Skip main navigation

Mismatch Profile Influences Outcome After Mechanical Thrombectomy

and on behalf of the FRAME Investigators*
Originally publishedhttps://doi.org/10.1161/STROKEAHA.120.031929Stroke. 2021;52:232–240

Abstract

Background and Purpose:

Mechanical thrombectomy (MT) is the recommended treatment for acute ischemic stroke caused by anterior circulation large vessel occlusion. However, despite a high rate of reperfusion, the clinical response to successful MT remains highly variable in the early time window where optimal imaging selection criteria have not been established. We hypothesize that the baseline perfusion imaging profile may help forecast the clinical response to MT in this setting.

Methods:

We conducted a prospective multicenter cohort study of patients with large vessel occlusion–related acute ischemic stroke treated by MT within 6 hours. Treatment decisions and the modified Rankin Scale evaluation at 3 months were performed blinded to the results of baseline perfusion imaging. Study groups were defined a posteriori based on predefined imaging profiles: target mismatch (TMM; core volume <70 mL/mismatch ratio >1.2 and mismatch volume >10 mL) versus no TMM or mismatch (MM; mismatch ratio >1.2 and volume >10 mL) versus no MM. Functional recovery (modified Rankin Scale, 0–2) at 3 months was compared based on imaging profile at baseline and whether reperfusion (modified Thrombolysis in Cerebral Infarction 2bc3) was achieved.

Results:

Two hundred eighteen patients (mean age, 71±15 years; median National Institutes of Health Stroke Scale score, 17 [interquartile range, 12–21]) were enrolled. Perfusion imaging profiles were 71% TMM and 82% MM. The rate of functional recovery was 54% overall. Both TMM and MM profiles were independently associated with a higher rate on functional recovery at 3 months Adjusted odds ratios were 3.3 (95% CI, 1.4–7.9) for TMM and 5.9 (95% CI, 1.8–19.6) for MM. Reperfusion (modified Thrombolysis in Cerebral Infarction 2bc3) was achieved in 86% and was more frequent in TMM and MM patients. Reperfusion was associated with a higher rate of functional recovery in MM and TMM patients but not among those with no MM.

Conclusions:

In this cohort study, about 80% of the patients with a large vessel occlusion–related acute ischemic stroke had evidence of penumbra, regardless of infarction volume. Perfusion imaging profiles predict the clinical response to MT.

Introduction

Mechanical thrombectomy (MT) has revolutionized the management of patients with an acute ischemic stroke (AIS) caused by an anterior circulation large vessel occlusion (LVO).1 Beyond 6 hours after onset, MT is reserved for patients with substantial penumbra and a small or moderate ischemic core, typically identified by the target mismatch (TMM) profile on multimodal imaging.2,3

Eight randomized controlled trials (RCTs) compared MT versus best medical treatment of LVO-related AIS, using various imaging selection criteria, in an early time window.4,5 Despite reperfusion rates close to 80%, >50% of participants treated by MT were dead or dependent at 3 months.1 In 2019, the Highly Effective Reperfusion Using Multiple Endovascular Devices collaborators reprocessed technically adequate perfusion imaging in a subset of 583 (33%) participants enrolled in 7 of the early-time-window RCTs.5 This subgroup analysis showed that almost every subject had substantial penumbra regardless of the core volume.5 This finding was likely related to the fact that many of these patients had penumbral imaging, which influenced patient selection. Infarct progression is highly variable even during the first hours after onset, and prior studies suggest that about 20% of patients complete their infarcts early.6 For example, in the EXTEND-IA study (Extending the Time for Thrombolysis in Emergency Neurological Deficits With Intra-Arterial Therapy) screening log, about 25% of the screened patients with an LVO-related AIS were not included due to the absence of TMM on baseline imaging performed within 6 hours after onset.7 Therefore, the prevalence of a favorable imaging profile in this setting may be less frequent than reported in many RCTs. Finally, international guidelines indicate that, in the early time window, patient selection based on TMM may exclude “a substantial proportion of patients who have the potential to respond favorably to reperfusion.”8,9 In conclusion, while every day several thousands of patients experiencing an LVO-related AIS are undergoing perfusion imaging before MT worldwide,10 the optimal baseline imaging profile that identifies patients who have the potential to respond favorably to endovascular reperfusion in the early time window remains unknown.

Based on these premises, we designed a study to assess a population of LVO-related AIS treated by MT within 6 hours and determine prevalence of different perfusion imaging profiles and their relationships with functional outcome at 3 months. Patients were managed according to current recommendations (ie, no specific advanced imaging selection criteria was required) by investigators blinded to perfusion imaging results. We had 2 hypotheses: first, the prevalence of a favorable imaging profile will be less frequent than previously reported in RCTs, and second, these profiles will predict the clinical response to endovascular reperfusion.

