Review Factors affecting clinical outcome in large-vessel occlusive ischemic strokes Michelle P. Lin1, Georgios Tsivgoulis2,3, Andrei V. Alexandrov4, and Jason J. Chang4,* Clinical outcome after large-vessel occlusive strokes depends on admitting clinical condition, successful recanalization, and robust collateral circulation. However, predicting successful recanalization and quantifying collateral status in the acute setting remain elusive. Successful recanalization has many predictive factors. Strong evidence supports increasing clot length being associated with poor recanalization. Current imaging techniques completed in the acute setting suggest that clot length can be estimated with a clot burden score. In vitro evidence suggests that clots with more red blood cells and less thrombin lyse more easily after systemic fibrinolysis. Clinical correlations with clot composition have been mixed, although one study suggested that clot composition could be predicted with computed tomography and correlate with successful recanalization. Finally, overwhelming proof shows that robust collateral circulation correlates with improved clinical outcome. Imaging modalities in the acute setting remain promising, with studies suggesting that collaterals can be quantified with computed tomography angiography and perfusion studies. Patients with large-vessel occlusive strokes have variable clinical responses to fibrinolysis and thrombectomy. Independent predictive variables that can possibly alter clinical outcome appear to be successful recanalization and robust collateral circulation. Future studies that allow for better prediction of successful recanalization and quantification of collateral status may help clinical decision-making when evaluating large-vessel occlusions. Key words: clot composition, clot length, collateral, ischemic stroke, outcome, recanalization
Introduction Studies consistently show that independent predictors in clinical outcome remain the patient’s admission clinical condition as
evaluated by baseline NIH Stroke Scale (NIHSS) scores (1–5) and ischemic lesion volume (6,7). Emphasis on lesion volume suggested that stroke mechanism, particularly lacunar strokes with their smaller ischemic volumes, could be an independent predictor for favorable clinical outcome (8–10). However, predicting clinical outcome with stroke mechanism, particularly in largevessel occlusions (LVO), remains elusive. As thrombus origin site (i.e., vein, artery, or heart) is thought to play a role in clot composition, cardioembolic and paradoxical strokes with their greater fibrin content and red cells were thought to be more susceptible to fibrinolysis than heterogeneous, chronic thrombotic arterial plaques formed in artery-to-artery strokes (11,12). However, studies continue to show that stroke mechanism does not influence recanalization or clinical outcome after intravenous rt-PA (IVtPA) (1,13–16). Use of IVtPA and endovascular thrombectomy have greatly improved recanalization with rates of 26·8–56·3% (1,17,18) and 71–83·6% (18,19), respectively. This has translated into improved clinical outcomes. However, as shown in ECASS I, as many as 37·5% of patients clinically deteriorate even after receiving IVtPA (20), and the presence of LVO is associated with higher mortality and poorer clinical outcome (21). Evolving observations and advances in brain and vasculature imaging have prompted interest in evaluating independent predictors of recanalization and clinical outcome beyond admission NIHSS scores and ischemic lesion volume. This review attempts to systematically classify and evaluate these factors contributing to recanalization and improved clinical outcome to facilitate clinical decision-making and predictive acumen when dealing with LVO.
Recanalization and improved clinical outcome Correspondence: Jason Chang*, Department of Neurology, University of Tennessee Health Science Center, 1325 Eastmoreland, Suite 460, Memphis, TN 38104, USA. E-mail: [email protected] 1 Department of Neurology, University of Southern California, Los Angeles, CA, USA 2 Department of Neurology, University of Athens School of Medicine, Athens, Greece 3 International Clinical Research Center, St. Anne’s University, Brno, Czech Republic 4 Department of Neurology, University of Tennessee Health Science Center, Memphis, TN, USA Received: 22 September 2014; Accepted: 22 September 2014; Published online 3 December 2014 Conflict of interest: None declared. Funding: M. P. L., A. V. A., and J. J. C. have no disclosures. G. T. is supported by a European Regional Development Fund – Project FNUSAICRC (No. CZ.l.05/1.1.00/02.0123). DOI: 10.1111/ijs.12406 © 2014 World Stroke Organization
The approval of IVtPA in 1996 allowed for the first Food and Drug Administration (FDA)-approved treatment in acute ischemic stroke (AIS) through fibrinolysis (14). Randomizedcontrolled trials have shown benefit with alteplase administered systemically up to 4·5 h after onset (22) and prourokinase intraarterially up to six-hours after onset (23). Recanalization is clearly associated with improved clinical outcome. Zangerle et al. reported that 58·3% of patients with complete recanalization had favorable 90-day Rankin scores compared with 10·0% for patients with partial recanalization and 5·6% for patients without recanalization (P < 0·001) (17). In a meta-analysis, Rha and Saver showed that recanalization significantly improved 90-day clinical outcome (OR 4·43, 95% CI 3·32– 5·91) and mortality (OR 0·24, 95% CI 0·7–17·4) (18). Table 1 summarizes the available data in four important studies (24–27) regarding early recanalization (within two-hours following rt-PA bolus) in patients with anterior LVO treated with IVtPA. Complete recanalization rates across different studies ranged from Vol 10, June 2015, 479–484
