Article Text
Abstract
Background Clot density (Hounsfield units, HU) and perviousness (post-contrast increase in the HU of clot) are thought to be associated with clot composition. We evaluate whether these imaging characteristics were associated with angiographic outcomes of aspiration and stent retriever thrombectomy in COMPASS: a trial of aspiration thrombectomy versus stent retriever thrombectomy as first-line approach for large vessel occlusion.
Methods Clot density and perviousness were measured by two independent operators who were blind to all the final angiographic and clinical outcomes. The association of clot density and perviousness with the Thrombolysis In Cerebral Infarction (TICI) scale after first pass was assessed using univariate and multivariate analysis.
Results Among all patients enrolled in COMPASS, 165 were eligible for the post-hoc analysis (81 patients in the aspiration first and 84 in the stent retriever first groups). Overall mean perviousness of clot was significantly higher in patient with mTICI 2b-3 after first pass (28.6±22.9 vs 20.3±19.2, p=0.017). Mean perviousness among patients who achieved TICI 2c/3 versus TICI 2b versus TICI 0-2a in the aspiration first group varied significantly (32.6±26.1, 35.3±24.4, and 17.7±13.1, p=0.013). The association of perviousness with first pass success was not significant in the stent retriever group. Using multivariate analysis, high perviousness (defined as cut-off >27.6) was an independent predictor of TICI 2b-3 (OR 3.82, 95% CI 1.10 to 13.19; p=0.034).
Conclusions Clot perviousness is associated with first pass angiographic success in patients treated with the aspiration first approach for thrombectomy.
- CT angiography
- thrombectomy
- stroke
Data availability statement
Data are available upon reasonable request. Requests for data availability should be addressed to COMPASS principal investigators.
Statistics from Altmetric.com
Introduction
Both aspiration and stent retriever thrombectomy techniques as first-line approach have proven to be safe and effective for the treatment of emergent large vessel occlusion (ELVO) in patients with acute ischemic stroke.1 2 Several studies have been conducted to identify imaging characteristics predictive of thrombectomy success with either treatment, such as clot density, clot burden score and location of occlusion on baseline imaging computed tomography (CT) or magnetic resonance imaging (MRI).3–5 Studies on the association of these clot characteristics with procedural and clinical outcomes so far have not been able to determine which patients are more likely to achieve successful recanalization, between the two types of thrombectomy techniques.
There are conflicting data on the association of clot density measured on non-contrast CT (NCCT) with angiographic outcomes of patients treated with various types of thrombectomy devices.6 More recently, clot perviousness (also referred to as permeability or porosity), which is defined as a difference in the clot density values between CT angiography (CTA) and NCCT, was proposed as a new determinant of acute stroke treatment.7–9 Clot perviousness was associated with higher red blood cell density and lower fibrin density.10 Current understanding of how the clot perviousness affects the success rates of modern thrombectomy devices for the treatment of ELVO is limited. Prior retrospective analyses of patients treated with stent retriever thrombectomy failed to predict the first-pass effect (FPE) or successful revascularization depending on clot perviousness values.11 The goal of this study was to examine whether clot perviousness is associate with angiographic outcomes of first pass aspiration and stent retriever thrombectomy using COMPASS: a trial of aspiration thrombectomy versus stent retriever thrombectomy as first-line approach for large vessel occlusion.1
Methods
COMPASS study design
The study protocol and consent were approved by the institutional review board at each participating center. The methodology and final results of the COMPASS trial have been previously published.12 Briefly, COMPASS was a prospective, multicenter, randomized, open-label, blinded outcome, core lab adjudicated, concurrent control, non-inferiority trial to assess whether patients treated with aspiration as first pass had non-inferior functional outcomes when compared to those treated with a stent retriever as first line. Patients presenting with acute ischemic stroke from ELVO within the first 6 hours of symptom onset, an Alberta Stroke Program Early CT Score (ASPECTS) ≥6, and occlusion of the internal carotid artery (ICA) or middle cerebral artery (MCA) M1 segment based on CTA or MR angiography (MRA) were selected for the trial. After participants were randomly assigned (1:1) by a central web-based system to treatment with either a direct aspiration as first pass thrombectomy (‘aspiration first’) or stent retriever (‘stent retriever