Quick search in Citations
This JournalAnywhereCitation
Advanced Search
Research Article
Originally Published 25 August 2015
Free Access

Predicting Collateral Status With Magnetic Resonance Perfusion Parameters: Probabilistic Approach With a Tmax-Derived Prediction Model

Mi Ji Lee, MD, Jeong Pyo Son, MS, Suk Jae Kim, MD, Sookyung Ryoo, MD, Sook-Young Woo, MS, Jihoon Cha, MD, Gyeong-Moon Kim, MD, PhD, Chin-Sang Chung, MD, PhD, Kwang Ho Lee, MD, PhD, and Oh Young Bang, MD, PhDAuthor Info & Affiliations
Stroke
Volume 46, Number 10
https://doi.org/10.1161/STROKEAHA.115.009828
Track CitationsAdd to favorites
PDF/EPUB

Abstract

Background and Purpose—

Good collateral flow is an important predictor for favorable responses to recanalization therapy and successful outcomes after acute ischemic stroke. Magnetic resonance perfusion–weighted imaging (MRP) is widely used in patients with stroke. However, it is unclear whether the perfusion parameters and thresholds would predict collateral status. The present study evaluated the relationship between hypoperfusion severity and collateral status to develop a predictive model for good collaterals using MRP parameters.

Methods—

Patients who were eligible for recanalization therapy that underwent both serial diffusion-weighted imaging and serial MRP were enrolled into the study. A collateral flow map derived from MRP source data was generated through automatic postprocessing. Hypoperfusion severity, presented as proportions of every 2-s Tmax strata to the entire hypoperfusion volume (Tmax≥2 s), was compared between patients with good and poor collaterals. Prediction models for good collaterals were developed with each Tmax strata proportion and cerebral blood volumes.

Results—

Among 66 patients, 53 showed good collaterals based on MRP-based collateral grading. Although no difference was noted in delays within 16 s, more severe Tmax delays (Tmax16–18 s, Tmax18–22 s, Tmax22–24 s, and Tmax>24 s) were associated with poor collaterals. The probability equation model using Tmax strata proportion demonstrated high predictive power in a receiver operating characteristic analysis (area under the curve=0.9303; 95% confidence interval, 0.8682–0.9924). The probability score was negatively correlated with the volume of infarct growth (P=0.030).

Conclusions—

Collateral status is associated with more severe Tmax delays than previously defined. The present Tmax severity–weighted model can determine good collaterals and subsequent infarct growth.

Introduction

Among patients with acute ischemic stroke, both antegrade flow and (delayed) retrograde collateral flow maintain cerebral perfusion within ischemic regions. Previous reports have suggested that an angiographic collateral grade determines the rate of recanalization, hemorrhagic transformation, and infarct growth after revascularization therapy.1–3 Patients with poor collateral flow show a low recanalization rate, regardless of the revascularization therapy mode or occlusion site.2 Even if recanalization is achieved after revascularization therapy, these patients often experience clinical deterioration because of symptomatic hemorrhagic transformation.3 Recanalization is related to a positive clinical outcome only if adequate collateralization prevents infarction until the vessel is recanalized.1,4,5 Therefore, a good collateral status could feasibly extend the time window for endovascular procedures.6,7
A recent trend is to visualize collaterals using noninvasive imaging methods such as computed tomography, angiography or perfusion, dynamic susceptibility contrast-enhanced magnetic resonance (MR) perfusion (DSC-MRP), and arterial spin labeling imaging.5,8–12 Tmax (deconvoluted time to peak) and cerebral blood volume (CBV) approaches are most widely used, allowing clinicians to identify mismatch.
The present study hypothesized that differential perfusion parameters and thresholds would predict collateral status. The current study evaluated the relationship between hypoperfusion severity (as measured by Tmax5), CBV, and collateral status (as measured by a DSC-MRP–based collateral map), while also developing a Tmax severity–weighted predictive model for good collaterals. We also evaluated clinical outcome (as measured by infarct growth at day 7 and modified Rankin Scale at 3 months) in relation with Tmax severity–weighted predictive model.

Patients and Methods

Patient Selection

Using data from a prospectively maintained registry, patients considered eligible for recanalization therapy for acute infarction within the middle cerebral artery (MCA) territory were identified. Data were retrospectively analyzed from consecutive patients who were treated from June 2005 to December 2012 at a university medical center. Inclusion criteria for this study were as follows: (1) subjects who presented within 6 hours of symptom onset; (2) subjects who underwent pretreatment and day 7 post-treatment brain magnetic resonance imaging (MRI), including diffusion-weighted imaging (DWI), MRP, and MR angiography (MRA); (3) subjects who had an National Institutes of Health Stroke Scale score of ≥4 points at admission; and (4) subjects with internal carotid artery or proximal MCA (M1 segment) occlusion associated with symptoms as determined by an admission MRA. The local institutional review board approved this study.

Evaluations

Patients were evaluated on demographic characteristics, medical history, vascular risk factors, routine blood tests, brain imaging, and cardiological assessments. Stroke mechanisms were subtyped using the Trial of Org10172 in Acute Stroke Treatment (TOAST) classification13 and were diagnosed by the consensus of 2 stroke neurologists (S.J.K. and S.R.). In patients who underwent conventional angiography, collateral grades were assessed by using the American Society of Interventional and Therapeutic Neuroradiology/Society of Interventional Radiology Collateral Flow Grading System.14

