Stroke ORIGINAL CONTRIBUTION Endovascular Treatment Effect Diminishes With Increasing Thrombus Perviousness Pooled Data From 7 Trials on Acute Ischemic Stroke Manon Kappelhof , MD; Manon L. Tolhuisen , MSc; Kilian M. Treurniet, MD; Bruna G. Dutra , MD; Heitor Alves, MD; Guang Zhang, MD; Scott Brown, PhD; Keith W. Muir, MD, PhD; Antoni Dávalos, MD, PhD; Yvo B.W.E.M. Roos, MD, PhD; Jeffrey L. Saver, MD, PhD; Andrew M. Demchuk, MD, PhD; Tudor G. Jovin, MD; Serge Bracard, MD, PhD; Bruce C.V. Campbell, MD, PhD; Aad van der Lugt, MD, PhD; Francis Guillemin, MD, PhD; Philip White, MD, PhD; Michael D. Hill, MD, PhD; Diederik W.J. Dippel, MD, PhD; Peter J. Mitchell, MD, PhD; Mayank Goyal, MD, PhD; Henk A. Marquering, PhD; Charles B.L.M. Majoie, MD, PhD; on behalf of the HERMES Collaborators BACKGROUND AND PURPOSE: Thrombus perviousness estimates residual flow along a thrombus in acute ischemic stroke, based on radiological images, and may influence the benefit of endovascular treatment for acute ischemic stroke. We aimed to investigate potential endovascular treatment (EVT) effect modification by thrombus perviousness. METHODS: We included 443 patients with thin-slice imaging available, out of 1766 patients from the pooled HERMES (Highly Effective Reperfusion Evaluated in Multiple Endovascular Stroke trials) data set of 7 randomized trials on EVT in the early window (most within 8 hours). Control arm patients (n=233) received intravenous alteplase if eligible (212/233; 91%). Intervention arm patients (n=210) received additional EVT (prior alteplase in 178/210; 85%). Perviousness was quantified by thrombus attenuation increase on admission computed tomography angiography compared with noncontrast computed tomography. Multivariable regression analyses were performed including multiplicative interaction terms between thrombus attenuation increase and treatment allocation. In case of significant interaction, subgroup analyses by treatment arm were performed. Our primary outcome was 90-day functional outcome (modified Rankin Scale score), resulting in an adjusted common odds ratio for a one-step shift towards improved outcome. Secondary outcomes were mortality, successful reperfusion (extended Thrombolysis in Cerebral Infarction score, 2B–3), and follow-up infarct volume (in mL). RESULTS: Increased perviousness was associated with improved functional outcome. After adding a multiplicative term of thrombus attenuation increase and treatment allocation, model fit improved significantly (P=0.03), indicating interaction between perviousness and EVT benefit. Control arm patients showed significantly better outcomes with increased perviousness (adjusted common odds ratio, 1.2 [95% CI, 1.1–1.3]). In the EVT arm, no significant association was found (adjusted common odds ratio, 1.0 [95% CI, 0.9–1.1]), and perviousness was not significantly associated with successful reperfusion. Follow-up infarct volume (12% [95% CI, 7.0–17] per 5 Hounsfield units) and chance of mortality (adjusted odds ratio, 0.83 [95% CI, 0.70–0.97]) decreased with higher thrombus attenuation increase in the overall population, without significant treatment interaction. CONCLUSIONS: Our study suggests that the benefit of best medical care including alteplase, compared with additional EVT, increases in patients with more pervious thrombi. GRAPHIC ABSTRACT: An online graphic abstract is available for this article. Key Words: cerebral infarction ◼ decision making ◼ reperfusion ◼ stroke ◼ thrombectomy ◼ tissue-type plasminogen activator Correspondence to: Manon Kappelhof, MD, Department of Radiology and Nuclear Medicine, G1-230 Amsterdam UMC, Meibergdreef 9, 1105 AZ Amsterdam. Email m.kappelhof@amsterdamumc.nl This article was sent to Scott E. Kasner, Guest Editor, for review by expert referees, editorial decision, and final disposition. The Data Supplement is available with this article at https://www.ahajournals.org/doi/suppl/10.1161/STROKEAHA.120.033124. For Sources of Funding and Disclosures, see page 3640. © 2021 The Authors. Stroke is published on behalf of the American Heart Association, Inc., by Wolters Kluwer Health, Inc. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution, and reproduction in any medium, provided that the original work is properly cited. Stroke is available at www.ahajournals.org/journal/str Stroke. 2021;52:3633–3641. DOI: 10.1161/STROKEAHA.120.033124 November 2021   3633 Kappelhof et al Thrombus Perviousness in Endovascular Treatment METHODS CLINICAL AND POPULATION Nonstandard Abbreviations and Acronyms Patients SCIENCES CTA computed tomography angiography The HERMES collaboration pooled individual patient data from EVT endovascular treatment 7 randomized controlled trials on EVT in acute ischemic stroke of the anterior circulation,9,13–19 totaling 1766 patients included FIV follow-up infarct volume between December 2010 and December 2014. Inclusion crite- HERMES Highly Effective Reperfusion Evaluated ria of the individual trials were reported previously.13–19 Pooling in Multiple Endovascular Stroke protocol, study selection, risk of bias, individual patient data IQR interquartile range acquisition, and data checks are described in the original pool- mRS modified Rankin Scale ing report.9 Control arm patients received best medical care, NCCT noncontrast computed tomography including alteplase if eligible. Intervention arm consisted of additional EVT. TAI thrombus attenuation increase Patients were included in the current study if thin-slice (≤2.5 mm) NCCT and CT angiography (CTA) images were I available (n=690). We excluded patients if NCCT and CTA n acute ischemic stroke, thrombi are often assumed images were on different scanners (n=11), or >30 minutes to completely occlude vessels, blocking all flow like apart (n=40). We excluded patients with scans that did not a cork on a bottle of wine. However, some thrombi cover the intracranial area of interest (n=33), were of insuf- are permeable, allowing for residual flow into and ficient quality (n=93; beam hardening: n=50, movement arti- through them.1–3 Thrombus perviousness estimates facts: n=32, contrast present on NCCT: n=6, scatter artifacts: residual flow through thrombi based on routine n=4, and venous phase CTA: n=1), or with incorrigible reg- noncontrast CT (NCCT) and single-phase CT angi- istration errors (n=76). Thrombi located too close to bone ography (CTA). 