Article Text

Original research
Computed tomography-based triage of extensive baseline infarction: ASPECTS and collaterals versus perfusion imaging for outcome prediction
  1. Rosalie McDonough1,
  2. Sarah Elsayed1,
  3. Tobias Djamsched Faizy2,
  4. Friederike Austein1,
  5. Peter B Sporns3,
  6. Lukas Meyer1,
  7. Matthias Bechstein1,
  8. Noel van Horn1,
  9. Marie Teresa Nawka1,
  10. Gerhard Schön4,
  11. Helge Kniep1,
  12. Uta Hanning1,
  13. Jens Fiehler1,
  14. Jeremy J Heit5,
  15. Gabriel Broocks1
  1. 1 Department of Diagnostic and Interventional Neuroradiology, University Medical Center Hamburg-Eppendorf, Hamburg, Germany
  2. 2 Radiology, Stanford University School of Medicine, Stanford, California, USA
  3. 3 Department of Neuroradiology, University Hospital Basel, Basel, Switzerland
  4. 4 Institute of Medical Biometry and Epidemiology, University Medical Center Hamburg-Eppendorf, Hamburg, Germany
  5. 5 Radiology, Neuroradiology and Neurointervention Division, Stanford University, Stanford, California, USA
  1. Correspondence to Dr Rosalie McDonough, Department of Diagnostic and Interventional Neuroradiology, University Medical Center Hamburg-Eppendorf, 20246 Hamburg, Germany; r.mcdonough{at}


Background Patients presenting with large baseline infarctions are often excluded from mechanical thrombectomy (MT) due to uncertainty surrounding its effect on outcome. We hypothesized that computed tomography perfusion (CTP)-based selection may be predictive of functional outcome in low Alberta Stroke Program Early CT Score (ASPECTS) patients.

Methods This was a double-center, retrospective analysis of patients presenting with ASPECTS≤5 who received multimodal admission CT imaging between May 2015 and June 2020. The predicted ischemic core (pCore) was defined as a reduction in cerebral blood flow (rCBF), while mismatch volume was defined using time to maximum (Tmax). The pCore perfusion mismatch ratio (CPMR) was also calculated. These parameters (pCore, mismatch volume, and CPMR), as well as a combined radiological score consisting of ASPECTS and collateral status (ASCO score), were tested in logistic regression and receiver operating characteristic (ROC) analyses. The primary outcome was favorable modified Rankin Scale (mRS) at discharge (≤3).

Results A total of 113 patients met the inclusion criteria. The median ischemic core volume was 74.1 mL (IQR 43.8–121.8). The ASCO score was associated with favorable outcome at discharge (aOR 3.7, 95% CI 1.8 to 10.7, P=0.002), while no association was observed for the CTP parameters. A model including the ASCO score also had significantly higher area under the curve (AUC) values compared with the CTP-based model (0.88 vs 0.64, P=0.018).

Conclusions The ASCO score was superior to the CTP-based model for the prediction of good functional outcome and could represent a quick, practical, and easily implemented method for the selection of low ASPECTS patients most likely benefit from MT.

  • CT
  • CT perfusion
  • stroke
  • thrombectomy
  • intervention

Data availability statement

Data are available upon reasonable request.

Statistics from

Request Permissions

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.


Mechanical thrombectomy (MT) is a known effective approach for the treatment of acute ischemic stroke (AIS) due to large vessel occlusion (LVO) and has shown promising results in several patient-level meta-analyses for certain patient subgroups.1 However, the effect of treatment on patients presenting with extensive baseline infarction, defined as an Alberta Stroke Program Early CT Score (ASPECTS) ≤5, remains unclear and is the current topic of multiple, ongoing clinical trials such as TENSION, TESLA, In Extremis, and SELECT 2.2–6 Data from the HERMES trial showed a trend towards better outcome in patients with low ASPECTS following thrombectomy versus best medical treatment alone, however the results were not statistically significant. Indeed, fewer than 10% of the included HERMES trial patients presented with an ASPECTS≤5, making it difficult to draw valid conclusions for this subgroup.1 Furthermore, many of the trials used heterogenous imaging protocols and the results were likely largely magnetic resonance imaging (MRI)-driven,7–10 with favorable outcome more often observed in patients with initial MRI imaging compared with those who underwent computed tomography (CT).11 12 Despite being the modality of choice for acute stroke imaging, it currently remains uncertain whether any CT-based parameters are suitable for outcome prediction in patients presenting with extensive baseline infarction.