MRP Methods and Image Analysis

MRI was performed using a 3T Philips Achieva MR scanner (Philips Medical Systems, Best, Netherlands). The identical MR scanner was used for all study subjects. Typical MRI sequences for acute stroke included at least DWI, DSC-MRP, fluid-attenuated inversion recovery, and MRA of the cervical and intracranial vessels (3-dimensional time-of-flight MRA and contrast-enhanced MRA, including the extracranial carotid and vertebral arteries). DWI was performed with 2 levels of diffusion sensitization (b values of 0 and 1000 s/mm2; 5- to 7-mm slice thickness; no gap). DSC-MRP was performed using gradient-echo and echo-planar imaging techniques after administration of intravenous gadolinium (Dotarem [gadoterate meglumine]; Guerbet, Aulnay-sous-Bois, France) with a repetition time of 1718 ms for a total acquisition time of ≈90 s with 20 to 22 slices. A contrast agent was injected at a dose of 0.1 mmol/kg body weight with a flow rate of 3 mL/s by a power injector into an antecubital vein via an 18-gauge intravenous cannula ≈7 s after beginning the acquisition. Other parameters for DSC-MRP were as follows: echo time=35 ms, flip angle=40°, acquisition matrix=128×128, field of view=24×24 cm2, section thickness=5 mm, and intersection gap=2 mm. In total, 1000 DSC-MRP raw images, composed of 50 time points per slice, were obtained.
Perfusion delay was defined based on the perfusion parameter Tmax. Tmax is the time to the peak of the residue function map generated by deconvolution of the tissue concentration over the time curve, using an arterial input function from the contralateral MCA.15 MRP postprocessing and data analysis were performed as described in our previous studies.16 MRI volume measurements were performed by an investigator (J.P.S.), who was blinded to the clinical information. Tmax volumes were stratified using a 2-s strata of Tmax thresholds, from Tmax2–4 s, Tmax4–6 s, … Tmax22–24 s, and Tmax>24 s. Because of too small volumes, Tmax12–16 s and Tmax18–22 s were merged into 4-s strata to avoid the possibility of under- or overestimation. To identify hypoperfusion severity distributions, the proportion of every single Tmax strata to the entire hypoperfusion volume (Tmax≥2s) was used in the analysis (ProportionTmax strata=VolumeTmax strata/VolumeTmax≥2s).
CBV values were determined in the lesion with Tmax ≥2 s using a computer-assisted volumetric analysis program (Medical Image Processing, Analysis and Visualization, version 3.0; National Institutes of Health, Bethesda, MD). Regions of interest (ROIs) of the Tmax ≥2s lesion were semiautomatically defined. Mirror ROIs were manually drawn over contralateral homologous regions and also manually corrected to avoid misregistration when the mirror ROI covered an area outside of the brain parenchyme. The same ROIs were used to calculate the relative CBV values. Relative CBV (rCBV) was defined as a ratio of the mean lesional CBV:the mean contralesional CBV in the mirror ROI.

Postprocessing Techniques to Generate an MRP-Derived Collateral Flow Map

Collateral flow maps were generated based on source data derived from DSC-MRP, as previously described.5 Collateral flow maps were automatically generated using in-house software—Fast Analysis SysTem for COLLaterals—developed using MATLAB (MathWorks, Natick, MA). All steps were typically completed within 5 minutes. Criteria for collateral flow map–based grades were chosen based on the concept of the American Society of Interventional and Therapeutic Neuroradiology/Society of Interventional Radiology scale14 with a classification of grade 0 (no visible collaterals), grade 1 (slow collaterals, visible only during the late phase toward the part of the occluded MCA territory with a persistence of some of the defect), grade 2 (rapid collaterals, visible during the mid to late phase toward the part of the occluded MCA territory with a persistent defect portion), grade 3 (slow but complete collateral flow to the ischemic bed), and grade 4 (complete and rapid collateral flow to the entire vascular bed in the occluded MCA territory). Six axial slices of the collateral flow maps were used to assess the leptomeningeal collateral grade. The striatocapsular region was not assessed owing to limitations of DSA-based collateral grading in this region. For statistical analysis, patients were divided into 2 groups according to collateral grade: grade 0 to 2 (poor) and grade 3 to 4 (good). Collateral flow images were reviewed independently by 2 investigators (S.J.K. and S.R.), both having knowledge of the symptomatic side and occlusion site. They reevaluated all collateral flow imaging data after 2 months. When a judgment between the 2 raters was inconsistent, a decision was made by consensus.

Outcome Measurements

In patients who underwent endovascular treatment, angiographic reperfusion was determined with the Thrombolysis in Cerebral Infarction grading system.14 Infarct growth was defined as the difference in DWI volume between day 7 and baseline. Modified Rankin scale was judged at 3 months after stroke onset.

Statistical Analysis

All data are presented as medians (25–75th percentile) or frequencies (percentages), unless otherwise specified. A Student t test or Mann–Whitney U test was used to compare continuous variables, and a Pearson χ2 test or Fisher exact test was used to compare categorical variables between groups. The Bonferroni method was used to correct P values for multiple comparisons.
Multiple ProportionTmax strata values were transformed because of skewed distributions. Square root or natural log transformations were applied in each ProportionTmax strata as shown below. The value 1 was added to enable the transformation. All ProportionTmax strata values were multiplied by 100, and rCBV values were multiplied by 10 to overcome their small units. Finally, transformations for each ProportionTmax strata (TransTmax strata) are as follows:
For predicting good (grades 3–4) collateral grades, a logistic regression analysis was performed using a stepwise method. Predictors were each TransTmax strata, whereas TransTmax2–4 was excluded because of high multicollinearity (model 1). A univariable logistic regression with TransrCBV (model 2) was performed, and a subsequent multivariable logistic regression with the addition of TransrCBV from model 1 (model 3) was established. Receiver operating characteristic (ROC) curves were used to evaluate the prediction performance of each model. The area under the curves (AUCs) were compared using a method from DeLong et al.17 In addition, a multinomial logistic regression analysis was performed to predict each collateral grade. Areas under the multiclass ROCs were calculated using a previously reported method18,19 and an open-source package pROC (version 1.7.2) for R. Correlations between the prediction model scores and infarct growth were assessed with Spearman test.
A value of P<0.05 was considered statistically significant. Statistical analyses were executed using SAS version 9.3 (SAS Institute, Cary, NC) and R 3.1.2 (Vienna, Austria; http://www.R-project.org/).

