Article Text
Abstract
Objective To investigate the associations between Alberta Stroke Program Early CT Score (ASPECTS) or distribution and sidedness of acute infarction and clinical outcomes following intervention with a direct aspiration first pass technique (ADAPT).
Methods A review was performed of patients who had undergone thrombectomy with ADAPT for emergent large vessel occlusions of the middle cerebral artery (MCA) between December 2012 and May 2015. Preintervention CT scans were reviewed by a blinded radiologist to calculate ASPECTS and determine the distribution of infarction. Clinical outcomes were compared for subsets of patients depending upon ASPECTS and regional infarction distribution (cortical, subcortical, or both).
Results One hundred and fifty-four patients (50% female, mean age 67) underwent thrombectomy using ADAPT for MCA emergent large vessel occlusion. The median presenting National Institute of Health Stroke Scale score was 15. Fifty-five per cent of patients had left-side occlusions. Similar good outcomes were achieved for patients with perfect and non-perfect ASPECTS (modified Rankin Scale (mRS) 0–2: 63% vs 51%, respectively; p=0.20). Similar outcomes were also achieved for patients with ‘poor’ ASPECTS (≤6) compared with those with ASPECTS >6 (mRS 0–2: 52% vs 53%, respectively; p=0.91). Regional distribution and sidedness of core infarction on preintervention CT also did not correlate with worse outcomes.
Conclusions Patients with moderate-sized core infarcts involving various distributions in either hemisphere can potentially achieve similar good outcomes compared with those with no core infarction at presentation. A treatment algorithm for acute ischemic stroke, which employs hardline ASPECTS thresholds or excludes patients with basal ganglia infarcts, might preclude patients who would potentially benefit from mechanical thrombectomy with ADAPT.
- Stroke
- Thrombectomy
Statistics from Altmetric.com
Introduction
Acute ischemic stroke (AIS) is a leading cause of morbidity and mortality, and its socioeconomic burden remains high.1 Recently, five major randomized trials have shown the superiority of mechanical thrombectomy over medical management for anterior circulation emergent large vessel occlusions (ELVO).2–6 The success of the recent randomized, controlled trials is, at least in part, due to the use of modern endovascular devices and improvements in prerandomization imaging protocols.7 Devices and techniques for acute stroke intervention have evolved rapidly over the past decade.8 Recently, a direct aspiration first pass technique (ADAPT) has gained in popularity and is now the most common technique used by interventionalists in the USA, as reported in a recent survey.9 However, despite recent dramatic technical and procedural improvements in thrombectomy, clinical outcomes after mechanical thrombectomy for anterior and posterior circulation ELVO have remained relatively static, with about 40% of patients having good neurological function at 90 days (modified Rankin Scale (mRS) 0–2).10 Optimizing triage strategies and enhancing patient selection for thrombectomy may be the next step in improving outcomes of patients with ELVO.
A common imaging tool for selecting candidates for intervention is the Alberta Stroke Program Early CT Score (ASPECTS), a scoring system associated with clinical stroke severity (National Institutes of Health Stroke Scale (NIHSS)), the likelihood of intracerebral hemorrhage, and outcomes after treatment with IV tissue plasminogen activator (IV tPA).11 Previous investigations of thrombolysis with IV tPA for AIS have suggested that ASPECTS <7 corresponds with a significant increase in morbidity and death.12 Another common selection criterion is based on the distribution of infarction, as prior studies have suggested that subcortical location or a combination of subcortical and cortical involvement is associated with worse outcomes after thrombectomy.13 ,14 In light of the increased speed and decreased hemorrhage rates associated with the new and most popular thrombectomy technique (ADAPT), this study re-investigated the associations of ASPECTS, territorial distribution of acute core infarction, and sidedness with clinical outcomes following thrombectomy.10
Patients and methods
Patients
A retrospective review of an institutional review board-approved, prospectively maintained database was performed to identify all consecutive patients with acute middle cerebral artery (MCA) ELVO who had undergone thrombectomy with ADAPT between December 2012 and May 2015 at a single tertiary referral center. Only patients with M1 and M2 segment occlusions were included in this analysis.
