values, 13-29 mL/100 g/min) but with asymmetrically lower flow in much of right hemisphere. There was corresponding increased CBV and MTT in the right hemisphere (not shown). (b) Transverse CBF map with balloon deflated shows that CBF is more symmetric but still globally low. (c) Transverse CBF map obtained with balloon inflated after acetazolamide injection shows multiple areas of steal. Because of these findings, patient underwent superficial temporal artery-to-MCA bypass before resection of metastases. (d) Transverse diffusion-weighted MR image (b = 1,000 sec/mm2) obtained after surgery shows graft has become occluded and patient has developed right MCA stroke (arrows).
Vasospasm
Vasospasm is a frequent complication in the early clinical course after aneurysmal SAH (12). Angiographic evidence of vasospasm is present in 60%–80% of patients with SAH, with approximately 32% of patients becoming symptomatic (44–46). In approximately 50% of cases, vasospasm is manifested by the onset of a neurologic deficit related to ischemia, with progression to infarction occurring in approximately half of the symptomatic cases (44). In addition, it is estimated that among patients with aneurysmal SAH who reach neurosurgical referral centers, 7% will be severely disabled and 7% will die as a result of vasospasm (46,47). Measurement of CBF can be useful in identifying patients at risk for cerebral ischemia by guiding therapeutic decisions and monitoring response to therapy (13,48–51). Various methods have been employed to measure cerebral perfusion, including PET, SPECT, xenon CT, and transcranial Doppler sonography. The latter has been the most widely used, but it is operator dependent, cannot quantify CBF at the tissue level, and may not be specific enough by itself to guide therapy (13,50). Studies with xenon CT have indicated that patients are at risk of ischemia or infarction if CBF decreases below 18–20 mL/100 g/min, while a PET study found CBF values of less than 12 mL/100 g/min to be a good predictor of subsequent infarction (13,48,51).
Perfusion CT has been used to monitor cerebral perfusion after SAH. Nabavi et al (13) reported that patients with delayed infarct after SAH, presumably due to vasospasm, had a lower mean CBF value than did patients with early or no infarct. Minimal CBF and CBV values occurred both 1–3 and 10–17 days after SAH, and mean CBF and CBV were significantly lower in patients with moderate to severe vasospasm, compared with those with absent to mild vasospasm. While their mean CBF, CBV, and MTT results were similar to those reported in studies with PET and xenon CT to assess patients after SAH, Nabavi et al could not define a clear-cut threshold for cerebral infarct by using perfusion CT. They attributed this to the use of relatively large ROIs, which resulted in volume averaging with normal and ischemic tissue. Further investigation may lead to the definition of a perfusion CT ischemic threshold in patients with aneurysmal SAH. In addition, the ability to assess CBV and MTT may help in understanding the impairment of autoregulation that is believed to occur in some patients after SAH (13) (Fig 10).
Figure 10a. Images in a 56-year-old woman with left hemiparesis 5 days after SAH and 1 day after clipping of posterior communicating artery aneurysm. (a) Transverse perfusion CT image reveals decreased CBF in bilateral ACA territories (arrows), with quantitative values of 7 and 17 mL/100 g/min on the right and left, respectively. (b, c) There is corresponding decrease in (b) CBV (arrows) and slight increase in (c) MTT (arrows) on transverse perfusion maps, findings that are of concern for infarction. (d) Subsequent oblique view from right internal carotid angiogram shows severe spasm involving ACAs (arrows). (e) Despite angioplasty, follow-up transverse diffusion-weighted MR image (b = 1,000 sec/mm2) demonstrates restricted diffusion in ACA territories (arrows) corresponding to changes on perfusion CT images.
