| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, Seoul National University College of Medicine, Seoul
National University Bundang Hospital, Seongnam-si, 463-707, Korea.
2 Department of Neurology, Seoul National University College of Medicine, Seoul
National University Bundang Hospital, Seongnam-si, Korea.
Received May 6, 2007;
accepted after revision March 28, 2008.
Address correspondence to J. H. Kim
(jaehkim{at}radiol.snu.ac.kr).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Fifty-eight patients who underwent 80-mm-coverage perfusion CT within 24 hours after stroke onset were included. Perfusion CT was performed using a 64-MDCT unit equipped with 40-mm-wide detector and the toggling table technique. Lesion detection by perfusion CT was analyzed using follow-up diffusion-weighted imaging and MR angiography as the reference standards. More conventional 20-mm-coverage perfusion CT was simulated by extracting data obtained at the basal ganglia level for comparison with 80-mm-coverage perfusion CT.
RESULTS. Fifty-one patients had acute infarctions and seven patients did not. For 80-mm-coverage perfusion CT, perfusion abnormality was detected in 42 of 51 patients (sensitivity, 82.4%; and specificity, 85.7%). When patients with small artery disease (small acute infarctions in the basal ganglia, thalamus, corona radiata, and pons) were excluded, sensitivity increased to 92.3%. As compared with 80-mm-coverage perfusion CT, 20-mm-coverage perfusion CT missed nine acute infarctions located above or below the level of the basal ganglia (p = 0.0039).
CONCLUSION. Perfusion CT with 80-mm-coverage was found to be useful as an initial imaging method in acute ischemic stroke, although it had low sensitivity for detecting small acute infarctions. In particular, this technique provided higher lesion detection than 20-mm-coverage perfusion CT.
Keywords: cerebral ischemia • CT • perfusion
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
With the development of MDCT tech nology, perfusion CT combined with CT angiography has become an easily accessible and accurate diagnostic tool in some institutes that provides an early vascular diagnosis useful for therapeutic decisions in acute ischemic stroke [12–15]. However, the most significant drawback of perfusion CT is its limited coverage of the brain in the z-axis. The widely available 4- and 16-MDCT units have a detector width of 20–24 mm and cannot provide sufficient information on entire brain perfusion [14–16]. To overcome this limitation, a toggling table technique with a single contrast injection has been attempted to double the coverage in the z-axis [17]. The recent advent of 64-MDCT equip ped with a wider detector (40 mm) would be expected to increase the brain coverage in the z-axis. In addition, the combined use of a wide detector and the toggling table technique can provide per fusion images of almost the whole brain. However, no clinical report has been issued on the combined use of a wide detector and the toggling table technique.
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
After CT, four perfusion parameter maps of cerebral blood volume (CBV), cerebral blood flow (CBF), mean transit time (MTT), and time to peak (TTP) were created and color-coded according to the range of values using a workstation (Extended Brilliance Workstation, Philips Medical Systems) (Fig. 2). CBV is inferred from a quantitative estimation of the partial volume averaging effect, which is completely absent in a reference pixel at the center of the large superior sagittal venous sinus [18]. Therefore, CBV was calculated using the ratio of two areas: the area under a tissue time–density curve and the area under a reference venous curve, normalized by a hematocrit factor (45%). The impulse function and the related MTT (width of the impulse function) resulted from a deconvolution (singular value decomposition) of the tissue time–density curve by a reference arterial curve [18], which was automatically measured by the software when placing the user-defined region of interest (ROI) around the ascending segment (segment A2) of the anterior cerebral artery. When defining the ROI, care was taken not to include an occluded or diseased vessel. CBF was calculated by dividing CBV by MTT, according to the central volume theorem. TTP was defined as the time required to reach peak enhancement.