Patient selection protocol
All patients presenting with stroke-like symptoms to our institution are rapidly triaged on arrival. Patients are immediately evaluated by the on-call neurologist and transferred for CT. The routine initial imaging examination for patients with suspected AIS includes non-enhanced CT brain scanning, CT angiography of the neck and brain, and CT perfusion of the brain. Patients with no contraindications, receive standard of care doses of IV tPA once the non-contrast CT scan has been reviewed and does not demonstrate hemorrhage. CT angiography and perfusion are then performed. These imaging sequences are reviewed immediately by the neurologist and on-call neurointerventionalist, and a joint decision to pursue thrombectomy is made. At our institution, candidacy for intervention is determined by CT perfusion imaging, irrespective of time of stroke onset.15 Patients with core infarcts but with regions of penumbra thought to contribute significantly to the presenting NIHSS score are eligible for recanalization. Patients who receive IV tPA are also candidates for mechanical thrombectomy.
Data collection
Patient demographic, radiographic, angiographic, and clinical data were collected. Angiographic images were reviewed to record the location of the ELVO, degree of occlusion, and any complications. The degree of occlusion before and after intervention was defined angiographically by the Thrombolysis in Cerebral Infarction (TICI) classification.16 Successful vessel revascularization was defined by a TICI score ≥2b. Procedure time was defined by groin needle stick access to vessel recanalization (greater than or equal to TICI 2b). The NIHSS score was determined by the stroke neurologist at admission and at discharge from the hospital. Clinical outcomes were determined by mRS at 90 days, obtained by the stroke neurologist. Good clinical outcome was defined as a mRS score of 0–2 at 90 days.
ASPECT scoring
The preintervention CT imaging was retrospectively reviewed by a single radiologist blinded to patient outcome to calculate the ASPECT score and to determine the regional distribution of infarction. The ASPECTS quantification was determined by the extent of hypodensity on CT and/or the area of decreased cerebral blood volume on CT perfusion.17 ,18 An ASPECTS of 10 is described as ‘perfect’ and implies an entirely reversible penumbra. A ‘non-perfect’ ASPECTS is ≤9 and corresponds with an area of core infarction. The distribution of the infarct was subdivided into ‘cortical’ (hemispheric cortex and insula regions) and ‘subcortical’ (lentiform nucleus, internal capsule, and caudate).
Outcome evaluation
Procedural and clinical outcomes were determined for subsets of patients depending upon the regional distribution of core infarction, ASPECTS, and sidedness (left vs right). For regional distribution of core infarction, patients were divided into those who had core infarctions involving the subcortical regions, cortical regions, and both subcortical and cortical regions. Patients were also subdivided into those with perfect ASPECTS (no core infarct), non-perfect ASPECTS (any core infarct), ASPECTS ≤6, and ASPECTS >6. To evaluate outcomes associated with sidedness, patients were subdivided into those with core infarctions involving the right cortex, left cortex, right subcortical structures, and left subcortical structures.
Statistical analysis
All data were represented as mean (SD) or frequency (percentage) for continuous or categorical data, respectively. To assess the association between the category and continuous or categorical data, a Student's t tests or χ2 test was used, respectively. Significance was assessed at an α level=0.05. All testing was performed using SAS V.9.4.
Results
Patient population and procedure
One hundred and fifty-four patients treated for MCA ELVO with aspiration thrombectomy using ADAPT met the inclusion criteria. The mean age of patients was 67.2±14.1 years (range 27–93), half were female (n=77), and the mean admission NIHSS score was 14.9±6.2 (range 0–33). Mean time from symptom onset to groin needle puncture (procedure initiation) was 454.3±344.7 min (range 25–1485). Mean procedure time (‘recanalization time’) was approximately 40.0±36.8 min from groin needle puncture access to vessel recanalization of at least TICI 2b (range 0–315). TICI 2b or 3 revascularization was obtained in 146 patients (95%). A second device (stent-retriever) was required after initial direct aspiration in 30 patients (30/154=19%). One patient (0.6%) had a significant intraprocedural complication: inadvertent intracranial internal carotid artery dissection that was successfully treated with a stent.
ASPECTS quantification
All but one of the 154 patients included received the full dedicated ischemic stroke CT protocol. The single patient who did not receive the full CT protocol did have CT angiography and non-enhanced CT, but there was a technical error which prevented acquisition of the CT perfusion. Both CT and CT perfusion findings were used in conjunction to calculate the ASPECTS for each patient (except for the one patient who did not have a CT perfusion study). The majority of patients (123/154; 79.9%) presented with a core infarct (‘non-perfect’ (≤9) ASPECTS) on the initial preintervention CT imaging (figure 1). Of the 122 patients with non-perfect ASPECTS, 76 (62%) had subcortical involvement, 93 (76%) showed cortical involvement, 45 (37.5%) showed isolated cortical involvement, 27 (22%) showed isolated subcortical involvement, and 48 (39%) showed both cortical and subcortical infarction. Patients with an entirely reversible penumbra (‘perfect’ ASPECTS of 10) did not significantly differ in baseline characteristics or demographics from patients with a non-perfect ASPECTS (table 1).