Results

Among 79 patients who met inclusion criteria during the study period, 66 (83.5%) were finally enrolled. Thirteen patients were excluded because of technically suboptimal imaging data: (1) an inability to generate a collateral flow map because of poor contrast bolus (n=2), (2) large perfusion defects because of chronic tissue loss (n=7), and (3) failure of MRI postprocessing for Tmax because of excessive patient motion or the absence of an identifiable, technically adequate arterial input function (n=4). Study profile is presented in Figure I in the online-only Data Supplement.
Of the 66 patients enrolled, 53 (80.3%) showed good (3–4) collaterals, and the other 20 patients demonstrated poor (1–2) collaterals. No patient showed a collateral grade=0. Typical cases are demonstrated in Figure 1. Patient characteristics are shown in Table 1. General characteristics, including vascular risk factors, stroke mechanisms, and occlusion sites, did not differ between groups. The median baseline National Institutes of Health Stroke Scale score was lower, and the median DWI volume was smaller, in the good collateral group (P=0.024 and P<0.001, respectively).
Table 1. Patient Characteristics Based on MR Perfusion–Based Collateral Grading
 Collateral GradeP Value
 Poor (n=13)Good (n=53)
Female sex, n (%)3 (23.1)18 (34.0)0.526
Age, y, median (IQR)63 (59–69)65 (52–72)0.699
Risk factors, n (%)
 Atrial fibrillation7 (53.8)22 (41.5)0.422
 Hypertension4 (30.8)26 (49.1)0.235
 Diabetes mellitus4 (30.8)11 (20.8)0.471
 Hyperlipidemia4 (30.8)14 (26.4)0.739
 Coronary artery disease1 (7.7)5 (9.4)>0.999
 Current smoking0 (0.0)10 (18.9)0.190
 Prior stroke4 (30.8)13 (24.5)0.645
Stroke mechanism, n (%)  0.426
 Atherosclerotic3 (25.0)21 (45.7) 
 Cardioembolic6 (50.0)19 (41.3) 
 Other2 (16.7)4 (8.7) 
 Undetermined1 (8.3)2 (4.3) 
Occlusion site, n (%)  0.413
 M11 (7.7)14 (26.4) 
 Distal ICA10 (76.9)30 (56.6) 
 Distal ICA+M12 (15.4)9 (17.0) 
Initial NIHSS score, median (IQR)16 (13–19)12 (10–16)0.024
Initial DWI lesion volume, mL, median (IQR)56.4 (23.2–74.1)7.0 (4.3–13.6)<0.001
Recanalization therapy, n (%)  0.114
 IV tPA only0 (0.0%)4 (7.5%) 
 Endovascular only1 (7.7%)16 (30.2%) 
 Both9 (69.2%)29 (54.7%) 
 No treatment3 (23.1%)4 (7.5%) 
Onset to MR perfusion, min, median (IQR)170 (145–220)165 (121–245)0.468
DWI–PWI mismatch volume, mL, median (IQR)*132.0 (102.0–159.7)118.2 (67.3–152.7)0.425
DWI indicates diffusion-weighted image; ICA, internal cerebral artery; IQR, interquartile range; IV tPA, intravenous tissue-type plasminogen activator; M1, main trunk of the middle cerebral artery; NIHSS, National Institutes of Health Stroke Scale; and PWI, perfusion-weighted image.
*Difference between PWI (Tmax >4 s) volume and DWI volume.
Figure 1. Four examples of representative perfusion imaging, collateral imaging, and the probability of good (3–4) collateral grades. Regions of decreased rCBV were marked with arrows. MR indicates magnetic resonance; and rCBV, relative cerebral blood volume.
Each ProportionTmax strata was compared between patients with good (n=53) and poor (n=13) collaterals (Figure 2). Each ProportionTmax strata within 16 s did not differ between the good and the poor collateral groups. Rather, patients with poor collaterals showed higher proportions of severely (>16 s) delayed perfusion, including ProportionTmax16–18, ProportionTmax18–22, ProportionTmax22–24, and ProportionTmax>24 (adjusted P=0.010, P<0.0001, P<0.0001, and P<0.0001, respectively).
Figure 2. Error bars comparing proportions of each Tmax strata to the entire hypoperfusion volume (Tmax ≥2 s). Values are represented as % of Tmax ≥2 s. *Statistically significant (P<0.05) after a Bonferroni correction for multiple comparisons.
Multivariable logistic regression was performed (Table 2). Models for predicting good collaterals were made by using Tmax proportions (model 1), CBV (model 2), and both (model 3).
Table 2. Multivariable Logistic Regression Analysis for Good Collaterals (Grade 3–4)
 β Coefficients*SEAdjusted OR95% CIP Value
Model 1
 TransTmax12–162.39290.983010.9451.594–75.1450.015
 TransTmax18–22−2.52361.06240.0800.010–0.6430.018
 TransTmax22–24−8.95663.5760<0.001<0.001–0.1430.012
Model 2
 TransrCBV0.42810.17120.5341.097–2.1460.012
Model 3
 TransTmax12–163.75831.582142.8761.930–952.6360.018
 TransTmax18–22−3.79381.58280.0230.001–0.5010.017
 TransTmax22–24−14.88685.8017<0.001<0.001–0.0300.010
 TransrCBV−0.66230.37980.5160.245–1.0850.081
CI indicates confidence interval; OR, odds ratio; and rCBV, relative cerebral blood volume.
*Obtained from stepwise logistic regression models adjusted for each Tmax strata proportion to the entire hypoperfusion volume (Tmax >2s; model 1), relative CBV (model 2), and both (model 3).
The ROC curve for the equation using Tmax (model 1) demonstrated high predictive power (R2=0.625, AUC=0.9303; 95% confidence interval, 0.8682–0.9924; Figure 3). The predictive power of CBV (model 2) was fair (AUC=0.7097; 95% confidence interval=0.5567–0.8627) but inferior to model 1 (P=0.001). The addition of CBV into the Tmax model (model 3) showed no additional accuracy above model 1 (AUC=0.9390; 95% confidence interval=0.8804–0.9977; P=0.534 for comparing ROC curves). The multinomial logistic regression analysis and prediction models for the individual collateral grades are presented in Table I in the online-only Data Supplement. Using the prediction equation, predicted collateral grades were consistent with observed grades in 44 patients (66.7%).
Figure 3. Receiver operating characteristic (ROC) curves for 3 models predicting good collateral grades based on Tmax (model 1), cerebral blood volume (CBV; model 2), and both Tmax and CBV (model 3).
Infarct growth at day 7 was negatively correlated with predicted probability scores (model 1, Spearman’s ρ=−0.268; P=0.030). In patients who underwent conventional angiography (n= 57), the association between probability scores (model 1) and infarct growth was similar to that between collateral grading and infarct growth (Figure 4). When stratified to the achievement of good recanalization (Thrombolysis in Cerebral Infarction 2b-3), a trend for negative correlation was seen between collateral probability scores and modified Rankin Scale at 3 months in the recanalized group (ρ=−0.392; P=0.051). Such correlations were not observed in nonrecanalized group.
Figure 4. Diffusion-weighted imaging lesion enlargement between baseline and day 7. A and B, Infarct growth by tertiles of collateral probability score using Tmax (model 1). C and D, Infarct growth by collateral grades assessed by conventional angiography.