4,5 Perviousness was associated with (n=6), partial occlusions (n=4), narrow and distal thrombi improved outcomes and recanalization rates after (n=3), and bilateral middle cerebral artery thrombi (n=1) were intravenous alteplase. 4 For endovascular treatment excluded. The remaining 443 patients were included in the (EVT) however, reported effects of perviousness vary final analysis (Figure 1). The HERMES data used for this study are available via the VISTA-Endovascular repository. We considerably.6,7 followed the PRISMA-IPD guideline for study execution and The current standard treatment of acute ischemic reporting (Methods in the Data Supplement). stroke due to anterior circulation large vessel occlu- sions consists of alteplase if patients are eligible, fol- lowed by EVT.8,9 Although EVT has greatly improved Thrombus Perviousness Measurements NCCT and CTA images were coregistered using rigid regis- outcomes, even after EVT more than half of patients tration with Elastix.20 Thrombus perviousness was quantified remain permanently disabled or die after their stroke.9 by measuring thrombus attenuation increase (TAI) between The added benefit of intravenous alteplase is inves- tigated in several recently completed and ongoing randomized trials.10 Alteplase benefit may vary case- by-case: large proximal occlusions are known to show little benefit,11 while 10% of EVT-eligible patients recanalize with alteplase before initiation of EVT.12 Selecting the treatment with the highest chance of benefit (alteplase, EVT, neither, or both) is, therefore, expedient and could enable a more targeted and effective use of treatment modalities. Currently, no radiological thrombus characteristics other than occlusion location are used in acute isch- emic stroke decision making.8 The possible association of thrombus perviousness with alteplase treatment suc- cess, either alone or combined with EVT, could play a role in improving stroke treatment selection.4,5 To assess the role of perviousness, a large data set is needed, enabling separate analysis of treatment approaches. We aimed to investigate the effect of thrombus pervious- Figure 1. Patient inclusion flowchart. ness on EVT results in the HERMES trial (Highly Effec- Other includes thrombus too close to bone (n=6), only partial occlusion (n=4), too narrow thrombus (n=3), bilateral middle cerebral artery tive Reperfusion Evaluated in Multiple Endovascular thrombus (n=1). CTA indicates computed tomographic angiography; Stroke) pooled data set of 7 large randomized trials on HERMES, Highly Effective Reperfusion Evaluated in Multiple EVT for stroke.9,13–19 Endovascular Stroke; and NCCT, noncontrast computed tomography. 3634   November 2021 Stroke. 2021;52:3633–3641. DOI: 10.1161/STROKEAHA.120.033124 Kappelhof et al Thrombus Perviousness in Endovascular Treatment NCCT and CTA images as described previously.4 For image TAI outliers (n=11) were identified with the outlier labeling CLINICAL AND POPULATION selection and ROI placement, open-source software ITK- rule with a multiplier of 2.2, marking them as outliers if their SNAP was used.21 difference with the median was ≥2.2× the interquartile range Thrombus attenuation was measured manually by placing 3 (IQR).26 For 10 patients (2%), one marker was adjusted due SCIENCES spherical ROIs with a radius of one millimeter in the proximal, to erroneous placement outside the vessel or thrombus. All middle, and distal parts of the thrombus (Figure 2). The average remaining outliers (n=4, 1%) were attributable to short pervious density (ρ in Hounsfield units [HU]) of all ROIs was calculated thrombi, random noise, or slice thickness differences between for NCCT and CTA (ρNCCT and ρCTA). TAI was calculated accord- NCCT and CTA, and were included in the final analysis. ing to: TAI=ρCTA−ρNCCT.5 Univariable and multivariable ordinal logistic regression Measurements were performed by one of 3 trained raters were used in the primary outcome variable analysis (mRS (Drs Kappelhof, Dutra, and Alves).22 Raters were blinded for shift). Univariable and multivariable binary logistic regression occlusion location, clinical information, and measured attenu- was used for dichotomized outcome variables. Associations ation values. Training was done with 25 cases from the MR between TAI and FIV were tested with linear regression. CLEAN Registry.23 Scan acquisition characteristics including Because FIV showed a right-skewed distribution, log-trans- NCCT and CTA slice thickness difference (CTA slice thickness formation was performed. Associations between TAI and out- minus NCCT slice thickness),24 scanner brand, and scanner comes are reported per 5 HU. kVp were recorded for all included patients. Regression analyses were adjusted for age, baseline National Institutes of Health Stroke Scale score, intravenous Outcome Measures alteplase, occlusion location, diabetes, stroke onset to random- ization time, slice thickness difference between NCCT and CTA, Our primary outcome was functional outcome (ordinal modi- and included random effects for allocated study and scanner fied Rankin Scale [mRS]) at 90 days. The mRS score ranges brand. Unadjusted and adjusted (common) odds ratios (u[c]OR from 0 to 6, with 0 indicating complete functional indepen- and a[c]OR) were reported with 95% CI. We added an inter- dence, and 6 indicating death. Secondary outcome measures action term between TAI and allocated treatment in separate, were dichotomized functional outcome (mRS score, 0–2 subsequent models. For successful reperfusion as outcome versus 3–6 indicating functional independence; mRS score, measure, an additional regression model included an interac- 5–6 versus 0–4 indicating poor outcome), mortality, success- tion term between TAI and alteplase treatment. If treatment ful reperfusion (extended Thrombolysis in Cerebral Infarction interaction was significant, subgroup analyses were performed. score grade, 2B–3) after EVT (intervention arm patients only), Exploratory subgroup analyses were performed using unad- and follow-up infarct volume (FIV) in mL as measured on fol- justed and adjusted ordinal logistic regression for EVT effect low-up NCCT.25 per quartile of TAI, for the primary outcome only. Statistical analyses were performed with R version 3.5.2 (R Statistical Analysis Foundation for Statistical Computing, Vienna, Austria). The sig- Statistical analyses were prespecified in a statistical analysis nificance level was set at P<0.05. plan (Methods in the Data Supplement). Baseline clinical, imag- ing, and follow-up variables were compared with the overall HERMES population and between every quartile of TAI. One- RESULTS way ANOVA was used to assess differences in normally distrib- uted numerical variables, Kruskal-Wallis and Mann-Whitney U Median age was 67 years (IQR, 59–76), median baseline tests for non-normal numerical variables, and Fisher exact tests National Institutes of Health Stroke Scale score was 18 for categorical variables. (interquartile range, 14–21), and 53% of patients were Figure 2. Thrombus attenuation increase measurement. Three regions of interest (red) are placed in a left middle cerebral artery thrombus on noncontrast computed tomography (CT; A) and CT angiography (B). To calculate thrombus attenuation increase as a measure for perviousness, the density (in Hounsfield Units) in each region of interest is measured. Subsequently, the average noncontrast CT attenuation value is deducted from the average CT angiography value. Stroke. 2021;52:3633–3641. DOI: 10.1161/STROKEAHA.120.033124 November 2021   3635 Kappelhof et al Thrombus Perviousness in Endovascular Treatment male (Table 1). In comparison to the overall HERMES Primary Outcome CLINICAL AND POPULATION population, mRS scores, occurrence of sICH, and FIV Ninety-day mRS was available for 438 of 443 patients. were slightly higher. Intravenous alteplase was adminis- Higher TAI corresponded to lower mRS (P<0.01; Fig- SCIENCES tered to 91% of the patients in the control arm and 85% ure 3). There was significant interaction between TAI and of the patients in the intervention arm. allocated treatment (P for interaction, 0.03; Table 2). In the TAI showed a slightly right-skewed distribution control arm, TAI was associated with improved outcomes with a mean of 4.5 HU (SD, 12.7) and median 3.2 (acOR, 1.22 [95% CI, 1.11–1.33]). In the intervention HU (IQR, −4.3 to 11.4; Figure I in the Data Supple- arm, no significant effect was found (acOR, 0.99 [95% CI, ment). Baseline characteristics did not significantly 0.88–1.11]). These results were consistent in an explor- differ between TAI quartiles (Table I in the Data Sup- atory analysis of alteplase-treated patients (390/443, plement). TAI values did not differ between patients 88%; Table III in the Data Supplement). Analysis per TAI with and without extracranial carotid tandem lesions quartile showed a nonsignificant benefit of EVT in the (P=0.47). Scan acquisition details are discussed in highest quartile (Table IV in the Data Supplement). Results in the Data Supplement. Table 1. Baseline and Follow-Up Characteristics Baseline clinical characteristics Current cohort (N=443) Known in HERMES total (N=1766) Known in P value Age, y Median (IQR) 67 (59–76) 442 68 (57–76) 1761 0.26 Male sex, n (%) Male 236 (53%) 442 929 (53%) 1762 0.78 Left hemisphere stroke, n (%) 198 (51%) 389 652 (51%) 1275 0.95 Baseline NIHSS, median (IQR) 18 (14–21) 440 17 (13–21) 1751 0.13 Treatment allocation, N (%) Intervention 210 (48%) 442 871 (49%) 1764 0.38 Control 232 (53%) 442 893 (51%) 1764 Intravenous alteplase, N (%) Intervention 179 (85%) 210 763 (88%) 871 0.23 Control 211 (91%) 232 809 (91%) 893 0.90 Onset to randomization time, min, median (IQR) 189 (148–252) 441 183 (140–245) 1756 0.62 Atrial fibrillation, n (%) 117 (30%) 390 424 (33%) 1286 0.14 Diabetes, n (%) 74 (17%) 442 287 (16%) 1756 0.82 Hypertension, n (%) 238 (54%) 442 988 (56%) 1757 0.25 Imaging Occlusion location, n (%) ICA 92 (21%) 443 442 (27%) 1648 0.80 M1 331 (75%) 443 1073 (65%) 1648 M2 19 (4.3%) 443 131 (7.9%) 1648 Other 1 (0.2%) 443 2 (0.1%) 1648 Carotid tandem lesion, n (%) Intervention 18 (8.6%) 210 122 (7.4%) 1648 0.71 Control 17 (7.3%) 232 NCCT thrombus density (HU), median (IQR) 55 (48–62) 443 NA Follow-up mRS at 90 days, n (%) 0 27 (6%) 438 156 (9%) 1743 0.04* 1 51 (12%) 438 244 (14%) 1743 2 68 (16%) 438 282 (16%) 1743 3 91 (21%) 438 274 (16%) 1743 4 89 (20%) 438 356 (20%) 1743 5 38 (9%) 438 150 (9%) 1743 Mortality (mRS score, 6), n (%) 74 (17%) 440 281 (16%) 1754 0.84 Post-EVT eTICI score 2B–3, n (%) 134 (74%) 180 548 (74%) 745 0.74 sICH, n (%) 22 (5%) 389 63 (4%) 1729 0.01* Follow-up infarct volume, mL, median (IQR) 51 (17–125) 425 41 (14–120) 1665 0.01* eTICI indicates extended Thrombolysis in Cerebral Infarction score; EVT, endovascular treatment; HERMES, Highly Effective Reperfusion Evaluated in Multiple Endovascular Stroke; HU, Hounsfield units; ICA, internal carotid artery; IQR, interquartile range; M1, segment 1 of medial cerebral artery; M2, segment 2 of medical cerebral artery; mRS, modified Rankin Scale; NA, not available; NCCT, noncontrast computed tomography; NIHSS, National Institutes of Health Stroke Scale; and sICH, symptomatic intracranial hemorrhage. *Statistical significance (P<0.05). 