Methods

Study Design

The FRAME study (French Acute Multimodal Imaging Study to Select Patients for Mechanical Thrombectomy) is a prospective multicenter cohort study involving 2 large French comprehensive stroke centers, Toulouse and Bordeaux. Investigating centers contracted with iSchema View to install a specific version of the RAPID software that did not process the diffusion weighted imaging (DWI)/perfusion weighted imaging (PWI)/computed tomography perfusion (CTP) maps on-site and automatically sent them to a secure server. Treatment decisions and patient management were made using available sequences: DWI/gradient echo/fluid attenuated inversion recovery and magnetic resonance angiography or noncontrast computed tomography and computed tomography angiography. A low Alberta Stroke Program Early CT Score on computed tomography or large DWI lesion was not an exclusion criterion. Referring centers typically did not use perfusion imaging. Patients were eligible for FRAME only if perfusion imaging was performed at the comprehensive center before MT. Investigators from both centers agreed to not use alternate perfusion software programs for treatment decisions in potentially eligible FRAME patients. The DWI/PWI and CTP maps were processed a posteriori after the end of the study. Mismatch (MM) and TMM status were determined by an independent investigator (S.C.) blinded to clinical history and outcome data.

Patient Population

Inclusion Criteria
  • Age ≥18 years.

  • DWI/PWI magnetic resonance imaging (MRI) or CTP performed before thrombectomy.

  • MT initiated within 6 hours after onset.

  • Occlusion of the internal carotid artery or the first or second segment of the middle cerebral artery, as seen on computed tomography or MR angiography.

Exclusion Criteria
  • Prestroke modified Rankin Scale (mRS) score estimated to be >1.

  • Evaluation of multimodal imaging profile for patient management.

  • Any terminal illness such that the patient would not be expected to survive >1 year.

  • Delay between imaging and femoral puncture >90 minutes.

  • Patients placed under guardianship or analogous situations.

Clinical Assessment

Certified neurologists evaluated the National Institutes of Health Stroke Scale (NIHSS) at baseline day 1, day 3, and 3 months. Independent certified physician evaluated mRS at 3 months.

Study Design Blinding

DWI, PWI, and CTP maps were stored on a secure server. They were not accessible on any clinical imaging system (emails, Picture Archiving and Communication System, etc). The clinical team, investigators, and independent mRS raters were blinded to these results until all clinical data had been collected and locked.

Imaging Assessment and Parameters

Maps were processed with the RAPID software, version 4.7.

TMM was defined according to the EXTEND-IA definition: mismatch ratio >1.2, mismatch volume >10 mL, and ischemic core lesion volume <70 mL.7 When perfusion imaging was not technically adequate on MRI, a magnetic resonance angiography core mismatch was used to define TMM: internal carotid artery/M1 occlusion with a core lesion <25 mL and M2 occlusion with a core lesion <15 mL as described previously.11 MM was defined by a mismatch ratio >1.2 and a mismatch volume >10 mL in patients with technically adequate perfusion imaging.

The site of vessel occlusion was defined on baseline computed tomography angiography, magnetic resonance angiography, and conventional angiography images. Reperfusion was defined as the modified Thrombolysis in Cerebral Infarction score of 2b, 2c, or 3 by an independent blinded investigator (M. Mazighi).

Primary Outcome
  • Rate of functional neurological recovery defined by an mRS score of 0 to 2 at 3 months.

Secondary Outcomes
  • Rate of complete recovery defined by an mRS score of 0 to 1 at 3 months.

  • Clinical response rate at 3 days defined by an improvement in the NIHSS score ≥8 points compared with the initial deficit or a score ≤1.

  • Mortality at 3 months.

  • Symptomatic intracranial hemorrhage defined as the occurrence of intracerebral hemorrhage within 3 days after the onset of AIS by a ≥4-point worsening of the NIHSS score measured at admission or the highest prehemorrhage score measured during the first 24 hours.

  • Infarct growth defined as the volume of the ischemic core at baseline (based on RAPID software) and the volume of infarction manually outlined on 24-hour follow-up scan on DWI (B1000) or noncontrast computed tomography.

Sample Size

Based on the EXTEND-IA screening log, we hypothesized that 75% of the clinically eligible patients will have a TMM on baseline imaging. As our selection criteria were less selective than those proposed in EXTEND-IA, we hypothesized a 60% functional recovery rate in TMM patients. We hypothesized a 35% rate of functional recovery for no TMM patients, which is similar to what was observed in the medical arm of the RCTs. Based on these assumptions, for 90% power using 5% significance with 2-sided testing, 195 patients (136 with TMM and 59 without) were required. We anticipated that 5% of perfusion imaging studies would be technically inadequate and an attrition rate of 5%, resulting in a sample size of 220.

