Clot Characteristics on Computed Tomography and Response to Thrombolysis in Acute Middle Cerebral Artery Stroke Mehmet A. Topcuoglu, MD,* Ethem Murat Arsava, MD,* and Erhan Akpinar, MD†

Background: Clinical and computer tomography angiography (CTA) correlates of hyperdense middle cerebral artery sign (HMCAS) and dot sign were revisited in patients treated for acute MCA stroke. Temporal evolution of these signs over 24 hours was assessed quantitatively by density (Hounsfield unit [HU]) measurements. Methods: Maximum pixel-sized HUs throughout proximal MCA and its insular fissure branches were determined in 131 patients with acute MCA stroke treated by intravenous thrombolysis and/or interventional thrombolysis/thrombectomy; 14 patients treated for vertebrobasilar stroke (VBS) and 42 nonstroke control subjects. Utility of visually determined HMCAS and dot sign, absolute HU of proximal and distal MCA, side-to-side HU ratio and difference, and hyperdense MCA burden score for the prediction of early dramatic recovery (EDR) and third-month favorable prognosis were evaluated. The clinical value of the changes in vessel hyperdensity over 24 hours was identified in subjects who received intravenous thrombolysis (99 MCA stoke and 11 VBS). A multivariate model with adjustment for age, baseline stroke severity (National Institutes of Health Stroke Scale [NIHSS]), and CTAbased modified clot burden score (mCBS) was used to determine independent predictors of short- and long-term clinical outcome. Results: The presence of HMCAS and dot sign, their density indices (maximum HU, ipsilateral-to-contralateral HU ratio, and difference), and changes in quantitative attenuation over 24 hours were not significantly associated with EDR and favorable third-month outcome in the multiple regression models, whereas NIHSS and mCBS were found to be significant independent ‘‘negative predictors’’ of both EDR and favorable prognosis, while age was a strong ‘‘negative indicator’’ only for 3-month good outcome. Average HU decrease over the first day was 5.7 HU in HMCAS (1) and 2.9 HU in dot sign (1) arteries. The densities of thrombi in MCA and insular branches were not different in subjects with and without cardioembolism. Conclusions: CTA provides dependable (high sensitivity and specificity) information regarding clot size and location, whereas hyperdense artery signs have low sensitivity and just acceptable specificity levels in this regard. However, the prognostic and diagnostic information generated by the presence of hyperdense artery signs and temporal change in attenuation can be useful in acute stroke settings where CTA is not readily available. Quantitative measures, rather than qualitative evaluation have a higher yield in determination of temporal change of the hyperdensity signs and its possible clinical correlates. Key Words: Hounsfield unit—middle cerebral artery—insular branch—attenuation— prognosis—emboli composition. Ó 2015 by National Stroke Association

From the *Department of Neurology and Neurological Intensive Care Unit; and †Department of Radiology, Hacettepe University Hospitals, Ankara, Turkey. Received August 13, 2014; revision received February 9, 2015; accepted February 21, 2015. Address correspondence to Mehmet A. Topcuoglu, MD, Neurological Intensive Care Unit, Department of Neurology, Hacettepe

University Hospitals, Sihhiye 06100, Ankara, Turkey. E-mail: mat@ hacettepe.edu.tr. 1052-3057/$ - see front matter Ó 2015 by National Stroke Association http://dx.doi.org/10.1016/j.jstrokecerebrovasdis.2015.02.017

Journal of Stroke and Cerebrovascular Diseases, Vol. -, No. - (---), 2015: pp 1-10

1

M.A. TOPCUOGLU ET AL.

2

Intravenous (IV) thrombolysis with recombinant tissue plasminogen activator (rtPA) is the key element in modern acute stroke management. Its efficacy is primarily determined by size and composition of the intravascular occluding thrombus. However, this critical information is not directly available to the treating physician in a timely manner in usual daily scenarios. The seminal randomized controlled trials demonstrating the success of IV rtPA as an acute stroke treatment modality have not used advanced imaging techniques, but just a plain head computerized tomography (CT).1,2 Considering that any further imaging study would postpone IV rtPA administration and probable tissue reperfusion, all contemporary guidelines follow the same path recommending ‘‘a plain head CT-based IV thrombolysis model’’ for regular acute stroke patients presenting during the first 4.5 hours.3,4 Fortunately, plain head CT conveys useful information about clot composition and burden especially in cases with the middle cerebral artery (MCA) strokes. The presence and length of the hyperdense middle cerebral artery sign (HMCAS) indicating M1 and proximal M2 segments thrombosis, and the MCA ‘‘dot’’ sign indicating thrombosis within insular branches (distal M2 and M3 segments), provide significant prognostic and diagnostic information.5,6 However, clinical utility of the hyperdense artery signs remains beyond consensus and being apparently underused. Moreover, clinical difference between HMCAS and MCA dot sign positive stroke patients has not been well defined.7 We herein present several previously un(der)sought features of these CT signs. First, we performed a comparative analysis on clinical correlates of hyperdense MCA and dot signs on admission CT. We then determined the change in qualitative and quantitative (density with Hounsfield unit [HU] measurement and burden with scoring hyperdense artery signs8) features of hyperdense MCA and dot signs from pretreatment CT obtained at admission to post-treatment CT obtained on the next day, and assessed their utility in predicting stroke etiology, response to lytic therapy, and short- and long-term prognosis.