479
Review
M. P. Lin et al.
Table 1 Distribution of early recanalization rates (≤ two-hours following IVtPA bolus) according to vessel location in different studies in patients with anterior large-vessel occlusions treated with systemic thrombolysis Variable
Saqqur et al. (24)
Bhatia et al. (25)
Kimura et al. (26)
Medonca et al. (27)
Imaging modality* Time from tPA-bolus Artery M1/MCA M2/MCA Terminal ICA Tandem ICA/MCA All
TCD Two-hours N CR 163 49 (30%) 113 50 (44%) 17 1 (6%) 22 6 (27%) 315 106 (34%)
TCD/DSA One to two-hours N CR 65 21 (32%) 13 4 (33%) 23 1 (4%) NA NA 101 26 (26%)
MRA Two-hours N CR 58 8 (14%) 37 6 (16%) 37 2 (5·4%) NA NA 132 16 (12%)
TCD One-hour N 251 194 61 NA 506
CR 26 (10%) 43 (22%) 3 (5%) NA 72 (14%)
*For assessment of recanalization. CR, complete recanalization; DSA, digital subtraction angiography; ICA, internal carotid artery; IVtPA, intravenous rt-PA; M1, proximal; M2, distal; MCA, middle cerebral artery; MRA, magnetic resonance angiography; NA, not available; TCD, transcranial Doppler.
Table 2 Pooled analysis of all studies evaluating early recanalization rates (≤ two-hours following IVtPA bolus) according to vessel location in patients with anterior large vessel occlusions treated with systemic thrombolysis
Vessel
Number of patients, n
Complete recanalization, % (95% CI)
M1 (MCA) M2 (MCA) Terminal ICA Tandem ICA/MCA
630 390 180 32
22% (19–25%), [n = 128] 29% (25–34%), [n = 114] 8% (5–14%), [n = 15] 19% (9–36%)*, [n = 6]
*95% CI were computed with adjusted Wald method. ICA, internal carotid artery; IVtPA, intravenous rt-PA; M1, proximal; M2, distal; MCA, middle cerebral artery.
12% to 34%. After pooling available data, Table 2 shows complete recanalization rates within two-hours after IVtPA according to vessel location: proximal (M1)/middle cerebral artery (MCA), 22%; distal (M2)/MCA, 29%; terminal internal carotid artery (ICA), 8%; and tandem ICA/MCA, 19%. With the failure of the Interventional Management of Stroke Trial 3 to show clinical outcome differences between endovascular retrieval devices and IVtPA (28), no evidence currently shows superiority of endovascular clot retrieval over IVtPA (29). However, endovascular device technology continues to evolve with newer devices suggesting greater recanalization success. Two factors have been shown to impact recanalization and stratify clinical outcome: (1) clot length and (2) clot composition.
Clot length Evidence regarding clot length and recanalization is somewhat contradictory, while clot density has largely been shown to be irrelevant in recanalization. Riedel et al. evaluated 138 patients presenting with MCA strokes, and determined that shorter clots (length < 5 mm) were more likely to lyse and that 6 points) emerged as an independent predictor of recanalization and favorable functional outcome at 90 days (36) (Table 3). A modified CBS was evaluated and shown via univariate analysis to be significantly associated with favorable 90-day clinical outcome (P < 0·001), but only showed a trend (P = 0·109) when evaluated in multivariable regression analysis (37). © 2014 World Stroke Organization
