Abstract
Intra-arterial therapy (IAT) for ischemic stroke aims to save brain tissue. Collaterals are thought to contribute to prolonged penumbra sustenance. In this study, we investigate the effect of collateral status on brain tissue salvage with IAT. In 500 patients randomized between IAT and standard care, collateral status was graded from 0 (absent) to 3 (good). Final infarct volumes (FIV) were calculated on post-treatment CT. FIVs were compared between treatment groups per collateral grade. Multivariable linear regression with interaction terms was performed to study whether collaterals modified IAT effect on FIV. Four-hundred-forty-nine patients were included in the analysis. Median FIV for the IAT group was significantly lower with 54.5 mL (95% IQR: 21.8–145.0) than for the controls with 81.8 mL (95% IQR: 40.0–154.0) (p = 0.020). Treatment effect differed across collateral grades, although there was no significant interaction (unadjusted p = 0.054; adjusted p = 0.105). For grade 3, IAT resulted in a FIV reduction of 30.1 mL (p = 0.024). For grade 2 and 1, this difference was, respectively, 28.4 mL (p = 0.028) and 28.4 mL (p = 0.29). For grade 0, this was 88.6 mL (p = 0.28) in favour of controls. IAT saves substantially more brain tissue as compared to standard care. We observed a trend of increasing effect of IAT with higher collateral grades.
Keywords: Acute ischemic stroke, collateral circulation, final infarct volume, intra-arterial therapy, treatment effect
Introduction
Recent advances in intra-arterial therapy (IAT) have resulted in a major change in the treatment of patients with acute ischemic stroke (AIS) due to a proximal large vessel occlusion in the anterior circulation. Recently, five large randomized controlled trials (RCTs) demonstrated high recanalization rates and improved functional outcome after IAT.1–5 Despite this new landmark in AIS therapy, IAT is not beneficial for everyone. A substantial amount of patients included in these RCTs did not reach functional independence at three months ranging from 33% to 71%, making patient selection for IAT essential in optimizing AIS management. Collaterals are anastomotic vessels providing alternative routes for blood flow, especially in patients with AIS. Their status has shown to be associated with IAT effect, and may help in the stratification of subjects eligible for IAT.6–9 Since good collateral status at baseline is related to an improved functional outcome, it is believed that collaterals are of major importance for sustaining the penumbra. Until now, the positive effect of a developed collateral network is commonly indirectly evaluated by assessment of a subject's functional ability and limitations, rated on relatively coarse scales such as the NIHSS or 90-day modified Rankin Scale (mRS). However, the direct relationship between collateral status and tissue outcome after IAT has not yet been investigated. This may provide prognostic information on the extent of brain tissue that might be saved with IAT, which may be useful in future patient selection models. In this post hoc analysis of MR CLEAN1 data, we investigate the direct relationship between collateral flow and the effect of IAT, as expressed by difference in final infarct volume (FIV).
Material and methods
Patient selection
The MR CLEAN trial protocol has been reported previously.10 In short, patients with AIS were randomized between IAT or standard care alone (control) if they showed a proximal anterior circulation arterial occlusion of the distal intracranial carotid artery (ICA), middle cerebral artery (M1 or M2), or anterior cerebral artery (A1 or A2) on CTA, which was treatable within 6 h after start of clinical symptoms. All patients or their legal representatives provided written informed consent before randomization. The study protocol was approved by the Medical Ethics Committee of Erasmus MC University Medical Center and the research board of each participating centre. This study is conducted in accordance with the principles of the Declaration of Helsinki of 1975 (revised 1983), and with the guidelines for Good Clinical Practice.
For this post hoc analysis, we included all patients with a five to seven day (range 3–9 days) follow-up non-contrast CT (NCCT), or one day follow-up NCCT if five to seven day follow-up NCCT was not available due to death or discharge. In case of hemicraniectomy, the last scan prior to surgery was selected for analysis. Patients with large haemorrhagic transformations with unclear boundaries and image data with extreme artefacts or insufficient scan quality were excluded since they precluded accurate infarct volume determination.