3636   November 2021 Stroke. 2021;52:3633–3641. DOI: 10.1161/STROKEAHA.120.033124 Kappelhof et al Thrombus Perviousness in Endovascular Treatment CLINICAL AND POPULATION SCIENCES Figure 3. Thrombus attenuation increase (TAI) values varied per modified Rankin Scale (mRS) score (P<0.001). Secondary Outcomes (P=0.47)—although only 31 patients (15%) in the inter- vention arm did not receive intravenous alteplase. Post- Functional Independence EVT eTICI scores were known in 26 of them (12% of all In the control arm, patients with 90-day mRS score of 0 intervention arm patients). to 2 (n=59/231; 26%) had more pervious thrombi than patients with mRS score of 3 to 6 (median, 8.5 [IQR, 1.1– 17.4]) versus 1.2 [IQR, −4.9 to 9.4], P≤0.01, respectively). Follow-Up Infarct Volume In the intervention arm, TAI values did not significantly A 5 HU increase in TAI was associated with a 9% differ between patients with mRS score of 0 to 2 and decrease in FIV (95% CI, 14%–5%). Interaction between mRS score of 3 to 6 (Figure II in the Data Supplement). TAI and treatment arm was not significant (P=0.11; Interaction between TAI and allocated treatment was Table 2; Figure IV in the Data Supplement). significant (P=0.046). Subgroup analysis showed a sig- nificantly positive association between TAI and functional independence in the control arm, but not in the interven- tion arm (Table 2). Figure 4 shows the adjusted prob- DISCUSSION ability of functional independence plotted against TAI, for The benefit of EVT as an addition to alteplase diminishes intervention and control arm patients. Treatment effect of in patients with more pervious thrombi, due to improved EVT in addition to best medical care (alteplase if eligible) outcomes with increasing perviousness in patients decreases with higher TAI, although there is no point receiving alteplase alone. Increased thrombus pervious- where the CIs are separated. Unadjusted results are pre- ness was associated with improved functional outcome sented in Figure III in the Data Supplement. in HERMES control arm patients, who, in a high propor- tion, were treated with intravenous alteplase. No signifi- cant effect of perviousness on functional outcome was mRS Score of 5 to 6 and Mortality found in patients included in the intervention arm, who Higher TAI was associated with a lower chance of poor received EVT in addition to best medical care. Results outcome in the control arm (P for interaction 0.03; of associations with secondary outcomes were similar. Table 2). Greater TAI also corresponded to a smaller No value for perviousness was observed where better chance of mortality in the overall population (aOR, 0.85 outcomes were associated with withholding EVT, though [95% CI, 0.75–0.97]), but treatment interaction was not the association with functional outcome was more pro- significant (P=0.58). nounced in the higher perviousness quartiles. Our findings generally agree with previous studies showing increased recanalization rates and improved Successful Endovascular Reperfusion outcomes in patients with more pervious thrombi after eTICI scores were available for 180/210 intervention arm intravenous alteplase treatment.4,5,7,27–29 These studies, patients (86%). Successful reperfusion was reached in however, compared intravenous alteplase to conserva- 134/180 cases (74%). We found no significant effect of tive treatment only4 or found no interaction with EVT, TAI on reperfusion (Table 2). Interaction between TAI and possibly due to lower patient numbers.5 Improved target alteplase was not significant for successful reperfusion vessel recanalization for pervious thrombi seems to lead Stroke. 2021;52:3633–3641. DOI: 10.1161/STROKEAHA.120.033124 November 2021   3637 Kappelhof et al Thrombus Perviousness in Endovascular Treatment Table 2. aOR for the Effect of TAI (per 5 HU) on Outcomes CLINICAL AND POPULATION Outcome measure aOR (95% CI)* Interaction with EVT aOR per arm (95% CI)* Ordinal mRS 1.10 (1.03–1.18)† P=0.03† Control: 1.22 (1.04–1.19)† SCIENCES Intervention: 0.96 (0.85–1.08) mRS score 0–2 1.15 (1.05–1.27)† P=0.05† Control: 1.33 (1.14–1.56)† Intervention: 1.04 (0.90–1.21) mRS score 5–6 0.92 (0.83–1.02) P=0.03† Control: 0.82 (0.70–0.95)† Intervention: 1.07 (0.90–1.27) Mortality 0.85 (0.75–0.97)† P=0.58 NA Final infarct volume (effect ratio) 0.91 (0.86–0.95)† P=0.11‡ Control: 0.87 (0.82–0.93)† Intervention: 0.97 (0.89–1.05) Intervention arm only: successful 0.95 (0.79–1.14) Interaction with IV NA reperfusion alteplase: P=0.47 Unadjusted results shown in Table II in the Data Supplement. Control arm: intravenous alteplase if eligible (n=232), intervention arm: additional EVT (n=210). aOR indicates adjusted odds ratio; CTA, computed tomography angiography; EVT, endovascular treat- ment; HU, Hounsfield units; mRS, modified Rankin Scale; NA, not applicable/available; NCCT, noncontrast computed tomography; and TAI, thrombus attenuation increase. *Adjusted for age (y), baseline NIHSS, alteplase, occlusion location, diabetes, stroke onset to randomization time, NCCT-CTA slice thickness difference; and including random effects for study and scanner brand. †Statistical significance (P<0.05). ‡Exploratory analysis in model with nonsignificant treatment interaction. to improved outcomes after intravenous alteplase,28 on could affect density measurements due to partial volume average leaving less opportunity for EVT to give addi- effect.24 Minor coregistration errors could minimally off- tional benefit. On the association