Figure 10b. Images in a 56-year-old woman with left hemiparesis 5 days after SAH and 1 day after clipping of posterior communicating artery aneurysm. (a) Transverse perfusion CT image reveals decreased CBF in bilateral ACA territories (arrows), with quantitative values of 7 and 17 mL/100 g/min on the right and left, respectively. (b, c) There is corresponding decrease in (b) CBV (arrows) and slight increase in (c) MTT (arrows) on transverse perfusion maps, findings that are of concern for infarction. (d) Subsequent oblique view from right internal carotid angiogram shows severe spasm involving ACAs (arrows). (e) Despite angioplasty, follow-up transverse diffusion-weighted MR image (b = 1,000 sec/mm2) demonstrates restricted diffusion in ACA territories (arrows) corresponding to changes on perfusion CT images.
Figure 10c. Images in a 56-year-old woman with left hemiparesis 5 days after SAH and 1 day after clipping of posterior communicating artery aneurysm. (a) Transverse perfusion CT image reveals decreased CBF in bilateral ACA territories (arrows), with quantitative values of 7 and 17 mL/100 g/min on the right and left, respectively. (b, c) There is corresponding decrease in (b) CBV (arrows) and slight increase in (c) MTT (arrows) on transverse perfusion maps, findings that are of concern for infarction. (d) Subsequent oblique view from right internal carotid angiogram shows severe spasm involving ACAs (arrows). (e) Despite angioplasty, follow-up transverse diffusion-weighted MR image (b = 1,000 sec/mm2) demonstrates restricted diffusion in ACA territories (arrows) corresponding to changes on perfusion CT images.
Figure 10d. Images in a 56-year-old woman with left hemiparesis 5 days after SAH and 1 day after clipping of posterior communicating artery aneurysm. (a) Transverse perfusion CT image reveals decreased CBF in bilateral ACA territories (arrows), with quantitative values of 7 and 17 mL/100 g/min on the right and left, respectively. (b, c) There is corresponding decrease in (b) CBV (arrows) and slight increase in (c) MTT (arrows) on transverse perfusion maps, findings that are of concern for infarction. (d) Subsequent oblique view from right internal carotid angiogram shows severe spasm involving ACAs (arrows). (e) Despite angioplasty, follow-up transverse diffusion-weighted MR image (b = 1,000 sec/mm2) demonstrates restricted diffusion in ACA territories (arrows) corresponding to changes on perfusion CT images.
Figure 10e. Images in a 56-year-old woman with left hemiparesis 5 days after SAH and 1 day after clipping of posterior communicating artery aneurysm. (a) Transverse perfusion CT image reveals decreased CBF in bilateral ACA territories (arrows), with quantitative values of 7 and 17 mL/100 g/min on the right and left, respectively. (b, c) There is corresponding decrease in (b) CBV (arrows) and slight increase in (c) MTT (arrows) on transverse perfusion maps, findings that are of concern for infarction. (d) Subsequent oblique view from right internal carotid angiogram shows severe spasm involving ACAs (arrows). (e) Despite angioplasty, follow-up transverse diffusion-weighted MR image (b = 1,000 sec/mm2) demonstrates restricted diffusion in ACA territories (arrows) corresponding to changes on perfusion CT images.
Tumors
Tumors are inherently associated with increased angiogenic activity and neovascularization that results in increased blood volume and hyperpermeability related to the immature vessels (6,7,52–54). Results of previous studies (52,53) have indicated that microvascular permeability increases with increasing biologic aggressiveness of tumors, while a reduction in permeability in response to antiangiogenic therapy correlates with decreased tumor growth. Results of initial studies (52,53) in which CBV and PS, a measure of microvascular permeability, were obtained with dynamic contrast agent–enhanced MR imaging indicates PS to be predictive of pathologic grade and to correlate with tumor mitotic activity.
Subsequently, dynamic contrast-enhanced CT has been investigated in animal models of brain tumors and used in human brain tumors to quantify CBF, CBV, and PS (6,8,12). The technique is modified from that used in cerebral ischemia to account for the extravasation of contrast material from the intravascular space to the extravascular space across the impaired blood-brain barrier and to allow measurement of PS (12). These calculations can be performed with commercially available software (CT Perfusion 2; GE Medical Systems). CBF, CBV, and PS were all higher in the tumor and in peritumoral areas than in normal tissue in animal models (12). Initial results obtained with this technique in humans revealed variable elevations in CBF and CBV in the tumor and a more conspicuous increase in PS (6,7). In addition, the elevated PS was evident only in the tumor and not in the surrounding tissues (6,7). CT may prove to be advantageous over MR imaging in the assessment of tumor angiogenesis, given the linear relationship between contrast agent concentration and attenuation changes, the lack of sensitivity to flow, the high spatial resolution, and the absence of susceptibility artifacts (7). However, the exposure to ionizing radiation, the potential for adverse reaction to the contrast agent, and the limited anatomic coverage are limitations of CT, compared with MR, for evaluation of the microvasculature (7).