Clinical and imaging assessment
In MR CLEAN, a central imaging committee assessed collateral status on baseline CTA. Image evaluators had more than 10 years of experience and were blinded to all clinical findings, except symptom side. All CTA images were independently graded by two neuroradiologists. A third reader resolved discrepancies between the initial two readers. Presence of collaterals was defined as visual appearance of vasculature distal to the proximal artery occlusion and was assessed on baseline CTA on a commonly used 4-point scale, with 0 for absent collaterals (0% filling of the occluded territory), 1 for poor collaterals (>0% and ≤50% filling of the occluded territory), 2 for moderate (>50% and <100% filling of the occluded territory), and 3 for good collaterals (100% filling of the occluded territory).11 Readers were allowed to use the non-ischemic hemisphere as normal reference. A mixture of CTA source images and maximum-intensity-projections were used for collateral grading, including all available slices. On the whole, if different slices expressed different collateral capacities, an average collateral score over all slices was determined. No fixed CTA protocols were used in MR CLEAN, and protocols varied per centre. Inter-observer reliability for collateral status assessment in MR CLEAN has previously been reported (kappa = 0.60).9
Whole-brain imaging was performed with thin-section acquisition on multi-slice CT scanners with at least 16 sections. FIV was defined as a hypodense lesion on follow-up NCCT. Adjacent hyperdense areas suspected for haemorrhagic transformation were considered part of the FIV. A previously validated method, which was optimized to work fully automatic without user input, was used to segment the lesions.12 This segmentation resulted in binary masks of the infarcted areas. All segmentations were inspected and adjusted if necessary by two experienced observers blinded to outcome (A.M.B and K.R.K.). Because this method was developed to identify infarcted areas after three days, manual adjustment of the masks in early follow-up cases was common. FIV was calculated in millilitres by multiplying the number of voxels of the segmented area with its voxel size
Statistical analysis
Dichotomous variables were presented as proportion of population. Continuous variables were tested for normality (Kolmogorov–Smirnov test) and presented as mean ± SD if normally distributed, or as median and interquartile range (IQR) otherwise. Differences in baseline parameters were assessed using the Chi-square test for categorical variables, or student's t-test or Mann–Whitney U test for continuous variables. Baseline differences were not used for adjustment of further analyses, as in MR CLEAN pre-specified prognostic variables were agreed upon for this purpose. Mann–Whitney U test was used to test for differences in the distribution of FIV between IAT patients and controls for the whole population and per collateral grade. Multivariable linear regression was performed to model the association between treatment allocation (IAT or standard care alone) and FIV per collateral grade separately, with and without adjusting for pre-specified prognostic variables. Interaction between treatment and collateral grades on tissue salvage was investigated for the population as a whole, by adding a multiplicative interaction term to the adjusted and unadjusted model. Results were reported as βs with 95% confidence intervals. FIV was log transformed to best satisfy the linear model (normal distribution of residuals and homoscedasticity). The exponent of β determines the relative difference in FIV for patients treated with IAT compared to standard care alone.
βs were adjusted for major pre-specified prognostic variables from the original MR CLEAN protocol statistical analysis plan: age, stroke severity (NIHSS) at baseline, time of onset to randomization, previous stroke, atrial fibrillation, diabetic status, and occlusion location. All p-values were calculated for two-sided tests. Statistical analysis was performed using SPSS v.22.0 (IBM Corp., Armonk, NY, USA). A p-value of <0.05 indicated statistical significance.
Results
Of the 500 patients included in the MR CLEAN trial, 472 patients underwent a follow-up NCCT (75.8% 5–7 day, 24.2% 1 day). Twenty-three patients were subsequently excluded because of hemicraniectomy with no pre-operative follow-up scan available (N = 4), insufficient scan quality (N = 8), large haemorrhagic transformations (N = 6), baseline magnetic resonance angiography instead of CTA (N = 2), insufficient vessel coverage (N = 2), or no baseline vessel imaging at all (N = 1), resulting in 449 patients for analysis. Baseline characteristics for all patients are displayed in Table 1. Mean age was 64.6 (SD ± 13.8) years, 40.4% were female, and median NIHSS at baseline was 18.0 (IQR: 14.0–22.0). Significant differences in distribution between collateral grades were seen for age, baseline NIHSS, Alberta Stroke Early CT Score (ASPECTS), use of statins and antiplatelet agents, history of ischemic stroke, hypertension, diabetes mellitus, peripheral artery disease, and hyperlipidaemia. All other baseline characteristics were evenly distributed between the groups.
Table 1.