between perviousness set marker placement in NCCT or CTA. Random (Gauss- and EVT, results vary: in contrast to our results, a recent ian) noise could cause affect HU values on either NCCT study in EVT-treated patients reported a significantly or CTA. These factors unlikely affect TAI measurement positive association between perviousness and func- validity: random noise and small coregistration flaws tional outcome in the adjusted analyses.7 Perviousness are likely random, we adjusted for slice thickness differ- was reported to be associated with higher chances of ences, and our results are reproduceable in studies using first-pass success of aspiration thrombectomy.30 the same methods.4,7 An effect of differences in scanner Previous studies have dichotomized TAI, to classify voltage on HU values was not supported by our data. thrombi into pervious and nonpervious.4,5 The optimal cut- off value varied between studies and data sets. However, these values were subsequently used by other study groups Limitations without testing validity in their data.31 Therefore, and since We excluded patients due to unavailable thin-slice imag- dichotomization was not necessary for our research ques- ing. Alternative methods like qualitative assessment of tion, we did not dichotomize TAI. Following this reasoning, residual flow is possible using thick-slice NCCT but less the TAI values in the per-quartile analysis should be inter- precise.24 Almost all centers acquire thin-slice imaging preted with care. Though interesting, the subgroup analysis initially but do not store such images: perviousness could is not meaningfully powered with ≈110 patients per group: be measured for many more patients if thin-slice imaging TAI cutoff values may not be applicable to other data sets. is preserved, and a practical, fast, (automatic) measure- The observed effect of perviousness may be partly ment method is available. explained by thrombus histopathology. A study on In our sample, we had no information on alteplase thrombi retrieved during EVT showed a positive cor- administration before scan acquisition. In patients receiv- relation between perviousness and fibrin/platelet frac- ing thrombolysis before transfer to an EVT-performing tions.31 However, conflicting results exist: a recent study center, alteplase may have been started before scan reported opposite results.32 RBC-rich thrombi were acquisition, already exposing some thrombi to alteplase. found to be more responsive to thrombolytic therapy.33 In The effects of alteplase administration on thrombus relation to EVT, fibrin-rich thrombi were associated with perviousness are currently unknown. Patients may have longer intervention times and a higher chance of second- recanalized and recovered after alteplase and before ran- ary embolisms.34 Thrombus perviousness may also be domization, leading to a possible inclusion bias, though associated with stroke cause, though published results start of EVT was not delayed by waiting for the effect of on this topic are inconsistent.35,36 alteplase (except for REVASCAT, n=206/1766).17 These Interestingly, we observed some negative values for patients may have had more pervious thrombi; thereby TAI. Slice thickness differences between NCCT and CTA the current study may under-represent patients with 3638   November 2021 Stroke. 2021;52:3633–3641. DOI: 10.1161/STROKEAHA.120.033124 Kappelhof et al Thrombus Perviousness in Endovascular Treatment CLINICAL AND POPULATION SCIENCES Figure 4. Probability of functional independence (90-d modified Rankin Scale [mRS] score, 0–2) vs thrombus attenuation increase (TAI). Results are adjusted for prespecified variables age, baseline National Institutes of Health Stroke Scale, alteplase, occlusion location, diabetes, stroke onset to randomization time, and noncontrast computed tomography-computed tomography angiography slice thickness difference. Control arm: best medical care (intravenous alteplase if eligible; n=232). Intervention arm: endovascular treatment in addition to best medical care (n=210). very pervious thrombi and underestimate the effect of our results do not provide definitive evidence to withhold perviousness. Likewise, we did not know the time from treatment based on a certain value of perviousness. alteplase administration to CTA acquisition. Thrombus Further research could focus on combinations of perviousness might increase if alteplase has had more thrombus characteristics, like perviousness, length, and time to act on the thrombus. location, in relation to alteplase and EVT response. Small- Our data consisted of pooled data from multiple trials. volume, short, pervious, distal M2-thrombi may benefit from We cannot exclude an uneven distribution of the number alteplase, whereas proximal, large-volume, long, impervi- of included patients per trial. However, since part of the ous ICA terminus occlusions may not. Larger thrombi may imaging was requested from trials individually, and scan- respond better to intravenous tenecteplase.38 Addition- ning protocols were heterogeneous, we are reasonably ally, the effect of perviousness on outcomes of combined sure that patients from each trial were included. In addi- alteplase and EVT versus direct EVT only, or on EVT device tion, we included a random effect for allocated study in outcomes could be studied further. In the future, thrombus all adjusted regression analyses. perviousness may support patient selection for alteplase Technical aspects of image acquisition can influence alone, combined alteplase and EVT, or EVT alone, and sup- TAI measurements. Slice thickness of the images we port endovascular treatment modality choice. used varied from 0.6 to 2.5 mm. Thicker slices give a lower measured density on NCCT due to volume aver- aging with