Clinical outcomes
Functional outcomes were available for 134 patients using the mRS at 90 days or best available. Overall, 53% of patients had a good functional outcome with an mRS of 0–2 and 47% achieved an mRS of 3–6. Fourteen patients (10%) died (mRS of 6) during the follow-up period. Ten patients (6.5%) had symptomatic hemorrhagic conversion.
ASPECTS and outcomes
Overall, patients with a core infarct (ASPECTS <10) were not significantly more likely to have a poor clinical outcomes (mRS 3–6) (p=0.20), death (p=0.90), or hemorrhagic conversion (p=0.39) than patients who had entirely reversible penumbras with perfect ASPECTS of 10. Patients with ‘poor’ ASPECTS (≤6) had equivalent good outcomes to those with more optimal scores (mRS 0–2: 52% vs 53%, respectively; p=0.91; table 2) despite a higher rate of postprocedure hemorrhage compared with patients who had ASPECTS >6 (16% vs 4%, respectively; p=0.015).
Region infarct distribution and outcomes
Patients with subcortical infarcts were not significantly more likely to have poor clinical outcomes (p=0.18), death (p=0.76), or hemorrhagic conversion (p=0.85) than those who had a perfect ASPECTS (table 3). Patients with cortical infarcts were not significantly more likely to have bad clinical outcomes (p=0.34), death (p=0.84), or hemorrhagic conversion (p=0.38) than those who had a perfect ASPECTS (table 4). Patients with both cortical and subcortical infarcts were not significantly more likely to have poor clinical outcomes (p=0.73), death (p=0.39), or hemorrhagic conversion (p=0.17) than those who had core infarctions involving only subcortical regions or only cortical regions (table 5).
Core infarct sidedness and outcomes
Patients with left-sided infarcts involving cortex were not significantly more likely to have bad clinical outcomes (p=0.26), death (p=0.88), or hemorrhagic conversion (p=0.86) than those who had right-sided cortical infarcts (table 6). Patients with left-sided infarcts involving subcortical regions were not significantly more likely to have bad clinical outcomes (p=0.07), death (p=0.83), or hemorrhagic conversion (p=0.83) than those who had right-sided subcortical infarcts (table 7).
Discussion
Identifying appropriate thrombectomy candidates is of paramount importance when developing algorithms for acute stroke care aimed at optimizing patient outcomes. Selection of candidates has broadened over time at our institution subsequent to recent improvements in devices, safety profile, and techniques (ie, ADAPT). Our institutional protocol for determining thrombectomy candidacy does not employ a strict ASPECTS threshold, and we do not strictly exclude patients based upon infarct distribution or sidedness. Specifically we do not stringently exclude patients with subcortical involvement or ASPECTS below that used as an inclusion criterion in the major thrombectomy randomized controlled trials (ASPECTS ≤6). Instead of using hardline threshold criteria, we rely on the clinical judgment of the treating physicians (neurointerventionalist and stroke neurologist) and the results of CT and CT perfusion to determine if there is brain tissue at risk that would significantly contribute to outcome (improved NIHSS), if salvaged.19
This retrospective study of the first 154 patients treated with the ADAPT approach for MCA ELVO at our institution was intended to identify patients who might not benefit from thrombectomy, and thus to refine our approach to candidate selection. These data demonstrate that all patients selected for thrombectomy, regardless of size, extent, sidedness, and region of core infarct, had equivalent outcomes, suggesting that the selection strategy we are employing is safe and effective.
Some institutions use ASPECTS thresholds as part of a treatment algorithm because prior studies with IV thrombolysis for AIS demonstrated an association between ASPECTS and outcomes.11 ,20 However, our study showed that similar good outcomes could be achieved in patients with non-perfect ASPECTS and those who had no visualized core infarction at presentation (ASPECTS=10). Also, some institutions exclude patients with a ‘poor’ ASPECTS (≤6) as some studies with IV thrombolysis suggested that the risk of death and bleeding was significantly worse compared with patients with higher ASPECTS (≥7).12 Our study demonstrates that those patients with ‘poor’ ASPECTS (≤6) can potentially achieve equal benefit from aspiration thrombectomy with ADAPT. Indeed, of the patients in our cohort with ‘poor’ ASPECTS, five had no residual symptoms at 90 days (mRS 0). It should be emphasized that not every patient with a large core infarct should be offered thrombectomy, but rather that some patients can achieve a satisfactory outcome if chosen carefully and they must be considered on a case-by-case basis by multidisciplinary team evaluation.