Statistical Analyses

Baseline characteristics were described by mean and SD or median and interquartile range (IQR) for quantitative variables and by number and percentage for qualitative variables. Descriptive analyses were performed for all study participants and by baseline imaging profile. Baseline characteristics were compared between imaging profile (TMM versus noTMM and MM versus no MM [noMM]) with χ2 test (or Fisher exact test as appropriate) for qualitative variables and with Wilcoxon–Mann Whitney U test for quantitative variables.

A logistic regression model was performed to assess whether the baseline imaging profile was associated with the primary outcome (mRS 0–2 at 3 months). The model was adjusted on prognostic factors such as age, reperfusion, NIHSS, delay from onset to end of procedure, and occurrence of reperfusion. If the verification of the comparability of the groups at inclusion revealed an imbalance between the two groups (P<0.2), the imbalance factors were also introduced as adjustment variables. The same analysis strategy of the primary end point was used for the secondary outcomes.

To assess the association between baseline imaging profile and reperfusion, a crude logistic regression model was performed. In this model, reperfusion was considered as the response variable and baseline profile as the explanatory variable.

Finally, a logistic regression model was performed for each baseline imaging profile to know whether the effect of reperfusion on the mRS at 3 months was the same according to the profile. Next, the 2 odds ratios (ORs) obtained were compared with a Breslow-Day test. Then, to have an average result on the profile, the same logistic regression model was performed on all patients but stratified on baseline imaging profile. This last methodology was also used with the MM profile and for the rate of complete recovery.

For all logistic regression models, if a subgroup had a size equal to zero, a Firth penalized maximum likelihood estimation was used. All tests were 2 sided and considered significant at α level of 0.05. All statistical analyses were conducted using SAS Software, version 9.4.

The trial protocol was approved by the French Ethical Committee (Comité de Protection des Personnes Sud-Ouest et Outre Mer III [CPP SOOM III]) on October 5, 2016, and was authorized by the French Health Authority. Every patient or his legal representative signed a written informed consent at inclusion.

Any data not published within the article are available, and anonymized data will be shared by request from any qualified investigator.

Results

Overall, 1107 patients were treated by MT in Toulouse and Bordeaux during the enrollment period from January 2017 to February 2019. Among these, 808 patients were treated by MT within 6 hours after the onset of an AIS in the anterior circulation. Five hundred eighty-eight of these patients were not enrolled in FRAME. For these patients, the mean age 72±14 years and baseline NIHSS score 17 (IQR, 12–22) were equivalent to the patients enrolled (see below). Reasons for not being enrolled in FRAME are summarized in the flowchart (Figure 1). The primary reason for noninclusion was the absence of multimodal imaging among transfer patients. These patients were typically taken directly to MT without additional imaging at the comprehensive center. An alternative software was used to process perfusion maps in 8 potentially eligible patients who were not enrolled in the study. Screened but not treated patients were not logged. Two hundred twenty patients were enrolled. Two patients were excluded from all analyses (one with no social security affiliation, and the second withdrew consent). No patient was lost of follow-up.

Figure 1.

Figure 1. Flowchart. Between January 2017 and February 2019, 808 patients were treated for large vessel occlusion (LVO)–related anterior circulation acute ischemic stroke. Reasons for noninclusion are listed in the figure. DWI indicates diffusion weighted imaging; MM, mismatch; MRA, magnetic resonance angiography; mRS, modified Rankin Scale; MT, mechanical thrombectomy; PWI, perfusion weighted imaging; and TMM, target mismatch. *In 8 patients, perfusion imaging was processed with an alternative software program, which was not blinded. Two hundred twenty were enrolled in FRAME (French Acute Multimodal Imaging Study to Select Patients for Mechanical Thrombectomy). Two patients were excluded from analyses (1 withdrew consent and 1 had no social security affiliation).

Among the 218 study participants, 208 had baseline MRI and 10 had CTP. MR perfusion imaging was not available in 2 and technically inadequate in 5. For those 7 patients, TMM status was based on the magnetic resonance angiography/core volume definition, and MM status could be assessed in 211 participants. When necessary, the core laboratory performed manual volumetric corrections to exclude artifacts: the arterial input function required manual selection in 12 patients. Artifacts were removed from 49 Tmax lesions (median artifact volume removed, 9 mL; IQR, 3–25). This resulted in 2 patients changing from MM to noMM profiles. Cerebellar artifacts were removed from 44 DWI lesions (median artifact volume removed, 20 mL; IQR, 8–30). This resulted in 4 cases changing from noMM to MM profiles.