surrounding brain tissue.24 In addition, CTA Conclusions scan timing can influence TAI, by slightly differing scan- In patients treated in the control arm of HERMES, of ning phase. Dynamic or multiphase CTA incorporates the whom most were treated with alteplase, increased throm- time dimension by making CTA scans at multiple time bus perviousness was associated with improved func- points after contrast injection, which can avoid the issue tional outcome, decreased mortality, and reduced infarct of scanning too early after contrast bolus.37 volume. We found no significant association among Difficult delineation of thrombi can hamper measure- patients allocated to the EVT-arm. The benefit of EVT ments. The distal thrombus border is hard to discern as an addition to best medical care including alteplase in case of poor collaterals on CTA and an isodense diminishes in patients with more pervious thrombi, due thrombus on NCCT. However, this occurred only in nine to improved outcomes with increasing perviousness in patients in our data set. Measurement in dynamic CTA patients receiving alteplase alone—though no value for may show the distal thrombus border more accurately in perviousness was observed where withholding EVT was patients with poor collateral flow.37 associated with better outcomes. Finally, it is important to note that although we found significant treatment interaction and decreasing addi- tional benefit of EVT with increasing perviousness, we ARTICLE INFORMATION did not find a TAI value where control arm patients did Received October 16, 2020; final revision received April 16, 2021; accepted May significantly better than intervention arm patients. Thus, 6, 2021. Stroke. 2021;52:3633–3641. DOI: 10.1161/STROKEAHA.120.033124 November 2021   3639 Kappelhof et al Thrombus Perviousness in Endovascular Treatment Affiliations 3. Brass LF, Wannemacher KM, Ma P, Stalker TJ. Regulating thrombus growth CLINICAL AND POPULATION Radiology and Nuclear Medicine (M.K., M.L.T., K.M.T., B.G.D., H.A., G.Z., H.A.M., and stability to achiev optimal response to injury. J Thromb Haemost. C.B.L.M.M.), Biomedical Engineering and Physics (M.L.T., B.G.D., H.A., H.A.M.); 2011;9:66–75. doi: 10.1111/j.1538-7836.2011.04364.x Neurology (Y.B.W.E.M.R.), Amsterdam University Medical Center, University of 4. Santos EM, Dankbaar JW, Treurniet KM, Horsch AD, Roos YB, Kappelle SCIENCES Amsterdam, the Netherlands. Altair Biostatistics, St Louis Park, MN (S. Brown). LJ, Niessen WJ, Majoie CB, Velthuis B, Marquering HA; DUST Investiga- Neuroscience & Psychology, University of Glasgow, Queen Elizabeth University tors. Permeable thrombi are associated with higher intravenous recom- Hospital, United Kingdom (K.W.M.). Neuroscience, Hospital Germans Trias i Pu- binant tissue-type plasminogen activator treatment success in patients jol, Universitat Autònoma de Barcelona, Spain (A.D.). Neurology, Comprehensive with acute ischemic stroke. Stroke. 2016;47:2058–2065. doi: 10.1161/ Stroke Center, David Geffen School of Medicine, University of California, Los STROKEAHA.116.013306 Angeles (UCLA) (J.L.S.). Clinical Neurosciences (A.M.D., M.D.H.) and Radiology 5. Santos EM, Marquering HA, den Blanken MD, Berkhemer OA, Boers AM, (M.G.), University of Calgary, Alberta, Canada. Neurology, University of Pittsburgh Yoo AJ, Beenen LF, Treurniet KM, Wismans C, van Noort K, et al; MR Medical Center, PA (T.G.J.). Diagnostic and Interventional Neuroradiology (S. CLEAN Investigators. Thrombus permeability is associated with improved Bracard) and Epidemiology (F.G.), University of Lorraine, University Hospital of functional outcome and recanalization in patients with ischemic stroke. Nancy, France. Medicine and Neurology (B.C.V.C.) and Radiology (P.J.M.), Royal Stroke. 2016;47:732–741. doi: 10.1161/STROKEAHA.115.011187 Melbourne Hospital, University of Melbourne, Parkville, Australia. Radiology and 6. Borst J, Berkhemer OA, Santos EMM, Yoo AJ, den Blanken M, Roos YBWEM, Nuclear Medicine (A.v.d.L.) and Neurology (D.W.J.D.), Erasmus Medical Center, van Bavel E, van Zwam WH, van Oostenbrugge RJ, Lingsma HF, et al; MR Rotterdam, the Netherlands. Institute of Neuroscience, Newcastle University, CLEAN investigators. Value of thrombus CT characteristics in patients with Newcastle upon Tyne, United Kingdom (P.W.). acute ischemic stroke. AJNR Am J Neuroradiol. 2017;38:1758–1764. doi: 10.3174/ajnr.A5331 Acknowledgments 7. Dutra BG, Tolhuisen ML, Alves HCBR, Treurniet KM, Kappelhof M, Yoo AJ, Jansen IGH, Dippel DWJ, van Zwam WH, van Oostenbrugge RJ, et al; MR We thank the HERMES (Highly Effective Reperfusion Evaluated in Multiple En- CLEAN Registry Investigators†. Thrombus imaging characteristics and out- dovascular Stroke) collaborators. comes in acute ischemic stroke patients undergoing endovascular treatment. Sources of Funding Stroke. 2019;50:2057–2064. doi: 10.1161/STROKEAHA.118.024247 8. Powers WJ, Rabinstein AA, Ackerson T, Adeoye OM, Bambakidis NC, The HERMES (Highly Effective Reperfusion Evaluated in Multiple Endovascular Becker K, Biller J, Brown M, Demaerschalk BM, Hoh B, et al; American Heart Stroke) collaboration was funded by a grant to the University of Calgary from Association Stroke Council. 2018 Guidelines for the early management of Medtronic LLC. patients with acute ischemic stroke: a guideline for healthcare professionals Disclosures from the American Heart Association/American Stroke Association. Stroke. 