British Journal of Radiology (2004) 77, S20-S26
© 2004 British Institute of Radiology
doi: 10.1259/bjr/29004097
Brain imaging using multislice CT: a personal perspective
S F S Halpin, MBBS, MRCP, FRCR
University Hospital of Wales, Heath Park, Cardiff CF14 4XW, UK
 |
Abstract
|
|---|
This article outlines the use of multislice CT of the brain with regard to routine scanning, as well as high speed and/or high resolution multiplanar reconstructions. It describes in detail the recent developments in cerebral perfusion scanning and CT arteriography. It will also outline the approach of the University Hospital of Wales to imaging with this new technology, emphasizing the differences between helical and non-helical techniques. Most of the images are taken from 16-slice machines, but the methods described are applicable to 4- and 8-slice scanners, and can readily be modified for the forthcoming 64-detector array equipment that will be available within the next few months.
|
Introduction
|
|---|
In general terms, multislice CT does not impact as heavily on routine brain scanning as it does on body imaging, simply because the brain moves little with respiration and does not provide a challenge in terms of spatial coverage. However, the general principle of "scan thin, view thick" applies as much in neuroradiology as it does elsewhere.
The first question to address is how to carry out the routine brain examination, and in particular, should helical or non-helical mode be used? Previous experience has taught that using helical mode in the head will lead to artefact at the brain–bone interface, resulting in loss of tissue definition or the false positive appearance of subdural collections. Even with modern multislice scanners, helical artefact can still be troublesome unless the slice collimation is kept to 1.25 mm or less. It is because of the potential for creating artefact that we stick to non-helical scans for our routine brain examinations, but even here there are several different options. It must be emphasized that the options we have chosen are the ones that best suit our case-mix and our local situation. Other alternatives may be just as valid in other departments.
Table 1
demonstrates some advantages and disadvantages of multislice CT, and Table 2 shows our protocol for a routine head scan using multislice CT.
|
Scanning and imaging technique
|
|---|
We use frontal and lateral localizers, and define three groups of slices, angled approximately parallel to the floor of the anterior cranial fossa — this usually equates to an angle rotated 15° from the radiographic baseline. Since the chance of a CT scan of the brain causing cataracts is zero, scans need not be angled to avoid the orbits. Furthermore, a conventional slice angle leads to better detection of pathology in the orbits, whilst assessment of Chiari malformation is difficult when the gantry is angled steeply from the anterior to the posterior fossa.
The first group of scans is through the posterior fossa, and starts just above C1, terminating at the lower third ventricle. Collimation is 1.25 mm, and images are reconstructed in contiguous 5 mm slices. Tube voltage (kV) and tube potential (mA) are set relatively high because the bone is relatively thick in this region, but also because we wish to retain the ability to retrospectively reconstruct the data into high quality 1.25 mm slices, should that prove necessary. The second group uses 3.75 mm slices, with images reconstructed at 7.5 mm, using lower kV and mA, and continues to the upper part of the lateral ventricles, where a third group consists of an identical slice profile, but lower mA. This scan protocol usually produces 22 images, which are printed on one sheet of film. Merging thin slices into larger ones dramatically reduces beam hardening artefact and produces images of far higher quality than those that would be produced by scans using 5 mm collimation.
Images through the posterior fossa are routinely reconstructed at 2.5 mm: these are not normally filmed but are available for soft-copy review should the radiologist wish to examine an area in more detail. A minor variation of this protocol would be to use the automated dose reduction features available on most manufacturers' equipment. This allows the user to set a noise index determined by practical experience. The scanner then calculates the lowest mA for each rotation that will give acceptable images. Given guidance from the manufacturer and some trial and error experience, these techniques can certainly reduce dose appreciably, at the expense, perhaps, of slightly grainy images.
Figure 1 demonstrates image quality through the temporal poles, traditionally a difficult area for CT.
By using 1.25 mm collimation in this way, it is possible to retrospectively reconstruct images through the posterior fossa to 1.25 mm, as long as raw data are retained. Typically in a busy unit, raw data for 2 days' scanning are kept on the console, and this is usually sufficient to allow thin slice reconstruction at the time of reporting.
An example of the use of multiple reconstructions in multislice CT is in the search for an acoustic schwannoma. In the past, the protocol was for a routine whole brain scan, followed by 3 mm images through the posterior fossa after injection of intravenous (iv) contrast medium, and then 1.5 mm slices through the skull base using a bony algorithm. Now, simply a routine pre- and post-contrast examination is performed, but the post-contrast images in the posterior fossa are reconstructed to 2.5 mm, and the skull base is imaged at 1.25 mm from a third reconstruction with a bone algorithm. Thin slice data from such a routine scan can then be used to make high quality reformats (Figure 2).
