Introduction This study investigates whether the Alberta Stroke Program Early CT Score (ASPECTS) quantification is associated with outcome following mechanical thrombectomy.
Objective To determine whether preintervention non-perfect ASPECT scores involving cortical or subcortical regions and the side of the non-perfect ASPECT score affects outcomes.
Methods A retrospective review of a prospectively maintained database of patients with acute ischemic stroke involving the anterior circulation who underwent thrombectomy between May 2008 and August 2012 at a single tertiary care center. The device for mechanical thrombectomy used was the penumbra aspiration system (Penumbra Inc, Alameda, California, USA) and the Solitaire stent retriever (ev3, Irvine, California, USA). A ‘blinded’ neuroradiologist obtained ASPECTS quantification and noted each region demonstrating early changes.
Results 149 patients (51.7% female, mean age 66.1±15.1 years) were included with an average National Institutes of Health Stroke Scale of 16.2±6.7. Patients with non-perfect ASPECT scores on pretreatment imaging were more likely to have a hemorrhagic conversion (p=0.04) evident on post-procedure CT. However, functional outcomes were the same. Patients with both cortical and basal ganglia non-perfect ASPECT scores were more likely to be in a persistent vegetative state or expire. No differences were identified in outcome among patients with left- versus right-sided infarcts affecting the basal ganglia or cortical regions.
Conclusions These findings support a strategy of selecting candidacy for thrombectomy that does not exclude patients with non-perfect ASPECT scores involving either the basal ganglia or cortical regions. Outcomes were identical among patients with no non-perfect ASPECT scores and those with cortical or subcortical infarcts, despite a higher incidence of hemorrhagic conversion found among those with non-perfect ASPECT scores.
Statistics from Altmetric.com
While the intravenous administration of thrombolytic agents (intravenous tissue plasminogen activator (IV tPA)) is the ‘gold standard’ for treatment of acute ischemic stroke (AIS),1 mechanical thrombectomy remains an option for those who are evaluated when treatment is not possible or otherwise do not qualify for IV tPA.2–5 Recent studies have shown clinical benefit for patients when mechanical recanalization can be achieved,6–11 underlining the importance of patient selection for intra-arterial therapy with imaging.4 Although fraught with inter-rater variability, the Alberta Stroke Program Early CT Score (ASPECTS) is a widely accepted scoring system that may help in the triaging of patients with AIS.12–16 ASPECTS has been shown to correlate with the National Institutes of Health Stroke Scale (NIHSS) on presentation and the likelihood of intracerebral hemorrhage and outcome after IV tPA administration.17
ASPECTS may be used in the evaluation of patients undergoing thrombectomy for AIS, but its predictive value remains to be proved. We investigated whether ASPECTS quantification with non-contrast and perfusion CT is associated with hemorrhagic conversion and outcome after mechanical thrombectomy. In addition, we examine whether infarct patterns on initial CT involving cortical or subcortical regions and the side of a non-perfect ASPECT score affects outcomes.
We retrospectively studied a prospectively maintained database of consecutive patients with AIS who underwent intra-arterial therapy between May 2008 and August 2012 at a high-volume tertiary care center (Medical University of South Carolina). The study was approved by the institutional review board of the Medical University of South Carolina. Mechanical thrombectomy was the preferred treatment for AIS. Candidacy for intervention was determined by CT perfusion imaging, irrespective of time of onset.5
Our protocol for determining candidacy does not employ strict ASPECTS criteria or thresholds. Rather, clinical judgment is employed and the treating interventionalist determines if recanalization in the presence of a non-perfect ASPECT score would improve the patient's eventual outcome. In general, patients with non-perfect ASPECT scores but with regions of penumbra thought to contribute significantly to the NIHSS were eligible for recanalization. This study aimed at investigating whether this approach was effective in identifying those patients most likely to benefit from recanalization. Patients who received IV tPA were also candidates for mechanical thrombectomy. During the study, the primary device for mechanical thrombectomy used was the Penumbra aspiration system (Penumbra Inc, Alameda, California, USA)10 and the Solitaire stent retriever (ev3, Irvine, California, USA). Posterior circulation strokes were excluded from the analysis.
Medical records, procedural records, and angiograms were reviewed. The intracranial vessel affected, recanalization status of the affected vessel as determined by the Thrombolysis in Cerebral Ischemia (TICI) scale, intraprocedural complications, and the presence of hemorrhage on post-procedure head CT were extracted. Functional outcome was assessed by 90-day follow-up modified Rankin scale (mRS) score.
A blinded neuroradiologist (HH) obtained ASPECTS quantification based on the CT perfusion blood volume maps. An admission head CT scan was reviewed in conjunction with the blood volume maps from the perfusion imaging dataset, and each region (‘cortical’ M1–6 and insula, ‘subcortical’ lentiform nucleus, internal capsule and caudate) showing early changes was denoted. The final ASPECT score was calculated based on the CT perfusion maps.