first’), operators were requested to do three passes with the assigned approach. In cases of persistent occlusion, they were allowed to use any therapy according to operator preference. The primary trial outcome was functional independence at 90 days measured by achieving the modified Rankin Scale (mRS) score of 0–2. Angiographic assessments of thrombectomy procedures in COMPASS were done by an independent core lab in a blinded manner using the Thrombolysis In Cerebral Infarction (TICI) grading scale (TICI 2b defined as 50–89% reperfusion; TICI 2c as 90–99% reperfusion, and TICI 3 as 100% reperfusion, respectively). Angiographic outcome after the first pass of the primary thrombectomy device was used for analysis. FPE was defined as TICI 2c/3 recanalization after a single (first) pass/use of the treatment device (aspiration catheter in the aspiration first group and stent retriever in the stent retriever first group).13 Modified FPE (mFPE) was defined as TICI 2b-3 recanalization after a single use of the treatment device.14
Clot density and perviousness
These measurements were not collected by the initial core lab. Clot density and perviousness were measured subsequently by two independent operators blinded to the knowledge of final angiographic or clinical outcome except for knowledge of stroke laterality as previously described.3 We excluded cases with artifact, poor contrast opacification on CTA or baseline MRA (figure 1). Regions with Hounsfield units (HU) values >100, which represented calcifications, were excluded.15 NCCT and CTA images were co-registered and aligned to accurately determine the location of the clot on both imaging modalities. Three regions of interest were placed in the proximal, middle and distal part of the thrombus to calculate the mean HU density value of the affected side (figure 2A). The contralateral HU mean clot density values were collected to correct for variability in hematocrit levels. Relative clot density (rHU) was calculated by dividing the mean HU value of the affected side by the mean HU value of the contralateral side.16 Clot perviousness was measured by subtracting the mean attenuation values of the thrombus on NCCT from CTA. Clot length was calculated by co-registering NCCT with CTA to identify the proximal and distal ends of the clot. The longest clot measurement was used for analysis in cases of clot extending into multiple M2 branches.
Statistical analysis
A descriptive analysis was performed. Categorical data were presented as percentage and proportion. We assessed the distribution of data using the Shapiro-Wilk normality test. Continuous data with normal distribution were presented as means and SD. We used median and interquartile ranges for data with skewed distribution. Univariate analysis was performed with independent sample t test (continuous data), χ2 test (categorical data), and Fischer exact test as appropriate. Analysis of variance (ANOVA) was used to compare means between three groups (3 degrees of reperfusion after first pass, that is, TICI 0-2a, 2b, 2 c-3). We performed receiver operator curve (ROC) analysis to estimate the predictive value of perviousness for FPE and mFPE. An appropriate cut-off was calculated using the Youden index (J) on the curve with significant area under the curve. The highest value of the Youden index (J) was used to dichotomize the data into two groups, that is, high perviousness and low perviousness. The association of high and low perviousness with FPE and mFPE was assessed for each of the two randomization groups. Multivariate regression analysis was carried out to assess the independent predictive value of high perviousness for FPE and mFPE. The statistical analysis was performed using SPSS version 25.
Results
Of the original 270 patients included in the COMPASS trial, 165 were eligible for the post-hoc analysis (81 patients in the aspiration first and 84 in the stent retriever first groups). Reasons for exclusion of the remainder 105 patients are shown in figure 1. Table 1 describes baseline demographic, clinical, imaging and procedural characteristics of the two treatment groups. There was no difference between the aspiration first and stent retriever first groups in regards to gender (female sex, 57% and 51%, p=0.47), age (71.1±13.1 and 70.5±14.1, p=0.76), baseline National Institutes of Health Stroke Scale (NIHSS) (16.4±5.8 and 17.0±6.5, p=0.51), location of large vessel occlusion (p=0.73), or comorbidities. The use of intravenous alteplase (64% and 67%, p=0.74), time from symptom onset to groin puncture (219±80 and 211±90 min, p=0.52), and use of a balloon guide (35% and 38%, p=064) was also similar between the two groups. The two treatment groups were also similar in regards to baseline clot HU on NCCT (49.8±8.2 vs 48.0±8.5, p=0.11), rHU (ratio of the clot HU and HU on contralateral site, 1.35±0.31 vs 1.3±0.29, p=0.20) and perviousness (26.8±21.8 vs 22.8±19.9, p=0.065).