We have also used perfusion CT to evaluate squamous cell carcinomas of the head and neck (Fig 11). Initial results revealed elevated PS, CBF, and CBV and a lower MTT in the primary tumor site, compared with those values in normal structures; however, further investigation is still necessary. This technique may provide a way to noninvasively measure tumor malignancy, guide biopsies to the most malignant portion of the tumor, and assess response to treatment (6,7,53).
Figure 11a. Aggressive tumor involving tongue base on the left side in a 59-year-old man. (a) Transverse CT image from perfusion scan. ROIs are placed in the following structures: 1 = carotid artery, 2 = jugular vein, 3 = tongue base cancer, 4 = normal contralateral genioglossus/geniohyoid complex, 5 = ipsilateral genioglossus/geniohyoid complex, 6 = sternocleidomastoid muscle, 7 = paraspinal muscle. (b) Transverse PS map shows increased permeability within tumor (arrows), compared with PS of surrounding tissue. (c) Transverse blood volume map shows increased volume within tumor (arrows), compared with that of adjacent tissue. (d) Transverse blood flow map shows increased flow within tumor (arrows), compared with that of surrounding tissue. (e) Transverse MTT map shows decreased MTT (arrows), compared with that of adjacent normal structures.
Figure 11b. Aggressive tumor involving tongue base on the left side in a 59-year-old man. (a) Transverse CT image from perfusion scan. ROIs are placed in the following structures: 1 = carotid artery, 2 = jugular vein, 3 = tongue base cancer, 4 = normal contralateral genioglossus/geniohyoid complex, 5 = ipsilateral genioglossus/geniohyoid complex, 6 = sternocleidomastoid muscle, 7 = paraspinal muscle. (b) Transverse PS map shows increased permeability within tumor (arrows), compared with PS of surrounding tissue. (c) Transverse blood volume map shows increased volume within tumor (arrows), compared with that of adjacent tissue. (d) Transverse blood flow map shows increased flow within tumor (arrows), compared with that of surrounding tissue. (e) Transverse MTT map shows decreased MTT (arrows), compared with that of adjacent normal structures.
Figure 11c. Aggressive tumor involving tongue base on the left side in a 59-year-old man. (a) Transverse CT image from perfusion scan. ROIs are placed in the following structures: 1 = carotid artery, 2 = jugular vein, 3 = tongue base cancer, 4 = normal contralateral genioglossus/geniohyoid complex, 5 = ipsilateral genioglossus/geniohyoid complex, 6 = sternocleidomastoid muscle, 7 = paraspinal muscle. (b) Transverse PS map shows increased permeability within tumor (arrows), compared with PS of surrounding tissue. (c) Transverse blood volume map shows increased volume within tumor (arrows), compared with that of adjacent tissue. (d) Transverse blood flow map shows increased flow within tumor (arrows), compared with that of surrounding tissue. (e) Transverse MTT map shows decreased MTT (arrows), compared with that of adjacent normal structures.
Figure 11d. Aggressive tumor involving tongue base on the left side in a 59-year-old man. (a) Transverse CT image from perfusion scan. ROIs are placed in the following structures: 1 = carotid artery, 2 = jugular vein, 3 = tongue base cancer, 4 = normal contralateral genioglossus/geniohyoid complex, 5 = ipsilateral genioglossus/geniohyoid complex, 6 = sternocleidomastoid muscle, 7 = paraspinal muscle. (b) Transverse PS map shows increased permeability within tumor (arrows), compared with PS of surrounding tissue. (c) Transverse blood volume map shows increased volume within tumor (arrows), compared with that of adjacent tissue. (d) Transverse blood flow map shows increased flow within tumor (arrows), compared with that of surrounding tissue. (e) Transverse MTT map shows decreased MTT (arrows), compared with that of adjacent normal structures.
