| ObjectiveTo investigate whether iDose4 iterative reconstruction can improve the image quality of low radiation whole-brain CT perfusion imaging in patients with ischemic stroke. And with 7 days follow-up CT plain scan as the standard, to evaluate the detection rate of ischemic lesions in low radiation whole-brain CT perfusion maps using iDose4 iterative reconstruction. In addition, to assess the existence of vascular stenosis, stenosis or occlusion position, etiology and severity in CT angiography images using iDose4 iterative reconstruction in patients in accordance with the inclusion criteria of ischemic stroke.Materials and methods1. Patient selectionThe Guangdong No.2 People’s Hospital’s institutional review board approved our study, and informed consent was obtained from the patients or their nearest relative(s) before multimode CT examination. From May 2014 to June 2015, patients with suspected ischemic stroke came to our hospital and were enrolled in our study. This study is a retrospective analysis.2. Image acquisition and ReconstructionAccording to inclusion criteria and exclusion criteria, the suitable candidates were selected. The "one-stop" multi-mode CT images acquisition of all cases were performed on a Brilliance 256-iCT scanner (Philips Healthcare, Best, the Netherlands), After ruling out cerebral hemorrhage and tumor lesions by NCCT, a whole brain CT perfusion scan was performed. After 5 min of CT perfusion scan, the neck and head CTA scanning was performed after injection of contrast agent. Finally, a follow-up NCCT was performed within 7 d. And within 7 days of follow-up CT scan. The canthomeatal line was as a standard positioning line, and the head of the patients was fixed with fixation band.The original CTP images from the CT console yielded 2 groups of databases from 2 reconstruction algorithms:FBP and iDose4. The original CTP images were reconstructed with iDose4. We used iDose4 (level 3) provided by the Philips vendor.3. Image postprocessingThe 2 datasets were postprocessed on a workstation (Version 5.0.1; Philips Intellispace Portal) for brain perfusion analysis using a deconvolution method to automatically obtain time-attenuation curves (TDC), CBV, CBF, and MTT color maps. We selected the anterior cerebral artery for arterial input function (AIF), and the superior sagittal sinus for venous output function (VOF) in the same plane.The CTA database after reconstruction were postprocessed automatically on a workstation (Version 5.0.1; Philips Intellispace Portal), used AVA software to postprocess, and obtained images of volume reconstruction (VR), maximum intensity projection (MIP) images and curved planar reconstruction (CPR) of neck vascular.4. Image analysis4.1 Quantitative comparison of imaging quality of Tmax between FBP and iDose4 reconstruction algorithmsNoise and signal to noise ratio of image was used to quantitative analysis, we selected the time point of maximal enhancement (Tmax) of the anterior cerebral artery based on TDC, and then manually selected 5 ROIs:frontal white matter, head of the caudate nucleus, lentiform nucleus, thalamus, and occipital white matter at the level where the ventricles and the basal ganglia could be seen optimally at Tmax. Average and SD values of all ROIs images derived using the 2 reconstruction algorithms were recorded. The noise (N) was the SD, and SNR was equal to the ratio of average values (AV) to SD:N= SD; SNR= AV/SD4.2 Qualitative analysis of perfusion imaging quality between iDose4 and FBPWe used a modified version of the scoring method proposed by Abels to perform the qualitative assessment. The observation index include quality of grey matter/ white matter differentiation in CBF and CBV maps, temporal grading, differentiation of ischemic and normal tissue, homogeneity, and compensation of artifacts; each index was graded on a 3-point scale (0= poor,1= fair,2= adequate) resulting in a maximum score of 8. Scores of>6 were considered to be of high quality (category 1), scores of<6 and>3 were considered to be of moderate quality (category 2) and scores of<3 were considered to be of low quality (category 3).4.3 Analysis of the ability of the perfusion parameters based on the reconstruction algorithm of iDose4 to detect lesions in ischemic strokeThe results of baseline non-contrast CT(NCCT) and within 7 days of follow-up NCCT images were analysed. And results of the follow-up NCCT performed within 7 d was as the standard, the ability of the perfusion parameters (MTT, CBV, and CBF) using iDose4 on admission to detect lesions in ischemic stroke was analyzed. The lesion was seen on both perfusion parameter maps and within 7 days of follow-up NCCT image, which was considered true positive. The lesion was not seen on both perfusion parameter maps and within 7 days of follow-up NCCT image, which was considered true negative. The lesion was seen on within 7 days of follow-up NCCT image but not on perfusion parameter maps, which was considered false negative. The lesion was seen on perfusion parameter maps but not on within 7 days