Patients and Methods Patients A total of 131 consecutive patients (age, 65 6 13 years; 63 women) with acute MCA territory stroke treated with IV and/or interventional thrombolysis/thrombectomy over a period of 8 years were included into this single center study. The study was restricted only to patients with pretreatment plain CT technically sufficient enough for quantitative and qualitative density evaluations. The vascular status in all included subjects was evaluated with at least 1 pretreatment angiography modality (CT angiography [CTA] in 119; catheter angiography in 9;

and magnetic resonance angiography in 3). CTAs are performed soon after completion of plain CT in the same session. Admission angiography documented occlusion of the M1 and/or proximal M2 in 114 and distal M2 and/ or M3 in the 17 study subjects. Right MCA was found to be occluded in 53, whereas the left one was occluded in 77. One bilateral MCA stroke case was excluded. Ipsilateral internal carotid artery (ICA) was occluded in 23 enrolled patients. Treatment modality was standard IV thrombolysis in 99 cases, whereas direct interventional in 14 and combined in 18. Analysis of changes with respect to frequency of the hyperdense artery sign and absolute HU values on follow-up CT (obtained 25 6 7 hours after onset, at average) was confined to 99 subjects treated with IV thrombolysis. Fourteen patients (age, 68 6 13 years; female, 6; National Institutes of Health Stroke Scale [NIHSS] at admission, 19.5 6 13.1) undergoing recanalization attempt (IV thrombolysis in 11; interventional in 2; and combined approach in 1) for acute vertebrobasilar territory stroke (VBS) served as ‘‘vascular control.’’ For temporal density evolution analysis, 11 VBS cases treated with IV rtPA were analyzed. In addition, 42 nonstroke subjects (age, 59 6 15 years and female, 20) with normal brain magnetic resonance imaging and CTA, evaluated in emergency services during the year 2011, were included as the ‘‘control’’ group. Stroke severity was quantified by NIHSS at admission, at 24 hours, and at discharge.9 The modified Rankin Scale was used for the evaluation of functional status at admission, at discharge, and at the end of the third month.10 ‘‘Early dramatic recovery (EDR)’’ or ‘‘favorable response to lysis’’ was defined as a decrease in NIHSS score to less than 4 and/or improvement of NIHSS score by at least 10 points at the end of first 24 hours.11 Stroke etiological classification was performed using ‘‘the causative classification of stroke system.’’12 The causative classification of stroke subgroups were limited to ‘‘cardioembolism,’’ ‘‘large-artery atherosclerosis,’’ and ‘‘other causes’’ which also included the undetermined group. Of note, all patients underwent a predefined etiological work-up including transthoracic echocardiography and 24-hour Holter monitoring. Relevant time points such as symptom-to-door, symptom-to-CT, symptom-to-needle, and time-to-control CT were noted. Admission hematocrit and serum glucose levels were also recorded.

Imaging Noncontrast (plain) pretreatment CT along with CTA and 24-hour follow-up plain CT were obtained using a multidetector row scanner (SOMATOM Emotion Duo or Sensation 16; Siemens, Erlangen, Germany). Image acquisition parameters of CT and CTA can be found in detail elsewhere.8 Of note, the routine slice thickness was 5 mm. The modified clot burden score (mCBS) was used

CLOT CHARACTERISTICS ON CT

to quantify the extent of vascular occlusion on pretreatment CTA.8 HMCAS was considered to be present when there was spontaneous visibility of complete or part of horizontal portion of MCA or higher attenuation of MCA compared with the peripheral parenchyma and other arteries and segments with no calcification. (Fig 1, A,B)13,14 The 10-point hyperdense MCA burden score (HMCABS) was used for quantification of the extent of the HMCAS.8 The dot sign was defined as hyperattenuation of an arterial structure in the Sylvian fissure relative to the surrounding parenchyma and to other arteries (Fig 1, B,C)6,15 The dot sign was not considered sufficient for a diagnosis of HMCAS, and its presence was not included in HMCABS scoring as well. The number of dot signs was also noted. Interobserver reliability for the detection of the dot sign was good, like HMCAS and HMCABS.8 For the determination of density in a particular vessel, a plenty of HU measurements were performed by placing single voxel-sized regions of interests throughout the MCA horizontal and insular portions bilaterally along with basilar artery tip (Fig 1, D). Images were zoomed as needed to find the point of highest attenuation. These points were detected first visually, and then, HU was measured in every point of suspicion. Afterward, the highest 2 measurements were averaged for HU value of that particular vessel. In MCA stroke patients, side-to-side HU ratio was obtained by dividing ipsilateral (symptomatic) HU to its contralateral (asymptomatic) counterpart. The side-to-side difference is calculated as subtraction of contralateral HU from ipsilateral HU measured. In control groups, side-to-side HU ratio was calculated by dividing the higher value to the lower ones. The side-to-side HU difference was calculated as again subtracting the lower value from the higher one in controls. Finally, prevalence and type (Fiorelli scale)16 of hemorrhagic complications were determined based on the 24-hour CT examinations. Symptomatic hemorrhage was considered to be present when a parenchymal hematoma type 2 was detected.