Review
M. P. Lin et al.
Table 3 Predictors of clinical outcome after large-vessel occlusions and supporting evidence
Predictors Recanalization Longer clot length Smaller clot burden score Thrombin-rich clot Hemoglobin-rich clot Hyperdense MCA sign Cardioembolic stroke* Paradoxical stroke* Collateral circulation
Favorable recanalization
Unfavorable recanalization
Unclear relationship to recanalization
Favorable outcome
n/a
n/a X X X
n/a
X
X
n/a n/a
Unfavorable outcome
X X n/a n/a
X X X X
Unclear relationship to outcome
n/a n/a X X
X X
*Compared with artery-to-artery occlusive stroke mechanisms. MCA, middle cerebral artery; n/a, not applicable.
Clot composition Several studies have attempted to classify clot composition and correlate this with response to IVtPA and recanalization and clinical outcome. In vitro and in vivo models Fibrinolysis can be limited due to excessive cross-linking within matured, calcified fibrin plaque. Clots retrieved from patients with AIS have variable proportions of fibrin and intermixed blood cells (38–40). Animal models of embolic stroke and femoral artery thrombosis have shown that clots composed of cross-linked fibrin and platelets displayed a relative resistance against fibrinolysis, whereas erythrocyte-rich clots showed better recanalization and response to fibrinolytics (41). Thrombin presence also affects recanalization with IVtPA and explains why fibrin-rich clots may still have limited success with IVtPA. Niessen et al. evaluated spontaneously forming fibrin clots and fibrin clots mixed with thrombin. These two types of clots were used to occlude rat MCAs. Significantly higher perfusion, which assumed better recanalization, was noted in the spontaneously forming clots (P < 0·05) after IVtPA. Thrombin may have hindered IVtPA by producing a denser and stiffer fibrin meshwork (41). Radiological surrogate for clot composition Numerous studies evaluating imaging markers of clot composition have assumed that HMCAS can represent a radiological correlate for a fibrin-rich thrombus. Kirchhof et al. conducted in vitro experiments in which they mixed platelet-enriched plasma and whole blood to produce samples with varying hematocrit levels ranging from 0 to 0·35. These artificially created clots were then evaluated by CTH at various time-points from 6 to 144 h. ‘Red thrombus’ with the highest hematocrit content had higher Hounsfield unit (HU) around 70. ‘White thrombus’ with the highest platelet content had a HU around 20 (42). Clinical studies utilizing hyperdense MCA sign Early studies suggested that HMCAS could be used as a marker for higher probability of recanalization and improved neurologic recovery after IVtPA (43). More recent studies have attempted to characterize clot composition as defined by HMCAS and have shown correlations with recanalization, but not with clinical outcome. © 2014 World Stroke Organization
Moftakhar et al. retrospectively evaluated 90 patients treated with IVtPA, intra-arterial tPA (IAtPA), and/or mechanical thrombectomy, and found that HU on initial CTH helped predict recanalization. The authors found strong correlations (IVtPA r = 0·69, IAtPA r = 0·72, P < 0·0001) between thrombus HU on pretreatment CTH and the Thrombolysis in Cerebral Infarction (TICI) score after IVtPA or IAtPA. Thrombectomy demonstrated a similarly significant correlation between HU and post-TICI score. Significantly higher HU was found in patients with TICI scores ≥ 2. For patients treated with IVtPA, mean corrected HU for patients with TICI ≥ 2 was 1·58 compared with 1·39 for patients with TICI < 2 (P = 0·01). Higher baseline HU was also noted in patients who had higher TICI scores (>2) after IAtPA or thrombectomy (19). Liebeskind et al. correlated HMCAS and blooming artifact (BA) on MRI gradient echo (GRE) sequence with histopathology from retrieved clots and clinical outcomes in 50 consecutive patients undergoing mechanical thrombectomy. They found that RBC-dominant, and to a lesser extent mixed clots, rather than fibrin-dominant clots were more commonly seen with HMCAS (100% vs. 67% vs. 20%, respectively. P = 0·016) and BA (100% vs. 63% vs. 25%, respectively. P = 0·002). However, presence of HMCAS or BA did not predict stroke severity or clinical outcome (40). More recently, Topcuoglu et al. quantified HMCAS density by HU, but found no relationship with 90-day clinical outcome (37).