Discussion

The main findings of the present study are as follows: (1) collateral status is more associated with the more severe hypoperfusion (Tmax>16 s) than a shorter perfusion delay; (2) a Tmax severity–weighted model predicts poor collaterals with high predictive power, whereas CBV data did not provide additional information for the prediction; and (3) a Tmax-derived collateral prediction model is negatively correlated with infarct growth.
Researchers have continuously sought to determine the optimal definition of a target mismatch pattern, as well as a mismatch ratio, MRP parameters, and the threshold for each parameter.20–25 However, few studies have demonstrated MRP parameters and thresholds for predicting collateral status (ie, a hypoperfusion intensity ratio).9 Besides the development of dedicated MRI sequences for collaterals, an effort to predict collateral status using Tmax is valuable for the following reasons. First, Tmax has several attractive properties for clinical use. In fact, Tmax has already been used in a number of clinical trials, including the Diffusion and Perfusion Imaging Evaluation for Understanding Stroke Evolution (DEFUSE) and Echoplanar Imaging Thrombolytic Evaluation Trial (EPITHET). Second, Tmax is the most widely studied MRP parameter for patients with acute ischemic stroke. Tmax has also been established as an accurate perfusion metric for discriminating both the infarct core and the penumbra. Therefore, comparing the Tmax thresholds for the penumbra versus collaterals may help our understanding of hemodynamic status. Finally, Tmax has the advantage of being able to measure the time to peak of the tracer fraction within a voxel at any given time. This is in comparison with the time the contrast bolus takes to transverse the voxel (eg, mean transit time) or the area under the deconvolved tissue curve (AUC) of the contrast bolus intensity (eg, CBV).
The present study revealed that collateral status is correlated with more severely delayed perfusion, and a Tmax severity–weighted model predicts poor collaterals with high accuracy. These results suggest that the more delayed perfusion (Tmax >16 s) was more closely correlated with collateral status than was a less delayed perfusion (Tmax 6–16 s). These findings indicate that delayed perfusion (Tmax >16 s) might be closer to the real infarct core than the Tmax threshold for the presence of a mismatch used in the EPITHET (≥2 s),26 DEFUSE (≥2 s),25 and DEFUSE-2 (>6 s).20 The present findings are in line with results from previous studies showing that collateral status on conventional angiography was not correlated with MR perfusion-diffusion mismatches and CBV/cerebral blood flow mismatches on a perfusion computed tomography.1,27 No significant difference was found between diffusion–perfusion mismatch (defined as Tmax ≥4 s) and pretreatment collaterals.1 The present results are also supported by previous studies revealing that a severe, but not mild, perfusion delay15 and low CBV are associated with hemorrhagic transformation, which is probably related to poor collaterals.28
It was speculated that high Tmax values seen in hyperacute stroke indicate poor delayed collateral supply. Furthermore, regions with long arrival delays, even if relatively well-perfused, are the most at risk if perfusion pressure further declines.29 However, there is increasing evidence that collateral status predicts tissue fate, regardless of the presence or absence of a target mismatch pattern and successful recanalization.1,5 The present data showed that Tmax thresholds for predicting collaterals are higher than those for the penumbra. This illustrates that collaterals and mismatches represent related, yet distinct, aspects of ischemic pathophysiology. First, most MRP studies have focused on the early phase (arterial and early capillary phase), but perfusion status at a later phase (late capillary or venous) may also influence tissue fate. It was recently reported that venous drainage identified by computed tomographic angiography was correlated with collateral status and infarct growth.30 Second, Tmax primarily reflects macrovascular features,29 but a severe Tmax change may reflect microvascular integrity supplied by collateral circulations. Microvascular features, including vasodilatory compensatory mechanisms, may be important for functional collaterals. Although the diameter of the leptomeningeal collateral needed to be >1.0 mm to provide adequate flow, the individual diameter in cadaver studies was ≈0.3 mm.31 Further studies are needed to better understand the physiological differences underlying these 2 conditions.
A few study limitations should be noted. First, the present findings should be interpreted with caution because of the limited sample size and data from a single center where the prevalence of intracranial atherosclerosis is high. Our results demonstrating a high prevalence of good collaterals may be because of the high prevalence of intracranial atherosclerosis. Pretreatment collateral status differs greatly among stroke subtypes. Almost two thirds the patients with intracranial atherosclerosis, but less than one third of patients with cardioembolism, show good collaterals.32,33 Further studies with different populations and larger cohorts are warranted. In this context, a multicenter study is being planned to evaluate the impact of collateral flow on stroke outcomes using this approach. Second, Tmax is sensitive to several hemodynamic effects, and its physiological interpretation is complex.29 Finally, further studies are needed to validate associations between clinical outcomes, Tmax-based collateral status and successful recanalization in a larger number of patients.
In conclusion, the present results indicate that collateral status is associated with a higher proportion of more severe Tmax delays than was previously defined. The Tmax severity–weighted model predicted good collaterals with high predictive power. Given that routine perfusion-weighted imaging data were used in the present study, this method could be feasible for clinical practice. However, these findings prompt larger, confirmatory studies. For instance, among clinical trials analyzing the Tmax-based mismatch pattern, additional analyses of Tmax-based collateral status may help elucidate possible differences regarding hemodynamic aspects and clinical impact between these 2 conditions.