Regional distribution of core infarction (cortical, subcortical, and both cortical and subcortical) on the preintervention CT scan was not significantly associated with worse clinical outcomes. Patients with subcortical infarctions are excluded from aspiration treatment at some institutions because of results from prior IV thrombolysis investigations. This study demonstrates that equally good outcomes can potentially be achieved in those patients when treated with ADAPT.
Boers et al21 recently described probabilities of infarcts in various topographic distributions following thrombectomy, demonstrating that deeper, end arteriole-supplied regions, such as the basal ganglia and insular cortex, are less likely to be spared than cortex after successful recanalization with mechanical thrombectomy. Even with this lack of benefit to these deeper regions, it might still often be beneficial to treat these patients because of the potentially salvageable cortex. Hemispheric lateralization (left vs right sidedness) of acute infarctions has been suggested as a predictor for clinical outcomes in patients treated endovascularly.22 This study demonstrated that patients with infarcts involving the left cortex and left subcortical regions can potentially achieve equal outcomes to those who have right-sided infarcts.
Intracranial hemorrhage is a feared complication of acute stroke intervention. Prior studies with older generation devices have demonstrated an increase in the number of hemorrhages for patients treated with mechanical thrombectomy for acute ischemic infarcts that involved the basal ganglia.14 ,15 However, we found no significant increase in hemorrhages for patients with subcortical infarctions. This difference in hemorrhage rates may be due to the atraumatic nature of the ADAPT approach compared with prior techniques. With ADAPT, the large aspiration catheter is carefully deployed at the clot face and aspiration is employed without placing undue torque on the vessel and the perforators arising from it. Therefore, unlike stent retrievers, there is less potential for shearing the endothelium or damaging the lenticulostriate perforators. For patients with larger core infarcts (‘poor’ ASPECTS ≤6) there was a significantly increased risk of hemorrhage (p=0.015). However, as described above, in general these patients are potentially able to achieve equal outcomes compared with those who have ASPECTS >6. Therefore, while patients with poor ASPECTS were more likely to demonstrate intracranial hemorrhage on CT after the procedure, these hemorrhages generally did not affect the final outcomes.
Study limitations
This study is limited by its retrospective nature. Over 12% of patients were lost to follow-up; 19% of patients with poor ASPECTS did not have 90-day mRS available compared with 11% of patients with ASPECTS ≥7. This might have skewed our conclusions erroneously to favor equivalence as patients with poor ASPECTS might be more likely to have worse outcomes, as has been suggested.
Conclusion
When chosen carefully, patients with core infarcts on the presenting CT imaging involving the basal ganglia or multiple territories (ASPECT ≤6) can achieve good clinical outcomes, similar to those of patients who had an entirely reversible penumbra at presentation. Also patients with left-sided infarctions can potentially achieve outcomes as good as those obtained by patients with right-sided infarctions. Thus, a treatment algorithm which excludes patients with basal ganglia infarcts, or one which employs hardline ASPECTS thresholds, might exclude patients who could potentially benefit from thrombectomy to a similar or equal degree as patients with less extensive core infarcts at presentation.
Acknowledgments
Alyssa Pierce assisted with the editing and revision of this manuscript.
References
Footnotes
Contributors Each author listed above should receive authorship credit based on the material contribution to this article, their revision of this article, and their final approval of this article for submission to this journal.
Competing interests AMS: Penumbra consulting, honorarium, speaker bureau; Pulsar Vascular consulting, honorarium, speaker bureau; MicroVention consulting, honorarium, speaker bureau, research; Stryker consulting, honorarium, speaker bureau. AST, RDT, and MIC: Codman consulting, honorarium, speaker bureau, research funding; Covidien consulting, honorarium, speaker bureau; Penumbra consulting, honorarium, speaker bureau, research grants; MicroVention consulting, honorarium, speaker bureau, research grants; Blockade stock, consulting, honorarium, speaker bureau; Pulsar Vascular stock, consulting, honorarium, speaker bureau, research; Medtronic consulting, honorarium, speaker bureau.
Ethics approval Medical University of South Carolina institutional review board.
Provenance and peer review Not commissioned; externally peer reviewed.