SS 211a - Stroke: aetiology, diagnosis and prognosis
Purpose: Plain computed tomography (CT) remains the most common imaging modality for acute ischaemic stroke. For mechanical thrombectomy, recommended selection criteria include an ASPECTS score of ≥ 6. However, detection of early ischaemic signs can be challenging, even for senior neuroradiologists. The purpose of our study was to assess the inter-rater agreement for the detection of early ischaemic signs in the ASPECTS regions between senior neuroradiologists.
Methods and Materials: Plain CT images from 132 patients with acute ischaemic stroke were included in this study. Three neuroradiologists, experienced with stroke imaging and evaluation of ASPECTS, independently scored all the cases. They were blinded to any clinical information. Cohen’s Kappa and Fleiss’ Kappa were computed both for 20 region-based agreement and for dichotomised ASPECTS (ASPECTS ≥6) as a measurement of agreement between the raters.
Results: At the regional level, paired Cohen’s Kappa were 0.35, 0.51, 0.34. The regional agreement was summarised as a Fleiss’ Kappa of 0.37. When considering dichotomised ASPECTS (≥6), paired Cohen’s Kappa were 0.36, 0.51, 0.25. Overall dichotomised ASPECTS was summarised as a Fleiss’ Kappa of 0.31.
Conclusion: We have found only a fair level of inter-rater agreement between neuroradiologists for the evaluation of early ischaemic changes. Our results are within the range of slight to moderate interrater agreement reported recently (Farzin et al., Neurology 2016). Clinicians should be encouraged to evaluate the inter-rater variability for evaluation of ASPECTS in their own center in order to design an optimal clinical pathway for acute stroke management.
Diagnostic accuracy for detection of intracranial haemorrhage in low-dose unenhanced head single-energy and dual-energy third generation dual-source computed tomography
Purpose: To determine diagnostic accuracy for detection of intracranial haemorrhage (ICH), image quality, and radiation dose of low-dose single-energy (SE) and dual-energy (DE) unenhanced head computed tomography (CT) on third-generation dual-source CT.
Methods and Materials: 123 patients with suspected ICH were examined on a 192-slice dual-source CT. Standard-dose SE (120-kV,Group A,n=36) and DE (80-/Sn150-kVp,Group B,n=30) images were compared with low-dose SE (Group C,n=32) and DE (Group D,n=25) using automated-tube-current-modulation (ATCM). Advanced-modeled-iterative-reconstruction (ADMIRE) was used for all protocols. Detection of ICH was performed by three blinded readers. Image quality was assessed quantitative and qualitative. Interobserver agreement was calculated using Fleiss’ Kappa. Radiation dose was assessed as dose-length product (DLP).
Results: Detection of ICH was excellent (sensitivity,94.9-100%; specificity,94.7-100%) in all protocols (p=1.00) with perfect interobserver agreement (0.83-0.96). Qualitative ratings showed significantly better ratings for both standard-dose protocols regarding gray-matter-to-white-matter (GM-WM) contrast (p≤0.014), whereas highest GM-WM CNR was observed in low-dose DE-CT (all,p≥0.057). Lowest posterior-fossa-artifact-index was measured for standard-dose DE-CT with significantly lower values compared to low-dose protocols (both p≤0.034). Delineation of ventricular margins and subarachnoidal spaces sharpness were rated excellent in all protocols (p≥0.096). Low-dose SE lowered radiation dose by 26% (DLP, 575.0±72.3mGy*cm vs. 771.4±146.8mGy*cm, p<0.001) and by 24% in DE (DLP, 587.0±103.2mGy*cm vs. 770.6±90.2mGy*cm, p<0.001). No significant differences were observed between low-dose protocols (p=1.00).
Conclusion: Low-dose unenhanced head SE and DE-CT using ATCM and ADMIRE provide excellent diagnostic accuracy for detection of ICH with good quantitative and qualitative image quality in third-generation dual-source CT while allowing for significant radiation dose reduction.