Baseline characteristics are summarized in the Table. Mean age was 71±15 years; median NIHSS score, 17 (IQR, 12–21); and median delay from onset to imaging, 154 minutes (IQR, 105–229). One hundred and forty-five (67%) patients received intravenous thrombolysis after a median delay of 155 minutes (IQR, 116–210). Reperfusion was achieved in 187 (86%) participants after a median delay of 265 minutes (IQR, 215–355).

Table. Baseline Characteristics According to TMM Profile

All participantsTMMnoTMMP value
n=218n=155n=63
Age, y71 (15)73 (13)66 (17)0.007W
Sex, female110 (51)80 (52)30 (48)0.59C
Prestroke mRS0 (0–0)0 (0–0)0 (0–0)0.76W
Hypertension127 (58)94 (61)33 (52)0.26C
Hyperlipidemia60 (28)49 (32)11 (17)0.03C
Diabetes34 (16)22 (14)12 (19)0.37C
Active smoker34 (16)19 (12)15 (24)0.03C
Previous myocardial infarction14 (6)7 (5)7 (11)0.12F
Atrial fibrillation52 (23)41 (27)11 (18)0.16C
Previous ischemic stroke29 (13)22 (14)7 (11)0.54C
Transfer43 (20)25 (16)18 (29)0.04C
Center: Toulouse159 (73)111 (72)48 (76)0.49C
Baseline NIHSS17 (12–21)16 (11–20)19 (13–25)0.02W
Baseline systolic blood pressure, mm Hg149 (24)149 (23)149 (28)0.93W
Baseline diastolic blood pressure, mm Hg83 (18)81 (14)87 (25)0.37W
Baseline glycemia, mmol/L7 (2)7 (2)8 (2)0.20W
Imaging
 Onset to imaging, min154 (105–229)151 (103–225)162 (112–259)0.29W
 MRI at baseline208 (95)145 (94)63 (100)0.07F
 Baseline ischemic core volume, mL18 (7–59)13 (5–22)102 (71–131)<0.0001W
 Baseline critical hypoperfusion volume, mL100 (63–143)92 (61–123)139 (78–183)0.002W
 Baseline ischemic core ASPECTS7 (6–9)8 (7–9)4 (2–6)<0.0001W
 Initial occlusion site
  ICA61 (28)43 (28)18 (29)0.90C
  M1102 (47)73 (47)29 (46)0.89C
  M255 (25)39 (25)16 (25)0.97C
Intravenous thrombolysis
 Intravenous alteplase145 (67)111 (72)34 (54)0.01C
 Onset to thrombolysis, min155 (116–210)154 (109–210)162 (135–219)0.50W
Thrombectomy
 General anesthesia89 (41)58 (38)31 (49)0.13C
 Onset to femoral puncture, min221 (175–295)218 (166–288)235 (180–320)0.26W
 Onset to end of procedure, min265 (215–355)262 (215–352)287 (215–360)0.49W
 Reperfusion-TICI 2bc3187 (86)139 (90)48 (76)0.01C

Data are mean (SD), median (IQR), or n (%). NIHSS is a standardized neurological examination for which the score ranges from normal (0) to death (42). ASPECTS reflects the extent of ischemic core outlined on either DWI or CTP maps generated by RAPID (10 is normal 0) involvement of the entire middle cerebral artery territory. M1, first segment of middle cerebral artery (prebifurcation); M2, second segment of middle cerebral artery (from bifurcation to the circular sulcus of the insula in the sylvian fissure). ASPECTS indicates Alberta Stroke Program Early CT Score; C, χ2 test; CTP, computed tomography perfusion; DWI, diffusion weighted imaging; F, Fisher test; ICA, internal carotid artery; IQR, interquartile range; MRI, magnetic resonance imaging; mRS, modified Rankin Scale; NIHSS, National Institutes of Health Stroke Scale; TICI, Thrombolysis in Cerebral Infarction; TMM, target mismatch; and W, Wilcoxon test.

Study Groups

On baseline imaging, 155 (71%) patients had TMM and 63 (29%) noTMM (Table). In the noTMM group, 48 (76%) had a core lesion volume >70 mL, 37 (59%) had noMM, and 23 (37%) had both features. One hundred and seventy-four (82%) patients had MM, and 37 (18%) had noMM. Twenty-three patients in each group had a core lesion volume >70 mL. Among patients with a core >70 mL, the median core lesion volume (IQR) was 116 mL (91–146) overall, 98 mL (80–119) in MM, and 144 mL (112–193) in noMM.