View larger version (73K):
[in this window]
[in a new window]
|
Figure 2. Sagittal reformat in a patient with craniopharyngioma, from a routine non-helical scan reconstructed at 1.25 mm. Note the fine detail of the aqueduct of Sylvius.
|
|
Similarly, if while reporting a brain scan the radiologist sees a possible lesion, for example in the orbit, then further reconstructions can be obtained with reformats in the axial and coronal plane, which will better define any pathology.
The GE 16-slice scanners have a minimum collimation of 0.625 mm. Using a 512 x 512 matrix, the resulting voxels are almost perfectly isotropic, yielding reformats that are hard to distinguish from native images. Using 16-slice CT at its most effective would suggest that we should scan the whole brain at this collimation, reconstructing images as 5 mm contiguous slices. Perhaps surprisingly, this would not result in a large increase in absorbed dose since this protocol uses its radiation very efficiently, making good use of the penumbra from adjacent slices. Our trials with brain scans at 0.625 mm suggest that the skin dose is only slightly higher than in our usual protocol, and this technique would produce exquisite reformats of, for example, middle ear anatomy in congenital or acquired deafness, or of the floor of the third ventricle in patients requiring third ventriculostomy.
It is worth noting that the optimal time for gaining post-contrast scans is at least 5 min after iv contrast administration. Since modern CT scans take only a few seconds, it is entirely possible to obtain post-contrast scans before the contrast medium has optimally entered the extracellular space of a potentially enhancing lesion. For this reason, for routine enhancement it is best to inject the contrast medium slowly, and to wait 3-4 min after the injection before obtaining the post-contrast scan.
Helical scans are occasionally used for brain imaging when rapid scans are needed in a restless or semi-conscious patient. Using 1.25 mm collimation, and a pitch of 0.9, a small child's head can be scanned in less than 5 s, considerably reducing the need for sedation. In a trauma victim, images of facial fractures are often needed in addition to the brain, which requires helical scanning, usually with a straight gantry, using 0.625 mm collimation. The facial bones are then reconstructed in three planes, with an additional volume rendered three-dimensional (3D) model, and reformatted 5 mm slices through the brain parallel to the floor of the anterior cranial fossa are printed. Figure 3 demonstrates typical image quality from a helical acquisition through the brain. Table 3 outlines our protocol for helical CT of the brain.
View larger version (50K):
[in this window]
[in a new window]
|
Figure 3. (a) Coronal reformat from 0.625 mm helical acquisition in a restless patient with a chronic subdural collection. Note the excellent grey–white differentiation. (b) Axial reformat from a helical acquisition in a sick patient with a left middle cerebral artery infarct. This was a 6-s scan for maximum speed, somewhat lessening final image quality compared with (a).
|
|
|
CT perfusion
|
|---|
CT perfusion (CTP) is becoming widely accepted as an important tool in the assessment of brain ischaemia, but it is not familiar to all radiologists.
The driving force behind the emergency assessment of brain ischaemia, often termed the "brain attack" protocol, has been the publication of the National Institute of Neurological Diseased Stroke (NINDS) [1] and European Cooperative Acute Stroke Study (ECASS) [2] studies, which indicated that iv rTPA can improve outcome if given within 3–6 h of ictus.
In Cardiff, we exclusively use CT for the investigation of acute stroke. MR diffusion/perfusion imaging is now reserved for research, and we believe that CT alone will provide all the necessary information to assess which patients should receive thrombolysis. The therapeutic target in "brain attack" is a patient with a short history of acute stroke, with no haemorrhage on a plain CT scan. There must be demonstrable embolic occlusion of a major cerebral artery, and cerebral perfusion studies should demonstrate salvageable brain tissue. After a plain scan for acute stroke patients, we carry out a CT angiogram of the circle of Willis, and this is followed by the perfusion study. Here, a 20 mm slab, usually through the basal ganglia, is selected and scanned at 1-s intervals for 50 s before, during and after a 40 ml bolus of iohexol 300. Scan data are then retrospectively reconstructed to two 10 mm slices with half second intervals, and the data are subjected to analysis with the semi-automated "CT Perfusion 3" software on an Advantage Windows workstation, resulting in parametric maps of cerebral blood flow (CBF), cerebral blood volume (CBV) and mean transit time (MTT). The entire protocol takes less than 15 min, including lifting the patient on and off the scanner. Table 4 outlines our routine protocol.