Follow-Up MRI
Time intervals between perfusion CT and follow-up MRI were as follows:
within 2 hours (n = 44; mean, 52 minutes), 2–6 hours
(n = 10; mean, 3 hours 20 minutes), and 6–12 hours (n
= 4; mean, 7 hours 35 minutes). MRI was conducted at 1.5 T (Intera, Philips
Medical Systems) using an acute stroke protocol, which consisted of DWI,
FLAIR, conventional gradient-echo, contrast-enhanced T1-weighted imaging in
the transverse plane, and 3D time-of-flight MRA of the intracranial region. DW
images were obtained by single-shot echo-planar imaging using the following
parameters: matrix, 128 x 128; slice thickness, 5 mm; TR range/TE range,
4000–5000/56–65; and b, 1,000 s/mm2. MRA parameters
were as follows: matrix, 256 x 256 interpolated to 512 x 512;
number of slabs, 5; total number of slices, 140; slice thickness, 0.5 mm; and
TR/TE, 23/6.9.
Image Interpretation
The lists of perfusion CT and MR images were randomly arranged for
retrospective interpretation. Two neuroradiologists first reviewed the
perfusion maps unaware of follow-up MRI findings. They determined visually the
presence or absence and location of perfusion abnormalities by consensus. To
recognize old tissue loss, including chronic infarction, unenhanced CT scans
were evaluated in conjunction with perfusion CT scans. The reviewers were
asked to qualitatively identify perfusion abnormalities in one or more of the
parameter maps (CBV, CBF, MTT, and TTP). To compare lesion detectability for
80-mm- and conventional 20-mm-coverage perfusion CT, we simulated
20-mm-coverage perfusion CT by extracting two slices of perfusion maps
obtained at the level of the basal ganglia from 80-mm-coverage perfusion CT.
Lesion detection was thus performed using these two slices.
After all of the perfusion CT data had been analyzed, the MRI scans (including DWI and MRA) were randomly ordered and analyzed in a separate session by the same two neuroradiologists. DWI and MRI were used to determine locations and sizes of acute infarctions and locations of stenoocclusive lesions.
Data Analysis
The 58 patients were arbitrarily classified into three groups: the large
artery occlusion group, the small artery disease group, and the undetermined
group. Patients with large artery occlusion on MRA, which correlated with the
location of acute infarctions, were allocated to the large artery occlusion
group. Those with large territorial infarctions without arterial occlusion on
MRA were also allocated to the large artery occlusion group because early
recanalization of the occluded artery was suspected. The small artery disease
group included those with small acute infarctions in the territories of the
perforator vessels (i.e., basal ganglia, thalamus, corona radiata, and pons)
without large artery occlusion on MRA. Lesions in the undetermined group were
characterized by the presence of small acute infarctions in the hemispheric
region without visualized arterial occlusion on MRA.
The diagnostic performance of 80-mm-coverage perfusion CT was compared with that of follow-up MRI using the following classifications: True-positive means the presence of hypoperfusion on perfusion CT and acute infarction on DWI in the same vascular territory; true-negative means no abnormality by either perfusion CT or DWI; false-positive means abnormality on perfusion CT but no acute infarction on DWI regardless of the presence of stenoocclusive lesions by MRA; and false-negative means normal findings on perfusion CT but acute infarction on DWI.
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
For 80-mm-coverage perfusion CT, perfusion abnormality was detected in 42 of 51 patients with acute infarctions (sensitivity, 82.4%; and specificity, 85.7%; p = 0.0007) (Figs. 3, 4, 5). Nine acute infarctions without perfusion abnormality (i.e., false-negative cases) were found at the corona radiata (n = 3), basal ganglia (n = 2), thalamus (n = 1), pons (n = 2), and superior cortex (n = 1). The size of acute infarctions in these false-negative cases ranged from 3 to 32 mm in maximal diameter. Six of the nine false-negative cases belonged to the small artery disease group, two belonged to the large artery occlusion group, and one, to the undetermined group. One false-positive case was found in a patient with chronic occlusion of the right internal carotid artery and symptoms of transient ischemic attack. All 14 patients who had IV thrombolytic therapy during the time between perfusion CT and MRI belonged to the true-positive group: None of these patients had a false-positive result that might suggest the occurrence of early reperfusion to rescue the tissue at risk.