Clinical baseline characteristics for the total population, and per collateral grade.
| Baseline Characteristics | Total population | Grade 0 | Grade 1 | Grade 2 | Grade 3 | p |
|---|---|---|---|---|---|---|
| Number of patients (N) | N = 449 | N = 20 | N = 126 | N = 180 | N = 123 | |
| Age – median (IQR) | 65.00 (55.0–76.0) | 72.50 (65.3–77.5) | 67.0 (57.0–76.8) | 65.0 (54.0–76.0) | 63.0 (52.5–73.0) | 0.049* |
| Male sex – n (%) | 267 (59.5) | 14 (70.0) | 79 (62.7) | 111 (61.7) | 63 (51.2) | |
| Left hemisphere (%) | 241 (53.7) | 11 (55.0) | 66 (52.4) | 95 (52.8) | 69 (56.1) | |
| NIHSS – median (IQR) | 18.0 (14.0–22.0) | 21.0 (17.8–23.0) | 20.0 (16.0–23.0) | 17.0 (14.0–21.0) | 16.0 (12.0–19.0) | <0.001* |
| Atrial fibrillation – n (%) | 117 (26.1) | 7 (35.0) | 33 (26.2) | 43 (23.9) | 34 (27.6) | |
| History of diabetes mellitus – n (%) | 54 (12.0) | 4 (20.0) | 22 (17.5) | 13 ( 7.2) | 15 (12.2) | 0.025* |
| History of hypertension – n (%) | 195 (43.4) | 8 (40.0) | 67 (53.2) | 63 (35.0) | 57 (46.3) | 0.014* |
| History of ischemic stroke – n (%) | 45 (10.0) | 1 ( 5.0) | 21 (16.7) | 14 ( 7.8) | 9 ( 7.3) | 0.048* |
| History of myocardial infarction – n (%) | 66 (14.7) | 4 (20.0) | 24 (19.0) | 21 (11.7) | 17 (13.8) | |
| History of peripheral artery disease – n (%) | 20 ( 4.5) | 0 ( 0.0) | 12 ( 9.5) | 5 ( 2.8) | 3 ( 2.4) | 0.027* |
| History of hyperlipidaemia – n (%) | 107 (23.8) | 3 (15.0) | 48 (38.1) | 33 (18.3) | 23 (18.7) | <0.001* |
| History of smoking – n (%)a | 127 (28.3) | 4 (20.0) | 38 (30.2) | 47 (26.1) | 38 (30.9) | |
| Current statin use – n (%) | 119 (26.5) | 4 (20.0) | 52 (41.3) | 33 (18.3) | 30 (24.4) | <0.001* |
| Current anticoagulant use – n (%) | 32 ( 7.1) | 1 ( 5.0) | 12 ( 9.5) | 8 ( 4.4) | 11 ( 8.9) | |
| Current antiplatelet use – n (%) | 121 (26.9) | 5 (25.0) | 46 (36.5) | 43 (23.9) | 27 (22.0) | 0.044* |
| Pre-stroke modified Rankin Scale score – n (%) | ||||||
| 0 | 367 (81.7) | 15 (75.0) | 97 (77.0) | 155 (86.1) | 100 (81.3) | |
| 1 | 44 ( 9.8) | 3 (15.0) | 12 ( 9.5) | 19 (10.6) | 10 (8.1) | |
| ≥2 | 38 (8.4) | 2 (10.0) | 17 (13.5) | 6 (3.6) | 13 (10.6) | |
| Systolic blood pressure – mean mmHg (SD)b | 144.0 (24.1) | 150.1 (25.8) | 145.1 (23.0) | 142.4 (24.8) | 144.1 (23.9) | |
| Treatment with IV alteplase – n (%) | 401 (89.3) | 17 (85.0) | 112 (88.9) | 164 (91.1) | 108 (87.8) | |
| Onset to IV alteplase in min–median (IQR) | 85.0 (65.0–110.0) | 80.0 (73.0–95.0) | 85.0 (69.8–107.8) | 81.5 (65.0–111.5) | 90.0 (63.8–119.3) | |
| ASPECTSc | <0.001* | |||||
| ASPECTS 0–4 | 27 ( 6.1) | 3 (15.0) | 13 (10.4) | 10 ( 5.6) | 1 ( 0.8) | |
| ASPECTS 5–7 | 80 (17.9) | 5 (25.0) | 23 (18.4) | 35 (19.6) | 17 (13.9) | |
| ASPECTS 8–10 | 339 (76.0) | 12 (60.0) | 89 (71.2) | 134 (74.9) | 104 (85.2) | |
| Level of occlusion – n (%) | ||||||
| ICA | 4 ( 0.9) | 0 ( 0.0) | 0 ( 0.0) | 1 ( 0.6) | 3 ( 2.4) | |
| ICA-T | 127 (28.3) | 5 (25.0) | 45 (35.7) | 51 (28.3) | 26 (21.1) | |
| M1 | 281 (62.6) | 12 (60.0) | 75 (59.5) | 112 (62.2) | 82 (66.7) | |
| M2 | 35 ( 7.8) | 3 (15.0) | 6 ( 4.8) | 15 ( 8.3) | 11 ( 8.9) | |
| A2 | 2 ( 0.4) | 0 ( 0.0) | 0 ( 0.0) | 1 ( 0.6) | 1 ( 0.8) | |
| Onset to randomization in min–median (IQR)d | 201.0 (150.0–259.5) | 198.5 (158.0–241.0) | 195.5 (148.0–248.5) | 195.5 (148.8–261.3) | 219.0 (159.0–271.0) | |
| Onset to reperfusion in min–median (IQR)e | 342.0 (274.0–397.5) | 307.5 (283.0–332.0) | 353.0 (264.0–418.0) | 338.5 (260.8–386.5) | 340.5 (274.0–394.8) | |
IQR: interquartile range; NIHSS: National Institutes of Health Stroke Scale range 0 to 42, higher scores indication more severe neurological deficits; SD: standard deviation; IV: intravenous; ASPECTS: Alberta Stroke Program Early CT Score, range 0 to 10, higher scores indicate less early ischemic changes; ICA: internal carotid artery (intracranial segment); ICA-T: internal carotid artery with involvement of the M1 segment; M1/2: middle cerebral artery segments.