There is considerable literature on the measurement and assessment of CBF, and a brief review is necessary to understand the parameters and quantitative data that the technique provides.
In 1981, Jones et al [3] used a primate model to examine the effects of restricted blood flow by clamping middle cerebral arteries to varying degrees and for different time periods before examining the brain histologically for infarction. Their findings form the basis for the practical use of perfusion studies in humans today.
- 15–30 min of complete ischaemia produced only microscopic infarcts.
- 2–3 h of ischaemia produced moderate to large infarcts.
- CBF of less than 23 ml 100 g–1 min–1 caused reversible paralysis.
- CBF < 10–12 ml 100 g–1 min–1 for 2–3 hours, or CBF <17–18 ml 100 g–1 min–1 permanently, resulted in irreversible paralysis.
In 1985, Powers et al [4] carried out a PET study of human ischaemia and found the minimum value of CBF where there was no clinical deficit to be 19 ml 100 g–1 min–1. Mayer et al [5] in 2001 found that all patients with a 70% reduction in CBF had an infarct, whilst half of patients with a 40–70% reduction in CBF had an infarct. In a CTP study of human ischaemia, Eastwood et al [6] found a strong correlation between a CBF of less than 15 ml 100 g–1 min–1 and infarction, a finding that correlates well with xenon CT, MRI and PET perfusion studies.
These figures have been paraphrased to the following: CBF of greater than 25 ml 100 g–1 min–1 is unlikely to lead to permanent stroke; between 20 ml 100 g–1 min–1 and 25 ml 100 g–1 min–1 there is likely to be a reversible deficit; a CBF of 15–20 ml 100 g–1 min–1 for more than 3–6 h will cause permanent loss of function; and a CBF of less than 15 ml 100 g–1 min–1 will cause permanent damage in less than 3 h. In addition, and in particular if there is any doubt regarding the quantitative data, then the relative reduction in CBF between the affected lobe and the normal one should be calculated.
These figures are only of use if CTP produces accurate quantitative data. The GE perfusion software is based on work by Cenic et al [7], who described a deconvolution analysis that means it is not necessary to know the size of the arterial input to calculate the true CBF.
There have been several studies that have validated CTP against other imaging studies, including PET, SPECT, xenon CT and MRI. Ezzeddine et al [8] found that admission CT arteriography (CTA)/CTP imaging significantly improved accuracy over that of initial clinical assessment and CT imaging alone in the determination of infarct localization, site of vascular occlusion and Oxfordshire classification in acute stroke patients.
Wintermark et al [9] have also published a comparison between the CBF defect as measured by CT (CT-CBF) and MRI, in which they found a strong correlation between CT-CBF and diffusion weighted (DW) MRI; and a strong correlation between the ischaemic penumbra and perfusion weighted (PW) MRI. In 2002 the same authors found that CT-CBF matched the diffusion defect in DW MRI at 3 days if there had been no recanalization, but that the DW lesion was smaller than the CT-CBF defect if successful recanalization had occurred [10]. Wintermark et al [11] found a strong correlation between CTP measurements of CBF with xenon CT.
Eastwood et al [6] reported a good correlation between CT-CBF and MRI, PET and xenon CT in the level of blood flow below which infarction will inevitably occur, but a lesser correlation between the CBF of normal grey matter, which CT-CBF tended to overestimate compared with xenon CT and MRI values.
An interesting model of acute stroke is a patient who is undergoing a test occlusion of an internal carotid artery (ICA). These patients are managed by me by placing a non-detachable balloon in the ICA and then transporting them from the vascular room into CT. Once on the table, the balloon is inflated and the perfusion scans are acquired. The balloon is then deflated and the patient is returned to the vascular room. The perfusion data are analysed before returning to inflate the balloon again to perform a 40-min clinical occlusion test only if the patient has "passed" the CTP test. In these patients, the perfusion study shows a rapid reduction in CBF, and a prolongation in MTT, but no change in CBV. In contrast, in established stroke there are matched defects in CBF and CBV.