Figure 11e. Aggressive tumor involving tongue base on the left side in a 59-year-old man. (a) Transverse CT image from perfusion scan. ROIs are placed in the following structures: 1 = carotid artery, 2 = jugular vein, 3 = tongue base cancer, 4 = normal contralateral genioglossus/geniohyoid complex, 5 = ipsilateral genioglossus/geniohyoid complex, 6 = sternocleidomastoid muscle, 7 = paraspinal muscle. (b) Transverse PS map shows increased permeability within tumor (arrows), compared with PS of surrounding tissue. (c) Transverse blood volume map shows increased volume within tumor (arrows), compared with that of adjacent tissue. (d) Transverse blood flow map shows increased flow within tumor (arrows), compared with that of surrounding tissue. (e) Transverse MTT map shows decreased MTT (arrows), compared with that of adjacent normal structures.
CONTROVERSIES
Although quantitative values can be acquired with perfusion CT, the accuracy of the flow values obtained has not been fully validated. It has not been determined if normal and disease thresholds as measured with PET or xenon CT can be applied in perfusion CT. Perfusion CT uses an intravascular tracer to measure intravascular CBF, which likely reflects a different physiologic mechanism than that of PET and xenon CT, which employ a diffusible tracer (55). The authors of one study (28) in which perfusion CT was compared with xenon CT in a small heterogeneous group of patients concluded that in regions excluding major vessels (Fig 12), quantitative results with perfusion CT are in agreement with those of xenon CT. The mean CBF value obtained in the basal ganglia by using perfusion CT in a group of control patients in another study (16) was greater than that reported in the literature for the basal ganglia when PET was used and for the cortical gray matter when xenon CT was used. However, authors of a different study (2) reported systematically low values for CBF as measured with perfusion CT, compared with the values reported for xenon CT, while CBV values obtained with perfusion CT were in good agreement with CBV values obtained by using MR techniques.
Figure 12a.(a-c) Because fewer major vessel branches (from a to c) are included in the ROI on transverse CBF maps, CBF values decrease.
Figure 12b.(a-c) Because fewer major vessel branches (from a to c) are included in the ROI on transverse CBF maps, CBF values decrease.
Figure 12c.(a-c) Because fewer major vessel branches (from a to c) are included in the ROI on transverse CBF maps, CBF values decrease.
Further studies comparing perfusion CT with more established methods of measuring cerebral perfusion are needed to fully validate the quantitative value of this technique (2,16). Other than exclusion of areas that contain major blood vessel branches, there are no standardized techniques or guidelines for placing ROIs on the perfusion maps, which may make comparing results between different investigators difficult. Larger ROIs may result in greater volume averaging of gray and white matter and, thus, lower quantitative values for CBF, compared with the results obtained when smaller ROIs centered in the cortex are used.
Uncertainties also exist regarding how the quantitative values should be calculated. The choice of input artery clearly plays a role, with different quantitative results depending on the artery chosen (Fig 13). It is probably more accurate to use an input artery from the normal side, although further investigation is still necessary (29). In some clinical situations, the normal side may not be known at the time of the examination or there may be bilateral or diffuse disease, as is often the case with atherosclerosis or vasospasm. In animal and human studies, extracranial arteries have been used as the arterial input (4,13,17). The animal studies showed highly significant correlates for CBF, CBV, and MTT values calculated with extracranial artery and internal carotid artery inputs (4). In human studies, there was good correlation between perfusion CT–derived data with the radial artery and other clinical and neuroradiologic findings (4,13). With the radial artery as the arterial input, however, the use of a specially designed holder for the patient’s forearm was necessary (4,13). We are currently investigating the superficial temporal artery as the arterial input in cases of bilateral or diffuse disease; however, one drawback is that this artery may not always be identifiable. The size of the arterial ROI may also affect results (55).