As a result, the decision to treat is often made based on the combined consideration of individual patient and imaging characteristics, further influenced by clinician experience and clinical routine. These include age, comorbidities, elapsed time between symptom onset and imaging, expectation for patient outcome, and the extent of infarcted tissue.13 Perfusion imaging has also emerged as the standard modality for therapy selection, allowing for the theoretical identification of salvageable tissue in relation to the volume of infarcted core.14 Those with a larger mismatch between core and penumbra are thought to be better candidates for MT.14 15 However, the determination of these parameters has been repeatedly shown to be both threshold and software-dependent, requiring most centers to develop their own algorithms.16 Nevertheless, a recent study using MRI perfusion (MRP)-based selection of patients with low ASPECTS (0–6) and a diffusion-weighted imaging (DWI)-defined core >70 mL reported promising results; in their study cohort, patients with a threshold-based large mismatch profile had a higher chance of favorable outcome following MT than those with smaller mismatch ratios.15

The goal of this study was to determine whether CT perfusion (CTP) parameters can also be employed for outcome prediction in patients with low ASPECTS. We hypothesized that in addition to other image-based parameters, CTP imaging increases the accuracy of outcome prediction following MT in patients with large baseline infarctions.


Study design and patient collective

This investigation represents a double-center, retrospective, observational cohort study. All patients who presented with AIS due to LVO of the anterior circulation between May 2015 and June 2020 were consecutively screened for inclusion with the following criteria: (1) age >18 years, (2) multimodal CT (non-contrast CT (NCCT), CT angiography (CTA), and CTP) imaging on admission, (3) an ASPECTS≤5, and (4) documented National Institutes of Health Stroke Scale (NIHSS). Baseline patient characteristics, as well as functional outcome (assessed by the modified Rankin Scale, mRS), were collected at discharge and at the 3-month follow-up timepoint (mRS90), if available.

Study protocols and procedures were conducted in compliance with the Declaration of Helsinki and in accordance with local ethical guidelines. Due to the retrospective study design and the use of anonymized patient data, the need for informed consent was waived by the ethics committee of the Hamburg Chamber of Physicians as well as the Stanford Institutional Review Board.

Imaging acquisition and interpretation

All patients received multimodal stroke imaging at admission with NCCT, CTA, and CTP performed in equal order on 256 or 384 dual-slice scanners (Philips iCT 256; Siemens Somatom Force, Erlangen, Germany). NCCT: 120kV, 280 to 340mA, 5.0 mm slice reconstruction, 1 mm increment; CTA: 100kV, 260 to 300mA, 5.0 mm slice reconstruction, 1 mm increment, 80 mL highly iodinated contrast medium and 50 mL NaCl flush at 4 mL/s; CTP: 80kV, 200 to 250mA, 5 mm slice reconstruction (maximum 10 mm), slice sampling rate 1.50 s (minimum 1.33 s), scan time 45 s (maximum 60 s), biphasic injection with 30 mL (maximum 40 mL) of highly iodinated contrast medium with 350 mg iodine/mL (maximum 400 mg/mL) injected with at least 4 mL/s (maximum 6 mL/s) followed by 30 mL sodium chloride chaser bolus, whole-brain coverage of 12 cm. All perfusion datasets underwent quality control and were excluded in case of severe motion artifacts.

ASPECTS was determined on the admission NCCT scans and was rated separately by two experienced neuroradiologists with subsequent consensus reading. Raw perfusion data were analyzed on a Siemens workstation using Syngo VPCT Neuro software (Siemens Healthcare, Erlangen, Germany) or with RAPID automated software (iSchemaView Inc, Menlo Park, CA, USA). For the former, quantitative maps of relative cerebralblood flow (rCBF), cerebral blood volume (CBV), mean transit time (MTT), and time to maximum (Tmax) were generated using a delay-insensitive algorithm. RAPID employs a rCBF threshold of ≤30% of the maximum for core volume determination; for comparability, we selected a predicted ischemic core (pCore) threshold of rCBF ≤20% for the Syngo software, as this has been described to have the highest level of agreement with RAPID-generated rCBF-based volumes.17 Hypoperfusion volume was determined to be the volume of tissue with a prolonged Tmax of at least 6 s (Tmax >6 s). Mismatch (penumbral) volume was defined as the difference between this volume and pCore. Finally, the pCore to perfusion mismatch ratio (CPMR) was calculated to be the hypoperfusion volume divided by the pCore volume.