It is often difficult to precisely time the onset of an embolus — many patients wake in the morning with a clinical stroke, or cannot remember the onset of symptoms. If, however, we believe there is a realistic chance of thrombolysis, an acute stroke patient in my hospital will have CTP as soon as possible. If the CBF defect is larger than the defect shown on the corresponding CBV map, then I believe that this is analogous to perfusion–diffusion mismatch in MRI, demonstrating salvageable brain tissue. Loss of CBV seems to be a terminal or near terminal event in the cascade from ischaemia to infarction. MTT maps tend to overestimate the ischaemic area: adequate oxygenated blood may take longer to arrive when travelling by a collateral route but still provide adequate supply. I believe that CBF values of 20–25 ml 100 g–1 min–1 are a better indicator of ischaemic but probably viable tissue. Figure 4 demonstrates a typical patient with a basilar thrombosis.
View larger version (95K):
[in this window]
[in a new window]
|
Figure 4. Parametric maps of (a) cerebral blood flow (CBF), measured in ml 100 g–1 min–1, (b) cerebral blood volume (CBV), measured in ml and (c) mean transit time (MTT), measured in s, in a patient who collapsed in the hospital while visiting a relative. A plain CT at 1 h was normal. The maps show typical reduction of CBF and prolongation of MTT in the territory supplied by the basilar artery, with preserved CBV.
|
|
It seems probable that, in the face of ischaemia, viable tissue attempts to autoregulate its perfusion, perhaps by vasodilatation, sometimes resulting in a slightly raised CBV. In the context of acute stroke, a raised CBV represents viable brain: a normal CBV could represent ischaemic brain on the brink of failing to autoregulate; or completely viable tissue: neither precludes thrombolysis. If, however, the CBF and CBV maps are matched, then it is likely that there is no possibility of salvage. Autoregulation has failed, or has no more purpose, and the risks of thrombolysis may well outweigh the benefits. The ideal patient for thrombolysis will have a 3–6 h history; a CBF defect larger than the CBV defect; and quantitative data suggesting survivable tissue. Intra-arterial thrombolysis is performed in nearly all patients who have CBF/CBV mismatch and regional CBF above 15 ml 100 mg–1 min–1. Intra-arterial thrombolysis should be avoided if the CBF and CBV maps matched, and if the CBF values are below 15 ml 100 g–1 min–1 with a history longer than, or approaching, 3 h. Exceptions may occur and clinical judgement is paramount. It is better to note the perfusion data than to rely on the clinical history of ischaemic time alone.
Critics of CT perfusion often point to the relatively small amount of brain covered in the 2 cm slab, especially compared with the whole brain coverage of diffusion/perfusion MRI. This limitation is recognized and brain attack protocols are aimed at major vessel occlusions, all of which will be detected by CTP and CTA. However, PW MRI will detect small lacunar infarcts below the threshold of detection of CTP, and other distal cortical infarcts beyond the CTP slab. Nevertheless, CTP is more than adequate to deal with the great majority of patients within our therapeutic target.
|
CT arteriography
|
|---|
With the advent of fast multislice CT scanners, it has become possible to acquire a CTA data set using sub-millimetre and sub-second scanning. Table 5 outlines the University Hospital of Wales' protocol, which is reliable and accurate enough to completely replace digital subtraction angiography (DSA) in the investigation of aneurysmal subarachnoid haemorrhage (SAH). The images are obtained in approximately 8 s, from C1 to the lateral ventricles. Since the scan ascends from the skull base, the basal carotid artery is scanned before the cerebral veins have filled with contrast medium, acheiving true CTA. In turn, this means that the cavernous segment of the carotid artery is not obscured by venous filling: a significant step forward in cerebral CTA.
To achieve this, it is necessary to deliver a short bolus of concentrated contrast medium and to time accurately the CTA acquisition with the arrival of the bolus in the head. There are two methods to achieve this. First, most manufacturers have a variant of the "SmartPrep" technique used by GE. In this method, a region of interest (ROI) is placed over a carotid artery and is monitored by the scanner for Hounsfield unit changes. When the density crosses above a pre-set threshold, the CTA acquisition is triggered a second or so later. Unfortunately the target arteries are small and minor patient movements can ruin the automated detection of the contrast medium bolus, so a test bolus technique is preferable. This requires cine-mode scans over 30 s just above or just below the skull base before and during iv administration of a 20 ml bolus of iohexol 350. Ideally this should be power injected at the same rate as the real injection (i.e. 4 ml s–1) and followed by a chaser of normal saline to best mimic the actual acquisition. The arrival of the bolus can be timed manually or by using a ROI, and the result is used to inform the CTA study. It is best to use a vein in the right antecubital fossa; power injections into the left arm veins can reflux into the jugular system, spoiling the bolus delivered to the brain.