Collateral circulation Three principal anatomical features allow for important collateral circuits (44,45): (1) large artery communications between the extracranial and intracranial circulations (46), (2) completeness of Circle of Willis (47,48), and (3) leptomeningeal anastomoses providing cortical surface perfusion (49,50). Robust collaterals are clearly linked to improved clinical outcome, reduced infarct volume (40,50,51), and decreased hemorrhagic transformation following thrombolysis (52). In addition, some evidence shows robust collateral circulation to be predictive for recanalization (48). Several imaging modalities – conventional angiography, noninvasive angiography, and perfusion studies – have been used to Vol 10, June 2015, 479–484
481
Review quantify and characterize the robustness of collateral circulation. Each has benefits and limitations. While there is no consensus on the utility of noninvasive imaging for acute stroke algorithm, and the concept of penumbral image-guided thrombolytic therapy remains to be validated in large clinical trials, compelling data demonstrate the pivotal role of collateral flow in predicting clinical outcomes in LVO. Conventional cerebral angiography: gold standard Characterization of collateral circulation with conventional angiography supports the correlation between robustness of collaterals, recanalization, and clinical outcome. Bang et al. conducted a prospective cohort study with 222 consecutive patients to evaluate the relationship between baseline angiographic collateral grades and recanalization rates after endovascular therapy. Complete revascularization occurred in 14%, 25%, and 42% in patients with poor, good, and excellent collaterals, respectively (P = 0·001) (51). Similarly in a retrospective, consecutive case series of 111 patients, Kucinski et al. found after IVtPA that significant univariate predictors of favorable outcome were occlusion location (P < 0·01), recanalization (P < 0·01), and collaterals (P < 0·01). After adjusting for covariates including age, gender, early ischemic signs, occlusion location, and recanalization, the model showed a significant relationship between good collaterals and favorable clinical outcome (OR 5·9, 95% CI 1·3 to 26·7, P = 0·02) (50). Conventional angiography, considered the gold standard for evaluation of collateral blood flow, has the most robust data (53) and high inter/intraobserver agreement (52–54). However, the plethora of data have also resulted in numerous different types of grading scales. MEDLINE and Embase searches by McVerry et al. resulted in 41 different criteria for grading collaterals (53). In addition, its applicability is limited by its invasive nature, relatively long acquisition and procedural times, and low accessibility for general use (54,55). Noninvasive angiography and perfusion studies may be more practical for evaluating collateral flow in an acute setting. Noninvasive angiography Several studies have established CTA’s role in predicting clinical outcome through grading of collateral status. Menon et al. described a novel method of scoring leptomeningeal collaterals (rLMC score) on CTA based on the extent of opacification of pial and lenticulostriate arteries distal to an M1/MCA ± intracranial ICA occlusion in six Alberta Stroke Program Early CT score (ASPECTS) regions plus ACA and basal ganglia regions. They used this classification scheme to quantify collateral circulation in a retrospective study of 138 patients with acute proximal M1/MCA and intracranial ICA occlusions. In multivariable analysis, the rLMC score remained an independent predictor for favorable (≤2) 90-day Rankin score (OR 16·7, 95% CI 2·9–97·4%) (56). Similarly, Lima et al. conducted a prospective cohort of 196 patients and found high leptomeningeal collateral grade to be an independent predictor of good clinical outcome (OR 1·93, 95% CI 1·06–3·34; P = 0·03) (57). And finally, the CT-based collated score (CS), graded from 0 to 3 (with 0 indicating absent collaterals and 3 indicating 100% collateral supply of the occluded MCA terri-
482
Vol 10, June 2015, 479–484
M. P. Lin et al. tory), predicted final infarct size and recanalization status in a series of patients with LVO treated with IVtPA (35). CT-based techniques provide a more accessible and expedient assessment for grading collateral flow in the acute setting than conventional angiography. Particularly when combined with computed tomography perfusion (CTP), CTA can rapidly generate quantitative and qualitative parameters that enable discrimination between normal, penumbra, and infarcted core (45). In addition, reliability has been shown to be moderate to excellent. However, CTA lacks the dynamic, temporal information which could result in overestimation of collateral flow or failure to reveal complete occlusions that are filled distally by retrograde collateral perfusion (53). Little