Statistics Values are given as mean 6 standard deviation, percentage and 95% confidence intervals (CIs) and median (interquartile range) as appropriate. Mann–Whitney U/Student t and chi-square/Fisher exact tests were used properly to compare subgroups in terms of numerical and categorical variables, respectively. Spearman test was used for correlation analysis. Receiver operator characteristics (ROC) curves were used to determine relevant thresholds of HU values to diagnose HMCAS and dot sign as well as to predict EDR and favorable 3-month prognosis. Positive and negative likelihood ratios (1LR and 2LR) were calculated. Comparison across groups

3

with respect to area under the curve (AUC) was performed with analysis of variance followed by Student– Newman–Keuls test for pairwise comparisons. Multiple regression analyses were performed to identify independent predictors of EDR and late favorable response to lysis. For these analyses, age (per decade), admission NIHSS (per point), mCBS (per point), presence and burden score of HMCAS, presence and number of the dot sign, and categories of HU (absolute ipsilateral HU and ratio of HU) for MCA proximal and distal portions were used as independent and early/late prognosis as dependent variables. Inter-rater reliability of the indices measured was determined in randomly assigned 20 subjects in the study population. P value lower than .05 was set as the statistically significance level. SPSS 15.0 (Chicago, Illinois, US) statistical package programs were used for analyses.

Results The dot sign was observed in 35 (26.7%) patients with acute MCA stroke treated with lysis, whereas HMCAS was present in 52 (39.7%) of them. CTA showed MCA horizontal segment occlusion in 94% of the patients with HMCAS and in 89% of those with dot sign (P . .05). In 114 patients in whom CTA documented MCA horizontal segment occlusion, 43% had HMCAS and 27% had dot sign. In 17 patients in whom CTA documented MCA insular branch occlusion, 18% had HMCAS and 24% had dot sign. In other words, HMCAS was correlated, albeit borderline, with proximal occlusion (P 5 .046), but the dot sign was not correlated with insular branch occlusion (P . .05). Eighteen (13.7%) patients had concomitant HMCAS and the dot sign. These patients had higher NIHSS at admission, at 24th hour, and at discharge compared with those with only dot sign (n 5 17), with only HMCAS (n 5 34), and with no hyperdense artery sign (n 5 62; Table S1 in Appendix). In addition, they had longer in-hospital stay, higher in-hospital mortality, and less (17%) favorable 3-month clinical outcome. Of note, 44% of patients with both HMCAS and dot signs had ICA occlusion, which was significantly more prevalent compared with other patients (P 5 .006, Table S2 in Appendix). Patients having dot sign but not HMCAS (‘‘isolated dot sign group’’) had the lowest NIHSS at the specified time points, in-hospital mortality, and lysis-associated cerebral hemorrhage rates. The 90-day favorable outcome was the highest (65%) in patients with isolated dot sign. These subjects had lower mCBS (mean, 2.05) in comparison with those with HMCAS (4.1 in isolated HMCAS and 5.4 in combined hyperdensity group, mean) (Tables S1 and S2 in Appendix). The patients with HMCAS but not dot sign (‘‘isolated HMCAS group’’) showed intermediate stoke severity

4

M.A. TOPCUOGLU ET AL.

Figure 1. (A) A typical example of HMCAS is visible both on pretreatment (on the left) and post-treatment CT (on the right). (B) Combined HMCAS and dot sign is present on the left hemisphere (closed arrows). (C) A representative example of the dot sign is shown (open arrows). Lower panel includes magnified and focused images. (D) Measurement of HU is shown; (a) a short segment HMCAS is recognizable on the left side; (b) the box is zoomed here for better demonstration of short segment HMCAS. (c) Pixel-sized density measurement was performed in all relevant visually determined points through the artery, and the maximum 2 were averaged for each artery. Measurement points were reduced for easier understanding in this example. Abbreviations: CT, computed tomography; HMCAS, hyperdense middle cerebral artery sign.

CLOT CHARACTERISTICS ON CT

5

Figure 1. (Continued)

based on NIHSS values. No correlation was found between presence of hyperdense artery sign and age, gender, stroke etiology, hematocrit, and glucose levels at the time of CT. The time intervals (from onset to arrival, then to CT, and then to IV rtPA administration) were also not different (Tables S1 and S2 in Appendix). HU measured in the ipsilateral proximal MCA of the HMCAS positive MCA stroke cases (54.3 6 7.3) was significantly higher in comparison with those of HMCAS negative MCA stroke cases (42.9 6 5.1), contralateral proximal MCA (42.1 6 4.8), and basilar artery tip (41.5 6 6.2) in all MCA stroke cases, proximal MCA of patients with VBS stroke (right: 42.8 6 4.1 and left:

42.5 6 4.7) and control subjects (right: 42.9 6 5.7 and left: 41.9 6 5.8; Table S2 in Appendix). Optimal ipsilateral HU, ipsilateral-to-contralateral HU difference, and ratio to define HMCAS were 46, 5, and 1.1, respectively. Hundred percent specific cutoffs for same parameters were 60, 13, and 1.38, respectively. Discriminative power of these Hounsfeld indices to diagnose HMCAS was good (lower limit of 95% CI of the AUCs were in the range of .870-.898; Table S3A in Appendix). HU measured in the ipsilateral insular branch(es) of the dot sign positive MCA stroke cases (44.5 6 5.7) was significantly higher in comparison with those of dot sign negative MCA stroke cases (35.5 6 5.1), contralateral

