Magnetic Resonance Arterial Spin Labeling And Its Application In Carotid Artery Stenosis

Part 1Objectives Perfusion territory region of interest(ROI)is commonly used in clinical research of cerebrovascular diseases.However,currently available flow territory maps were mostly based on anatomy atlas,among which we only found one open-access flow territory mask online.The purpose of this study was to develop a gray matter perfusion territory mask base on the Chinese population using territory arterial spin labeling(tASL).Methods Fifty-four healthy volunteers without cerebrovascular disease were included in this study.All subjects underwent three-dimensional time-of-flight magnetic resonance angiography(3D TOF MRA),tASL,ASL and isotropic three-dimensional T1 weighted images.The tASL perfusion weighted images were divided into the non-variant type,unilateral A1 missing type,unilateral fetal type posterior cerebral artery(FTP)and bilateral FTP type according to the circle of Willis(CoW)on 3D TOF MRA.Since the labeling efficiency is not consistent with each vessel in tASL,the mean CBF of each perfusion territory should be adjusted to the same level.Then the three rescaled tASL CBF maps of each subject were combined to a whole CBF map,and co-registered to MNI space with T1 anatomical image.We assumed bilateral arterial flow territories were symmetrical so ipsilateral right and left variant type are identical.?The smaller internal carotid artery(ICA)territory from a subject with missing A1 segment was considered as the middle cerebral artery(MCA)territory.?Similarly,the larger ICA territory on the contralateral side represented the combination of ipsilateral MCA and bilateral anterior cerebral artery(ACA)flow territories.The ACA flow territory can be extracted by deducting the contralateral ICA territory.?Subjects without fetal type posterior cerebral arteries were used for making flow territory mask for posterior circulation.Here we got the initial perfusion territories.To determine the boundary between each feeding artery flow territory,it was defined that,on the scaled territory CBF map,a voxel belonged to the flow territory with the highest CBF on it.The perfusion boundary of the subcortical gray matter was delineated manually according to the anatomy and the initial perfusion territory.Results We developed and published the tASL based gray matter flow territory mask.Conclusion The tASL based arterial flow territory mask will provide a new choice for ROIs in cerebrovascular imaging study.Part 2Objective Arterial spin labeling is a noninvasive MRI perfusion imaging method using arterial blood flow as an endogenous contrast agent.Our study has two purposes,the first one is to investigate the relationship between the level of collateral circulation and normalized perfusion territories after carotid endarterectomy(CEA),assessed by 3D ASL based arterial transit artifacts(ATA)and territorial arterial spin labeling(tASL)respectively;the second purpose is to investigate the cognitive difference between the perfusion territory normalization group and non-normalization group.Methods This study enrolled 22 patients with severe carotid stenosis that underwent CEA and 54 volunteers without significant carotid stenosis.All patients were scanned with ASL and tASL within one month before and one week after CEA.We use tASL images of healthy subjects to create normal probabilistic perfusion territory maps in different circle of Willis types.Two neuro-radiologists assessed collateral circulation on preoperative ASL images based on the presence of ATA and hypoperfusion and the Alberta Stroke Program Early Computed Tomography Score locations.The postoperative flow territories were considered normalized if it conformed to the probabilistic perfusion territory map in healthy population with the same circle of Willis type,by which patients were divided into flow territory normalized group and non-normalized group.Neuropsychological tests were performed on patients before and within seven days after surgery,and on twenty age and sex-matched healthy controls in two weeks interval.Significant cognitive improvement was considered as greater than mean+2×standard deviation of the control group.Results Good agreement between the two readers on ASL collateral grading(?=0.762)was observed.ATA based collateral score in preoperative ASL was significantly higher in flow territory normalized group(n=11,50%)after CEA(P<0.0001).The MoCA and MMSE test scores showed a trend of postoperative improvement in the flow territory normalized group(1.36>1.35,1.18>1.02),no improvement was found in flow territory non-normalized group(0.8<1.35,0<1.02).Other tests exhibited no significant statistical difference between the two groups.Conclusions The study demonstrated that sufficient collateral flow in carotid stenosis patients is associated with normalization of tASL perfusion territories after CEA.Patients with better perfusion territory normalization may show cognitive improvement after CEA.Part 3Objectives Cerebral hyperperfusion is a rare,serious complication after carotid revascularization,which could cause a disastrous outcome if not appropriately managed.The aim of this study was to investigate the association between post-carotid revascularization cerebral hyperperfusion and preoperative image markers such as arterial spin labeling(ASL)perfusion features and circle of Willis(CoW)pattern.Methods Forty-eight consecutive carotid stenosis patients who underwent carotid endarterectomy(CEA)or carotid artery stenting(CAS)were enrolled.All patients had single post-label delay(PLD)ASL,territory-ASL,and three-dimensional time-of-flight MR angiography(3D TOF MRA)within two weeks before surgery and within three days after surgery.Spatial coefficient variation of cerebral