of follow-up NCCT image, which was considered false positive.4.4 Analysis of vascular conditionsAccording to a variety of postprocessing images, we simple analysed of CTA images, to assess the presence of vascular stenosis or occlusion, the causes of stenosis or occlusion, location and the degree of stenosis. Secondly, combined with the results of CT perfusion and within 7 days of follow-up NCCT, the responsibility of vascular condition was analyzed4.5 Method of measurement of vascular stenosis and occlusionThe stenosis degree of extracranial carotid artery was analysed using the North American symptomatic carotid artery resection operation test (NASCETF) method, the degree of stenosis= (diameter of normal distal of the stenosis vascular-the narrowest diameter)/diameter of normal distal of the stenosis vascular x 100%. Normal vascular diameter measurement was selected the physiological enlargement of internal carotid artery and distal arteries plaque in the arteries in the distal end of the bifurcation, if internal carotid artery was throughout stenosis or occlusion, we selected the corresponding part of the contralateral internal carotid artery. The most obvious narrow was as the measurement area. Stenosis severity grading standard:0% to 9% was considered no stenosis,10% to 29% for mild stenosis,30% to 69%was considered moderate stenosis,70% to 99% for severe stenosis,100% for occlusion.The stenosis degree of intracranial vascular was analysed using the Warfarin-Aspirin Symptomatic Intracranial Disease Study(WASID)method, the degree of stenosis= (diameter of normal distal of the stenosis vascular-the narrowest diameter)/diameter of normal distal of the stenosis vascular x 100%. The measurement of the normal vessel diameter was the widest point in the normal artery of the narrow distal end of the target artery. If the target artery segment is narrow or occlusion, the superior artery of the superior artery which is supplied by the artery segment is selected, the most obvious narrow was as the measurement area. Classification criteria:0% to 29% no stenosis,30% to 49% for mild stenosis,50% to 79% for moderate stenosis,80% to 99% for severe stenosis,100% for occlusion.5. Calculation of radiation doseThe dose length product (DLP) was recorded, and the effective radiation dose was calculated according to the formula:Effective radiation dose (E)= DLP × κ [κ= 0.0023, κ is a constant region-specific conversion coefficient, mSv/(mGyxcm)].6. Statistical AnalysisStatistical analysis was conducted using the Statistical Package for the Social Sciences software (version 20.0, SPSS, Chicago, Illinois).Image noise, signal to noise first performed normality test, if accorded with normal distribution, using the paired t-test to compare, if not, the non parametric Wilcoxon rank test was used and the normality test used one sample Kolmogorov Smirnov analysis to test. Using non parametric Wilcoxon rank test to compare the subjective score of image quality and chi square test to compare the qualitative categorization of CTP maps derived from 2 different reconstruction algorithms.The sensitivity of the detection of acute ischemic lesions in perfusion maps (MTT, CBV, CBF) was calculated according to the formula: sensitivity= true positive/(true positive+false negative).The degree of stenosis was calculated according to the formula:the degree of stenosis= (diameter of normal distal of the stenosis vascular-the narrowest diameter)/diameter of normal distal of the stenosis vascular x 100%. The Kappa test was used to analyze consistency between the radiologists; kappa> 0.75 showed high consistency, kappa= 0.40-0.75 showed moderate consistency, and kappa< 0.40 showed poor consistency. Noise and SNR was expressed as mean ± SD, the subjective scores were expressed as Median (four point spacing, F25-F75). A value of p< 0.05 indicated a statistically significant result.Result35 patients were included in our study, which 19 cases were male,16 female, mean age for (57.2±10.1) years (range from 41 to 76 years old), onset to imaging time for (4.16±22) h (range:4-72h), and all were anterior circulation stroke.1. Quantitative comparison of imaging quality of Tmax between FBP and iDose4 reconstruction algorithmsThe noise and signal-to-noise ratio of all the regions of interest of the 2 reconstruction methods were in accord with normal distribution (P value is> 0.05). Noise from all ROIs reconstructed using iDose4 was significantly lower than noise in those reconstructed using FBP, and SNR of all ROIs reconstructed by iterative reconstruction (IR) was significantly superior to SNR in those reconstructed by FBP(P value is<0.05).2. Consistent qualitative analysis of perfusion images between iDose4 and FBPThe qualitative scores of CBF was 3.00-6.00 for FBP,5.00～6.00 for iDose4, z was -2.784, P was 0.005; CBV was 5.00～6.00 for FBP,6.00～7.00 for iDose4, z was -3.674, P was 0.000; MTT was3.00～5.00for FBP,4.00～6.00 for iDose4, z was -3.394, P was 0.001. Compared to FBP, the subjective score of MTT, CBV, and CBF by iDose4 was significantly