6

insular fissure branches (34.6 6 5.2) of MCA stroke subjects, both side insular MCA branches of patients with VBS stroke (right: 37.1 6 4.1 and left: 35.2 6 4.8), and dot sign negative control subjects (right: 38.1 6 5.7 and left: 35.6 6 5.9; Table S2 in Appendix). Optimal ipsilateral HU, ipsilateral-to-contralateral HU difference, and ratio to define dot sign were 38, 5, and 1.17, respectively. Hundred percent specific cutoffs for same parameters were 47, 10, and 1.30, respectively. Discriminative power of these Hounsfield indices to diagnose dot sign was lower than those of HMCAS but still acceptable (lower limit of 95% CI of the AUCs were in the range of .811-.888; Table S3B in Appendix). No significant difference was detected in attenuation quantification indices of proximal and distal arteries of different stroke etiology groups (cardioembolic versus others, Table S4 in Appendix). Changes in hyperdense artery frequency and vessel density from pretreatment CT to post-treatment 24th hour CT were evaluated in 99 MCA stroke patients treated with IV rtPA and 11 VBS stroke treated with IV rtPA (Table S5 in Appendix). HMCAS disappeared in 18 (43.9%) of 41 HMCAS (1) MCA stroke subjects. During the first day, ipsilateral HU decreased by 5.7 on average (from 53.5 to 47.8), symptomatic-to-asymptomatic HU ratio by .11 (from 1.27 to 1.15), and difference by 5.1 (from 11.2 to 5.9) in subjects with HMCAS, whereas no significant HU change was seen in follow-up CT at the contralateral MCA and basilar tip in HMCAS (1) MCA cases; ipsilateral MCA, contralateral MCA, and basilar tip in the HMCAS (-) subjects and those with VBS stroke (Table S5A in Appendix). HMCAS disappearing subjects had higher favorable outcome at discharge and third month along with lesser frequency of associated ICA occlusion (Table S5B in Appendix). Of note, 7% (4 of 58) of patients with no HMCAS on admission CT showed HMCAS on follow-up CT. There was no significant relationship between HMCAS sign clearance and Fiorelli’s PH2 type hemorrhage and early recovery. Of note, in 6 patients complicated with PH2 hemorrhage, no HU decrease was observed. The dot sign disappeared in 12 (44.4%) of 27 dot sign (1) MCA stroke cases treated by only IV rtPA. Ipsilateral insular vessel HU decreased by 2.9 on average (from 44.4 to 40.8), symptomatic-to-asymptomatic HU ratio by .21 (from 1.35 to 1.14), and difference by 6.2 (from 11.2 to 4.9) in dot sign (1) subjects, whereas no significant HU change was seen in follow-up CT at the contralateral insular MCA branches in dot sign (1) MCA cases; ipsilateral and contralateral M3 segment MCA in the subjects without dot sign, and those with VBS stroke (Table S5A in Appendix). Two patients with no dot sign on admission CT were categorized to have dot sign on follow-up CT. Dot sign disappearance and persistence were not correlated with outcome (Table S5B in Appendix). EDR was evaluated in 99 patients treated with IV rtPA. EDR-1, defined as NIHSS decrease higher or equal to 10,

M.A. TOPCUOGLU ET AL.

was observed in 13%, whereas EDR-2 (EDR-1 criterion and/or 24th hour NIHSS less than 4) was noted in 24% (Table S6 in Appendix). No parameters were significantly correlated to the EDR-1 on multivariate analysis. However, NIHSS at admission (P 5 .043) and mCBS (P 5 .034) were inversely related to EDR-2, that is, indicated a higher chance of EDR in patients with less severe stroke in advance. Several CT characteristics of proximal MCA such as presence of HMCAS and higher ipsilateral M1 HU indices including absolute HU value or symptomatic-to-asymptomatic side HU ratio and difference showed a trend of negative correlation with EDR-2 with univariate analysis (.05 , P , .1), but no relationship persisted on multivariate analysis. Of note, neither the presence of the dot sign nor its HU indices did affect EDR-1 and EDR-2 frequency (Table S6 in Appendix). Third month favorable prognosis was found to be negatively affected by increased age (P 5 .0027 per decade), admission NIHSS (P 5 .0001 per point), and mCBS (P 5 .001, per point) in patients treated with IV rtPA (Table 1) and also in the whole study population (Table S7 in Appendix). None of the CT-based characteristics of the occluding clot in proximal and/or insular portion of the MCA such as HMCAS and dot sign, HMCA clot burden score, dot sign positive vessel number, ipsilateral absolute HU value or side-to-side HU ratio, and HU difference had an independent value in terms of prognosis determination on the multivariate analysis, which included CTA-based clot burden score. When this parameter was excluded from the model, in other words, when adjusted separately for age (decade) and NIHSS (point), several indices of the CT characteristics of the proximal, but not insular, MCA clot were found to be significant independent prognostic determiners. These independent negative predictors of 3-month good recovery were HMCA clot burden score (P 5 .006); symptomatic MCA HU values (P 5 .0019 for absolute value or P 5 .006 for being higher than 46); and symptomatic-toasymptomatic side ratio higher than 1.1 (P 5 .0039) and difference (P 5 .0204 for absolute; P 5 .0071 for being higher than 5). Finally, we calculated the sensitivity and specificity of HMCAS and dot sign on an artery basis. When there was evidence for an artery occlusion on CTA, HMCAS had a fair sensitivity (40% [95% CI: 31%-49%]) but good specificity (93% [95 CI: 88%-96%]). The diagnostic power of the dot sign was slightly lower (sensitivity 27% [95% CI: 19%-35%] and specificity 87% [82%-92%]). Negative and positive predictive value and likelihood ratios of these hyperdense artery signs are beyond perfect (see Table S8 in Appendix for details). It is important to note that measured density of hyperdense arteries (and their hyperdense signs), analyzed in the control group, showed positive correlation to hematocrit values (r 5 1.5760; 95% CI, .4136-.7028; P , .0001; for correlation graph see Figure S1 in Appendix).