Collaterals were assessed independently of clinical information on 20 mm single-phase CTA maximum intensity projections (MIP) and assigned a score of 0–4 according to Souza et al.18 Finally, a combined ASPECTS and collateral (ASCO) score was defined by the addition of the two (eg, ASPECTS 5 and collateral score 3 translated to an ASCO score of 8), with a theoretical range of 0–9.

Statistical analysis

The primary outcome was defined as mRS at discharge. This parameter has been shown to be associated with final clinical outcome and was chosen due to the fact that the 3-month follow-up was not as comprehensively available.19 To this end, a subanalysis testing for differences between the two timepoints was performed (see online supplemental material). For logistic regression analyses, mRS was binarized at ≤3 vs >3, as patients presenting with extensive baseline infarctions are known to have relatively poor outcomes.11 The results are presented as odds ratios (ORs) with 95% confidence intervals (CIs).

Univariable logistic regression was employed to examine the relationship between the primary outcome and the individual aforementioned perfusion parameters, as well as to identify statistically relevant clinical and imaging patient characteristics. These variables were also tested in univariable receiver operating characteristic (ROC) curve analyses with favorable outcome (mRS ≥3) as the dependent variable. Multivariable logistic regression with backward selection was then performed. Variables were assessed for collinearity and each model was adjusted for age, sex, NIHSS score, and successful MT to minimize potentially confounding effects. Three fully adjusted models representing CTP parameters (pCore, mismatch volume, CPMR, ‘Model 1’), radiographic features (ASCO score, ‘Model 2’), and a nested clinical model incorporating age, sex, NIHSS, and successful MT (‘Model 3’) were constructed and compared in multivariable ROC curve analyses. Differences in the area under the curve (AUC) were assessed using Delong’s test.

Finally, because reperfusion status has a known influence on outcome, a subanalysis restricted to patients with Thrombolysis in Cerebral Infarction scale (TICI) 2b/3 (n=53) analogous to the above was also performed (see online supplemental material and online supplemental figure II).

All analyses were performed with the R statistics program (v.3.5.2R; Core Team 2019, Vienna, Austria; RStudio IDE v. 1.1.463; Boston, MA, USA).20 Figures were created using the ggplot2 grammar of graphics. Normally distributed variables are displayed as mean and standard deviation (SD). Non-normally distributed data are displayed as median and interquartile range (IQR). Categorical variables are reported as proportions. P values<0.05 were considered significant.


Patient collective

A total of 113 patients fulfilled the inclusion criteria. Eighty-one (71%) underwent MT, of which 55 (68%) achieved a TICI score of 2b/3. Fifty-eight patients (51%) received intravenous (IV) recombinant tissue plasminogen activator (rtPA). No differences were observed in the time from symptom onset to admission (mean 3 vs 2.75 hours, P=0.120), nor in the vessel or side of occlusion between the two treatment groups (successful (TICI 2b/3) vs unsuccessful MT/no treatment). Patients selected for thrombectomy were slightly younger (mean 73 vs 79 years, P=0.018) and had higher ASPECTS scores (5 (IQR 4–5) vs 4 (IQR 2–5), P=0.014). Collateral status also tended to be better, however this was not significant (score ≥2 47.3% vs 29.3%, P=0.053). The ASCO score was in turn higher for the successfully recanalized group, with a median value of 6 (4.5–6.5) vs 5 (3–6) (P=0.011) (figure 1A). The average age of the total cohort was 76.2 (range 34–98) years and 55.3% were female (table 1).

Figure 1

Boxplots illustrating the relationship between (A) ASCO (Alberta Stroke Program Early CT Score (ASPECTS) and collateral) score and modified Rankin Scale (mRS) at discharge and (B) predicted ischemic core (pCore) volume and mRS at discharge.

Table 1

Patient characteristics and outcome by reperfusion status

The median pCore volume was 74.1 mL (IQR 43.8–121.8) and did not differ significantly between patients who underwent successful MT and those who did not (94.7 (IQR 45.5–133) vs 70.3 mL (IQR 37.3–110), respectively, P=0.185) (figure 1B). Mismatch volume (and, in turn, hypoperfusion volume), however, was larger in the successfully recanalized group (85.2 (IQR 57–147) vs 52.6 mL (IQR 24.9–88.9), P<0.001), as was CPMR (2.08 (IQR 1.58–3.54) vs 1.6 (IQR 1.29 vs 3.29), P=0.033) (table 1).

mRS90 was documented for 84 patients (74.3%) and mRS at discharge for 96 (85%). Favorable outcome (mRS 0–3) at discharge was only observed in 14% of all cases; however, slightly higher rates were observed in the successfully recanalized group (18.9% vs 4.65%, P=0.013). This trend held true for the later follow-up time point (mRS90), with 16.7% favorable outcome in those achieving vessel patency versus 2.78% in those who did not (P<0.001). Higher rates of complication (eg, intracranial hemorrhage) were also observed in the endovascularly treated group, however this was not significant (22.5% vs 13.5%, P=0.392). The overall mortality rate was 41.7%.