References

1.
Bang OY, Saver JL, Buck BH, Alger JR, Starkman S, Ovbiagele B, et al.; UCLA Collateral Investigators. Impact of collateral flow on tissue fate in acute ischaemic stroke. J Neurol Neurosurg Psychiatry. 2008;79:625–629. doi: 10.1136/jnnp.2007.132100.
Crossref
PubMed
Google Scholar
2.
Bang OY, Saver JL, Kim SJ, Kim GM, Chung CS, Ovbiagele B, et al. Collateral flow predicts response to endovascular therapy for acute ischemic stroke. Stroke. 2011;42:693–699. doi: 10.1161/STROKEAHA.110.595256.
Crossref
PubMed
Google Scholar
3.
Bang OY, Saver JL, Kim SJ, Kim GM, Chung CS, Ovbiagele B, et al.; UCLA-Samsung Stroke Collaborators. Collateral flow averts hemorrhagic transformation after endovascular therapy for acute ischemic stroke. Stroke. 2011;42:2235–2239. doi: 10.1161/STROKEAHA.110.604603.
Crossref
PubMed
Google Scholar
4.
Kucinski T, Koch C, Eckert B, Becker V, Krömer H, Heesen C, et al. Collateral circulation is an independent radiological predictor of outcome after thrombolysis in acute ischaemic stroke. Neuroradiology. 2003;45:11–18. doi: 10.1007/s00234-002-0881-0.
Crossref
PubMed
Google Scholar
5.
Kim SJ, Son JP, Ryoo S, Lee MJ, Cha J, Kim KH, et al. A novel magnetic resonance imaging approach to collateral flow imaging in ischemic stroke. Ann Neurol. 2014;76:356–369. doi: 10.1002/ana.24211.
Crossref
PubMed
Google Scholar
6.
Ribo M, Flores A, Rubiera M, Pagola J, Sargento-Freitas J, Rodriguez-Luna D, et al. Extending the time window for endovascular procedures according to collateral pial circulation. Stroke. 2011;42:3465–3469. doi: 10.1161/STROKEAHA.111.623827.
Crossref
PubMed
Google Scholar
7.
Galimanis A, Jung S, Mono ML, Fischer U, Findling O, Weck A, et al. Endovascular therapy of 623 patients with anterior circulation stroke. Stroke. 2012;43:1052–1057. doi: 10.1161/STROKEAHA.111.639112.
Crossref
PubMed
Google Scholar
8.
Campbell BC, Christensen S, Tress BM, Churilov L, Desmond PM, Parsons MW, et al.; EPITHET Investigators. Failure of collateral blood flow is associated with infarct growth in ischemic stroke. J Cereb Blood Flow Metab. 2013;33:1168–1172. doi: 10.1038/jcbfm.2013.77.
Crossref
PubMed
Google Scholar
9.
Bang OY, Saver JL, Alger JR, Starkman S, Ovbiagele B, Liebeskind DS; UCLA Collateral Investigators. Determinants of the distribution and severity of hypoperfusion in patients with ischemic stroke. Neurology. 2008;71:1804–1811. doi: 10.1212/01.wnl.0000335929.06390.d3.
Crossref
PubMed
Google Scholar
10.
Robson PM, Dai W, Shankaranarayanan A, Rofsky NM, Alsop DC. Time-resolved vessel-selective digital subtraction MR angiography of the cerebral vasculature with arterial spin labeling. Radiology. 2010;257:507–515. doi: 10.1148/radiol.092333.
Crossref
PubMed
Google Scholar
11.
McVerry F, Liebeskind DS, Muir KW. Systematic review of methods for assessing leptomeningeal collateral flow. AJNR Am J Neuroradiol. 2012;33:576–582. doi: 10.3174/ajnr.A2794.
Crossref
PubMed
Google Scholar
12.
Goyal M, Demchuk AM, Menon BK, Eesa M, Rempel JL, Thornton J, et al.; ESCAPE Trial Investigators. Randomized assessment of rapid endovascular treatment of ischemic stroke. N Engl J Med. 2015;372:1019–1030. doi: 10.1056/NEJMoa1414905.
Crossref
PubMed
Google Scholar
13.
Adams HP, Bendixen BH, Kappelle LJ, Biller J, Love BB, Gordon DL, et al.; and the TOAST Investigators. Classification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. Stroke. 1993;24:35–41. doi: 10.1161/01.STR.24.1.35.
Crossref
PubMed
Google Scholar
14.
Higashida RT, Furlan AJ, Roberts H, Tomsick T, Connors B, Barr J, et al.; Technology Assessment Committee of the American Society of Interventional and Therapeutic Neuroradiology; Technology Assessment Committee of the Society of Interventional Radiology. Trial design and reporting standards for intra-arterial cerebral thrombolysis for acute ischemic stroke. Stroke. 2003;34:e109–e137. doi: 10.1161/01.STR.0000082721.62796.09.
Crossref
PubMed
Google Scholar
15.
Ostergaard L, Weisskoff RM, Chesler DA, Gyldensted C, Rosen BR. High resolution measurement of cerebral blood flow using intravascular tracer bolus passages. Part I: Mathematical approach and statistical analysis. Magn Reson Med. 1996;36:715–725. doi: 10.1002/mrm.1910360510.
Crossref
PubMed
Google Scholar
16.
Kim JH, Bang OY, Liebeskind DS, Ovbiagele B, Kim GM, Chung CS, et al.; UCLA-Samsung Stroke Collaborators. Impact of baseline tissue status (diffusion-weighted imaging lesion) versus perfusion status (severity of hypoperfusion) on hemorrhagic transformation. Stroke. 2010;41:e135–e142. doi: 10.1161/STROKEAHA.109.563122.
Crossref
PubMed
Google Scholar
17.
DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics. 1988;44:837–845. doi: 10.2307/2531595.
Crossref
PubMed
Google Scholar
18.
Hand DJ, Till RJ. A simple generalisation of the area under the ROC curve for multiple class classification problems. Machine Learn. 2001;45:171–186. doi: 10.1023/A:1010920819831.
Crossref
Google Scholar
19.
Robin X, Turck N, Hainard A, Tiberti N, Lisacek F, Sanchez JC, et al. pROC: an open-source package for R and S+ to analyze and compare ROC curves. BMC Bioinformatics. 2011;12:77. doi: 10.1186/1471-2105-12-77.
Crossref
PubMed
Google Scholar
20.
Lansberg MG, Straka M, Kemp S, Mlynash M, Wechsler LR, Jovin TG, et al.; DEFUSE 2 Study Investigators. MRI profile and response to endovascular reperfusion after stroke (DEFUSE 2): a prospective cohort study. Lancet Neurol. 2012;11:860–867. doi: 10.1016/S1474-4422(12)70203-X.
Crossref
PubMed
Google Scholar
21.
Kidwell CS, Alger JR, Saver JL. Beyond mismatch: evolving paradigms in imaging the ischemic penumbra with multimodal magnetic resonance imaging. Stroke. 2003;34:2729–2735. doi: 10.1161/01.STR.0000097608.38779.CC.
Crossref
PubMed
Google Scholar
22.
Toth G, Albers GW. Use of MRI to estimate the therapeutic window in acute stroke: is perfusion-weighted imaging/diffusion-weighted imaging mismatch an EPITHET for salvageable ischemic brain tissue? Stroke. 2009;40:333–335. doi: 10.1161/STROKEAHA.108.525683.
Crossref
PubMed
Google Scholar
23.
Kakuda W, Lansberg MG, Thijs VN, Kemp SM, Bammer R, Wechsler LR, et al.; DEFUSE Investigators. Optimal definition for PWI/DWI mismatch in acute ischemic stroke patients. J Cereb Blood Flow Metab. 2008;28:887–891. doi: 10.1038/sj.jcbfm.9600604.
Crossref
PubMed
Google Scholar
24.
Nagakane Y, Christensen S, Ogata T, Churilov L, Ma H, Parsons MW, et al.; EPITHET Investigators. Moving beyond a single perfusion threshold to define penumbra: a novel probabilistic mismatch definition. Stroke. 2012;43:1548–1555. doi: 10.1161/STROKEAHA.111.643932.
Crossref
PubMed
Google Scholar
25.
Albers GW, Thijs VN, Wechsler L, Kemp S, Schlaug G, Skalabrin E, et al.; DEFUSE Investigators. Magnetic resonance imaging profiles predict clinical response to early reperfusion: the diffusion and perfusion imaging evaluation for understanding stroke evolution (DEFUSE) study. Ann Neurol. 2006;60:508–517. doi: 10.1002/ana.20976.
Crossref
PubMed
Google Scholar
26.
Davis SM, Donnan GA, Parsons MW, Levi C, Butcher KS, Peeters A, et al.; EPITHET Investigators. Effects of alteplase beyond 3 h after stroke in the Echoplanar Imaging Thrombolytic Evaluation Trial (EPITHET): a placebo-controlled randomised trial. Lancet Neurol. 2008;7:299–309. doi: 10.1016/S1474-4422(08)70044-9.
Crossref
PubMed
Google Scholar
27.
Khatri R, Rodriguez GJ, Suri MF, Vazquez G, Ezzeddine MA. Leptomeningeal collateral response and computed tomographic perfusion mismatch in acute middle cerebral artery occlusion. J Vasc Interv Neurol. 2011;4:1–4.
PubMed
Google Scholar
28.
Hermitte L, Cho TH, Ozenne B, Nighoghossian N, Mikkelsen IK, Ribe L, et al. Very low cerebral blood volume predicts parenchymal hematoma in acute ischemic stroke. Stroke. 2013;44:2318–2320. doi: 10.1161/STROKEAHA.113.001751.
Crossref
PubMed
Google Scholar
29.
Calamante F, Christensen S, Desmond PM, Ostergaard L, Davis SM, Connelly A. The physiological significance of the time-to-maximum (Tmax) parameter in perfusion MRI. Stroke. 2010;41:1169–1174. doi: 10.1161/STROKEAHA.110.580670.
Crossref
PubMed
Google Scholar
30.
Parthasarathy R, Kate M, Rempel JL, Liebeskind DS, Jeerakathil T, Butcher KS, et al. Prognostic evaluation based on cortical vein score difference in stroke. Stroke. 2013;44:2748–2754. doi: 10.1161/STROKEAHA.113.001231.
Crossref
PubMed
Google Scholar
31.
Phan TG, Hilton J, Beare R, Srikanth V, Sinnott M. Computer modeling of anterior circulation stroke: proof of concept in cerebrovascular occlusion. Front Neurol. 2014;5:176. doi: 10.3389/fneur.2014.00176.
Crossref
PubMed
Google Scholar
32.
Kim SJ, Seok JM, Bang OY, Kim GM, Kim KH, Jeon P, et al. MR mismatch profiles in patients with intracranial atherosclerotic stroke: a comprehensive approach comparing stroke subtypes. J Cereb Blood Flow Metab. 2009;29:1138–1145. doi: 10.1038/jcbfm.2009.38.
Crossref
PubMed
Google Scholar
33.
Lee MJ, Bang OY, Kim SJ, Kim GM, Chung CS, Lee KH, et al. Role of statin in atrial fibrillation-related stroke: an angiographic study for collateral flow. Cerebrovasc Dis. 2014;37:77–84. doi: 10.1159/000356114.
Crossref
PubMed
Google Scholar