CLOT CHARACTERISTICS ON CT

7

Table 1. Multivariate analysis in IV rtPA patients

Parameter

Favorable

Nonfavorable

P

Model 1 (all in)

N Age (per decade) Female/male Cardioembolism Admission NIHSS (per point) mCBS (per point) HMCAS HMCABS Dot sign HMCAS 1 dot sign M1 HU ipsilateral M1 HU ipsilateral .46 M1 HU difference M1 HU difference .5 M1 HU ratio M1 HU ratio .1.1 M3 HU ipsilateral M3 HU ipsilateral .38 M3 HU difference M3 difference .5 M3 HU ratio M3 HU ratio .1.17

40 61 6 13 17/23 32 (80%) 12.5 6 4.4 1.75 6 1.13 9 (23%) .65 6 1.41 9 (23%) 2 (5%) 44.9 6 5.9 11 (28%) 3.5 6 4.1 8 (20%) 1.08 6 0.1 8 (20%) 37.7 6 6.7 15 (38%) 3.4 6 6.3 9 (23%) 1.1 6 .19 10 (25%)

59 69 6 12 32/27 38 (64%) 18.0 6 4.5 3.56 6 2.14 30 (51%) 2.03 6 2.31 18 (31%) 12 (20%) 49.7 6 8.2 37 (63%) 7.0 6 7.1 32 (54%) 1.17 6 .17 31 (53%) 38.3 6 6.4 27 (46%) 4.0 6 5.5 19 (32%) 1.13 6 .18 18 (31%)

.003 .252 .070 ,.001 ,.001 .005 .001 .380 .032 .002 .001 .006 .001 .007 .001 .650 .373 .609 .351 .527 .637

Coefficient, P 2.0747; P 5 .0247 NS NS 2.3455; P 5 .00001 2.8178; P 5 .0002 NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS NS

Model 2 (age and NIHSS adjusted) Coefficient, P — NS — — NS 2.0597, P 5 .006 NS 2.017; P 5 .0019 2.2394; P 5 .006 2.0158; P 5 .0204 2.2359; P 5 .0071 NS 2.2517; P 5 .0039

Abbreviations: HMCABS, hyperdense middle cerebral artery burden score; HMCAS, hyperdense middle cerebral artery sign; HU, Hounsfield unit; IV rtPA, intravenous recombinant tissue plasminogen activator; mCBS, modified clot burden score; NIHSS, National Institutes of Health Stroke Scale; NS, not significant.

Discussion Noncontrast CT examinations revealed HMCAS in 2 of every 5 and dot sign in 1 of every 4 patients with acute MCA occlusion. These frequencies stand in the midrange of the reported incidence in the germane literature.6,7,14,15,17-19 The reported incidence of hyperdense artery shows significant variability, which is largely because of time to scanning and the slice thickness used.20 We herein report that both of these hyperdense artery signs are correlated with occlusion of M1 and/or proximal M2 (or horizontal) segments of the MCA. In other words, location of hyperdensity as HMCAS and dot signs in plain CT do not specify the site of arterial occlusion. In a previous angiographic correlation study, sensitivity of the dot sign was 38% against distal M2 and/or M3 occlusion, similar to 39%, the value reported for HMCAS against M1 and/or proximal M2 occlusion. Contrary to this low sensitivity, both signs were quite specific (100% for dot sign and 95% for HMCAS).17 This finding is not in agreement with our observation indicating that most of the dot signs are correlated to occlusion of more proximal, or horizontal, portion of the MCA instead of specific insular branches occlusion. One of the explanations for this discrepancy can be that perhaps at least in some instances dot sign represents distal flow stagnation in the symptomatic territory rather than presence of in-situ

thrombus. The lower HU values measured in our study for the dot sign (44.8 HU) compared with those of HMCAS (54.2 HU) might also support the flow stagnation hypothesis, whereas it should be mentioned that contamination from partial volume effect could have contributed to this difference as well. In our patients with acute MCA occlusion, 14% had ‘‘combined’’ hyperdense artery; in other words, HMCAS plus dot sign. The previous studies reported this frequency between 2% and 18%.6,7,17 However, no specific analysis was pursued for this highly particular group in those studies. We found that these patients had significantly more frequent ICA occlusion (44%). Albeit not appeared as an independent indicator of prognosis on models with multiple adjustments, presence of combined hyperdense artery sign were associated with more severe strokes, higher morbidity (3-month favorable prognosis chance was 17%), and mortality rates in our cohort. Furthermore, these patients had a higher burden of intraluminal thrombus. All these findings lead us to hypothesize that the presence of HMCAS and the dot sign together might signify a potential of resistance to recanalization by IV rtPA, and it would be wise to plan for interventional modalities in such cases. Failure to recanalization by IV rtPA in patients with high clot burden (such as longer or tandem clots) has been a