Leptomeningeal score (LMs) on computed tomography angiography (CTA) and effect of endovascular reperfusion (ER) on clinical outcome in patients with acute ischaemic stroke (AIS)
Purpose: Aim of this study was to review the importance of LMs on the outcome after ER and to evaluate a reliable score system for grading the collateral state.
Methods and Materials: In our database, we identified 125 consecutive AIS patients (mean age 59,8 y, April 2009 - September2016), with classic on-set, that undergoing to ET after basal CT/CTA. We assessed the baseline ASPECT on non-enhanced (NCCT), CTA images, collaterals, clot burden score (CBS) on CTA and degree of reperfusion after ET (TICI 0-III) with clinical data. Two readers, blinded to clinical informations, evaluated the images in parallel and LMs was based on scoring pial and lenticulostriate collaterals (0, no; 1, less; 2, equal or more prominent compared with matching region in opposite hemisphere) in 6 ASPECTS regions (M1-6) plus anterior region and basal ganglia. Good clinical outcome (GCO) was defined as mRS≤2 at 90 days.
Results: TICI III-IIb (OR 27.50; 95%-CI 9.34-81.63) and good collaterals (OR 9.76; 95%-CI 3.28-56.72) were independent predictors of favourable outcome, such as female sex (OR 0.27; 95%-CI 3.45-8.63), younger age (OR 75.30; 95%-CI 7.44-87.31) and higher NCCT ASPECT (OR 2.56; 95%-CI 1.34-6.33). On multivariate analysis GCO well matched with good LMs and TICI III-IIb, but there was no statistical association between LMs and reperfusion (p=0.5).
Conclusion: LMs alone appear to be as a strong imaging parameter for GCO after ET in AIS patients and it might be suitable for imaging based selection.The effect of reperfusion on outcome seems to be strong as the LMs.
Purpose: Intracranial large vessel occlusion (LVO) in patients with acute ischemic stroke (AIS) has multiple pathogenic causes. Identification of the underlying cause of LVO may have major implications for clinical decisions, treatment, and outcome. The aim of this study is to identify the most likely cause of intracranial LVO.
Methods and Materials: Data (n=500) from the MR CLEAN trial which investigated the effectiveness of intra-arterial treatment were reanalysed. Patients with AIS are classified in one of the four pre-defined categories of presumed causes of AIS: carotid dissection, carotid atherosclerosis, cardiac embolism, or undefined. Categorization was based on clinical data (atrial fibrillation (AF) and myocardial infarction (MI)), electrocardiogram (ECG), and CT angiography (CTA). Atherosclerotic carotid disease was defined by the presence of an occlusion, stenosis (>50%), plaque ulcerations or severe calcifications (upper quartile of volume of calcifications at bifurcation) on the symptomatic side.
Results: Data of 426 patients could be analysed. Carotid dissection was identified in 21 patients (4,9%) and cardiac embolism in 126 (29,6%) patients (AF (n=92) and MI (n=47)). Carotid atherosclerosis was present in 154 (36,2%) patients subdivided in occlusion (n=61), significant stenosis (n=49), ulcerations (n=25), and severe calcifications (n=19). The cause of LVO could not be classified in 125 patients (29,3%) according to our criteria.
Conclusion: Based on clinical information, ECG and CTA the most likely cause of AIS due to LVO can be classified in more than 70% of the patients. Carotid vessel atherosclerosis and cardiac embolism are major causes of LVO in patients with AIS.
Acute ischaemic stroke CT imaging and clinical score in patients with isolated intracranial distal artery occlusion
Purpose: Evaluation of CT characteristics in patients with acute ischemic stroke by isolated intracranial distal artery occlusion in correlation with acute and subacute neurological symptoms due to the final infarction area.
Methods and Materials: 55 consecutive stroke patients with an acute distal cerebral artery occlusion were retrospectively analysed. The following criteria were evaluated: presence of infarct demarcation at stroke onset in the non-contrast CT (NECT), localisation of the thrombus and collaterals in CT-angiography [M2- and M3-segment of the middle cerebral artery (n =46), A2-segment of the anterior cerebral artery (n=5), P2-segment of the posterior cerebral artery (n=1) and superior cerebellar artery (n=1)], presence of mismatch in the CT-perfusion and final infarction in the follow up NECT after 24 hours. NIHSS scores of the first and the 4th day and type of therapy were evaluated.