eLetters(0)

eLetters should relate to an article recently published in the journal and are not a forum for providing unpublished data. Comments are reviewed for appropriate use of tone and language. Comments are not peer-reviewed. Acceptable comments are posted to the journal website only. Comments are not published in an issue and are not indexed in PubMed. Comments should be no longer than 500 words and will only be posted online. References are limited to 10. Authors of the article cited in the comment will be invited to reply, as appropriate.

Comments and feedback on AHA/ASA Scientific Statements and Guidelines should be directed to the AHA/ASA Manuscript Oversight Committee via its Correspondence page.

Sign In to Submit a Response to This Article

Information & Authors

Information

Published In

Go to Stroke
Go to Stroke

The image is taken from an article in this issue, “Are We Barking Up the Wrong Vessels?: Cerebral Microcirculation After Subarachnoid Hemorrhage” by Terpolilli et al (Stroke. 2015;46:3014–3019).

The image is Figure 2.

Stroke
Volume 46 • Number 10 • October 2015
Pages: 2800 - 2807
PubMed: 26306754

Copyright

© 2015 American Heart Association, Inc.

Versions

You are viewing the most recent version of this article.

History

Received: 24 April 2015
Revision received: 19 July 2015
Accepted: 21 July 2015
Published online: 25 August 2015
Published in print: October 2015

Permissions

Request permissions for this article.