M.A. TOPCUOGLU ET AL.

8 21,22

well-demonstrated and widely accepted concept. In concordance with this concept, our data show that combined hyperdense artery signs might represent a quite practical marker of high clot burden detectable on plain CT. Albeit presence of ‘‘isolated’’ dot sign seemed to be correlated with better prognosis (3-month good outcome was 65%) on univariate comparisons, this correlation did not come out as an independent predictor of 3-month favorable prognosis on multivariate analysis. This observation disagrees at the first glance, but actually not, with the largely acknowledged notion stating a connection between good outcome and the dot sign that has actually been demonstrated in its first description.6 Subsequent articles were also supportive of this notion but none of them used a multivariate adjustment in IV thrombolysis patients like ours.7,17,23,24 Our adjustment with age, NIHSS, and mCBS obscured its predictive value for prognosis. We detected that isolated dot sign group had lower pretreatment NIHSS (3 points at average) and lower mCBS (2.05 versus 4.1) compared with ones with isolated HMCAS or even those without hyperdense artery. The dot sign seems, therefore, to be an epiphenomena reflecting a milder clinical deficit and lower clot burden. As a result, nonsignificance on the models, which incorporate information on NIHSS and mCBS can be justifiable. We herein quantified luminal attenuation for not only HMCAS, but also the dot sign. Using several type of control vessels, we determined 100% specific and optimally sensitive thresholds for diagnosis of HMCAS and dot sign (see Table S3 in Appendix). As mentioned previously, absolute HU of the dot sign is significantly less than those of HMCAS (100% specific cutoffs: 60 and 47 for HMCAS and dot sign, respectively) but ipsilateralto-contralateral difference (13 versus 10) and ratio (1.38 versus 1.30) look pretty similar to each other. Utility of these reference values to diagnose hyperdense artery sign was excellent for HMCAS (AUC of ROC curves range, .916-.942) and good-to-excellent for dot sign (AUC of ROC curves range, .867-.935). In concordance with prior studies stating that up to half of the hyperdense artery sign being cleared with IV rtPA,7,25 44% of HMCAS and 45% of dot sign resolved after IV thrombolysis in our cohort. From pretreatment to post-treatment 24-hour CT, HU decrease was 5.7 and 3 HU at average for HMCAS and dot sign, respectively. Disappearance of a hyperdense artery sign on follow-up CT is usually attributed to recanalization of an occluded vessel.26 Accordingly, disappearance is associated with favorable outcome. Likewise, persistence of hyperdense artery sign on follow-up CT scan after rtPA is associated with poorer outcome.27 Our data verified this connection and extended the concept in several aspects: First, disappearance rate seems to be less in cases with ICA occlusion. Second, de novo appearance of HMCAS on

follow-up CT occurs in some patients (7% of HMCAS negative cases). This phenomenon has not been mentioned in the literature. We do not have a certain explanation but think that clot retraction can contribute to this event. Although there might be a certain level of error in this observation, we believe that this is minimal as the diagnoses were based on HU measurements. Third, we detected that disappearance of a dot signs does not convey prognostic information unlike HMCAS. Finally, no connection appeared between HMCAS clearance and significant hemorrhagic complications. In contrast, HMCAS persisted without any decrease in HU in all patients with Fiorelli’s PH2 type hemorrhage. At this point, it is important to note that adjacent parenchymal hypodensity complicates visual recognition of hyperdense artery sign28-30 and can lead to false-positive diagnosis. We suggest to use the quantification of intraluminal attenuation at least for follow-up studies. EDR was herein diagnosed when there was at least 10-point NIHSS decrease (EDR-1) and/or decrease in NIHSS below 4 (EDR-2) during the first day after thrombolysis. We failed to demonstrate any independent indicator of EDR-1, whereas clot burden (mCBS) and severity of stroke (NIHSS) appeared as significant ‘‘negative’’ indicators of EDR-2. Some previous studies highlighted an inverse relationship between HMCAS and EDR with different descriptions such as NIHSS improvement more than 8.7 In our study, presence of hyperdense arteries and their quantification indices showed only a trend of negative effect (more pronounced for HMCAS than dot sign) on EDR-2. Three factors, age, clinical severity (NIHSS) before treatment, and CTA-determined clot burden (mCBS), were found to be significant and independent determiners of 3-month prognosis in our study. The presence of any of these parameters indicated worse prognosis both in the entire study population and IV rtPA only group. Age and NIHSS are well-established indicators of not only therapeutic benefit on functional outcome but also rtPA-associated hemorrhage in all studies including those producing prognostic scores or instruments for the acute stroke period.31-36 The presence of HMCAS was included in some of these scores33,34 but ‘‘objective’’ clot burden parameters were not included in any of them. In our study, CTA determined mCBS appeared the strongest prognosis definer in the setting of thrombolysis. Furthermore, mCBS possibly shadowed the effect of other parameters containing information pertaining to clot size and location. We therefore performed additional analyses via submodels excluding this parameter and adjusting only against age and clinical severity, which was indeed a better reflection of ‘‘the routine CT-based thrombolysis strategy.’’ This analysis disclosed that routine plain CT-based clot burden score (HMCABS) was connected to worse prognosis, supporting several previous thin-slice CT studies using more