Results: Initial versus follow up NECT showed no significant differences regarding rates of no infarct (49% vs 41.3%), partial (20% and 20%) and complete demarcation (30.9% vs 37.7%). Nevertheless, an overall significant improvement of the neurological symptoms was observed [mean initial NIHSS:11.2±5.6 versus NIHSS on day 4: 3.5±0.86(P=.017)]. Patients with an initial perfusion mismatch had no significant better outcome than those with an initial match. No significant better outcomes were found in the group treated with iv thrombolysis versus without iv thrombolysis or mechanical thrombectomy (P>0.05).
Conclusion: Our data suggest that distal intracranial occlusions without concomitant proximal stenosis or occlusion have a relative benign course poorly improved by the use of iv thrombolysis or mechanical thrombectomy.
Prediction of malignant cerebellar oedema development after acute ischaemic stroke using multiparametric CT
Purpose: Malignant cerebellar oedema (MCE) is a life-threatening complication after acute cerebellar stroke. Aim of this study was to identify imaging predictors using multiparametric CT including whole-brain CT perfusion (WB-CTP).
Methods and Materials: We consecutively selected all subjects with cerebellar WB-CTP alterations and follow-up-confirmed infarction from a cohort of 3254 patients who underwent multiparametric CT. Follow-up imaging was assessed for the presence of MCE, measured using an established 10-point scale, of which scores ≥4 are considered malignant. posterior-circulation-Acute-Stroke-Prognosis-Early-CT-Score (pc-ASPECTS) was determined to assess ischemic changes on non-contrast CT (NCCT), CT angiography (CTA), and on parametric WB-CTP maps (cerebral blood flow, CBF; cerebral blood volume, CBV; mean transit time, MTT; time to drain, TTD). Chi-square, Mann-Whitney-U tests and receiver operating characteristics (ROC) analyses were performed for statistical analyses.
Results: Forty-seven patients were included. 38 patients (80.9%) were categorized as MCE- and 9 (19.1%) as MCE+. MCE+ patients had larger CBF, CBV, MTT and TTD deficit volumes (each with p<0.001) and showed significantly lower median pc-ASPECTS assessed using WB-CTP (CBF: 5 vs. 8; CBV: 8 vs. 9; MTT: 5 vs. 8; TTD: 5 vs. 8; each with p<0.001) compared to MCE- patients, while median pc-ASPECTS on NCCT was not significantly different (9 vs. 10, p=0.097). ROC analyses yielded the largest area-under-the-curve values for CBF (0.977) and CBV deficit volumes (0.952) and pc-ASPECTS on CBF (0.934), whereas pc-ASPECTS on NCCT (0.681) and CTA (0.712) provided significantly less diagnostic value.
Conclusion: WB-CTP provides added diagnostic value regarding the prognosis of MCE development after stroke.
Purpose: To evaluate the effects of γ-secretase inhibitor treatment after transient cerebral ischaemia using diffusion magnetic resonance imaging (d-MRI).
Methods and Materials: 20 middle cerebral artery occlusion (MCAO) rats were divided into two groups: control group (n = 10) and treated group (n = 10), receiving γ-secretase inhibitor (N-[N- (3,5-difluorophenacetyl)-1-alanyl]-S-phenylglycine t-butylester, DAPT) at 3 days after ischemia. All rats were performed with behavioral test, MRI and immunofluorescence staining (IF) at 24h and 7d after MCAO. Percent changes of all values of striatum (STR) were measured and compared between the two groups, that including MR T2 signal intensity (T2 SI), fractional anisotropy (FA), mean diffusivity (MD), mean kurtosis (MK), as well as intensity fluorescence of microglia/NICD.