Statistically significant difference (p < 0.05).
Current smoking status was missing in 23 patients in MR CLEAN.
Systolic blood pressure at baseline was missing in one patient in MR CLEAN.
ASPECTS was not available in one patient in MR CLEAN.
Randomization time was missing in two patients in MR CLEAN.
Onset to reperfusion (TICI 2B-3) was only available for patients treated with IAT (n = 214).
Final infarct volume
In the total population, FIV for patients allocated to IAT was significantly lower with 54.5 mL (IQR 21.8–145.0; N = 214) than for patients allocated to standard care alone with 81.8 mL (IQR 40.0–154.0; N = 235) (p = 0.020). Differences in FIV between treatment groups ordered by collateral grade are displayed in Table 2 and Figure 1. There was substantial heterogeneity in the effect of IAT on FIV across collateral grades. A significant difference in FIV was observed for grade 3 with 24.8 mL (IQR 10.3–59.9) versus 54.9 mL (IQR 22.1–94.6); p = 0.024 and for grade 2 collaterals 42.4 mL (IQR 15.4–94.9) versus 70.8 mL (IQR 31.9–109.3); p = 0.028 between, respectively, IAT and control group. For grades 0 and 1, the difference was smaller, 119.4 mL (IQR 60.3–199.5) versus 147.8 mL (IQR 70.7–279.5); p = 0.29 for grade 1 and 338.9 mL (IQR 205.4–416.1) versus 250.3 mL (IQR 59.6–396.5); p = 0.28 for grade 0. The extent of infarct volumes is illustrated in Figure 2, in which the frequency of infarction, stratified stratified by treatment and collateral grade, is shown as colour-coded overlays.
Table 2.
Median infarct volumes per collateral grade.
| Collateral score | Number of patients (N) | Final infarct volume – median mL (IQR) | Intra-arterial therapya | Number of patients N (%) | Final infarct volume – median mL (IQR) | p |
|---|---|---|---|---|---|---|
| Grade 0 (absent) | 20 | 271.0 (105.8–411.8) | Yes | 7 (35.0) | 338.9 (205.4–416.1) | 0.275 |
| No | 13 (65.0) | 250.3 (59.6 -396.5) | ||||
| Grade 1 (poor) | 126 | 139.8 (63.9–228.1) | Yes | 64 (50.8) | 119.4 (60.3–199.5) | 0.294 |
| No | 62 (49.2) | 147.8 (70.7–279.5) | ||||
| Grade 2 (moderate) | 180 | 55.9 (27.4–107.8) | Yes | 80 (44.4) | 42.4 (15.4–94.9) | 0.028* |
| No | 100 (55.6) | 70.8 (31.9–109.3) | ||||
| Grade 3 (good) | 123 | 36.9 (15.4–87.6) | Yes | 63 (50.8 | 24.8 (10.3–59.9) | 0.024* |
| No | 61 (49.2) | 54.9 (22.1–94.6) |
Note: Additionally, per collateral grade differences in infarct volumes for patients treated with IAT and controls are given.
IQR: interquartile range.
Statistically significant difference (p < 0.05).
Treated with intra-arterial therapy with/or without standard care (intravenous thrombolysis).
Figure 1.