CLOT CHARACTERISTICS ON CT

sophisticated methods quantifying the length and volumes of the HMCAS. All former studies indicated that high clot burden is a critical indicator of worse prognosis, resistance to IV thrombolysis, and other recanalization attempts.18,37 In this context, HMCABS seems to be an ‘‘easier’’ method for quantification of the clot burden, or the extent of vascular obliteration, in stroke. In our study, neither the presence of the dot sign nor its quantitative attenuation indices were found to be independent predictors of outcome. Furthermore, the presence of HMCAS was only found to be a significant predictor in only univariate but not in multivariate analyses. However, dichotomy using optimal cutoffs of hyperdense MCA attenuation (pixel-sized HU measurement) parameters (MCA HU $46; ipsilateral-tocontralateral HU difference $5; and MCA ratio $1.1) contributed to estimation of worse prognosis in ageand NIHSS-adjusted submodels. Therefore, probably because of improvement in the reliability of evaluating HMCAS, measurements of attenuation can be suggested to estimate prognosis in the absence of pretreatment CTA. We have previously highlighted that despite its high specificity, HMCAS has limited utility in decisionmaking process of acute MCA stroke thrombolysis and early prognostification perhaps because of its low sensitivity.8 Now, we herein see that the same is valid for the dot sign. Like HMCAS, the dot sign can also be seen in nonstroke subjects albeit usually in more than 1 vessel. Compared with those of HMCAS, its sensitivity (27% versus 40%) and specificity (87% versus 93%) were even less. In addition, we demonstrated that density quantification could not add much to treatment decision making in patients with acute MCA occlusion confirmed by CTA because clot hyperattenuation did not reflect its composition and therefore response to recanalization attempt. This is valid for both HMCAS and the dot sign. However, we found that MCA hyperattenuation extent scores in CT with routine thickness were acceptably correlated to CTA-based qualitative clot extent scoring systems. For sure, the information regarding clot size can more easily and dependably be obtained with direct measurement of the length of the nonopacified portion (clot) of the vessel in CTA. Still, these plain CT clot presence and burden parameters could provide useful additional information in the absence of pretreatment vascular imaging.

Supplementary Data Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.jstrokecerebrovas dis.2015.02.017.

References 1. Tissue plasminogen activator for acute ischemic stroke. The National Institute of Neurological Disorders and

9

2.

3.

4.

5. 6.

7.

8.

9.

10.

11.

12.

13.

14.

15. 16.

17.

18.

19.

Stroke rt-PA Stroke Study Group. N Engl J Med 1995; 333:1581-1587. Hacke W, Kaste M, Bluhmki E, et al. Thrombolysis with alteplase 3 to 4.5 hours after acute ischemic stroke. N Engl J Med 2008;359:1317-1329. Practice advisory: thrombolytic therapy for acute ischemic stroke–summary statement. Report of the Quality Standards Subcommittee of the American Academy of Neurology. Neurology 1996;47:835-839. American College of Emergency P, American Academy of N. Clinical policy: use of intravenous tPA for the management of acute ischemic stroke in the emergency department. Ann Emerg Med 2013;61:225-243. Gacs G, Fox AJ, Barnett HJ, et al. CT visualization of intracranial arterial thromboembolism. Stroke 1983;14:756-762. Barber PA, Demchuk AM, Hudon ME, et al. Hyperdense sylvian fissure MCA ‘‘dot’’ sign: a CT marker of acute ischemia. Stroke 2001;32:84-88. Li Q, Davis S, Mitchell P, et al. Proximal hyperdense middle cerebral artery sign predicts poor response to thrombolysis. PLoS One 2014;9:e96123. Topcuoglu MA, Arsava EM, Kursun O, et al. The utility of middle cerebral artery clot density and burden assessment by noncontrast computed tomography in acute ischemic stroke patients treated with thrombolysis. J Stroke Cerebrovasc Dis 2014;23:e85-e91. Brott T, Adams HP Jr, Olinger CP, et al. Measurements of acute cerebral infarction: a clinical examination scale. Stroke 1989;20:864-870. Banks JL, Marotta CA. Outcomes validity and reliability of the modified Rankin scale: implications for stroke clinical trials: a literature review and synthesis. Stroke 2007; 38:1091-1096. Labiche LA, Al-Senani F, Wojner AW, et al. Is the benefit of early recanalization sustained at 3 months? A prospective cohort study. Stroke 2003;34:695-698. Ay H, Furie KL, Singhal A, et al. An evidence-based causative classification system for acute ischemic stroke. Ann Neurol 2005;58:688-697. Puig J, Pedraza S, Demchuk A, et al. Quantification of thrombus Hounsfield units on noncontrast CT predicts stroke subtype and early recanalization after intravenous recombinant tissue plasminogen activator. AJNR Am J Neuroradiol 2012;33:90-96. Leys D, Pruvo JP, Godefroy O, et al. Prevalence and significance of hyperdense middle cerebral artery in acute stroke. Stroke 1992;23:317-324. Shetty SK. The MCA dot sign. Radiology 2006;241: 315-318. Fiorelli M, Bastianello S, von Kummer R, et al. Hemorrhagic transformation within 36 hours of a cerebral infarct: relationships with early clinical deterioration and 3-month outcome in the European Cooperative Acute Stroke Study I (ECASS I) cohort. Stroke 1999; 30:2280-2284. Leary MC, Kidwell CS, Villablanca JP, et al. Validation of computed tomographic middle cerebral artery ‘‘dot’’sign: an angiographic correlation study. Stroke 2003;34: 2636-2640. Kamalian S, Morais LT, Pomerantz SR, et al. Clot length distribution and predictors in anterior circulation stroke: implications for intra-arterial therapy. Stroke 2013;44: 3553-3556. Aries MJ, Uyttenboogaart M, Koopman K, et al. Hyperdense middle cerebral artery sign and outcome after intravenous thrombolysis for acute ischemic stroke. J Neurol Sci 2009;285:114-117.