Statistical analyses were performed using SAS V.9.2 (SAS Institute, Cary, North Carolina, USA). A group of 149 subjects was analyzed using descriptive statistics to characterize demographics, ASPECT score, infarct type, and other clinical variables describing treatment, complications, and outcomes. Differences between groups were tested using the Student t test for continuous measures and a χ2 test for categorical measures. Differences between groups were tested using Fisher’s exact test for categorical measures with expected cell sizes <5. All tests were two-sided and assessed at a significance level of 0.05.
One hundred and forty-nine patients (51.7% female, mean age 66.1±15.1 years) underwent thrombectomy for AIS involving the anterior circulation during the study. Patients presented with an average NIHSS of 16.2±6.7 and were treated 10.5±12.1 h from symptom onset, on average.
The majority of patients treated (n=100; 67%) had a non-perfect ASPECT score (ASPECTS ≤ 9) before treatment. Among those with a non-perfect ASPECT score, 35 showed cortical changes, 29 subcortical changes and 36 demonstrated early changes in both cortical and subcortical regions. Patients presenting with a region of non-perfect ASPECT score compared with those presenting with entirely reversible penumbra (ASPECTS 10) did not differ in their baseline demographics, including the side of the stroke, time from symptom onset, recanalization time, or procedural complications (table 1).
We present the distribution of ASPECT scores in table 2.
Overall ASPECTS and outcome
Patients with non-perfect ASPECT scores on pretreatment imaging were more likely (p<0.0001) to manifest a hyperdense middle cerebral artery sign and have a hemorrhagic conversion (p=0.04) evident on post-procedure CT than those without any non-perfect ASPECT score. However, functional outcome at 90 days did not differ between the two groups. Patients with an ASPECT score ≤6 were more were likely to have a bad outcome when measured by an mRS of 5–6 at 90 days (p=0.0002). Patients with an ASPECT score ≥7 had a significantly lower rate of death (mRS of 6) at 90 days (12.5%) than those with an ASPECT score ≤6 (36%) (table 3).
Non-perfect ASPECT scores and outcome
Patients with perfect ASPECTS (ie, 10) were compared with those with non-perfect ASPECT scores involving the basal ganglia (BG) (‘subcortical’, n=65). Patients with non-perfect ASPECT scores involving the BG were more likely to have a hemorrhagic conversion than those without a non-perfect ASPECT score (p=0.04), but no differences in outcome at 90 days were identified (table 4).
Patients with perfect ASPECT scores were compared with those with non-perfect ASPECT scores involving the cortex (‘cortical’, n=71). Patients with non-perfect ASPECT scores involving the cortex were more likely to have a hemorrhagic conversion than those without non-perfect ASPECT score (p=0.05), but no differences in outcome were identified (table 5).
Lastly, all patients with non-perfect ASPECT scores were grouped according to the presence of cortical regions only, BG only, or both cortical and BG regions involved (‘both’) (table 6).
Hemorrhagic conversion rates were similar between the three groups but patients with both cortical and BG non-perfect ASPECT scores were more likely to be in a persistent vegetative state or died at 90 days.
Right versus left hemisphere non-perfect ASPECT scores and outcome
The impact of the side of the of occlusion and non-perfect ASPECT scores was also investigated. Patients with left- versus right-sided non-perfect ASPECT scores involving the BG (table 7) as well as non-perfect ASPECT scores involving the cortical left- versus right-sided regions (table 8) were compared. No differences in outcome were identified.
ASPECTS quantification in acute stroke assessment
The ASPECT score is a metric under investigation for quantifying early ischemic changes on non-contrasted head CT and predicting outcomes in patients undergoing evaluation of AIS.17–20 In recent years its use has been investigated in CT perfusion imaging and MRI diffusion-weighted imaging (DWI).12 ,14 ,21–23 While the results of DWI-ASPECTS are promising, the speed and wide availability of CT make it more generalizable and applicable. At our institution, CT perfusion is the preferred imaging modality in the triaging of AIS. ASPECTS determined from a non-contrast CT alone can be subject to poor inter-rater reliability, thereby limiting its use. However, we believe the use of the blood volume map from CT perfusion makes it more objective.12 ,23
Importance of patient selection
Mechanical thrombectomy devices have improved over the past 5 years allowing disruption and removal of acute occlusions for reperfusion of an ischemic penumbra.7–9 ,24–27 With newer devices, it is now commonplace to achieve TICI 2B and three revascularizations in the majority of cases (>75%) compared with <30% with earlier devices.3 ,4 ,26 However, despite these promising advances in thrombectomy technique, patient outcomes have only moderately improved, suggesting that patient selection is crucial. Our center has employed CT perfusion imaging in acute stroke triaging to assess the degree of penumbra versus non-perfect ASPECT score. Analysis of the efficacy of this strategy showed that patient outcomes are identical for those presenting well beyond the traditional time point5 and identifies those patients with sufficient collaterals to still benefit from recanalization.