When the 12 patients belonging to the small artery disease group were excluded from the entire population, the sensitivity and specificity of lesion detection for 80-mm-coverage perfusion CT were 92.3% and 85.7%, respectively (p < 0.0001). When the 34 patients belonging to the large artery occlusion group were excluded, sensitivity and specificity were 58.8% and 85.7%, respectively (p = 0.0778).
Twenty-millimeter-coverage perfusion CT detected abnormal perfusion in 33 of 51 patients with acute infarctions (sensitivity, 64.7%; and specificity, 85.7%; p = 0.0164). As compared with 80-mm-coverage perfusion CT, 20-mm-coverage perfusion CT missed nine acute infarctions located above or below the level of the basal ganglia: that is, superior cortex (n = 4), periventricular white matter (n = 4), and pons (n = 1) (Fig. 5). The lesion detection rate of 80-mm-coverage perfusion CT was significantly higher than that of 20-mm-coverage perfusion CT (p = 0.0039).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In this study, false-negative cases on perfusion CT usually occurred in patients with small acute infarctions, and were probably caused by the greater thickness of perfusion CT (10 mm) as compared with DWI (5 mm). Regardless of the risk of partial volume averaging, 10-mm slice thickness was adopted to increase signal-to-noise ratios [16–19]. Most (67%, 6/9 cases) of the false-negative cases had small artery disease and thus were not candidates for intraarterial thrombolytic therapy. On the other hand, most true-positive cases (76%, 32/42 cases) had large artery occlusion and therefore were primary candidates for intraarterial thrombolytic therapy [3]. It has been suggested that stroke subtypes are not a consideration during triage for IV thrombolytic therapy [20], but further discussion concerning thrombolytic therapy type (intraarterial and IV) with respect to the type of stroke is beyond the scope of this study. Nevertheless, the perfusion CT–based selection of candidates for intraarterial thrombolytic therapy would not appear to be substantially hampered by its low sensitivity for small infarctions.
DWI at the hyperacute stage of stroke is not the best reference standard to reflect the final patient outcome. In this study, DWI was performed immediately after perfusion CT in most patients. Early DWI might reflect hemodynamic abnormalities seen on perfusion CT with a smaller chance of re perfusion or secondary ischemic events com pared with DWI performed later. Never theless, early reperfusion, either spontan eously or by IV thrombolytic therapy, might salvage the ischemic tissue completely, and occlusion might worsen during the short period between perfusion CT and follow-up MRI. One patient with false-positive results who experienced a transient ischemic attack in association with preexisting chronic occlusion of the right internal carotid artery, recovered completely without thrombolytic therapy. In this study, none of the 14 patients who underwent IV thrombolytic therapy had a false-positive result, which suggests that the tissue at risk had been completely rescued. On the other hand, secondary is chem ic events, such as clot progression, clot fragmentation and migration, and new emboli, that develop between perfusion CT and follow-up MRI present a challenge because they might cause infarctions larger than the initial perfusion abnormalities or additional new infarctions. However, on the basis of visual analysis, no patient appeared to have worsening of the occlusion.
Most acute infarctions confined to the superior cortex were detected by 80-mm-coverage perfusion CT and not by 20-mm-coverage perfusion CT, the latter of which was usually performed at the level of the basal ganglia. Twenty-millimeter-coverage perfusion CT did detect most large territorial infarctions but provided less accurate information as to lesion extent. Although the clinical usefulness of 20-mm-coverage perfusion CT has been well established [14–16], its limited coverage of the z-axis is a serious technical limitation. A recent study reported the use of the toggling table technique to extend anatomic coverage in perfusion CT, but only 40 mm of the z-axis was scanned because a 20-mm-width detector was used [17]. Another study used a dual-injection protocol, which consisted of two 20-mm-coverage perfusion CT exam inations at an interval of 5 minutes to double coverage length [14]. The advent of 64-MDCT equipped with a 40-mm-wide detector in combination with the toggling table technique enables 80-mm-coverage perfusion CT, which allows examinations of almost the entire brain, except for the lower posterior fossa and the superior frontoparietal cortices.