data exist comparing magnetic resonance angiography (MRA) with CTA in evaluating collateral status. One case series had 10 consecutive patients with intracranial occlusions (seven MCA, two ACA, one bilateral vertebral arteries) who underwent both CTA and MRA with conventional angiography serving as a gold standard. In all patients, occlusive disease was well visualized in CTA, but not MRA (58). Perfusion imaging Miteff et al. evaluated a case series of 92 patients with proximal intracranial occlusion with symptom onset within six-hours. They uniquely evaluated CTP mismatch ratio and hypothesized that collateral circulation status would influence progression of tissue from penumbral to ischemic. Collateral circulation status was dichotomized into ‘good’ vs. ‘bad’. Several mismatch ratios defined as [volume of mean transit time tissue/volume of cerebral blood volume tissue] were compared. They found that lower mismatch ratios – with a cutoff value of
Introduction Studies consistently show that independent predictors in clinical outcome remain the patient’s admission clinical condition as
evaluated by baseline NIH Stroke Scale (NIHSS) scores (1–5) and ischemic lesion volume (6,7). Emphasis on lesion volume suggested that stroke mechanism, particularly lacunar strokes with their smaller ischemic volumes, could be an independent predictor for favorable clinical outcome (8–10). However, predicting clinical outcome with stroke mechanism, particularly in largevessel occlusions (LVO), remains elusive. As thrombus origin site (i.e., vein, artery, or heart) is thought to play a role in clot composition, cardioembolic and paradoxical strokes with their greater fibrin content and red cells were thought to be more susceptible to fibrinolysis than heterogeneous, chronic thrombotic arterial plaques formed in artery-to-artery strokes (11,12). However, studies continue to show that stroke mechanism does not influence recanalization or clinical outcome after intravenous rt-PA (IVtPA) (1,13–16). Use of IVtPA and endovascular thrombectomy have greatly improved recanalization with rates of 26·8–56·3% (1,17,18) and 71–83·6% (18,19), respectively. This has translated into improved clinical outcomes. However, as shown in ECASS I, as many as 37·5% of patients clinically deteriorate even after receiving IVtPA (20), and the presence of LVO is associated with higher mortality and poorer clinical outcome (21). Evolving observations and advances in brain and vasculature imaging have prompted interest in evaluating independent predictors of recanalization and clinical outcome beyond admission NIHSS scores and ischemic lesion volume. This review attempts to systematically classify and evaluate these factors contributing to recanalization and improved clinical outcome to facilitate clinical decision-making and predictive acumen when dealing with LVO.
Recanalization and improved clinical outcome Correspondence: Jason Chang*, Department of Neurology, University of Tennessee Health Science Center, 1325 Eastmoreland, Suite 460, Memphis, TN 38104, USA. E-mail: [email protected] 1 Department of Neurology, University of Southern California, Los Angeles, CA, USA 2 Department of Neurology, University of Athens School of Medicine, Athens, Greece 3 International Clinical Research Center, St. Anne’s University, Brno, Czech Republic 4 Department of Neurology, University of Tennessee Health Science Center, Memphis, TN, USA Received: 22 September 2014; Accepted: 22 September 2014; Published online 3 December 2014 Conflict of interest: None declared. Funding: M. P. L., A. V. A., and J. J. C. have no disclosures. G. T. is supported by a European Regional Development Fund – Project FNUSAICRC (No. CZ.l.05/1.1.00/02.0123). DOI: 10.1111/ijs.12406 © 2014 World Stroke Organization
The approval of IVtPA in 1996 allowed for the first Food and Drug Administration (FDA)-approved treatment in acute ischemic stroke (AIS) through fibrinolysis (14). Randomizedcontrolled trials have shown benefit with alteplase administered systemically up to 4·5 h after onset (22) and prourokinase intraarterially up to six-hours after onset (23). Recanalization is clearly associated with improved clinical outcome. Zangerle et al. reported that 58·3% of patients with complete recanalization had favorable 90-day Rankin scores compared with 10·0% for patients with partial recanalization and 5·6% for patients without recanalization (P < 0·001) (17). In a meta-analysis, Rha and Saver showed that recanalization significantly improved 90-day clinical outcome (OR 4·43, 95% CI 3·32– 5·91) and mortality (OR 0·24, 95% CI 0·7–17·4) (18). Table 1 summarizes the available data in four important studies (24–27) regarding early recanalization (within two-hours following rt-PA bolus) in patients with anterior LVO treated with IVtPA. Complete recanalization rates across different studies ranged from Vol 10, June 2015, 479–484