Keywords

  1. collateral circulation
  2. diffusion magnetic resonance imaging
  3. magnetic resonance imaging
  4. ROC curve
  5. stroke

Subjects

  1. Ischemic Stroke

Authors

Affiliations

Mi Ji Lee, MD
From the Departments of Neurology (M.J.L., S.J.K., S.R., G.-M.K., C.-S.C., K.H.L., O.Y.B.) and Radiology (J.C.), Samsung Medical Center, School of Medicine (M.J.L., S.J.K., S.R., G.-M.K., C.-S.C., K.H.L., O.Y.B., J.C.) and Department of Health Sciences and Technology, Samsung Advanced Institute for Health Sciences and Technology (J.P.S., O.Y.B.), Sungkyunkwan University, Seoul, Korea; and Biostatistics Team, Samsung Biomedical Research Institute, Seoul, Korea (S.-Y.W.).
View all articles by this author
Jeong Pyo Son, MS
From the Departments of Neurology (M.J.L., S.J.K., S.R., G.-M.K., C.-S.C., K.H.L., O.Y.B.) and Radiology (J.C.), Samsung Medical Center, School of Medicine (M.J.L., S.J.K., S.R., G.-M.K., C.-S.C., K.H.L., O.Y.B., J.C.) and Department of Health Sciences and Technology, Samsung Advanced Institute for Health Sciences and Technology (J.P.S., O.Y.B.), Sungkyunkwan University, Seoul, Korea; and Biostatistics Team, Samsung Biomedical Research Institute, Seoul, Korea (S.-Y.W.).
View all articles by this author
Suk Jae Kim, MD
From the Departments of Neurology (M.J.L., S.J.K., S.R., G.-M.K., C.-S.C., K.H.L., O.Y.B.) and Radiology (J.C.), Samsung Medical Center, School of Medicine (M.J.L., S.J.K., S.R., G.-M.K., C.-S.C., K.H.L., O.Y.B., J.C.) and Department of Health Sciences and Technology, Samsung Advanced Institute for Health Sciences and Technology (J.P.S., O.Y.B.), Sungkyunkwan University, Seoul, Korea; and Biostatistics Team, Samsung Biomedical Research Institute, Seoul, Korea (S.-Y.W.).
View all articles by this author
Sookyung Ryoo, MD
From the Departments of Neurology (M.J.L., S.J.K., S.R., G.-M.K., C.-S.C., K.H.L., O.Y.B.) and Radiology (J.C.), Samsung Medical Center, School of Medicine (M.J.L., S.J.K., S.R., G.-M.K., C.-S.C., K.H.L., O.Y.B., J.C.) and Department of Health Sciences and Technology, Samsung Advanced Institute for Health Sciences and Technology (J.P.S., O.Y.B.), Sungkyunkwan University, Seoul, Korea; and Biostatistics Team, Samsung Biomedical Research Institute, Seoul, Korea (S.-Y.W.).
View all articles by this author
Sook-Young Woo, MS
From the Departments of Neurology (M.J.L., S.J.K., S.R., G.-M.K., C.-S.C., K.H.L., O.Y.B.) and Radiology (J.C.), Samsung Medical Center, School of Medicine (M.J.L., S.J.K., S.R., G.-M.K., C.-S.C., K.H.L., O.Y.B., J.C.) and Department of Health Sciences and Technology, Samsung Advanced Institute for Health Sciences and Technology (J.P.S., O.Y.B.), Sungkyunkwan University, Seoul, Korea; and Biostatistics Team, Samsung Biomedical Research Institute, Seoul, Korea (S.-Y.W.).
View all articles by this author
Jihoon Cha, MD
From the Departments of Neurology (M.J.L., S.J.K., S.R., G.-M.K., C.-S.C., K.H.L., O.Y.B.) and Radiology (J.C.), Samsung Medical Center, School of Medicine (M.J.L., S.J.K., S.R., G.-M.K., C.-S.C., K.H.L., O.Y.B., J.C.) and Department of Health Sciences and Technology, Samsung Advanced Institute for Health Sciences and Technology (J.P.S., O.Y.B.), Sungkyunkwan University, Seoul, Korea; and Biostatistics Team, Samsung Biomedical Research Institute, Seoul, Korea (S.-Y.W.).
View all articles by this author
Gyeong-Moon Kim, MD, PhD
From the Departments of Neurology (M.J.L., S.J.K., S.R., G.-M.K., C.-S.C., K.H.L., O.Y.B.) and Radiology (J.C.), Samsung Medical Center, School of Medicine (M.J.L., S.J.K., S.R., G.-M.K., C.-S.C., K.H.L., O.Y.B., J.C.) and Department of Health Sciences and Technology, Samsung Advanced Institute for Health Sciences and Technology (J.P.S., O.Y.B.), Sungkyunkwan University, Seoul, Korea; and Biostatistics Team, Samsung Biomedical Research Institute, Seoul, Korea (S.-Y.W.).
View all articles by this author
Chin-Sang Chung, MD, PhD
From the Departments of Neurology (M.J.L., S.J.K., S.R., G.-M.K., C.-S.C., K.H.L., O.Y.B.) and Radiology (J.C.), Samsung Medical Center, School of Medicine (M.J.L., S.J.K., S.R., G.-M.K., C.-S.C., K.H.L., O.Y.B., J.C.) and Department of Health Sciences and Technology, Samsung Advanced Institute for Health Sciences and Technology (J.P.S., O.Y.B.), Sungkyunkwan University, Seoul, Korea; and Biostatistics Team, Samsung Biomedical Research Institute, Seoul, Korea (S.-Y.W.).
View all articles by this author
Kwang Ho Lee, MD, PhD
From the Departments of Neurology (M.J.L., S.J.K., S.R., G.-M.K., C.-S.C., K.H.L., O.Y.B.) and Radiology (J.C.), Samsung Medical Center, School of Medicine (M.J.L., S.J.K., S.R., G.-M.K., C.-S.C., K.H.L., O.Y.B., J.C.) and Department of Health Sciences and Technology, Samsung Advanced Institute for Health Sciences and Technology (J.P.S., O.Y.B.), Sungkyunkwan University, Seoul, Korea; and Biostatistics Team, Samsung Biomedical Research Institute, Seoul, Korea (S.-Y.W.).
View all articles by this author
Oh Young Bang, MD, PhD
From the Departments of Neurology (M.J.L., S.J.K., S.R., G.-M.K., C.-S.C., K.H.L., O.Y.B.) and Radiology (J.C.), Samsung Medical Center, School of Medicine (M.J.L., S.J.K., S.R., G.-M.K., C.-S.C., K.H.L., O.Y.B., J.C.) and Department of Health Sciences and Technology, Samsung Advanced Institute for Health Sciences and Technology (J.P.S., O.Y.B.), Sungkyunkwan University, Seoul, Korea; and Biostatistics Team, Samsung Biomedical Research Institute, Seoul, Korea (S.-Y.W.).
View all articles by this author

Notes

The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.115.009828/-/DC1.
Correspondence to Oh Young Bang, MD, PhD, Department of Neurology, Samsung Medical Center, Sungkyunkwan University, 50 Irwon-dong, Gangnam-gu, Seoul 135–710, South Korea. E-mail [email protected]

Disclosures

None.