There were differences in baseline characteristics between the TMM and the noTMM groups. Patients with TMM were older, had a lower baseline NIHSS, and had a smaller median core lesion volume. TMM patients were more likely to receive tPA (tissue-type plasminogen activator) and experienced a higher rate of reperfusion after MT. There were also differences in baseline characteristics between the MM and the noMM groups summarized in Table I in the Data Supplement. Both TMM and MM profiles were associated with an increased rate of reperfusion, for example, 89% among MM patients versus 70% among noMM (OR, 3.5 [95% CI, 1.5–8.1]; P=0.004).

Functional Outcome at 3 Months

TMM Versus noTMM

One hundred and seventeen (54%) patients achieved functional recovery at 3 months. The rate of functional recovery was 61% in the TMM group versus 35% in the noTMM group (OR, 3.3 [95% CI, 1.4–7.9]; P=0.007) after prespecified adjustment for age, baseline NIHSS, occurrence of reperfusion, delay from onset to reperfusion, and any imbalance in baseline characteristics (P<0.2) between the 2 groups (Figure 2A). The rate of complete recovery (mRS score, 0–1) at 3 months was 41% in the TMM group versus 17% in the noTMM, with an adjusted OR of 3.0 ([95% CI, 1.2–7.6] P=0.02).

Figure 2.

Figure 2. Functional outcome at day 90 according to target mismatch (TMM) profile.A, Functional outcome at day 90 according to the TMM profile on baseline imaging. B, Functional outcome at day 90 according to TMM profile on baseline imaging, stratified by reperfusion (Thrombolysis in Cerebral Infarction [TICI] 2bc3). mRS indicates modified Rankin Scale.

The rate of functional recovery associated with reperfusion in the TMM group was 64% in reperfusers (R+) versus 38% in nonreperfusers (R−; OR, 3.0 [95% CI, 1.02–8.7]; P=0.05). In patients with noTMM, these rates did not differ, 35% R+ versus 33% R− (OR, 1.1 [95% CI, 0.32–3.7]; P=0.88). The difference between these ORs was not significant. The OR for functional recovery after reperfusion stratified by TMM profile was 1.96 ([95% CI, 0.87–4.43] P=0.10; Figure 2B).

The rate of complete recovery associated with reperfusion in the TMM group was 45% in R+ versus 0% in R− (OR, 27.4 [95% CI, 1.5–507.4]; P=0.03) and 15% in R+ versus 27% in R− in the noTMM group (OR, 0.5 [95% CI, 0.12–1.9]). The difference between the two ORs was significant (P=0.0004), and the OR of having a complete recovery after reperfusion stratified by TMM profile was 3.4 ([95% CI, 1.1–10.2] P=0.03).

MM Versus noMM

The rate of functional recovery at 3 months in patients with MM on baseline imaging was 58% compared with 38% in noMM patients (adjusted OR, 5.9 [95% CI, 1.8–19.6]; P=0.004) after prespecified adjustment for age, baseline NIHSS, occurrence of reperfusion, delay from onset to reperfusion, and any imbalances in baseline characteristics (P<0.2) between the two groups (Figure 3A). Thirty-six percent of the MM patients versus 22% of the noMM had a complete recovery (adjusted OR, 3.1 [95% CI, 0.9–10.2]; P=0.07).

Figure 3.

Figure 3. Functional outcome at day 90 according to mismatch (MM) profile.A, Functional outcome at day 90 according to MM profile on baseline imaging. B, Functional outcome at day 90 according to MM profile on baseline imaging, stratified by reperfusion (Thrombolysis in Cerebral Infarction [TICI] 2bc3). mRS indicates modified Rankin Scale.

Reperfusion was associated with an increased rate of functional recovery in the MM group, 60% in R+ versus 32% in R− (OR, 3.3 [95% CI, 1.2–9.3]; P=0.02), but not in the noMM group, 35% in R+ versus 45% in R− (OR, 0.64 [95% CI, 0.15–2.7]; P=0.54; Figure 3B). The P of the difference between the ORs was 0.06. The OR for functional recovery after reperfusion stratified on MM profile was 1.98 ([95% CI, 0.88–4.46] P=0.10). Reperfusion was associated with an increased rate of complete recovery in the MM groups, 41% in R+ versus 0% in R− (OR, 26.8 [95% CI, 1.5–485.6]; P=0.03), but not in the noMM group, 15% in R+ versus 36% in R− (OR, 0.32 [95% CI, 0.06–1.6]; P=0.17). The P of the difference between the ORs was 0.0002 and the OR of having a complete recovery after reperfusion stratified on MM profile was 3.4 ([95% CI, 1.13–10.26] P=0.03).

Initial Clinical Response

One hundred and twenty-three of 209 assessed patients (59%) experienced a favorable clinical response. The adjusted OR for favorable clinical response in patients with TMM versus noTMM was 2.9 ([95% CI, 1.3–6.5] P=0.007). The adjusted OR for favorable clinical response in patients with MM versus noMM was 3.4 ([95% CI, 1.2–9.4] P=0.02).