We have found CTA to be better than DSA in the pre-treatment assessment and detection of brain aneurysms. Several studies have demonstrated the efficacy of the technique [12–23], and CTA is far preferable to 2D DSA in this regard. In our department, every patient with non-traumatic SAH will undergo a cerebral CTA study. Maximum intensity projection (MIP) images in three planes are provided immediately to the neurosurgeons, and a volume rendered model is made available for study on the workstation [21]. Our as yet unpublished data have demonstrated that this approach does not miss any intracranial aneurysms, as long as the study is of diagnostic quality. We have compared DSA and CTA in 50 patients. 13 patients had normal studies. 56 aneurysms were detected by both techniques, and all studies were rated to be of diagnostic quality. 30 of the aneurysms were under 5 mm, and 5 were less than 2 mm in diameter. In three patients it was difficult to distinguish with CTA between a true posterior communicating artery aneurysm and an infundibulum, but all of these had larger aneurysms elsewhere responsible for the SAH. Although the literature suggests that CTA is not good at detecting aneurysms smaller than 2 mm [20], this is not our experience and we believe that such small aneurysms are often more clearly seen on CTA because of the ease of manipulation of the volume rendered data set. These results have been confirmed by other reports [12, 24]. As a result of our studies and experience, we have abandoned DSA in the investigation of aneurysmal SAH, preferring to plan treatment using CTA. We continue to carry out 4-vessel DSA as a preliminary step in the embolisation of the causative aneurysm, but we have yet to discover a new aneurysm by doing so. If the CTA study is negative, we perform DSA, because we believe CTA may miss a small arteriovenous malformation or dural fistula, but again we have yet to see such a lesion after normal CTA.
Figures 5 and 6 demonstrate the elegant images provided by CTA.
View larger version (66K):
[in this window]
[in a new window]
|
Figure 5. Steep occipitomental projection from a digital subtraction angiography (DSA) examination (a), and corresponding CT arteriography (CTA) image in the same patient (b). Note that CTA much better demonstrates the duplication of the anterior communicating artery, in association with a small anterior communicating artery aneurysm, thus aiding treatment planning.
|
|
View larger version (161K):
[in this window]
[in a new window]
|
Figure 6. Volume rendered image from a CT arteriography data set. An aneurysm clip is displayed against the terminal carotid artery. Note the lack of distortion of neighbouring vessels.
|
|
|
Conclusion
|
|---|
Multislice CT has re-invigorated CT scanning in neuroradiology. With the advent of new and faster scanning techniques it is now possible to separate arterial from venous flow; to scan small children without the need for sedation or anaesthesia; to demonstrate complex anatomy in multiple planes; and to complete the "brain attack" examination in less than 15 min.
Received for publication December 1, 2003.
Revision received June 24, 2004.
Accepted for publication July 1, 2004.
|
References
|
|---|
- Intracerebral hemorrhage after intravenous t-PA therapy for ischemic stroke. The NINDS t-PA Stroke Study Group. Stroke 1997;28:2109–18.[Abstract/Free Full Text]
- Hacke W, Kaste M, Fieschi C, Toni D, Lesaffre E, von Kummer R, et al. Intravenous thrombolysis with recombinant tissue plasminogen activator for acute hemispheric stroke. The European Cooperative Acute Stroke Study (ECASS). JAMA 1995;274:1017–25.[Abstract/Free Full Text]
- Jones TH, Morawetz RB, Crowell RM, Marcoux FW, Fitzgibbons SJ, DeGirolami U, et al. Thresholds of focal cerebral ischemia in awake monkeys. J Neurosurg 1981;54:773–82.[Medline]
- Powers WJ, Grubb RL Jr, Darriet D, Raichle ME. Cerebral blood flow and cerebral metabolic rate of oxygen requirements for cerebral funciton and viability in humans. J Cereb Blood Flow Metab 1985;5:600–8.[Medline]
- Mayer TE, Hamann GF, Baranczyk J, Rosengarten B, Klotz E, Wiesmann M, et al. Dynamic CT perfusion imaging of acute stroke. Am J Neuroradiol 2000;21:1441–9.[Abstract/Free Full Text]