Distribution of final infarct volume (FIV) per collateral grade for patients treated with intra-arterial therapy (IAT) and controls. Patients treated with IAT are depicted in orange, controls in grey. FIV for IAT vs. controls using the Mann–Whitney U test was significantly different at 24.8 vs. 54.9 mL for grade 3 (p = 0.024), and at 42.4 vs. 70.8 mL for grade 2 (p = 0.028). No significant FIV difference was found for grade 1 collaterals, at 119.4 vs. 147.8 mL (p = 0.29), and for grade 0, at 338.9 vs. 250.3 mL (p = 0.28). *Indicates statistical significance for p < 0.05.
Figure 2.
Infarction overlay plot of all final infarct volumes (FIVs) of patients allocated to standard care (control) and intra-arterial therapy (IAT), ordered by collateral grade. Individual follow-up non-contrast CTs have been coregistered into the same coordinate space of a healthy subject using a linear affine registration. The sum of FIVs is superimposed on a reference CTA image to show the voxel-wise frequency of infarction. The colour bar indicates the percentage of infarctions of the grouped population.
After adjustment for pre-specified prognostic variables, the same trend was visible (Table 3). Distinct effects were observed for grade 2 and 3, while no significant association between IAT and FIV was found for grade 0 and 1. For grade 2 (N = 180), treatment with IAT led to relative difference in FIV of 30% (Exp. β 0.70; p = 0.058), if patients were treated with IAT. For collateral grade 3 (N = 123), treatment with IAT led to a 31% difference in FIV (Exp. β 0.69; p = 0.094). For grade 1 and 0, the βs were respectively Exp. β = 0.92 (p = 0.69) and Exp. β = 1.97 (p = 0.210), but the interaction between collateral grade and treatment fell short of significance (unadjusted p = 0.054, adjusted p = 0.105).
Table 3.
Results of the linear regression analysis for the association between collateral score and final infarct volume, adjusted for pre-specified prognostic variables.
| Collateral score | Number of patients (N) | Adjusteda | β log transformed (95%CI)b | Exp. (β)c | p |
|---|---|---|---|---|---|
| Grade 0 (absent) | 20 | Yes | 0.678 (−0.442 to 1.798) | 1.97 | 0.210 |
| No | 0.636 (−0.576 to 1.849) | 1.89 | 0.285 | ||
| Grade 1 (poor) | 126 | Yes | −0.081 (−0.480 to 0.318) | 0.92 | 0.688 |
| No | −0.010 (−0.408 to 0.387) | 0.99 | 0.959 | ||
| Grade 2 (moderate) | 180 | Yes | −0.357 (−0.727 to 0.013) | 0.70 | 0.058 |
| No | −0.390 (−0.759 to −0.022) | 0.68 | 0.038* | ||
| Grade 3 (good) | 123 | Yes | −0.364 (−0.791 to 0.063) | 0.69 | 0.094 |
| No | −0.425 (−0.862 to 0.012) | 0.65 | 0.057 | ||
| Interaction termd | 449 | Yes | −0.212 (−0.469 to 0.044) | n/a | 0.105 |
| No | −0.260 (−0.524 to 0.004) | n/a | 0.054 |
Exp. (β): the exponent of β; CI: confidence interval.
Statistically significant difference (p < 0.05).
Adjusted for pre-specified prognostic variables.
Due to the non-normal distribution of final infarct volume, a log+1 transformation was performed to best fit the assumptions associated with the linear regression model.
Exponent of β was calculated, to determine the relative difference of final infarct volume when treated with intra-arterial therapy.
Multiplicative interaction term used in the regression model, to determine the modification by collaterals of intra-arterial therapy effect on final infarct volume.
Discussion
Among AIS patients with a proximal arterial occlusion, IAT is associated with smaller FIVs as compared to standard care alone. We found a trend towards increasing treatment effect with increasing collaterals. Our studies showed that in patients supported by an adequate collateral circulation, an average of one-third additional brain tissue can be saved with IAT. Our results indicate the importance of collaterals in the maintenance of tissue viability.
Identifying the group of patients with no treatment benefit is the fundamental principle of patient selection, in order to prevent unnecessary risk and expense of treatment. Unfortunately, this analysis is limited by the small number of patients in the absent collateral grade group, amounting to only 20 subjects. However, from a physiological perspective, it is likely that IAT does not lead to substantial brain salvage in these patients due to the fast progression of irreversible brain damage.13 Consistent with this idea, we found no effect in patients with collateral grade 0. Therefore, it might be that the proposed treatment window of <6 h is too long to save brain in these patients. Further research is needed to explore the relationship between collateral status and infarct progression in time.