In the entire cohort of 165 patients, clot density measured with HU on NCCT was similar between patients with FPE or no FPE (48.6±8.9 vs 49.2±8.0, p=0.70) or with mFPE versus no mFE (49.1±8.2 vs 48.8±8.6, p=0.82). The same trend was seen for aspiration first (FPE vs no FPE; 49.3±8.61 vs 50.3±7.85, p=0.59, and for mFPE vs no mFPE; 49.4±7.8 vs 50.9±8.7, p=0.43) and stent retriever first cohorts (FPE vs no FPE; 48.0±9.3 vs 48.1±8.0, p=0.94, and for mFPE vs no mFPE; 48.6±8.7 vs 47.4±8.3, p=0.50). rHU in patients with FPE using aspiration or stent retriever first was similar when compared with patients with no FPE (1.28±0.31 vs 1.39±0.31, p=0.12 for aspiration first; 1.29±0.24 vs 1.29±0.32, p=0.98 for stent retriever first). Similarly, no difference was seen for mFPE within both groups (1.32±0.33 vs 1.39±0.27, p=0.40 for aspiration first, 1.27±0.23 vs 1.31±0.35, p=0.60 for stent retriever first).
Overall mean clot perviousness was significantly higher in patients with mFPE compared with patients who did not have mFPE (28.6±22.9 vs 20.3±19.2, p=0.017). Mean perviousness for patients with and without FPE was 27.5±23.2 versus 23.8±20.9 (p=0.28). There was a significant difference in mean perviousness values among patients who achieved TICI 2 c/3 versus TICI 2b versus TICI 0-2a in the aspiration first group (32.6±26.1, 35.3±24.4, and 17.7±13.1, respectively; p=0.013; table 2 and figure 2B). However, there was no difference between clot HU density on NCCT among these three groups (mean HU values of 49.3±8.6, 49.3±6.6, and 51.1±8.8, p=0.64). In the stent retriever first group, there was no difference in either perviousness or HU density of clot in patients with TICI 2c/3, TICI 2b or TICI 0-2a after first pass (mean perviousness 22.3±18.7, 22.7±17.3, and 22.0±22.4, p=0.99; mean HU on NCCT 48.0±9.3, 50.4±6.8 and 47.4±8.3, p=0.55; table 2 and figure 2B).
The area under the curve (AUC) for clot perviousness as a predictor of FPE was not significant (AUC 0.54, 95% CI 0.45 to 0.63; p=0.35). For mFPE the AUC was significant at 0.61 (95% CI 0.52 to 0.70; p=0.017). Using the Youden index (J), we calculated the optimal perviousness cut-off to be 27.6 HU. Based on this value, 53 cases (53/165, 31.1%) had clots with a perviousness of >27.3 and were categorized as highly pervious. There was no significant difference between aspiration first and stent retriever first groups in regard to the number of patients with highly pervious clots (27/81 (33.3%) vs 26/84 (30.9%), p=0.74).
The distribution of TICI values achieved after first pass in patients with low (≤27.6) and high (>27.6) perviousness groups is shown in figure 2C and table 2. Perviousness was significantly associated with TICI 2b-3 successful reperfusion after first pass (mFPE) in the aspiration first group (p=0.012, figure 2D). This association was borderline significant for the stent retriever first approach (p=0.054). The association of perviousness was not significant for FPE in the aspiration first or stent retriever first group (p=0.20 and 0.39, respectively; figure 2D).
There was a weak negative correlation between clot length and perviousness (r=−0.104, p=0.23). In the aspiration first group, clot length values were similar in patients with and without FPE (9.3±3.9 mm and 10.3±4.6 mm, p=0.32) and in patients with and without mFPE (9.7±4.4 mm and 10.3±4.1 mm, p=0.57). In the stent retriever first group, there was also no significant difference in clot length according to recanalization success (11.64±4.4 mm with FPE and 11.4±4.9 mm without FPE, p=0.82, and 10.5±4.3 mm with mFPE and 13.0±4.9 mm without mFPE, p=0.055, respectively).
A 30 day mRS was available for all patients included in the study. We did not find a significant difference in mean perviousness between favorable and unfavorable outcome groups at 30 days (25.5±24.1 vs 25.15, p=0.92). A 90 day mRS was available for 158 (95.7%) patients. There was no significant difference in the mean perviousness between favorable and unfavorable outcome groups (24.8±23.4 vs 26.5±20.7, p=0.63). There was no significant association of high perviousness (>27.6 HU) with favorable clinical outcome (mRS 0–2) in the aspiration first (p=0.87) or stent retriever first groups (p=0.2) at 30 days. A similar trend was observed for 90 day outcomes with p=0.63 and 0.055 for aspiration first and stent retriever first groups, respectively.