Maruya et al. [21] found that the detection rate of perfusion deficit in acute ischemic stroke by 20-mm-coverage perfusion CT was 65.5% overall—100% for 10 large territorial infarctions and 47.4% for 19 small lacunar and subcortical infarctions. False-negative cases were attributed to small lesion size (three patients) and to inadequate coverage (seven patients) in their study. In our study, the sensitivity of 20-mm-coverage perfusion CT with respect to lesion detection was 64.7%, which concurs with this previous study. In our study, however, 80-mm-coverage perfusion CT detected nine more lesions located above or below the basal ganglia level, resulting in an overall sensitivity of 82.4%, which was significantly higher than that of 20-mm-coverage perfusion CT. Thus, 80-mm-coverage perfusion CT is likely to be a potent initial imaging technique in acute ischemic stroke.
The perfusion CT scanning interval used in our study was rather long (4 seconds) as compared with 1–2 seconds required for conventional perfusion CT without a toggling table. Because of the physical limitation imposed by the toggling table technique, the minimum scanning interval was 4 seconds. Wintermark et al. [16] reported that a scanning interval longer than 1 second leads to overestimation of CBV, CBF, and TTP and underestimation of MTT, but they also found that scanning intervals of up to 4 seconds (with a slight increase of contrast volume) did not alter the quantitative accuracy of perfusion CT. Therefore, we think that the slight quantitative inaccuracies introduced during the calculation of perfusion par a meters did not greatly influence our visual analysis. Nevertheless, the effect of a longer scanning interval on the detection of small acute infarctions has not been studied. If we had used a dual-injection technique instead of a toggling table with a 40-mm-wide detector to cover 80 mm, the scanning inter val could have been shortened to 1 second, but at the expense of a doubling of contrast volume and a prolonged total scanning time. Further technical improve ments, such as a detector wider than 40 mm or a faster toggling table technique, could shorten the scanning interval.
The effective doses of radiation administered during diagnostic CT exam inations are typically estimated to be in the range of 1–10 mSv, and the standard brain CT dose, to be 2–2.5 mSv [22, 23]. The effective dose of radiation administered during 80-mm-coverage perfusion CT in our study was 2.1 mSv. Although coverage in the z-axis was extended, the longer scanning interval (4 seconds) used reduced scan repetition (15 scans) as compared with previous studies, and the lower kVp and lower mAs values we used resulted in an acceptable level for the effective dose. Imaging protocols used in previous perfusion CT studies have been somewhat variable, with different radiation doses [16, 24]. Because of the increasing use of CT for diagnostic workup, physicians need to be aware of the dual effect of radiation dose on image quality and lifetime radiation risk.
Several limitations of our study should be considered. First, the study population does not accurately reflect the general patient population in terms of the proportions of the various types of acute ischemic stroke. Moreover, we did not perform perfusion CT for all consecutive patients with symptoms of acute ischemic stroke during the study period, and there was a slight tendency to perform perfusion CT when a large infarction was suspected or when MRI was not immediately available. Second, no quant itative evaluation of perfusion para meters was conducted. Previous studies have reported that perfusion CT can allow differentiation between ischemic core and penumbra with optimal thresholds of parameters, and in particular found that CBV is the best parameter to predict final infarction [25, 26], although this is still challenging in an emergent clinical setting. Third, CT angiography was not conjoined with perfusion CT although some patients during the latter period of the study did undergo CT angio graphy. Recently, the combined use of perfusion CT and CT angiography has been reported to improve the assessment of acute stroke [26, 27]. Therefore, 80-mm-coverage perfusion CT with quantitative analysis and CT angiography appears to offer a robust imaging technique for emergent treatment planning, particularly in the CT-based selec tion of candidates for intraarterial throm bolytic therapy.
In conclusion, 80-mm-coverage perfusion CT was found to be both feasible and useful for assessing acute ischemic stroke in an emergent setting despite its low sensitivity for detecting small acute infarctions. In particular, this technique provided higher lesion detection than 20-mm-coverage perfusion CT. At the expense of a few extra minutes of examination time after un en hanced CT, 80-mm-coverage perfusion CT provides useful perfusion data on almost the entire brain and thus aids emergent treatment planning in patients with acute ischemic stroke.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