Sources of Funding

This study was supported by the Korean Health Technology Research and Development Project, Ministry of Health & Welfare, Republic of Korea (HI14C15310000).

Metrics & Citations

Metrics

Citations

Download Citations

If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Select your manager software from the list below and click Download.

    • Ting-Yu Chang,
    • Soren Christensen,
    • Michael Mlynash,
    • Jeremy J. Heit,
    • Michael P. Marks,
    • Sarah Lee,
    • Margy E. McCullough-Hicks,
    • Lili Velickovic Ostojic,
    • Stephanie Kemp,
    • Gregory W. Albers,
    • Aditya Srivatsan,
    • Tsong-Hai Lee,
    • Maarten G. Lansberg,
    Perfusion Profiles May Differ Between Asymptomatic Versus Symptomatic Internal Carotid Artery Occlusion, Journal of Stroke, 26, 1, (108-111), (2024).https://doi.org/10.5853/jos.2023.02768
    Crossref
    • Jiangbo Ding,
    • Xuying Chang,
    • Yong Shen,
    • Peiyu Ma,
    • Ruoyu Zhang,
    • Guangwu Yang,
    • Linjie Mu,
    • Xingkui Zhang,
    • Zhigao Li,
    • Jinwei Tang,
    • Zhiwei Tang,
    Prediction of cerebral infarction after bypass surgery in adult moyamoya disease: combing parameters on 4D perfusion CT with clinical related factors, Acta Neurochirurgica, 166, 1, (2024).https://doi.org/10.1007/s00701-024-06373-8
    Crossref
    • Rui He,
    • Jia Zhou,
    • Xiaoyu Xu,
    • Xiaoer Wei,
    • Feng Wang,
    • Yuehua Li,
    RETRACTED: Comparing the predictive value of quantitative magnetic resonance imaging parametric response mapping and conventional perfusion magnetic resonance imaging for clinical outcomes in patients with chronic ischemic stroke, Frontiers in Neuroscience, 17, (2023).https://doi.org/10.3389/fnins.2023.1177044
    Crossref
    • Sukhdeep Singh Bal,
    • Fan-pei Gloria Yang,
    • Nai-Fang Chi,
    • Jiu Haw Yin,
    • Tao-Jung Wang,
    • Giia Sheun Peng,
    • Ke Chen,
    • Ching-Chi Hsu,
    • Chang-I Chen,
    Core and penumbra estimation using deep learning-based AIF in association with clinical measures in computed tomography perfusion (CTP), Insights into Imaging, 14, 1, (2023).https://doi.org/10.1186/s13244-023-01472-z
    Crossref
    • Chun-Xia Li,
    • Frank Tong,
    • Doty Kempf,
    • Leonard Howell,
    • Xiaodong Zhang,
    Longitudinal evaluation of the functional connectivity changes in the secondary somatosensory cortex (S2) of the monkey brain during acute stroke, Current Research in Neurobiology, 5, (100097), (2023).https://doi.org/10.1016/j.crneur.2023.100097
    Crossref
    • Sarah Emhemed Abousrafa,
    • Grant Mair,
    MRI for collateral assessment pre-thrombectomy and association with outcome: a systematic review and meta-analysis, Neuroradiology, 65, 6, (1001-1014), (2023).https://doi.org/10.1007/s00234-023-03127-8
    Crossref
    • Seyed Ali Nabavizadeh,
    • Ronald L. Wolf,
    Clinical Applications of MR Perfusion Imaging, Functional Neuroradiology, (119-160), (2023).https://doi.org/10.1007/978-3-031-10909-6_5
    Crossref
    • Brian Tsui,
    • May Nour,
    • Iris Chen,
    • Joe X. Qiao,
    • Banafsheh Salehi,
    • Bryan Yoo,
    • Geoffrey P. Colby,
    • Noriko Salamon,
    • Pablo Villablanca,
    • Reza Jahan,
    • Gary Duckwiler,
    • Jeffrey L. Saver,
    • David S. Liebeskind,
    • Kambiz Nael,
    MR Angiography in Assessment of Collaterals in Patients with Acute Ischemic Stroke: A Comparative Analysis with Digital Subtraction Angiography, Brain Sciences, 12, 9, (1181), (2022).https://doi.org/10.3390/brainsci12091181
    Crossref
    • A. Potreck,
    • E. Scheidecker,
    • C.S. Weyland,
    • U. Neuberger,
    • C. Herweh,
    • M.A. Möhlenbruch,
    • M. Chen,
    • S. Nagel,
    • M. Bendszus,
    • F. Seker,
    RAPID CT Perfusion–Based Relative CBF Identifies Good Collateral Status Better Than Hypoperfusion Intensity Ratio, CBV-Index, and Time-to-Maximum in Anterior Circulation Stroke, American Journal of Neuroradiology, 43, 7, (960-965), (2022).https://doi.org/10.3174/ajnr.A7542
    Crossref
    • Smit D. Patel,
    • David Liebeskind,
    Collaterals and Elusive Ischemic Penumbra, Translational Stroke Research, 14, 1, (3-12), (2022).https://doi.org/10.1007/s12975-022-01116-2
    Crossref
  11. See more

View Options

View options

PDF and All Supplements

Download PDF and All Supplements

PDF/EPUB

View PDF/EPUB
Login options

Check if you have access through your login credentials or your institution to get full access on this article.

Personal login Institutional Login
Save for later Item saved, go to cart
Purchase Options

Purchase this article to access the full text.

Purchase access to this article for 24 hours

Predicting Collateral Status With Magnetic Resonance Perfusion Parameters
Stroke
  • Vol. 46
  • No. 10

Purchase access to this journal for 24 hours

Stroke
  • Vol. 46
  • No. 10
Restore your content access

Enter your email address to restore your content access:

Note: This functionality works only for purchases done as a guest. If you already have an account, log in to access the content to which you are entitled.

Media

Figures

Other

Tables

Share

Share

Share article link

Share

View full text|Download PDF

Comment Response