Three variables (perviousness >27.6, age, and symptom onset to hospital arrival time) showed significant association with FPE or mFPE and thus were included in the multivariate regression model. High perviousness was an independent predictor of mFPE with an odds ratio of 3.82 (95% CI 1.10 to 13.19, p=0.034; table 3). We did not find any significant predictors of successful reperfusion (TICI 2b-3) after first pass in the stent retriever first group (table 3). We could not identify an independent predictor of FPE in the aspiration first or stent retriever first groups (table 3).
Discussion
Our main finding is that clot perviousness is associated with first pass angiographic success in patients treated with the aspiration first approach for thrombectomy, whereas the treatment success of the stent retriever first approach is less dependent on clot perviousness. We have shown that clot perviousness but not clot density measurement on NCCT may help distinguish ELVO patients who are more likely to have successful recanalization with aspiration thrombectomy alone. Because perviousness indicates how much iodine contrast may penetrate the clot, this imaging characteristic may provide information about clot composition.6 Highly pervious clots are shown to respond more effectively to intravenous alteplase, which is likely explained by an increased surface of lytic interaction within the clot. Such association was confirmed by several independent studies.9 17 18 However, establishing the association between clot imaging properties such as density or perviousness and the efficacy of mechanical thrombectomy has been elusive.6 8 11
Our findings are intriguing because using clot perviousness as a baseline imaging marker may allow the operator to select the technique with the highest chance of first pass successful recanalization. Currently, both aspiration and stent retriever first approaches share equal safety and efficacy profiles, and the choice of a particular technique is often based on individual operator preferences. The ability to predict whether the aspiration fist approach may or may not lead to successful recanalization has several potential advantages. Our data provide the first indication that highly pervious clots may result in higher treatment success after first pass using the aspiration first approach. However, according to our study, clots with low perviousness are more resistant to either thrombectomy approach, and further research is needed to determine the most effective treatment approach in this population of ELVO patients. Currently, there is no method of prompt assessment of clot perviousness; the available techniques are time consuming and need co-registration of NCCT and post-contrast CTA. Once the value of perviousness is validated in further studies, its automated measurement may become an integral part of rapid evaluation of patients with acute ischemic stroke to guide the neurointerventionalists on selecting the most effective treatment strategy. Automated ELVO detection, ASPECTS, and advanced perfusion quantitative analyses to predict clinical outcomes after thrombectomy are just some examples of swift technological advances in machine-learning imaging tools that have become an essential part of modern clinical practice.
The shorter procedure duration time has been associated with favorable clinical outcomes. With each thrombectomy attempt, there is evidence for thrombus compression, potentially making each further thrombectomy pass more challenging and less successful.19 Additionally, there are substantial financial advantages of the aspiration alone approach by avoiding the use of stent retrievers.1 Our findings may also be applied to future thrombectomy device innovations. As clots with low perviousness are less likely to demonstrate the first pass effect, improvement tailored towards this subgroup of ELVO is likely to have a more meaningful clinical impact.
Study limitations
Our study has several limitations that need to be taken into consideration when interpreting our findings. First, patients with symptom onset >6 hours were excluded from randomization into COMPASS. Thrombectomy in patients with symptom onset >6 hours was proven to be highly effective in two separate randomized controlled trials and has become the standard therapy for ELVO in this patient population.20 21 There is evidence of time-dependent changes of intracranial thrombi with regard to its density and area; therefore further research is needed to study imaging clot characteristics in patients treated with thrombectomy at a later time window.22
Second, we excluded patients with isolated ICA or tandem ICA and MCA occlusions from our analysis because of challenges in accurate measurement of density and perviousness of clots within ICA. It remains to be determined whether the association between perviousness and aspiration thrombectomy remains significant in this population of patients. Prior research indicates that isolated ICA and tandem ICA and MCA occlusions are associated with a significantly higher red blood cell and lower fibrin composition in comparison to isolated MCA occlusions.19
Third, due to a relatively small sample size, we may not have recognized the effect of perviousness on stent retriever thrombectomy. As shown in figure 2B, non-pervious clots (with perviousness values of 10 or less) showed a trend towards higher recanalization with first pass using stent retrievers; however, this did not reach statistical significance. Although prior studies failed to identify any significant association between stent retriever success and overall clot perviousness, further research using a large sample size and perviousness thresholds such as more or less than 10 may reveal additional important findings.11
Finally, other imaging characteristics such as clot volume and clot burden score, intracranial vessel tortuosity and angulation, and a particular thrombectomy device design (mode and type of aspiration, stent retriever design) may have various degrees of effect on the efficacy of thrombectomy depending on the value of perviousness. These factors were not addressed in our current study.