2018;49:e46–e110. doi: 10.1161/STR.0000000000000158 Dr Kappelhof reports MRCLEAN-NOIV coordinator. Dr Brown reports personal 9. Goyal M, Menon BK, van Zwam WH, Dippel DW, Mitchell PJ, Demchuk AM, (University of Calgary, Medtronic). Dr Muir reports personal (Boehringer Ingelheim, Dávalos A, Majoie CB, van der Lugt A, de Miquel MA, et al; HERMES col- Bayer, Daiichi Sankyo, Biogen, ReNeuron). Drs Roos and Marquering reports share- laborators. Endovascular thrombectomy after large-vessel ischaemic stroke: holder (Nico-Lab). Dr Saver reports personal (Medtronic). Dr Demchuk reports per- a meta-analysis of individual patient data from five randomised trials. Lancet. sonal (Medtronic, Boehringer Ingelheim), patent (Circle NVI). Dr Jovin received per- 2016;387:1723–1731. doi: 10.1016/S0140-6736(16)00163-X sonal (Ceronovus, Codman Neurovascular, Neuravi, Stryker, Fundacio Ictus), grants 10. Yang P, Zhang Y, Zhang L, Zhang Y, Treurniet KM, Chen W, Peng Y, Han H, (Stryker Neurovascular, GE Healthcare), shares (Silk Road, Anaconda, Route 92, Wang J, Wang S, et al; DIRECT-MT Investigators. Endovascular thrombec- Blockade), other (Anaconda, Route92, Vizai, FreeOx, Corindus, Methinks, Contego tomy with or without intravenous alteplase in acute stroke. N Engl J Med. Medical). Dr Campbell reports the following institutional (National Health and Medi- 2020;382:1981–1993. doi: 10.1056/NEJMoa2001123 cal Research Council, Royal Australasian College of Physicians, Royal Melbourne 11. De Silva DA, Brekenfeld C, Ebinger M, Christensen S, Barber PA, Butcher Hospital Foundation, National Heart Foundation, National Stroke Foundation of KS, Levi CR, Parsons MW, Bladin CF, Donnan GA, et al; Echoplanar Imag- Australia, Medtronic). Dr van der Lugt received grants (GE Healthcare, Siemens ing Thrombolytic Evaluation Trial (EPITHET) Investigators. The benefits of Engineers, Thrombolytics Science Inc, Ceronovus, Penumbra, Stryker, Medtronic, intravenous thrombolysis relate to the site of baseline arterial occlusion in AngioCare BV, Covidien/EV3, Medac GmbH/Lamepro, Dutch Heart Foundation, the Echoplanar Imaging Thrombolytic Evaluation Trial (EPITHET). Stroke. Thrombolytic Science Inc). Dr White reports the following institutional (Stryker, 2010;41:295–299. doi: 10.1161/STROKEAHA.109.562827 Penumbra, Medtronic), personal (Microvention). Dr Hill reports the following institu- 12. Mueller L, Pult F, Meisterernst J, Heldner MR, Mono ML, Kurmann R, tional (NoNO Inc, Boehringer Ingelheim), patent (Systems and Methods for Assist- Buehlmann M, Fischer U, Mattle HP, Arnold M, et al. Impact of intra- ing in Decision-Making and Triaging for Acute Stroke Patients), director (Canadian venous thrombolysis on recanalization rates in patients with stroke Neuroscience Federation, Canadian Stroke Consortium, Circle Neurovascular). Dr treated with bridging therapy. Eur J Neurol. 2017;24:1016–1021. doi: Dippel reports the following institutional (Dutch Heart Foundation, AngioCare BV, 10.1111/ene.13330 Covidien/EV3, Medac GmbH/Lamepro, Penumbra, Top Medical/Concentric, Stryk- 13. Saver JL, Goyal M, Bonafe A, Diener HC, Levy EI, Pereira VM, Albers GW, er, Ceronovus, Brain Foundation Netherlands, the Netherlands Organisation for Cognard C, Cohen DJ, Hacke W, et al; SWIFT PRIME Investigators. Stent- Health Research and Development, Health Holland Top Sector Life Sciences and retriever thrombectomy after intravenous t-PA vs. t-PA alone in stroke. N Health, Medtronic). Dr Mitchell reports the following institutional (Medtronic, Stryker, Engl J Med. 2015;372:2285–2295. doi: 10.1056/NEJMoa1415061 Codman, Johnson & Johnson). Dr Goyal reports personal (Medtronic, Stryker, Mi- 14. Goyal M, Demchuk AM, Menon BK, Eesa M, Rempel JL, Thornton J, Roy D, crovention), patent (Systems of acute stroke diagnosis). Dr Majoie received grants Jovin TG, Willinsky RA, Sapkota BL, et al; ESCAPE Trial Investigators. Ran- (CardioVasculair Onderzoek Nederland/Dutch Heart Foundation, European Com- domized assessment of rapid endovascular treatment of ischemic stroke. N mission, Dutch Health Evaluation Program, Stryker, Toegepast Wetenschappelijk Engl J Med. 2015;372:1019–1030. doi: 10.1056/NEJMoa1414905 Instituut voor Neuromodulatie [TWIN] Foundation). 15. Berkhemer OA, Fransen PS, Beumer D, van den Berg LA, Lingsma HF, Yoo AJ, Schonewille WJ, Vos JA, Nederkoorn PJ, Wermer MJ, et al; MR CLEAN Supplemental Materials Investigators. A randomized trial of intraarterial treatment for acute ischemic Expanded Methods (Statistical Analysis Plan) stroke. N Engl J Med. 2015;372:11–20. doi: 10.1056/NEJMoa1411587 Expanded Results (Scan Acquisition Effects) 16. Campbell BC, Mitchell PJ, Kleinig TJ, Dewey HM, Churilov L, Yassi N, Yan B, Online Tables I–IV Dowling RJ, Parsons MW, Oxley TJ, et al; EXTEND-IA Investigators. Endo- Online Figures I–IV vascular therapy for ischemic stroke with perfusion-imaging selection. N Engl J Med. 2015;372:1009–1018. doi: 10.1056/NEJMoa1414792 17. Jovin TG, Chamorro A, Cobo E, de Miquel MA, Molina CA, Rovira A, San Román L, Serena J, Abilleira S, Ribó M, et al; REVASCAT Trial Investiga- REFERENCES tors. Thrombectomy within 8 hours after symptom onset in ischemic stroke. 1. Voronov RS, Stalker TJ, Brass LF, Diamond SL. Simulation of intrathrom- N Engl J Med. 2015;372:2296–2306. doi: 10.1056/NEJMoa1503780 bus fluid and solute transport using in vivo clot structures with single 18. Muir KW, Ford GA, Messow CM, Ford I, Murray A, Clifton A, Brown MM, platelet resolution. Ann Biomed Eng. 2013;41:1297–1307. doi: 10.1007/ Madigan J, Lenthall R, Robertson F, et al; PISTE Investigators. Endovascular s10439-013-0764-z therapy for acute ischaemic stroke: the Pragmatic Ischaemic Stroke Throm- 2. Laurens N, Koolwijk P, de Maat MP. Fibrin structure and wound healing. J Thromb bectomy Evaluation (PISTE) randomised, controlled trial. J Neurol Neurosurg Haemost. 2006;4:932–939. doi: 10.1111/j.1538-7836.2006.01861.x Psychiatry. 2017;88:38–44. doi: 10.1136/jnnp-2016-314117 3640   November 2021 Stroke. 