Figure 13a. Depending on the choice of input artery, (a) ACA, (b) right MCA, or (c) left MCA, the quantitative values for CBF vary, as seen on transverse CBF maps.
Figure 13b. Depending on the choice of input artery, (a) ACA, (b) right MCA, or (c) left MCA, the quantitative values for CBF vary, as seen on transverse CBF maps.
Figure 13c. Depending on the choice of input artery, (a) ACA, (b) right MCA, or (c) left MCA, the quantitative values for CBF vary, as seen on transverse CBF maps.
The reproducibility of perfusion CT has also not been fully validated. Software to analyze the perfusion CT data is commercially available and relatively easy to use, although training is required. The amount of training necessary to create reliable perfusion maps is not known (55). Results of an initial investigation (56) indicate that the findings are reproducible between different operators. This study used experienced radiologist investigators to create the perfusion maps. The reproducibility between experienced and inexperienced radiologists, between radiologists and technologists, or between the same radiologist or technologist on different days has not been evaluated (55).
Another limitation of perfusion CT is its restricted anatomic coverage. A toggling-table technique has been described that allows coverage at two distinct table positions (57). While this does allow greater coverage, it is at the expense of temporal resolution (5 seconds). Alternatively, if an abnormality is clinically suspected distant from the level of the basal ganglia, such as in the posterior fossa, the examination can be tailored to the suspected area (Fig 14).
Figure 14a. Images in a 72-year-old man who became confused and developed left-sided ataxia during hepatic angiography. Suspicion of a posterior fossa pathologic condition lead to perfusion CT performed through the level of the cerebellum. (a-c) Transverse perfusion CT maps show (a) decreased CBF (arrows), (b) decreased CBV (arrows), and (c) increased MTT (arrows) in medial left cerebellum. (d) Transverse diffusion-weighted MR image (b = 1,000 sec/mm2) shows corresponding area of restricted diffusion (arrows).
Figure 14b. Images in a 72-year-old man who became confused and developed left-sided ataxia during hepatic angiography. Suspicion of a posterior fossa pathologic condition lead to perfusion CT performed through the level of the cerebellum. (a-c) Transverse perfusion CT maps show (a) decreased CBF (arrows), (b) decreased CBV (arrows), and (c) increased MTT (arrows) in medial left cerebellum. (d) Transverse diffusion-weighted MR image (b = 1,000 sec/mm2) shows corresponding area of restricted diffusion (arrows).
Figure 14c. Images in a 72-year-old man who became confused and developed left-sided ataxia during hepatic angiography. Suspicion of a posterior fossa pathologic condition lead to perfusion CT performed through the level of the cerebellum. (a-c) Transverse perfusion CT maps show (a) decreased CBF (arrows), (b) decreased CBV (arrows), and (c) increased MTT (arrows) in medial left cerebellum. (d) Transverse diffusion-weighted MR image (b = 1,000 sec/mm2) shows corresponding area of restricted diffusion (arrows).
Figure 14d. Images in a 72-year-old man who became confused and developed left-sided ataxia during hepatic angiography. Suspicion of a posterior fossa pathologic condition lead to perfusion CT performed through the level of the cerebellum. (a-c) Transverse perfusion CT maps show (a) decreased CBF (arrows), (b) decreased CBV (arrows), and (c) increased MTT (arrows) in medial left cerebellum. (d) Transverse diffusion-weighted MR image (b = 1,000 sec/mm2) shows corresponding area of restricted diffusion (arrows).
In summary, perfusion CT is clearly a viable alternative to other modalities used to measure cerebral perfusion. This technique is fast and available for most standard spiral CT scanners equipped with the appropriate software. Perfusion CT can be used to assess not only patients with acute stroke but also a wide range of patients with other cerebrovascular diseases. It may also be helpful in the diagnosis and subsequent treatment response in patients with a variety of tumors. Further investigations are necessary to determine the accuracy, reliability, and reproducibility of the quantitative results.
No comments:
Post a Comment