479
Review
M. P. Lin et al.
Table 1 Distribution of early recanalization rates (≤ two-hours following IVtPA bolus) according to vessel location in different studies in patients with anterior large-vessel occlusions treated with systemic thrombolysis Variable
Saqqur et al. (24)
Bhatia et al. (25)
Kimura et al. (26)
Medonca et al. (27)
Imaging modality* Time from tPA-bolus Artery M1/MCA M2/MCA Terminal ICA Tandem ICA/MCA All
TCD Two-hours N CR 163 49 (30%) 113 50 (44%) 17 1 (6%) 22 6 (27%) 315 106 (34%)
TCD/DSA One to two-hours N CR 65 21 (32%) 13 4 (33%) 23 1 (4%) NA NA 101 26 (26%)
MRA Two-hours N CR 58 8 (14%) 37 6 (16%) 37 2 (5·4%) NA NA 132 16 (12%)
TCD One-hour N 251 194 61 NA 506
CR 26 (10%) 43 (22%) 3 (5%) NA 72 (14%)
*For assessment of recanalization. CR, complete recanalization; DSA, digital subtraction angiography; ICA, internal carotid artery; IVtPA, intravenous rt-PA; M1, proximal; M2, distal; MCA, middle cerebral artery; MRA, magnetic resonance angiography; NA, not available; TCD, transcranial Doppler.
Table 2 Pooled analysis of all studies evaluating early recanalization rates (≤ two-hours following IVtPA bolus) according to vessel location in patients with anterior large vessel occlusions treated with systemic thrombolysis
Vessel
Number of patients, n
Complete recanalization, % (95% CI)
M1 (MCA) M2 (MCA) Terminal ICA Tandem ICA/MCA
630 390 180 32
22% (19–25%), [n = 128] 29% (25–34%), [n = 114] 8% (5–14%), [n = 15] 19% (9–36%)*, [n = 6]
*95% CI were computed with adjusted Wald method. ICA, internal carotid artery; IVtPA, intravenous rt-PA; M1, proximal; M2, distal; MCA, middle cerebral artery.
12% to 34%. After pooling available data, Table 2 shows complete recanalization rates within two-hours after IVtPA according to vessel location: proximal (M1)/middle cerebral artery (MCA), 22%; distal (M2)/MCA, 29%; terminal internal carotid artery (ICA), 8%; and tandem ICA/MCA, 19%. With the failure of the Interventional Management of Stroke Trial 3 to show clinical outcome differences between endovascular retrieval devices and IVtPA (28), no evidence currently shows superiority of endovascular clot retrieval over IVtPA (29). However, endovascular device technology continues to evolve with newer devices suggesting greater recanalization success. Two factors have been shown to impact recanalization and stratify clinical outcome: (1) clot length and (2) clot composition.
Clot length Evidence regarding clot length and recanalization is somewhat contradictory, while clot density has largely been shown to be irrelevant in recanalization. Riedel et al. evaluated 138 patients presenting with MCA strokes, and determined that shorter clots (length < 5 mm) were more likely to lyse and that 6 points) emerged as an independent predictor of recanalization and favorable functional outcome at 90 days (36) (Table 3). A modified CBS was evaluated and shown via univariate analysis to be significantly associated with favorable 90-day clinical outcome (P < 0·001), but only showed a trend (P = 0·109) when evaluated in multivariable regression analysis (37). © 2014 World Stroke Organization
Review
M. P. Lin et al.
Table 3 Predictors of clinical outcome after large-vessel occlusions and supporting evidence
Predictors Recanalization Longer clot length Smaller clot burden score Thrombin-rich clot Hemoglobin-rich clot Hyperdense MCA sign Cardioembolic stroke* Paradoxical stroke* Collateral circulation
Favorable recanalization
Unfavorable recanalization
Unclear relationship to recanalization
Favorable outcome
n/a
n/a X X X
n/a
X
X
n/a n/a
Unfavorable outcome
X X n/a n/a
X X X X
Unclear relationship to outcome
n/a n/a X X
X X
*Compared with artery-to-artery occlusive stroke mechanisms. MCA, middle cerebral artery; n/a, not applicable.