Mortality at 3 Months

Overall, 29 (13%) patients died within the first 3 months, including 19 (12%) TMM and 10 (16%) noTMM. There was no difference in mortality rate based on imaging profile (TMM versus noTMM, P=0.48 or MM versus noMM, P=0.27).

Symptomatic Intracerebral Hemorrhage

Overall, 9 (4%) patients experienced a symptomatic intracranial hemorrhage, including 3 (2%) TMM and 6 (10%) noTMM. TMM on baseline imaging was associated with a lower rate of symptomatic intracranial hemorrhage (OR, 0.18 [95% CI, 0.05–0.8]; P=0.02) that was not significant after adjustment. There was no relationship between the occurrence of a symptomatic intracranial hemorrhage and MM versus noMM imaging profiles at baseline (P=0.20).

Infarct Growth

Infarct growth was measured after a median delay of 26 hours (IQR, 23–30) between baseline and follow-up imaging. This delay was equivalent in both groups. Median growth was 6.2 mL (IQR, −0.2 to 34). It was 5.4 (IQR, 0.28–24.3) in TMM versus 13.1 (IQR, −2.6 to 77.6) in noTMM (P=0.14). Infarct growth was also not different between MM and noMM (P=0.13).

Discussion

In this prospective cohort of patients experiencing an LVO-related AIS treated by MT within 6 hours and managed by investigators blinded to the baseline perfusion imaging profiles, our results show that (1) the prevalence of the favorable MM and TMM profiles was slightly less frequent than previously reported; (2) these profiles are independent predictors of functional outcome at 3 months; (3) patients with noMM profile on baseline imaging, which suggests a completed infarction, did not appear to benefit from endovascular reperfusion.

In FRAME, the investigators were blinded to the perfusion imaging results, and a low Alberta Stroke Program Early CT Score or large DWI lesion was not an exclusion criterion. These two factors likely explain the much broader patient population enrolled in FRAME compared with the penumbral imaging meta-analysis performed by Highly Effective Reperfusion Using Multiple Endovascular Devices collaborators.5 The prevalence of TMM was lower (71% versus 90%); the proportion of patients who had a large ischemic core lesion (>70 mL) was higher (22% versus 8%) as was the rate of noMM (18% versus <1%).5

Our results show for the first time in patients treated by MT within 6 hours that baseline imaging profiles are independent predictors of functional outcome at 3 months after adjustment for predefined factors including the occurrence of reperfusion. They explain, in part, why, in the early time window, the EXTEND-IA trial, which required the presence of TMM as a selection criterion for all patients, showed the largest treatment effect and the highest good outcome rate in both treatment arms compared with other RCTs that did not require systematic perfusion imaging selection.7

We also investigated the relationships between baseline imaging profile, reperfusion rates, and outcome. Interestingly, MM profile was associated with a 20% higher rate of reperfusion by comparison with noMM. One explanation for this finding is that better collaterals are associated with the MM profile and are also predictors of successful reperfusion during MT.12 In addition, the occurrence of reperfusion was associated with an increased rate of functional and complete recovery in the MM but not in the noMM group (P for the difference between the ORs was 0.06 and 0.0002, respectively). Therefore, in the early time window, MM profile appears to be a predictor of both reperfusion rates and the chance of a favorable clinical outcome following reperfusion. Our results suggest that these relationships contribute to the heterogenous clinical response to MT, despite a reperfusion rate close to 90% achieved within 4.6 hours after onset.

Our study has several limitations.

First, the major limitation was the absence of a medical management control group. Thrombectomy has previously been shown to be safe and effective in RCTs using various types of imaging inclusion criteria. It was not ethically acceptable to propose an RCT to test our hypotheses. By blinding the results of the perfusion studies and enrolling a substantial proportion of patients with a large infarction, we were able to document the underestimated prevalence of unfavorable imaging profiles in the early time window and describe their impact on patient outcome after MT. Less than 1% of patients with documented completed infarction were enrolled in previous trials.5 Therefore, a dedicated RCT testing MT versus best medical treatment is necessary to clarify the efficacy of MT in this subgroup of patients that accounted for about 20% of our study participants.