higher.The qualitative categorization of CBF yielded 34.3% poor quality perfusion maps for FBP,11.4% poor quality maps for iDose4, χ2 was 7.036, P was 0.030; CBV yielded 11.4% poor quality perfusion maps for FBP,2.9% poor quality maps for iDose4, χ2 was 7.485, P was 0.024, MTT yielded 28.6% poor quality perfusion maps for FBP,11.4% poor quality maps for iDose4, χ2 was 5.318, P was 0.070. the proportion of the poor quality of CBV and CBF maps was significantly lower, but the proportion of the poor quality of MTT maps was insignificantly lower using iDose4.There was high consistency between the 2 radiologists, the κ value of CBF, CBV, MTT of FBP vs iDose4 were 0.851 vs 0.938,0.815 vs 0.836,0.874 vs 0.862, respectively.3. The ability of detect lesions on baseline CT and within 7 days follow-up NCCTThere are 16 patients with ischemic infarction lesions among 35 patients on baseline CT scan results, including 7 cases of small or lacunar infarction(diameter<2cm). The follow-up results showed that there were infarctions or ischemic lesions in all 35 cases, including 12 cases of small or lacunar infarction(diameter<2cm).4. Analysis of the ability of perfusion parameters based on the reconstruction algorithm of iDose4 to detect lesions in ischemic strokeThere was high consistency between the 2 radiologists (κ value= 0.893). The cases of true positive of MTT, CBV, CBF was 18,10,15, respectively, and the cases of false negative of MTT, CBV, CBF was17,25,20, respectively. The sensitivities of MTT, CBV, and CBF in detecting stroke were 51.4%,28.6%, and 42.9%, respectively. But the detection rate of small and lacunar infarction in low-dose CTP was 0.5. CTA image analysis resultsamong CTA of 35 patients, vascular abnormality was found in 4 cases; internal carotid artery or at the end of carotid artery has unstable plaque (including soft plaque and mixed plaque) and with local moderate stenosis in 1 case, and with local severe stenosis in 2 cases; there was calcified plaque in the intracranial segment of the internal carotid artery, and with local no stenosis in 4 cases, with local mild stenosis in 3 cases, with moderate stenosis in 1 case; Simple middle cerebral artery was moderate stenosis in 4 cases, severe stenosis in 3 cases, and occlusion in 4 cases; anterior cerebral artery A2 segment was moderate stenosis in 2 cases; internal carotid artery or at the end of carotid artery has unstable plaque (including soft plaque and mixed plaque) and with ipsilateral middle cerebral artery moderate stenosis in 2 cases, severe stenosis in 2 cases, and occlusion in 3 cases.Combined with follow-up NCCT results, among the CTA results of 12 cases of small infarct or lacunar infarction, there was no abnormality vascular in 4 cases, there was calcified plaque in the intracranial segment of the internal carotid artery with local no stenosis or mild stenosis in 7 cases, there was mixed plaque in the origin of internal carotid artery with local moderate stenosis in 1 case; there are more than moderate stenosis in the other 23 cases. Combined with CT perfusion maps, we found MTT perfusion abnormalities in 7 cases, CBV perfusion abnormalities in 6 cases, CBF perfusion abnormalities in 7 cases among 7 cases with middle cerebral artery occlusion; MTT perfusion abnormalities in 4 cases, CBV perfusion abnormalities in 1 case, CBF perfusion abnormalities in 3 cases among 5 cases with middle cerebral artery severe stenosis; MTT perfusion abnormalities in 5 cases, CBV perfusion abnormalities in 2 cases, CBF perfusion abnormalities in 3 cases among 6 cases with middle cerebral artery moderate stenosis; MTT perfusion abnormalities in 1 case, CBV perfusion abnormalities in 0 case, CBF perfusion abnormalities in 1 case with intracranial internal carotid artery moderate stenosis; MTT perfusion abnormalities in 1 case, CBV perfusion abnormalities in 1 case, CBF perfusion abnormalities in 1 case among 2 cases with extracranial carotid artery severe stenosis; there was no abnormal in the perfusion maps in 2 cases with A2 segment of anterior cerebral artery moderate stenosis,1 case with simple carotid artery stenosis,7 cases with carotid artery no or mild stenosis,4 cases without abnormal blood vessels. There was high consistency between the 2 radiologists. The kappa value for the assessment on whether Presence of carotid artery stenosis was 0.936, on vascular stenosis etiology and site was 1.000, on the degree of stenosis was 0.913.6. Radiation DoseThe DLP of NCCT, whole brain CT perfusion, CTA were 912.5,955.8385.7 mGyxcm, the effective radiation dose were 2.1,2.2,0.9mSv.ConclusioniDose4 iterative reconstruction technique could improve the image quality of low-radiation-dose CTP, low-radiation-dose CTP in the diagnostic value of ischemic stroke value is higher, the CTA can be detect vascular stenosis or occlusion position, etiology and severity in patients with ischemic cerebral stroke. The radiation dose of multimodal CT examination based on iDose4 was lower. And clinical diagnosis value was high in ischemic stroke. |
PDF Full Text Request