Results: At 7 days after stroke, percent changes of all values in the treated group were more closer to the baseline compared to the control group (treated vs control: T2 SI, 19.67±7.23% vs 33.33±2.52%, P=0.037; FA, -24.67±6.02% vs -44.67± 5.51%, P=0.039; MD, -6.7±4.04% vs 16.67±8.51%, P=0.033; MK, 49±12.53% vs 123.67±40.15%, P=0.037). Corroborated by IF, microglia and Notch activation were markedly reduced in treated group (treated vs control: micrglia, 57±3.61% vs 125.67±6.43%, P=0.001; NICD, 33±7.51% vs 166.33±12.22%, P=0.000). Significant therapeutic DAPT effects on functional recovery were detected at 7 days after stroke (treated vs control: 4.25±0.5 vs 7.0±0.816, P=0.001).
Conclusion: Our data demonstrate that d-MRI can dynamically identify and quantitatively characterize white matter recovery after DAPT treatment. And DKI parameters in treated group, be closer to the baseline, were in agreement with reduced activation of microglia/NICD.
Diagnostic importance of brain CT perfusion 4D in the detection of acute suptratentorial infarctions
Purpose: CT brain perfusion 4D aids with assessing the certain part of the brain and evaluating perfusion deficits by displaying 4D-DSA views of blood flow in vessels and 3D perfusion maps. It can be performed rapidly and aids in the detection of salvageable tissue (penumbra) from the unsalvageable core infarct.
Methods and Materials: We included 258 patients with symptoms of acute ischemic stroke admitted to Emergency Room for the last 10 months CT perfusion was performed on 160-slice scanner (Toshiba Aquilion Prime) using 50 mL of nonionic iodinated contrast media. Total coverage was 80mm. This acquisition also allows generation of intracranial angiographic data as well as perfusion maps. Whole brain non-contrast CT (NCCT) was performed before CTP.
Results: The study was designed as a retrospective study and included 258 patients. We compared brain tissue involvement on NCCT with CTP (T max, cerebral blood volume [CBV], and cerebral blood flow [CBF]). The sensitivity of CTP (92,1%) was significantly higher compared with that of NCCT (15%). CTP detected significantly more ischemic lesions.
Conclusion: Perfusion CT has proven to be a valuable tool in the diagnosis of acute ischemic stroke. The use of CTP and the advances of this technology will hopefully lead to an extended time window for the use of drugs like tPA for the revascularization of the stroke-affected tissue.
Purpose: For differentiation of blood-clot-types in stroke-workups, we tested if dual-layer computed tomography (DLCT) can characterise predefined clots of varying red-blood-count (RBC) in native and contrast-enhanced scans.
Methods and Materials: Five clots of three ovine blood-clot-types (n=15) containing histologically confirmed RBC composition of 0% (fibrin-rich), 36% (intermediate) and 99% (RBC-rich) were provided by Neuravi-Ltd. (Galway, Ireland). The clots were scanned in a DLCT (Philips, Amsterdam, Netherlands) with 250 mAs, 120 kVp and 0.6 mm slice-thickness and three different settings; a) in a tube containing saline, b) 5 minutes and c) 3 days after exposure to a 1:50 dilution of Accupaque-350 (GE-Healthcare, Boston, USA). The ROI-density of the clots was measured in Hounsfield-units in regular CT datasets as well as Virtual-Non-Contrast (VNC) reconstructions of DLCT.
Results: In native saline-environment the clots differed in density (mean HU±SD) (fibrin-rich 23.6±1.1, intermediate 34.9±1.6, RBC-rich 46.7±1.6). Blood-clots did not show any overlap of density in the native scans and VNC-reconstructions at different time-points (p<0.0001 for each setting and clot-type).However, they could not be differentiated in the regular early contrast-enhanced scan (b) (fibrin-rich 108.5±7.8, intermediate 105.3±3.5, RBC-rich 104.8±3.8). In the late scan, the fibrin-rich clots showed a further increase of density due to uptake of contrast-medium (fibrin-rich 163.6±3.6, intermediate 138.3±4.1, RBC-rich 109.6±5.4).
Conclusion: DLCT allows differentiating the blood-clot-types both in native as well as contrast enhanced scans.The fibrin content of the clots is associated with uptake of contrast medium after prolonged exposure to contrast agents which can be used as an independent predictor of clot type.