- Eastwood JD, Lev MH, Azhari T, Lee TJ, Baboriak DP, Delong DM, et al. CT perfusion scanning with deconvolution analysis: pilot study in patients with acute middle cerebral artery stroke. Radiology 2002;222:227–36.[Abstract/Free Full Text]
- Cenic A, Nabavi DG, Craen RA, Gelb AW, Lee TY. Dynamic CT measurement of cerebral blood flow: a validation study. Am J Neuroradiol 1999;20:63–73.[Abstract/Free Full Text]
- Ezzeddine MA, Lev MH, McDonald CT, Rordorf GA, Oliveira-Filho J, Aksoy FG, et al. CT angiography with whole brain perfused blood volume imaging: added clinical value in the assessment of acute stroke. Stroke 2002;33:959–66.[Abstract/Free Full Text]
- Wintermark M, Reichhart M, Thiran JP, Maeder P, Chalaron M, Schnyder PJ, et al. Prognostic accuracy of cerebral blood flow measurement by perfusion computed tomography, at the time of emergency room admission, in acute stroke patients. Ann Neurol 2002;51:417–32.[CrossRef][Medline]
- Wintermark M, Reichhart M, Thiran JP, Maeder P, Chalaron M, Schnyder PJ, et al. Comparison of admission perfusion computed tomography and qualitative diffusion- and perfusion-weighted magnetic resonance imaging in acute stroke patients. Stroke 2002;33:2025–31.[Abstract/Free Full Text]
- Wintermark M, Thiran JP, Maeder P, Schnyder PJ, Meuli R. Simultaneous measurement of regional cerebral blood flow by perfusion CT and stable xenon CT: a validation study. Am J Neuroradiol 2001;22:905–14.[Abstract/Free Full Text]
- Zouaoui A, Sahel M, Marro B, Clemenceau S, Dargent N, Bitar A, et al. Three-dimensional computed tomographic angiography in detection of cerebral aneurysms in acute subarachnoid hemorrhage. Neurosurgery 1997;41:125–30.[CrossRef][Medline]
- Young N, Dorsch NW, Kingston RJ, Markson G, McMahon J. Intracranial aneurysms: evaluation in 200 patients with spiral CT angiography. Eur Radiol 2001;11:123–30.[CrossRef][Medline]
- van Gelder JM. Computed tomographic angiography for detecting cerebral aneurysms: implications of aneurysm size distribution for the sensitivity, specificity, and likelihood ratios. Neurosurgery 2003;53:597–606.[Medline]
- Shen WC. CT angiography for the diagnosis of aneurysm in the bifurcation of the middle cerebral artery. Eur J Radiol 1999;32:129–35.[Medline]
- Seo JB, Im JG, Goo JM, Chung MJ, Moon WK, Lee KH, et al. Comparison of contrast-enhanced CT angiography and gadolinium-enhanced MR angiography in the detection of subsegmental-sized pulmonary embolism. An experimental study in a pig model. Acta Radiol 2003;44:403–10.[CrossRef][Medline]
- Rohnert W, Hanig V, Hietschold V, Abolmaali N. Detection of aneurysm in subarachnoid hemorrhage — CT angiography vs. digital subtraction angiography. Aktuelle Radiol 1998;8:63–70. [In German.][Medline]
- Rajagopal KV, Lakhkar BN, Acharya DK. Three-dimensional CT angiography in the evaluation of cerebral arteries in acute hemorrhage. Neurol India 2003;51:206–7.[Medline]
- Prokop M. Multislice CT angiography. Eur J Radiol 2000;36:86–96.[CrossRef][Medline]
- Pedersen HK, Bakke SJ, Hald JK, Skalpe IO, Anke IM, Sagsveen R, et al. CTA in patients with acute subarachnoid haemorrhage. A comparative study with selective, digital angiography and blinded, independent review. Acta Radiol 2001;42:43–9.[Medline]
- Moore EA, Grieve JP, Jager HR. Robust processing of intracranial CT angiograms for 3D volume rendering. Eur Radiol 2001;11:137–41.[CrossRef][Medline]
- Matsumoto M, Endo Y, Sato M, Sato S, Sakuma J, Konno Y, et al. Acute aneurysm surgery using three-dimensional CT angiography without conventional catheter angiography. Fukushima J Med Sci 2002;48:63–73.[Medline]
- Lenhart M, Bretschneider T, Gmeinwieser J, Ullrich OW, Schlaier J, Feuerbach S. Cerebral CT angiography in the diagnosis of acute subarachnoid hemorrhage. Acta Radiol 1997;38:791–6.[Medline]
- Villablanca JP, Jahan R, Hooshi P, Lim S, Duckwiler G, Patel A, et al. Detection and characterization of very small cerebral aneurysms by using 2D and 3D helical CT angiography. Am J Neuroradiol 2002;23:1187–98.[Abstract/Free Full Text]
Top
Abstract
Introduction