10 20. Riedel CH, Jensen U, Rohr A, et al. Assessment of thrombus in acute middle cerebral artery occlusion using thin-slice nonenhanced Computed Tomography reconstructions. Stroke 2010;41:1659-1664. 21. Rubiera M, Ribo M, Delgado-Mederos R, et al. Tandem internal carotid artery/middle cerebral artery occlusion: an independent predictor of poor outcome after systemic thrombolysis. Stroke 2006;37:2301-2305. 22. Gralla J, Burkhardt M, Schroth G, et al. Occlusion length is a crucial determinant of efficiency and complication rate in thrombectomy for acute ischemic stroke. AJNR Am J Neuroradiol 2008;29:247-252. 23. Man S, Hussain MS, Wisco D, et al. The location of pretreatment hyperdense middle cerebral artery sign predicts the outcome of intraarterial thrombectomy for acute stroke. J Neuroimaging 2015;25:263-268. 24. Somford DM, Nederkoorn PJ, Rutgers DR, et al. Proximal and distal hyperattenuating middle cerebral artery signs at CT: different prognostic implications. Radiology 2002; 223:667-671. 25. Kharitonova T, Thoren M, Ahmed N, et al. Disappearing hyperdense middle cerebral artery sign in ischaemic stroke patients treated with intravenous thrombolysis: clinical course and prognostic significance. J Neurol Neurosurg Psychiatry 2009;80:273-278. 26. Tsao JW, Gean AD, Glenn OA, et al. Hyperdense MCA resolved after tPA. Neurology 2002;58:1512. 27. Paliwal PR, Ahmad A, Shen L, et al. Persistence of hyperdense middle cerebral artery sign on follow-up CT scan after intravenous thrombolysis is associated with poor outcome. Cerebrovasc Dis 2012;33:446-452.

M.A. TOPCUOGLU ET AL. 28. Jha B, Kothari M. Pearls & oy-sters: hyperdense or pseudohyperdense MCA sign: a Damocles sword? Neurology 2009;72:e116-e117. 29. Maramattom BV, Wijdicks EF. A misleading hyperdense MCA sign. Neurology 2004;63:586. 30. Heckmann JG. A note on the hyperdense middle cerebral artery sign. Emerg Med J 2013;30:345. 31. Saposnik G, Guzik AK, Reeves M, et al. Stroke Prognostication using Age and NIH Stroke Scale: SPAN-100. Neurology 2013;80:21-28. 32. Lou M, Safdar A, Mehdiratta M, et al. The HAT Score: a simple grading scale for predicting hemorrhage after thrombolysis. Neurology 2008;71:1417-1423. 33. Strbian D, Meretoja A, Ahlhelm FJ, et al. Predicting outcome of IV thrombolysis-treated ischemic stroke patients: the DRAGON score. Neurology 2012;78:427-432. 34. Strbian D, Engelter S, Michel P, et al. Symptomatic intracranial hemorrhage after stroke thrombolysis: the SEDAN score. Ann Neurol 2012;71:634-641. 35. Ntaios G, Faouzi M, Ferrari J, et al. An integer-based score to predict functional outcome in acute ischemic stroke: the ASTRAL score. Neurology 2012;78:1916-1922. 36. Whiteley WN, Thompson D, Murray G, et al. Targeting recombinant tissue-type plasminogen activator in acute ischemic stroke based on risk of intracranial hemorrhage or poor functional outcome: an analysis of the third international stroke trial. Stroke 2014;45:1000-1006. 37. Shobha N, Bal S, Boyko M, et al. Measurement of length of hyperdense MCA sign in acute ischemic stroke predicts disappearance after IV tPA. J Neuroimaging 2014; 24:7-10.