Clot composition Several studies have attempted to classify clot composition and correlate this with response to IVtPA and recanalization and clinical outcome. In vitro and in vivo models Fibrinolysis can be limited due to excessive cross-linking within matured, calcified fibrin plaque. Clots retrieved from patients with AIS have variable proportions of fibrin and intermixed blood cells (38–40). Animal models of embolic stroke and femoral artery thrombosis have shown that clots composed of cross-linked fibrin and platelets displayed a relative resistance against fibrinolysis, whereas erythrocyte-rich clots showed better recanalization and response to fibrinolytics (41). Thrombin presence also affects recanalization with IVtPA and explains why fibrin-rich clots may still have limited success with IVtPA. Niessen et al. evaluated spontaneously forming fibrin clots and fibrin clots mixed with thrombin. These two types of clots were used to occlude rat MCAs. Significantly higher perfusion, which assumed better recanalization, was noted in the spontaneously forming clots (P < 0·05) after IVtPA. Thrombin may have hindered IVtPA by producing a denser and stiffer fibrin meshwork (41). Radiological surrogate for clot composition Numerous studies evaluating imaging markers of clot composition have assumed that HMCAS can represent a radiological correlate for a fibrin-rich thrombus. Kirchhof et al. conducted in vitro experiments in which they mixed platelet-enriched plasma and whole blood to produce samples with varying hematocrit levels ranging from 0 to 0·35. These artificially created clots were then evaluated by CTH at various time-points from 6 to 144 h. ‘Red thrombus’ with the highest hematocrit content had higher Hounsfield unit (HU) around 70. ‘White thrombus’ with the highest platelet content had a HU around 20 (42). Clinical studies utilizing hyperdense MCA sign Early studies suggested that HMCAS could be used as a marker for higher probability of recanalization and improved neurologic recovery after IVtPA (43). More recent studies have attempted to characterize clot composition as defined by HMCAS and have shown correlations with recanalization, but not with clinical outcome. © 2014 World Stroke Organization
Moftakhar et al. retrospectively evaluated 90 patients treated with IVtPA, intra-arterial tPA (IAtPA), and/or mechanical thrombectomy, and found that HU on initial CTH helped predict recanalization. The authors found strong correlations (IVtPA r = 0·69, IAtPA r = 0·72, P < 0·0001) between thrombus HU on pretreatment CTH and the Thrombolysis in Cerebral Infarction (TICI) score after IVtPA or IAtPA. Thrombectomy demonstrated a similarly significant correlation between HU and post-TICI score. Significantly higher HU was found in patients with TICI scores ≥ 2. For patients treated with IVtPA, mean corrected HU for patients with TICI ≥ 2 was 1·58 compared with 1·39 for patients with TICI < 2 (P = 0·01). Higher baseline HU was also noted in patients who had higher TICI scores (>2) after IAtPA or thrombectomy (19). Liebeskind et al. correlated HMCAS and blooming artifact (BA) on MRI gradient echo (GRE) sequence with histopathology from retrieved clots and clinical outcomes in 50 consecutive patients undergoing mechanical thrombectomy. They found that RBC-dominant, and to a lesser extent mixed clots, rather than fibrin-dominant clots were more commonly seen with HMCAS (100% vs. 67% vs. 20%, respectively. P = 0·016) and BA (100% vs. 63% vs. 25%, respectively. P = 0·002). However, presence of HMCAS or BA did not predict stroke severity or clinical outcome (40). More recently, Topcuoglu et al. quantified HMCAS density by HU, but found no relationship with 90-day clinical outcome (37).