Logistic regression

Univariable analysis

Univariable logistic regression analyses of the entire cohort showed that unfavorable outcome was significantly associated with higher age, lower ASPECTS, female sex, poor collateral score, and lower ASCO score (table 2). Furthermore, early neurological improvement (ENI, defined by the percent change in NIHSS between admission and 24 hours21) was positively associated with favorable outcome (P=0.004). None of the individual perfusion parameters had a significant impact on outcome at any time point, regardless of treatment category. Linear regression of pCore and ASCO score showed a significant association (P<0.0001); for every increase in ASCO, there is an approximately 16 mL decrease in pCore. ASCO score was also significantly associated with outcome (P<0.0001), while no such relationship was observed between pCore and mRS at discharge (P=0.382) (online supplemental figure I).

Table 2

Univariable and multivariable receiver operating characteristic curve analyses

Multivariable analysis

A multivariable logistic regression model of the above-named variables was assembled; due to collinearity, collateral status was excluded. Only ASCO (aOR 3.6, 95% CI 1.5 to 11.6, P=0.012) and ENI (aOR 1.02, 95% CI 1.0 to 1.05, P=0.03) were observed to be significant and remained so following adjustment for successful MT (figure 2). Backward selection revealed sex, ASCO score, and ENI to be the most relevant candidate variables. In this reduced model, all three were significantly associated with outcome (female sex: aOR 0.06, 95% CI 0.004 to 0.4, P=0.01; ASCO score: aOR 3.7, 95% CI 1.8 to 10.7, P=0.002; ENI: aOR 1.02, 95% CI 1.0 to 1.05, P=0.012).

Figure 2

Probability curves for favorable outcome (defined as mRS discharge ≥3), stratified according to successful recanalization and ASCO score. ASCO, Alberta Stroke Program Early CT Score (ASPECTS) and collateral; mRS, modified Rankin Scale; TICI, Thrombolysis in Cerebral Infarction.

ROC analysis

Univariable analysis

In univariable ROC analyses, ASCO (AUC 0.88, 95% CI 0.80 to 0.95) and collaterals (AUC 0.80, 95% CI 0.71 to 0.89) had the highest AUC values of the variables found to be significantly associated with favorable outcome, followed by ASPECTS (AUC 0.77, 95% CI 0.68 to 0.86), ENI (AUC 0.76, 95% CI 0.61 to 0.90), age (AUC 0.71, 95% CI 0.57 to 0.86), and sex (AUC 0.70, 95% CI 0.58 to 0.82, table 2)). According to Delong’s test, the diagnostic power of ASCO was statistically different than collateral status (P=0.01). Using an ASCO cut-off of ≤5, sensitivity and specificity were 100% and 61.2%, respectively. The CTP parameters, conversely, had the lowest AUC values of 0.45 (95%CI 0.27 to 0.62), 0.65 (95%CI 0.46 to 0.83), and 0.62 (95%CI 0.43 to 0.82), for pCore, mismatch volume, and CPMR, respectively.

Multivariable analyses

The AUC of a model consisting of the ASCO score and one of pCore, mismatch volume, and CPMR were 0.88 (95%CI 0.80 to 0.95) and 0.64 (95%CI 0.45 to 0.84), respectively. These models were then each adjusted for age, sex, and NIHSS and again tested for their ability to predict favorable outcome. The AUC for the adjusted Model 1 was 0.87, 95% CI 0.74 to 1.00), while the AUC of the nested Model 3 was 0.82 (95%CI 0.68 to 0.97) (table 2, figure 3). Indeed, Delong’s test revealed that the two models did not differ significantly from one another (P=0.182). Conversely, the AUC for the adjusted Model 2 was 0.94 (95%CI 0.87 to 0.99) and differed significantly from both the adjusted Model 1 and Model 3 (P=0.0175 and P=0.0276, respectively, figure 3).