Second, some baseline characteristics of the FRAME participants were slightly different than those of the Highly Effective Reperfusion Using Multiple Endovascular Devices meta-analysis population: 25% of the patients had proximal M2 occlusions, and this, along with the higher rate of reperfusion in FRAME, may explain why the favorable outcome rate was slightly higher in our study.1

Third, in contrast to most previous studies, 95% of the enrolled patients were evaluated with MR perfusion rather than computed tomography perfusion. We used the rapid MR protocol that was developed for DEFUSE 3 (Endovascular Therapy Following Imaging Evaluation for Ischemic Stroke 3), which led to door-to-randomization times in this study similar to those obtained with CTP.2 Direct comparisons have demonstrated that both CTP and MRI processed with RAPID software can identify regions of ischemic core and critical hypoperfusion with good accuracy, and previous RCTs using both techniques did not report any significant outcome differences in MR versus CTP selected patients.2,13 In some centers, MRI is associated with treatment delays. In FRAME, the median delay from door to the end of procedure was 132 minutes. This is similar to the results of a recent survey in the United States that reported a median delay from door to first pass of 125 minutes.14

Fourth, only 20% of the eligible patients for MT were enrolled in FRAME. The main reason for excluding eligible patients was the absence of perfusion imaging in transfer patients. Age and NIHSS in excluded patients were similar to those of the patients enrolled in the study. Interestingly, among the FRAME participants, the prevalence of patients with completed infarction on baseline imaging was associated with longer delay between onset to imaging (Table I in the Data Supplement). We, therefore, suspect that the selection bias against transfer patients might have resulted in an underrepresentation of the noMM profile in our dataset.

Fifth, in FRAME, we assessed the relationships between clinical outcomes and 2 predefined baseline imaging profiles, TMM and MM. TMM was used for our sample size calculation based on the EXTEND-IA results.7

We made this choice for 2 reasons: (1) EXTEND-IA was the only study that used systematic TMM selection in the early time window; (2) this TMM definition was less stringent than its counterpart used in DEFUSE 3. As many experts felt that current imaging selection criteria might exclude some patients who may benefit from endovascular reperfusion, we, therefore, chose the less stringent definition available.

Assessment of patients with mismatch was supported by previous experience demonstrating the benefit of endovascular reperfusion in patients with large infarction with mismatch.6 Several substudies evaluating the impact of the topography of the stroke, other imaging modalities, and thresholds are ongoing to improve the definition of an optimal imaging profile in the era of personalized medicine.

In 2017, a comparative cohort study found that in patients with a large core and substantial penumbra, 25% of the participants treated by MT had a functional recovery versus none in the medical arm.15 The same finding was observed in the Highly Effective Reperfusion Using Multiple Endovascular Devices meta-analysis among the 50 patients with a large core who typically had a mismatch on CTP.5 In FRAME, half of the 46 (22%) participants with a large infarction (>70 mL) had a mismatch. Interestingly 42% of these patients experienced a functional recovery 3 months after a successful MT versus none in the absence of reperfusion (Figure I in the Data Supplement). Altogether, these findings suggest that the core size limitation (<70 mL) used to define the TMM profile likely excludes some patients who could benefit from endovascular reperfusion while the MM profile might identify patients with a large infarction who may benefit from MT. The DEFUSE 2 results showed that the benefit of endovascular reperfusion in patients with a TMM was independent from the delay from stroke onset to imaging.16 This hypothesis was validated by the positive results of DEFUSE 3, but DEFUSE 3 was restricted to patients with a core volume <70 mL.2 Therefore, we speculate that an advanced imaging profile combining a documented large core and a mismatch could be an interesting selection criterion for an extended-time-window RCT testing MT versus the best medical treatment in this patient population.

In conclusion, despite achieving a reperfusion rate of almost 90% after a median delay of only 4.6 hours from symptom onset, only 54% of the FRAME patients were functionally independent at 3 months. Our results suggest that MM profile may help to identify the patients who have the potential to respond favorably to endovascular reperfusion in the early time window and inform future clinical trials.

Nonstandard Abbreviations and Acronyms

AIS

acute ischemic stroke

EXTEND-IA

Extending the Time for Thrombolysis in Emergency Neurological Deficits With Intra-Arterial Therapy

FRAME

French Acute Multimodal Imaging Study to Select Patients for Mechanical Thrombectomy

IQR

interquartile range

LVO

large vessel occlusion

MM

mismatch

MRI

magnetic resonance imaging

mRS

modified Rankin Scale

MT

mechanical thrombectomy

NIHSS

National Institutes of Health Stroke Scale

noMM

no mismatch

noTMM

no target mismatch

OR

odds ratio

RCT

randomized controlled trial

TICI

Thrombolysis in Cerebral Infarction

TMM

target mismatch

tPA

tissue-type plasminogen activator

Supplemental Materials

Table I

Figure I

List of FRAME Coinvestigators

Footnotes

Registration: URL: https://www.clinicaltrials.gov; Unique identifier: NCT03045146.

*A list of all FRAME Investigators is given in the Data Supplement.

The Data Supplement is available with this article at https://www.ahajournals.org/doi/suppl/10.1161/STROKEAHA.120.031929.