Collateral circulation Three principal anatomical features allow for important collateral circuits (44,45): (1) large artery communications between the extracranial and intracranial circulations (46), (2) completeness of Circle of Willis (47,48), and (3) leptomeningeal anastomoses providing cortical surface perfusion (49,50). Robust collaterals are clearly linked to improved clinical outcome, reduced infarct volume (40,50,51), and decreased hemorrhagic transformation following thrombolysis (52). In addition, some evidence shows robust collateral circulation to be predictive for recanalization (48). Several imaging modalities – conventional angiography, noninvasive angiography, and perfusion studies – have been used to Vol 10, June 2015, 479–484
481
Review quantify and characterize the robustness of collateral circulation. Each has benefits and limitations. While there is no consensus on the utility of noninvasive imaging for acute stroke algorithm, and the concept of penumbral image-guided thrombolytic therapy remains to be validated in large clinical trials, compelling data demonstrate the pivotal role of collateral flow in predicting clinical outcomes in LVO. Conventional cerebral angiography: gold standard Characterization of collateral circulation with conventional angiography supports the correlation between robustness of collaterals, recanalization, and clinical outcome. Bang et al. conducted a prospective cohort study with 222 consecutive patients to evaluate the relationship between baseline angiographic collateral grades and recanalization rates after endovascular therapy. Complete revascularization occurred in 14%, 25%, and 42% in patients with poor, good, and excellent collaterals, respectively (P = 0·001) (51). Similarly in a retrospective, consecutive case series of 111 patients, Kucinski et al. found after IVtPA that significant univariate predictors of favorable outcome were occlusion location (P < 0·01), recanalization (P < 0·01), and collaterals (P < 0·01). After adjusting for covariates including age, gender, early ischemic signs, occlusion location, and recanalization, the model showed a significant relationship between good collaterals and favorable clinical outcome (OR 5·9, 95% CI 1·3 to 26·7, P = 0·02) (50). Conventional angiography, considered the gold standard for evaluation of collateral blood flow, has the most robust data (53) and high inter/intraobserver agreement (52–54). However, the plethora of data have also resulted in numerous different types of grading scales. MEDLINE and Embase searches by McVerry et al. resulted in 41 different criteria for grading collaterals (53). In addition, its applicability is limited by its invasive nature, relatively long acquisition and procedural times, and low accessibility for general use (54,55). Noninvasive angiography and perfusion studies may be more practical for evaluating collateral flow in an acute setting. Noninvasive angiography Several studies have established CTA’s role in predicting clinical outcome through grading of collateral status. Menon et al. described a novel method of scoring leptomeningeal collaterals (rLMC score) on CTA based on the extent of opacification of pial and lenticulostriate arteries distal to an M1/MCA ± intracranial ICA occlusion in six Alberta Stroke Program Early CT score (ASPECTS) regions plus ACA and basal ganglia regions. They used this classification scheme to quantify collateral circulation in a retrospective study of 138 patients with acute proximal M1/MCA and intracranial ICA occlusions. In multivariable analysis, the rLMC score remained an independent predictor for favorable (≤2) 90-day Rankin score (OR 16·7, 95% CI 2·9–97·4%) (56). Similarly, Lima et al. conducted a prospective cohort of 196 patients and found high leptomeningeal collateral grade to be an independent predictor of good clinical outcome (OR 1·93, 95% CI 1·06–3·34; P = 0·03) (57). And finally, the CT-based collated score (CS), graded from 0 to 3 (with 0 indicating absent collaterals and 3 indicating 100% collateral supply of the occluded MCA terri-
482
Vol 10, June 2015, 479–484
M. P. Lin et al. tory), predicted final infarct size and recanalization status in a series of patients with LVO treated with IVtPA (35). CT-based techniques provide a more accessible and expedient assessment for grading collateral flow in the acute setting than conventional angiography. Particularly when combined with computed tomography perfusion (CTP), CTA can rapidly generate quantitative and qualitative parameters that enable discrimination between normal, penumbra, and infarcted core (45). In addition, reliability has been shown to be moderate to excellent. However, CTA lacks the dynamic, temporal information which could result in overestimation of collateral flow or failure to reveal complete occlusions that are filled distally by retrograde collateral perfusion (53). Little data exist comparing magnetic resonance angiography (MRA) with CTA in evaluating collateral status. One case series had 10 consecutive patients with intracranial occlusions (seven MCA, two ACA, one bilateral vertebral arteries) who underwent both CTA and MRA with conventional angiography serving as a gold standard. In all patients, occlusive disease was well visualized in CTA, but not MRA (58). Perfusion imaging Miteff et al. evaluated a case series of 92 patients with proximal intracranial occlusion with symptom onset within six-hours. They uniquely evaluated CTP mismatch ratio and hypothesized that collateral circulation status would influence progression of tissue from penumbral to ischemic. Collateral circulation status was dichotomized into ‘good’ vs. ‘bad’. Several mismatch ratios defined as [volume of mean transit time tissue/volume of cerebral blood volume tissue] were compared. They found that lower mismatch ratios – with a cutoff value of