Conclusions
Clot perviousness is associated with first pass angiographic success in patients treated with the aspiration first approach for thrombectomy, which may help distinguish ELVO patients who are more likely to have successful recanalization with aspiration thrombectomy alone. Our findings suggest that clot perviousness as a baseline imaging marker may aid the operator in selecting the technique with the highest chance of first pass successful recanalization.
Data availability statement
Data are available upon reasonable request. Requests for data availability should be addressed to COMPASS principal investigators.
References
Footnotes
Twitter @rdeleacymd, @Ansaar_Rai, @donfreimd, @AdamArthurMD, @italolinfante
Contributors MM, MW, JM, AHS, AT– study concept and design. MM, MW wrote the manuscript. MM and CTP – statistical analysis. All authors edited the manuscript and approved the final version.
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 MM: Grant NIH R21NS109575. Consultant; Medtronic, Canon medical, Cerenovus. Stock options: Serenity medical, Synchron, VICIS, Endostream. MW: None. JF: Consultant: Stryker, Penumbra, Microvention. Stock: Cerebrotech, The Stroke Project. RDL: Consultant: Penumbra, Cerenovus, Siemens. DF: Consultant: Penumbra, Cerenovus, Stryker, Genentech, Shape Memory Medical. EIL: 6; Consultant: Penumbra, NextGen Biologics, Rapid Medical, Cognition Medical, Three Rivers Medical, Stryker, MedX, Endostream Medical. KS Consultant: Penumbra, Canon Medical Systems, Medtronic, Jacobs Institute, Neurovascular Diagnostics. RH: Consultant: Penumbra, Endostream, Cerebrotech, Synchron, InNeuroCo, Medtronic, Microvention, Stryker, Cerenovus; Elum, Three Rivers. KW: Consultant: Penumbra. IC: Consultant: Medtronic, Microvention, Penumbra. Stock: Cerenovus, Serenity Medical, Cerebrotech, Three Rivers Medical, Q’apel. ATR: Consultant: Penumbra, Microvention, Stryker. DF: Consultant: Penumbra, Cerenovus, Stryker, Genentech, Shape Memory Medical, Siemens. JEDA: Consultant: Penumbra, Medtronic. MK: Consultant: Penumbra, Medtronic, Endostream. ASA: Consultant: Cerenovus, Medtronic, Microvention, Penumbra, Scientia, Siemens, Stryker. Stock: Bendit, Cerebrotech, Endostream, Magneto, Marblehead, Neurogami, Serenity, Synchron, Triad, Vascular Simulations. BWB: Consultant: Penumbra, Medtronic, Stryker, Cerenovus, Viz.ai, 880 medical, Route 92, Artio Medical. JE: Consultant: Penumbra, Medtronic, Stryker, Route 92 Medical. IL: Consultant: Penumbra, Medtronic, Stryker, Microvention, InNeuroCo, Three Rivers. KF: None. AT: Consultant: Cardinal Consulting, Cerenovus, Corindus, Medtronic, Siemens, 880 Medical. Stocks: Cerebrotech, Endostream. Imperative Care, Three Rivers Medical, Vastrax, Shape Memory, Synchron, Serenity Medical, Blink TBI, Echovate, RIST, Apama, Q’Apel, VizAi, Early Bird Medical, Rapid Medical, Spinnaker Medical. AHS: Stocks: Amnis Therapeutics, Apama Medical, BlinkTBI, Inc, Buffalo Technology Partners, Inc, Cardinal Health, Cerebrotech Medical Systems, Inc, Claret Medical, Cognition Medical, Endostream Medical, Ltd, Imperative Care, International Medical Distribution Partners, Rebound Therapeutics Corp, Silk Road Medical, StimMed, Synchron, Three Rivers Medical, Inc, Viseon Spine, Inc. Consultant/Advisory Board: Amnis Therapeutics, Boston Scientific, Canon Medical Systems USA, Cerebrotech Medical Systems, Inc, Cerenovus, Claret Medical, Corindus, Endostream Medical, Ltd, Guidepoint Global Consulting, Imperative Care, Integra, Medtronic, MicroVention. JM: Stock options: Cerebrotech, Imperative Care, Endostream, Viseon, BlinkTBI, Serenity, Cardinal Consulting, NTI, RIST, Viz.ai, Synchron. Consultant: Imperative Care, Cerebrotech, VIseon, Endostream, Vastrax, RIST, Synchron, Viz.ai, Perflow, CVAid.
Provenance and peer review Not commissioned; externally peer reviewed.