2021;52:3633–3641. DOI: 10.1161/STROKEAHA.120.033124 Kappelhof et al Thrombus Perviousness in Endovascular Treatment 19. Bracard S, Ducrocq X, Mas JL, Soudant M, Oppenheim C, Moulin T, in acute ischemic stroke. J Stroke Cerebrovasc Dis. 2020;29:104541. doi: CLINICAL AND POPULATION Guillemin F; THRACE investigators. Mechanical thrombectomy after intra- 10.1016/j.jstrokecerebrovasdis.2019.104541 venous alteplase versus alteplase alone after stroke (THRACE): a ran- 30. Mokin M, Waqas M, Fifi J, De Leacy R, Fiorella D, Levy EI, Snyder K, Hanel R, domised controlled trial. Lancet Neurol. 2016;15:1138–1147. doi: 10.1016/ Woodward K, Chaudry I, et al. Clot perviousness is associated with first pass SCIENCES S1474-4422(16)30177-6 success of aspiration thrombectomy in the COMPASS trial. J Neurointerv 20. Klein S, Staring M, Murphy K, Viergever MA, Pluim JP. elastix: a toolbox Surg. 2021;13:509–514. doi: 10.1136/neurintsurg-2020-016434 for intensity-based medical image registration. IEEE Trans Med Imaging. 31. Berndt M, Friedrich B, Maegerlein C, Moench S, Hedderich D, Lehm M, 2010;29:196–205. doi: 10.1109/TMI.2009.2035616 Zimmer C, Straeter A, Poppert H, Wunderlich S, et al. Thrombus perme- 21. Yushkevich PA, Piven J, Hazlett HC, Smith RG, Ho S, Gee JC, Gerig G. ability in admission computed tomographic imaging indicates stroke patho- User-guided 3D active contour segmentation of anatomical structures: sig- genesis based on thrombus histology. Stroke. 2018;49:2674–2682. doi: nificantly improved efficiency and reliability. Neuroimage. 2006;31:1116– 10.1161/STROKEAHA.118.021873 1128. doi: 10.1016/j.neuroimage.2006.01.015 32. Benson JC, Fitzgerald ST, Kadirvel R, Johnson C, Dai D, Karen D, Kallmes 22. Santos EM, Yoo AJ, Beenen LF, Berkhemer OA, den Blanken MD, Wismans DF, Brinjikji W. Clot permeability and histopathology: is a clot’s pervious- C, Niessen WJ, Majoie CB, Marquering HA; MR CLEAN investigators. ness on CT imaging correlated with its histologic composition? J Neurointerv Observer variability of absolute and relative thrombus density measurements Surg. 2020;12:38–42. doi: 10.1136/neurintsurg-2019-014979 in patients with acute ischemic stroke. Neuroradiology. 2016;58:133–139. 33. Puig J, Pedraza S, Demchuk A, Daunis-I-Estadella J, Termes H, Blasco G, doi: 10.1007/s00234-015-1607-4 Soria G, Boada I, Remollo S, Baños J, et al. Quantification of thrombus houn- 23. Jansen IGH, Mulder MJHL, Goldhoorn RB; MR CLEAN Registry investiga- sfield units on noncontrast CT predicts stroke subtype and early recanaliza- tors. Endovascular treatment for acute ischaemic stroke in routine clinical tion after intravenous recombinant tissue plasminogen activator. AJNR Am practice: prospective, observational cohort study (MR CLEAN Registry). J Neuroradiol. 2012;33:90–96. doi: 10.3174/ajnr.A2878 BMJ. 2018;360:k949. doi: 10.1136/bmj.k949 34. Sporns PB, Hanning U, Schwindt W, Velasco A, Buerke B, Cnyrim C, 24. Tolhuisen ML, Enthoven J, Santos EMM, Niessen WJ, Beenen LFM, Dippel Minnerup J, Heindel W, Jeibmann A, Niederstadt T. Ischemic stroke: his- DWJ, van der Lugt A, van Zwam WH, Roos YBWEM, van Oostenbrugge RJ, tological thrombus composition and pre-interventional CT attenuation are et al. The effect of non-contrast CT slice thickness on thrombus density and associated with intervention time and rate of secondary embolism. Cerebro- perviousness assessment. Lect Notes Comput Sci. 2017;10555:105–115. vasc Dis. 2017;44:344–350. doi: 10.1159/000481578 25. Boers AMM, Jansen IGH, Beenen LFM, Devlin TG, San Roman L, Heo 35. Boodt N, Compagne KCJ, Dutra BG, Samuels N, Tolhuisen ML, Alves JH, Ribó M, Brown S, Almekhlafi MA, Liebeskind DS, et al. Association HCBR, Kappelhof M, Lycklama À Nijeholt GJ, Marquering HA, Majoie of follow-up infarct volume with functional outcome in acute ischemic CBLM, et al; Coinvestigators MR CLEAN Registry. Stroke etiology and stroke: a pooled analysis of seven randomized trials. J Neurointerv Surg. thrombus computed tomography characteristics in patients with acute isch- 2018;10:1137–1142. doi: 10.1136/neurintsurg-2017-013724 emic stroke: a MR CLEAN registry substudy. Stroke. 2020;51:1727–1735. 26. Hoaglin DC, Iglewicz B. Fine-tuning some resistant rules for outlier labeling. doi: 10.1161/STROKEAHA.119.027749 J Am Stat Assoc. 1987;82:1147–1149. 36. Kufner A, Erdur H, Endres M, Nolte CH, Scheel M, Schlemm L. Associa- 27. Mishra SM, Dykeman J, Sajobi TT, Trivedi A, Almekhlafi M, Sohn SI, Bal S, tion between thrombus perviousness assessed on computed tomog- Qazi E, Calleja A, Eesa M, et al. Early reperfusion rates with IV tPA are deter- raphy and stroke cause. Stroke. 2020;51:3613–3622. doi: 10.1161/ mined by CTA clot characteristics. AJNR Am J Neuroradiol. 2014;35:2265– STROKEAHA.120.031148 2272. doi: 10.3174/ajnr.A4048 37. Santos EMM, d’Esterre CD, Treurniet KM, Niessen WJ, Najm M, Goyal M, 28. Menon BK, Al-Ajlan FS, Najm M, Puig J, Castellanos M, Dowlatshahi D, Demchuk AM, Majoie CB, Menon BK, Marquering HA; PRove-IT investigators. Calleja A, Sohn SI, Ahn SH, Poppe A, et al; INTERRSeCT Study Investi- Added value of multiphase CTA imaging for thrombus perviousness assess- gators. Association of clinical, imaging, and thrombus characteristics with ment. Neuroradiology. 2018;60:71–79. doi: 10.1007/s00234-017-1907-y recanalization of visible intracranial occlusion in patients with acute ischemic 38. Campbell BCV, Mitchell PJ, Churilov L, Yassi N, Kleinig TJ, Dowling RJ, stroke. JAMA. 2018;320:1017–1026. doi: 10.1001/jama.2018.12498 Yan B, Bush SJ, Dewey HM, Thijs V, et al; EXTEND-IA TNK Investigators. 29. Bilgic AB, Gocmen R, Arsava EM, Topcuoglu MA. The effect of clot volume Tenecteplase versus alteplase before thrombectomy for ischemic stroke. N and permeability on response to intravenous tissue plasminogen activator Engl J Med. 2018;378:1573–1582. doi: 10.1056/NEJMoa1716405 Stroke. 2021;52:3633–3641. DOI: 10.1161/STROKEAHA.120.033124 November 2021   3641