Our findings are consistent with a reported case series of 50 patients with large vessel occlusions who received IAT.14 Using a dichotomized collateral score (poor/good) and an ellipsoid approximation of FIV as outcome measure (using Kothari's ABC/2 method11 on 12–24 h MRI, diffusion-weighted imaging or follow-up NCCT), a good collateral status was found to be associated with smaller infarct volumes. By virtue of the randomized trial data, the present study extends these findings by demonstrating an IAT benefit for tissue salvage in the moderate and good collateral groups.
Multiple studies have demonstrated the importance of baseline collateral status assessment.3,9,15,16 A recently published substudy of MR CLEAN demonstrated increased IAT benefit in patients with better baseline collateral scores.9 Although FIV and functional outcome are closely related, it is important to note the differences between these outcome measures. The brain consists of many regions, with each its own implication in the patient's functional wellbeing. When exposed to ischemia, the severity of the injury does not only depend heavily on the amount of brain tissue involved, but also on the part of brain that is affected. Therefore, functional end-points are of direct relevance to the patient, but do not fully capture the effect of collaterals on brain tissue salvage by IAT. Besides these pathophysiological differences, functional outcome measures are prone to subjectivity of the assessor and scored on a relatively coarse ordinal scale. Moreover, follow-up imaging is acquired within days after stroke onset, unlike most accepted functional outcome measures, which are obtained after a considerable amount of time. FIV would therefore be less influenced by post-stroke rehabilitation, and non-stroke related morbidity and mortality. In the search of patient selection criteria for IAT, it is important to identify differences directly related to treatment. So, although our analysis is limited by the relatively small numbers, especially with grade 0 and 3, this study provides unique information regarding the extent of brain salvage by IAT, in relation to the strength of the collateral circulation.
Our study has several limitations: some of the FIV measurements were performed on follow-up NCCTs acquired as early as day 1. Despite the inclusion of patients with early follow-up imaging, we lost approximately 6% of the total population due to death before imaging could be obtained. There is a strong likelihood that these patients suffered from a relative large infarction. Furthermore, this rather broad time-window of follow-up imaging might have led to a bias in FIV assessment. In this study, FIV was identified as hypodense lesion on NCCT. The spatial extent of these hypodensities can vary with time due to brain swelling and mass effect caused by oedema. In addition, blood–brain barrier breakdown and capillary leakage can lead to haemorrhagic transformation, resulting in density increase of the infarcted tissue, which temporarily may become isodense to grey matter (fogging effect).17,18 These effects could have resulted in an over- or underestimation of infarct volumes. Moreover, infarct growth might still be present after 24 h.19,20 However, the amount of infarct growth after this time is likely to be small, as final infarct imaging at this time point is strongly associated with functional outcome.21 Further research on identifying patients with fast or slow FIV progression over time is warranted.
In addition, collateral status assessment can be prone to inter-observer variability.22 Despite the high agreement between observers reported by Tan et al.,23 a moderate agreement was reported in MR CLEAN. Nevertheless, a beneficial effect of IAT was found for patients with moderate and good collaterals. Approaches for collateral grading vary widely, but almost all use categorical scales.22 It is important to improve this assessment and classification of collateral flow. In MR CLEAN, the collateral status was assessed using single-phase CTA, which is known to have a lower temporal and spatial resolution than the gold standard four-vessel digital subtraction angiography (DSA). However, for collateral status assessment in the acute setting, DSA is not practical since it is an invasive modality and resource intensive.24 Where conventional single-phase CTA fails to capture delayed collateral flow enhancement with a single snapshot, CTA acquired by multiple timeframes may help overcome this limitation.25–28 Both multiphase CTA and time-invariant CTA are promising non-invasive techniques, but are as today only routinely used in few centres. Furthermore, conventional CT or MR angiographic collateral grading does not include the time component of bolus speed and arrival, which is encoded in perfusion parameters such as CBF, MTT, or T-max and also in conventional DSA. In other words, collateral grading in conventional CT or MR angiography will differentiate the abundance of vasculature, but not the velocity, and this may be relevant in acute stroke triage: plentiful collaterals with fast arrival may designate a patient with better outcome from IAT in contrast to a patient with plentiful but slow collaterals and late arrival. This aspect has been touched in a study that used contrast-enhanced and TOF MRA to predict tissue outcome.29 In future research, a more quantitative measure of collateral flow, especially in combination with new dynamic imaging techniques, may help to fully encompass its role in improving patient selection for IAT.
In conclusion, IAT saves substantially more brain tissue as compared to standard care. We observed a trend of increasing effect of IAT with higher collateral grades. Our results suggest that baseline collateral status might be useful as an imaging selection tool to identify patients in whom the benefit of IAT is expected to be very small.