blood flow(CBF-CoV),whole-brain perfusion volume,and territory perfusion volume ratio were calculated respectively from ASL and territory-ASL.Postoperative CoW was classified into two groups based on patency of the first segment of the anterior cerebral arteries(A1)and anterior communicating artery(AcomA).ASL perfusion features,type of CoW,and clinical characteristics were analyzed between hyperperfusion group and non-hyperperfusion group using Mann-Whitney U test and Fisher exact test to identify cerebral hyperperfusion risk factors.Results Higher of CBF-CoV(P=0.005),lower whole brain perfusion volume ratio(P=0.012),missing any of A1 or AcomA in CoW(P=0.002 for postoperative MRA and P=0.004 for preoperative MRA),and large artery stroke history(P=0.028)were significantly associated with higher risk of cerebral hyperperfusion.Two cases with cerebral hyperperfusion syndrome were also discussed:one patient had preoperative regional reduced CBF and higher signal heterogeneity on hemisphere of carotid stenosis side;another patient had A1 missing on the contralateral side of the stenosis carotid artery.Conclusions Single-PLD ASL and MRA might be useful non-invasive imaging tools to identify patients with a higher risk of cerebral hyperperfusion after carotid revascularization.Part 4Purpose The post-label delay(PLD)is an important parameter in pseudo-continuous arterial spin labeling(pCASL).To avoid the underestimation of cerebral blood flow(CBF)caused by long ATT,longer PLD is commonly used on the elderly and patients with steno-occlusive diseases.However,prolonged PLD may also reduce the signal-to-noise ratio of images.Long-label pCASL can address this limitation by increasing the amount of labeled blood.The purpose of the study was to compare the reliability and reproducibility of CBF measurement with standard-label pCASL and long-label pCASL in the whole brain,brain lobes,and arterial flow territories.Method Twenty subjects(13 subjects with carotid stenosis or intracranial artery stenosis,7 healthy volunteers,age,56.6± 17.2 years)were scanned twice on one 3.0T scanner by standard-label pCASL(label duration(LD)=1500ms,post-label delay(PLD)=2000ms)and long-label pCASL(LD=3500ms,PLD=2000ms).The interval between the test and retest was one hour.Intraclass correlation coefficient(ICC),within-subject coefficient of variation(wsCV),random noise and signal coefficient of variation(CoV)were calculated to assess CBF measurement reliability and reproducibility in the whole brain,brain lobes,and arterial flow territories.Measurement agreement and difference were compared in different brain regions using correlation coefficient plots and Bland-Altman plots respectively.Results Overall,the reliability and reproducibility of the two kinds of pCASL were good(ICC?0.88,wsCV?0.10).The ICC in occipital lobe and posterior cerebral artery flow territory obtained by the two pCASL were equal,except that the ICCs were higher with long-label pCASL;wsCV,noise,and CoV obtained by long-label pCASL was overall higher than standard-label pCASL in all ROIs.In all ROIs,the correlation analysis showed higher test-retest agreement and the Bland-Altman plots demonstrated lower measurement difference in long-label pCASL.In both ASL methods,anterior circulation flow territories had higher CBF reliability and reproducibility than posterior circulation flow territories.Conclusion We demonstrated good reliability and reproducibility of long-label pCASL and a more robust CBF measurement in anterior flow territories than in posterior flow territories.Part 5Purpose High intravascular signal,known as arterial transit artifact(ATA),can be seen in perfusion images of arterial spin labeling,indicating prolonged arterial transit time(ATT)compared with post-label delay(PLD).Multi-PLD ASL can address this limitation by estimating CBF along with ATT 1,2.Multi-PLD ASL calculates ATT and ATT corrected CBF from perfusion weighted images of multiple PLDs,which commonly acquired by single-shot accelerated stack of spiral 3D fast echo(single-shot FSE)using either sequential acquisitions or a more time-efficient encoding strategy-Hadamard-encoding.However,not all intravascular high signal can be explained by ATA.The purpose of this study was to first demonstrate that inflow fluctuations of arterial blood can cause artifactual ATA in perfusion images of each PLD in single-shot FSE multi-PLD ASL,which may corrupt the final ATT map and ATT corrected CBF map;second to demonstrate that the artifactual ATA can be suppressed by an optimized inflow-saturation method.Method Three healthy subjects(age 40.3±17.1 y/o)were scanned on a 3.0T scanner(Siemens Prisma)using a 32-channel head coil.Both sequential three-delay ASL(SASL)and Hadamard-encoded three-delay ASL(HASL)were acquired without inflow saturation or with an optimized serial saturation method.Transit time was calculated through a signal-weighted delay method.Perfusion weighted images(control-label)of each single acquisition at different PLDs and transit time maps with and without optimized inflow saturation were compared visually.Results Intravascular high signal and intravascular low signal were observed in 12 out of 15(80%)single acquisitions of non-inflow-saturated HASL.The same artifacts were observed in 14 out of 15(93.3%)acquisitions of non-inflow-saturated SASL.Most of these artifacts were reduced after using inflow saturation.An intravascular high signal was seen on PLD=1000ms perfusion weighted images in both inflow-saturated and non-inflow-saturated sequence of a sixty-year-old subject,thus considered to be ATA.Conclusion Inflow saturation can reduce or eliminate artifacts resembling ATA in both Hadamard-encoded and sequential multi-PLD ASL,and can be expected to improve clinical interpretation of multi-PLD ASL data.