Figure 3

Receiver operating characteristic (ROC) curve analysis of the perfusion and ASCO score models, compared with the base model of patient characteristics/therapy effect (Model 3). The area under the curve (AUC) of Model 2 was significantly better than that of Models 1 and 3, while Model 1 did not differ significantly from Model 3. ASCO, Alberta Stroke Program Early CT Score (ASPECTS) and collateral; MT, mechanical thrombectomy; NIHSS, National Institutes of Health Stroke Scale.


To the best of our knowledge, this is this first study that to investigate the role of CTP for outcome prediction following MT in patients with large baseline core infarctions. In a study by Sarraj et al,22 the influence of favorable imaging profiles (either NCCT or CTP-driven) on the decision to treat was investigated. Here, it was found that concordance of the two modalities (eg, ASPECTS≥6 and core lesion <70 mL) more often led the neurointerventionalist to proceed with thrombectomy, with resulting higher rates of favorable outcome. In discordant patient profiles with unfavorable CTP, however, higher rate of adverse outcome following MT were observed, suggesting a higher reliability of NCCT findings. Indeed, although commonly used for patient treatment selection, CTP-based determination of infarct core lesion size is often overestimated. This ‘ghost infarct core’ phenomenon could lead to the erroneous exclusion of patients who would have otherwise benefited from MT.23

Despite the generally poor outcome profile of the entire cohort, mRS at discharge and mRS90 of patients in which TICI 2b/3 recanalization was achieved were significantly better (figures 2 and 4), with no significant differences observed for feared thrombectomy-associated complications (eg, intracranial hemorrhage). This confirms observations from previous retrospective studies in which it was found that while still achievable in patients with ASPECTS≤5, the likelihood of a good outcome remains lower than in those with ASPECTS≥6.24 Nevertheless, when taken together, these results suggest that patients with large baseline lesions still have salvageable brain tissue.

Figure 4

Computed tomography imaging example of a patient with a large pCore volume, however favorable ASCO score (upper images). The follow-up imaging demonstrates minimal residual hypodensity following successful mechanical thrombectomy. In contrast, a patient with a relatively small pCore lesion and unfavorable ASCO profile (lower images). Despite successful recanalization, the patient had extensive infarction and died a few days later. ASCO, Alberta Stroke Program Early CT Score (ASPECTS) and collateral; CT, computed tomography; mRS, modifed Rankin Scale; NCCT, non-contrast CT; pCore, predicted ischemic core; rCBF, reductionin cerebral blood flow ; TICI, Thrombolysis in Cerebral Infarction.

Interestingly, the parameter that appeared to be the most robust for outcome prediction in our analysis was the ASCO score. Higher ASPECTS and good collateralization have been described to be predictive of MT recanalization status and are therefore associated with good outcome.25 Although perhaps intuitive, the exact nature of the physiological relationship between the two is not completely understood. ASPECTS represents the visual/radiological evidence of tissue hypoattenuation, which is directly related to early net water uptake of the hypoperfused tissue.26 Conversely, collateral status may depict a further dimension. Perhaps the combination of the two parameters could increase the temporal resolution of certain stages of infarct progression and therefore serve as a more reliable predictor of potential tissue salvage. This is perhaps particularly important when considering the relatively low interrater reliability for ASPECTS scoring, which could further skew the decision-making process when considered alone.27

Although patients with a low initial ASPECTS often have poor intracranial collaterals, it has been described that as many as 30% have a collateral score of 2–4, which has been shown to be favorable for treatment in patients with favorable ASPECTS.28 Furthermore, differentiation within the lower ASPECTS range may be of importance, as patients with very low ASPECTS (particularly 0–2) are unlikely to benefit from MT. Indeed, this is a topic of research in current randomized low ASPECTS trials.2 3 In another recent study by Desai et al,29 it was shown that while the clinical-core mismatch (CCM) decreased with both increasing time and decreasing ASPECTS, the prevalence of CCM within each ASPECTS group remained stable, supporting the use of ASPECTS as a robust imaging tool for the late time window. The proposed ASCO score incorporates both features and was highly predictive of functional outcome in this patient cohort. Despite the inclusion criterion of ASPECTS≤5, there was still a significant association between higher ASPECTS score and favorable outcome. Based on the ROC analysis threshold of ASCO≥5 from this study, patients with ASPECTS≤1 would not be expected to achieve improved outcome, while those with an ASPECTS of 2 would likely only profit from MT if very good collateralization is present.28 Conversely, patients with ASPECTS in the range 3–5 and intermediate to good collaterals would have a greater probability of favorable outcome.22

To better tease apart this relationship, it would be interesting to include an evaluation of quantitative net water uptake on the admission images, to test whether patients with low ASPECTS however low ischemic water uptake may particularly benefit from MT.30 This could potentially strengthen the predictive ability of the model and further aid in patient selection for treatment.