Supplementary Material
Acknowledgements
We thank Kilian Treurniet and Katinka van Kranendonk for their assistance in data acquisition and analysis.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Anna MM Boers and Ivo GH Jansen are supported by a personal grant from the Stichting Toegepast Wetenschappelijk Instituut voor Neuromodulatie (TWIN). The MR CLEAN trial was funded by the Dutch Heart Foundation and through unrestricted grants from AngioCare BV, Covidien/EV3®, MEDAC Gmbh/LAMEPRO and Penumbra Inc.
Declaration of conflicting interests
The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: AMM Boers and HA Marquering are co-founders and shareholders of Nico-lab BV. Dr Yoo received research grant (significant) from Penumbra Inc, and research grant (modest) from Neuravi Inc. Maastricht University MC, Erasmus MC University Medical Center, and Academic Medical Center Amsterdam received Speaker's bureau fee from Stryker Inc. for consultations by Drs Zwam, Dippel, and Majoie, respectively. The other authors report no conflicts.
Authors' contributions
Anna MM Boers: Study conception and design; acquisition of data; analysis and interpretation of data; drafting of manuscript.
Ivo GH Jansen: Study conception and design; acquisition of data; analysis and interpretation of data; drafting of manuscript.
Olvert A Berkhemer: Acquisition of data; critical revision.
Albert J Yoo: Analysis and interpretation of data; drafting of manuscript.
Hester F Lingsma: Study conception and design; drafting of manuscript.
Cornelis H Slump: Analysis and interpretation of data; critical revision.
Yvo BWEM Roos: Analysis and interpretation of data; critical revision.
Robert J van Oostenbrugge: Acquisition of data; critical revision.
Diederik WJ Dippel: Study conception and design; acquisition of data; critical revision.
Aad van der Lugt: Acquisition of data; analysis and interpretation of data; critical revision.
Wim H van Zwam: Acquisition of data; analysis and interpretation of data; critical revision.
Henk A Marquering: Study conception and design; analysis and interpretation of data; drafting of manuscript.
Charles BLM Majoie: Study conception and design; acquisition of data; analysis and interpretation of data; drafting of manuscript.
References
- 1.Berkhemer O, Fransen P, Beumer D, et al. A randomized trial of intraarterial treatment for acute ischemic stroke. New Engl J Med 2015; 372: 11–20. [DOI] [PubMed] [Google Scholar]
- 2.Campbell BCV, Mitchell PJ, Kleinig TJ, et al. Endovascular therapy for ischemic stroke with perfusion-imaging selection. N Engl J Med 2015; 372: 1009–18. [DOI] [PubMed] [Google Scholar]
- 3.Goyal M, Demchuk AM, Menon BK, et al. Randomized assessment of rapid endovascular treatment of ischemic stroke. N Engl J Med 2015; 372: 1019–30. [DOI] [PubMed] [Google Scholar]
- 4.Jovin TG, Chamorro A, Cobo E, et al. Thrombectomy within 8 hours after symptom onset in ischemic stroke. N Engl J Med 2015; 372: 1–11. [DOI] [PubMed] [Google Scholar]
- 5.Saver JL, Goyal M, Bonafe A, et al. Stent-retriever thrombectomy after intravenous t-PA vs. t-PA alone in stroke. N Engl J Med 2015; 372: 2285–95. [DOI] [PubMed] [Google Scholar]
- 6.Leng X, Fang H, Leung TWH, et al. Impact of collaterals on the efficacy and safety of endovascular treatment in acute ischaemic stroke: a systematic review and meta-analysis. J Neurol Neurosurg Psychiatry 2016; 87: 537–44. [DOI] [PubMed] [Google Scholar]
- 7.Bang OY, Saver JL, Kim SJ, et al. Collateral flow predicts response to endovascular therapy for acute ischemic stroke. Stroke 2011; 42: 693–699. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Liebeskind DS, Tomsick TA, Foster LD, et al. Collaterals at angiography and outcomes in the Interventional Management of Stroke (IMS) III trial. Stroke 2014; 45: 759–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Berkhemer OA, Jansen IGH, Beumer D, et al. Collateral status on baseline computed tomographic angiography and intra-arterial treatment effect in patients with proximal anterior circulation stroke. Stroke 2016; 47: 768–76. [DOI] [PubMed] [Google Scholar]