Impact of non-perfect ASPECT scores on outcome
Experience incorporating ASPECTS data in the setting of IV tPA administration has shown that patients with a smaller infarct burden at presentation (ASPECTS ≥8) had a greater likelihood of good functional outcome.17 ,28 There is evidence that patients with a baseline ASPECT score of >7 who receive intra-articular thrombolysis have better outcomes at 90 days.28 ,29 However, for ASPECTS, data on patient outcomes following mechanical thrombectomy are not well described.
There is general awareness that the overall ASPECTS value may not be as important as the specific regions that are involved to arrive at the value. For example, patients with BG non-perfect ASPECT scores with cortical sparing (eg, ASPECTS 7) may not be considered candidates for intervention at some centers but may be routinely treated at others. Another patient with the same ASPECTS of 7 involving three cortical areas may be considered a candidate. Given the lack of guidelines or consensus on determining candidacy for thrombectomy, the approach, imaging modality, and inclusion criteria of each center will vary. Our protocol for determining candidacy did not employ strict ASPECTS criteria or thresholds. Rather, clinical judgment was employed and the treating neurointerventionalist determined if recanalization in the setting of a non-perfect ASPECT score would improve the patient's outcome. Our clinical judgment was based on the patients’ comorbidities, their pre-stroke baseline performance, and was influenced by frank discussions with families and caregivers about patients’ probability of functional recovery. This study aimed at investigating whether this approach was effective in identifying those patients most likely to benefit from recanalization.
This study found that patients with non-perfect ASPECT scores involving either the cortical or subcortical regions benefited from recanalization to the same extent as those with entirely reversible penumbra (ASEPCTS 10) by achieving the same functional outcomes. In addition, patients with BG non-perfect ASPECT scores achieved the same outcomes as those involving the cortex. These findings suggest that patients with BG infarcts and cortical sparing should not be excluded from intra-articular recanalization treatment. Lastly, the side of the hemisphere involved was also not associated with differing outcomes.
We found that patients who had both cortical and subcortical infarcts had poorer outcomes. This represents a subset of patients in whom a modification of the selection criteria might yield improved outcomes overall and which, if verified in larger multicenter trials, might have implications for the design of future clinical trial involving the use of thrombectomy in AIS.
ASPECTS and intracerebral hemorrhage
Experience with the Mechanical Embolus Removal in Cerebral Ischemia (MERCI) device demonstrated significantly higher rates of intracerebral hemorrhage and worse functional outcomes among patients with BG infarctions as opposed to cortical infarctions.30 In this study, patients with any non-perfect ASPECT score (cortical or subcortical) were found to have a higher hemorrhagic conversion rate than those without non-perfect ASPECT scores. There were no differences in rates of hemorrhage between patients with BG, cortical, or both BG and cortical infarcts. However, the higher hemorrhage rates did not translate into any difference in outcome. This finding is probably explained by the fact that the hemorrhages almost universally occurred in the region of pretreatment non-perfect ASPECT score and thus did not add neurologic morbidity.
There are several important differences between the MERCI report and this study. First, the thrombectomy devices used in this study included the Penumbra separator system and the Solitaire stent retriever. It is possible that different mechanical forces associated with different thrombectomy devices may be associated with unique hemorrhagic rates. In addition, Loh et al report a higher initial NIHSS in patients with BG infarction,30 so it is unclear how much the hemorrhage as opposed to the more severe symptomatology on presentation affected the outcome. Lastly, the authors reported their outcomes as discharge mRS, while we based ours on 90-day mRS. It is likely that patients with poor mRS at discharge from our institution benefited from intensive rehabilitation and were able to recover some of their functionality.
This study is limited in that the patients were evaluated and treated at a single center. Although the database maintained is prospective, it was retrospectively evaluated for this study. TICI scores were assessed by the treating physician, which also introduces a potential source of bias. Only one neuroradiologist calculated the ASPECT scores, and thus we were unable to assess inter-rater reliability.
We have provided support for a strategy of selecting candidacy for thrombectomy in AIS using CT perfusion imaging which does not exclude patients with non-perfect ASPECT scores involving either the BG or cortical regions. Outcomes were identical among patients with perfect ASPECT scores and those with either cortical or subcortical infarcts, despite a higher incidence of hemorrhagic conversion found among those with non-perfect ASPECT scores. The patients with poorer outcomes were those with both BG and cortical infarcts on pretreatment imaging.
Contributors All authors contributed to the drafting and revision of the article, and gave their final approval for its submission to this journal.
Competing interests AMS received non-financial support from Pulsar Vascular, outside the submitted work. MIC received grants and non-financial support from Penumbra and Mircovention; non-financial support from Stryker and Covidien, outside the submitted work. RT received grants and non-financial support from Penumbra, Mircovention, and Stryker; non-financial support from Covidien, outside the submitted work. ASR received grants and non-financial support from Penumbra, Mircovention, Stryker; non-financial support from Covidien, Medpace, Siemens, and Boston Scientific, outside the submitted work.
Ethics approval The study was approved by the institutional review board of the Medical University of South Carolina.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.