Finally, the fact that we were unable to replicate the results of previous studies is, at least in part, likely due to the use of different modalities. MRP only provides relative values of blood flow or transit time, and time-based parameters such as Tmax and MTT tend to be overestimated.31 Furthermore, DWI determination of the infarct core has repeatedly been shown to more accurately represent the irreversibly damaged tissue when compared with threshold-based CTP parameters.32 Although it perhaps provides a more accurate picture of the extent of hypoperfusion, high DWI signal can also be partially or even completely reversible in early time windows. Together, these factors could potentially skew the ratio between penumbra and core, making the transferability of mismatch ratios to other modalities difficult. CTP, on the other hand, offers a quantitative approach and is much more widely available. This, combined with its speed, is the reason it has remained the modality of choice in the acute stroke setting.33 Nevertheless, many smaller hospitals are not equipped to perform perfusion analyses. The ASCO score provides a convenient, CTP-independent scoring measure for potential outcome prediction.

Another aspect is the chosen definition of low ASPECTS. Based on the current literature, we decided to set relatively stringent inclusion parameters to better understand the treatment effect, if any, in patients with severe baseline infarctions. The recently published MRI-based study defined low ASPECTS as ≤6, a 1-point change that could lead to substantial differences in outcome and its potentially associated predictor variables, particularly when comparing MRI and CT. Taken together, we feel that our study more accurately depicts both the patient population and the imaging modality used to determine the initial extent of ischemic stroke lesions.


This study has several limitations. First, despite being a double-center initiative, a relatively small number of patients were included. Few patients with low ASPECTS receive multimodal CT imaging and even fewer undergo MT (81 in this study). This is likely due to the uncertainty surrounding the efficacy and safety of treatment in this particular subgroup. Second, the recruited patients at the participating centers likely underwent a form of ‘pre-selection’ by the attending physician(s) and tended to have both better clinical and imaging profiles. Third, CTP volumes and ratios were calculated using two different software platforms, Syngo VPCT Neuro software and RAPID, which can lead to heterogeneity of the data. However, both offer delay-insensitive algorithms and have been shown to have a good level of agreement for the calculated core lesion volumes (and therefore mismatch profiles), following appropriate adjustment of the relevant parameters (eg, rCBF ≤20%).17 34 This situation also reflects real-world scenarios, where not all hospitals have access to the same software. Furthermore, much of the MRI-based study data are generated from different scanners and at different field strengths. In contrast to CT scanners, which are similarly calibrated, this can also lead to greater variation in the data. Finally, due to the recruitment of patients within the relatively recent past, mRS90 was not as comprehensively available, leading us to choose mRS at discharge as the primary outcome. It has been shown, however, that these two timepoints are well-correlated.19


Our study presents a quick, reliable, and easily implemented scoring system for outcome prediction following MT in patients with low ASPECTS. In contrast, no association between the RAPID software-based CTP volumes and functional outcome was observed. Their application as a treatment selection tool for MT in patients with large ischemic cores should therefore be considered with caution to avoid the wrongful exclusion of patients who would have otherwise benefited from revascularization. The determination of the CTP-independent ASCO score, consisting of ASPECTS and collateral status, not only avoids the known pitfalls of perfusion imaging, it also saves time and could be of particular use in smaller hospitals where perfusion imaging may not be available.

Data availability statement

Data are available upon reasonable request.

Ethics statements


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • Twitter @rosevmcd, @noelvanhorn, @JeremyHeitMDPHD

  • Contributors Conception/design of work: GB. Data collection: RM, SE, TDF, JJH, PBS, Fabian Flottmann, NvH, MTN, MB, LM. Data analysis and interpretation: RM, GB, HK, GS, FA, UH, LM. Drafting the article: RM. Critical revision of the article: RM, SE, TDF, FA, PBS, LM, MB, NvH, MTN, GS, HK, UH, JF, JJH, GB. Final approval of the version to be published: GB, JF. Agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved: RM, SE, TDF, FA, PBS, LM, MB, NvH, MTN, GS, HK, UH, JF, JJH, GB.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests JF reports personal fees as a consultant for Microvention, Stryker, Cerenovus, Acandis, Penumbra, and Medtronic outside the submitted work. He is a member of the executive boards of the German Society of Neuroradiology (DGNR) and the European Society of Minimally Invasive Neurological Therapy (ESMINT).

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.