- 10.Fransen PS, Beumer D, Berkhemer OA, et al. MR CLEAN, a multicenter randomized clinical trial of endovascular treatment for acute ischemic stroke in the Netherlands: study protocol for a randomized controlled trial. Trials 2014; 15: 343. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Boers AM, Marquering HA, Jochem JJ, et al. Automated cerebral infarct volume measurement in follow-up noncontrast ct scans of patients with acute ischemic stroke. Am J Neuroradiol 2013; 34: 1522–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Nakano S, Iseda T, Kawano H, et al. Correlation of early CT signs in the deep middle cerebral artery territories with angiographically confirmed site of arterial occlusion. Am J Neuroradiol 2001; 22: 654–659. [PMC free article] [PubMed] [Google Scholar]
- 13.Elijovich L, Goyal N, Mainali S, et al. CTA collateral score predicts infarct volume and clinical outcome after endovascular therapy for acute ischemic stroke: a retrospective chart review. J Neurointerv Surg 2015, pp. 1–4. [DOI] [PubMed] [Google Scholar]
- 14.Kothari RU, Brott T, Broderick JP, et al. The ABCs of measuring intracerebral hemorrhage volumes. Stroke A J Cereb Circ 1996; 27: 1304–1305. [DOI] [PubMed] [Google Scholar]
- 15.Nambiar V, Sohn SI, Almekhlafi MA, et al. CTA collateral status and response to recanalization in patients with acute ischemic stroke. Am J Neuroradiol 2014; 35: 884–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Menon BK, Qazi E, Nambiar V, et al. Differential effect of baseline computed tomographic angiography collaterals on clinical outcome in patients enrolled in the interventional management of stroke III trial. Stroke 2015; 46: 1239–1244. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Becker H, Desch H, Hacker H, et al. CT fogging effect with ischemic cerebral infarcts. Neuroradiology 1979; 18: 185–192. [DOI] [PubMed] [Google Scholar]
- 18.Simard JM, Kent TA, Chen M, et al. Brain oedema in focal ischaemia: molecular pathophysiology and theoretical implications. Lancet Neurol 2007; 6: 258–68. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Baird AE, Benfield A, Schlaug G, et al. Enlargement of human cerebral ischemic lesion volumes measured by diffusion-weighted magnetic resonance imaging. Ann Neurol 1997; 41: 581–589. [DOI] [PubMed] [Google Scholar]
- 20.Karonen JO, Vanninen RL, Liu Y, et al. Combined diffusion and perfusion MRI with correlation to single-photon emission ct in acute ischemic stroke: ischemic penumbra predicts infarct growth. Stroke 1999; 30: 1583–1590. [DOI] [PubMed] [Google Scholar]
- 21.Yoo AJ, Chaudhry ZA, Nogueira RG, et al. Infarct volume is a pivotal biomarker after intra-arterial stroke therapy. Stroke 2012; 43: 1323–30. [DOI] [PubMed] [Google Scholar]
- 22.McVerry F, Liebeskind DS, Muir KW. Systematic review of methods for assessing leptomeningeal collateral flow. Am J Neuroradiol 2012; 33: 576–582. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Tan IYL, Demchuk AM, Hopyan J, et al. CT Angiography clot burden score and collateral score: correlation with clinical and radiologic outcomes in acute middle cerebral artery infarct. Am J Neuroradiol 2009; 30: 525–531. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Latchaw RE, Alberts MJ, Lev MH, et al. Recommendations for imaging of acute ischemic stroke: a scientific statement from the American Heart Association. Stroke 2009; 40: 3646–3678. [DOI] [PubMed] [Google Scholar]
- 25.Smit EJ, Vonken E-J, Meijer E Al. Timing-invariant CT angiography derived from CT perfusion imaging in acute stroke: a diagnostic performance study. Am J Neuroradiol 2015; 36: 1834–1838. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Smit EJ, Vonken E, van Seeters T, et al. Timing-invariant imaging of collateral vessels in acute ischemic stroke. Stroke 2013; 44: 2194–2199. [DOI] [PubMed] [Google Scholar]
- 27.Menon BK, d'Esterre CD, Qazi EM, et al. Multiphase CT angiography: a new tool for the imaging triage of patients with acute ischemic stroke. Radiology 2015; 275: 510–520. [DOI] [PubMed] [Google Scholar]
- 28.Frölich AMJ, Wolff SL, Psychogios MN, et al. Time-resolved assessment of collateral flow using 4D CT angiography in large-vessel occlusion stroke. Eur Radiol 2014; 24: 390–396. [DOI] [PubMed] [Google Scholar]
- 29.Ernst M, Forkert ND, Brehmer L, et al. Prediction of infarction and reperfusion in stroke by flow- and volume-weighted collateral signal in MR angiography. Am J Neuroradiol 2015; 36: 275–282. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.