For Sources of Funding and Disclosures, see page 240.

Presented in part at the International Stroke Conference, Los Angeles, CA, February 19–21, 2020.

Correspondence to: Jean-Marc Olivot, MD, CHU Toulouse, Neurologie Vasculaire, 1 Pl du Dr Baylac, 31059 Toulouse, France. Email

References

  • 1. Goyal M, Menon BK, van Zwam WH, Dippel DW, Mitchell PJ, Demchuk AM, Dávalos A, Majoie CB, van der Lugt A, de Miquel MA, et al; HERMES Collaborators. Endovascular thrombectomy after large-vessel ischaemic stroke: a meta-analysis of individual patient data from five randomised trials.Lancet. 2016; 387:1723–1731. doi: 10.1016/S0140-6736(16)00163-XCrossrefMedlineGoogle Scholar
  • 2. Albers GW, Marks MP, Kemp S, Christensen S, Tsai JP, Ortega-Gutierrez S, McTaggart RA, Torbey MT, Kim-Tenser M, Leslie-Mazwi T, et al; DEFUSE 3 Investigators. Thrombectomy for stroke at 6 to 16 hours with selection by perfusion imaging.N Engl J Med. 2018; 378:708–718. doi: 10.1056/NEJMoa1713973CrossrefMedlineGoogle Scholar
  • 3. Nogueira RG, Jadhav AP, Haussen DC, Bonafe A, Budzik RF, Bhuva P, Yavagal DR, Ribo M, Cognard C, Hanel RA, et al; DAWN Trial Investigators. Thrombectomy 6 to 24 hours after stroke with a mismatch between deficit and infarct.N Engl J Med. 2018; 378:11–21. doi: 10.1056/NEJMoa1706442CrossrefMedlineGoogle Scholar
  • 4. Martins SO, Mont’Alverne F, Rebello LC, Abud DG, Silva GS, Lima FO, Parente BSM, Nakiri GS, Faria MB, Frudit ME, et al; RESILIENT Investigators. Thrombectomy for stroke in the Public Health Care System of Brazil.N Engl J Med. 2020; 382:2316–2326. doi: 10.1056/NEJMoa2000120CrossrefGoogle Scholar
  • 5. Campbell BCV, Majoie CBLM, Albers GW, Menon BK, Yassi N, Sharma G, van Zwam WH, van Oostenbrugge RJ, Demchuk AM, Guillemin F, et al; HERMES Collaborators. Penumbral imaging and functional outcome in patients with anterior circulation ischaemic stroke treated with endovascular thrombectomy versus medical therapy: a meta-analysis of individual patient-level data.Lancet Neurol. 2019; 18:46–55. doi: 10.1016/S1474-4422(18)30314-4CrossrefMedlineGoogle Scholar
  • 6. Olivot JM, Sissani L, Meseguer E, Inoue M, Labreuche J, Mlynash M, Amarenco P, Mazighi M. Impact of initial diffusion-weighted imaging lesion growth rate on the success of endovascular reperfusion therapy.Stroke. 2016; 47:2305–2310. doi: 10.1161/STROKEAHA.116.013916LinkGoogle Scholar
  • 7. Campbell BC, Mitchell PJ, Kleinig TJ, Dewey HM, Churilov L, Yassi N, Yan B, Dowling RJ, Parsons MW, Oxley TJ, et al; EXTEND-IA Investigators. Endovascular therapy for ischemic stroke with perfusion-imaging selection.N Engl J Med. 2015; 372:1009–1018. doi: 10.1056/NEJMoa1414792CrossrefMedlineGoogle Scholar
  • 8. Turc G, Bhogal P, Fischer U, Khatri P, Lobotesis K, Mazighi M, Schellinger PD, Toni D, de Vries J, White P, et al. European Stroke Organisation (ESO) - European Society for Minimally Invasive Neurological Therapy (ESMINT) guidelines on mechanical thrombectomy in acute ischaemic strokeendorsed by Stroke Alliance for Europe (SAFE).Eur Stroke J. 2019; 4:6–12. doi: 10.1177/2396987319832140CrossrefMedlineGoogle Scholar
  • 9. Powers WJ, Rabinstein AA, Ackerson T, Adeoye OM, Bambakidis NC, Becker K, Biller J, Brown M, Demaerschalk BM, Hoh B, et al; American Heart Association Stroke Council. 2018 guidelines for the early management of patients with acute ischemic stroke: a guideline for Healthcare Professionals from the American Heart Association/American Stroke Association.Stroke. 2018; 49:e46–e110. doi: 10.1161/STR.0000000000000158LinkGoogle Scholar
  • 10. Kansagra AP, Goyal MS, Hamilton S, Albers GW. Collateral effect of Covid-19 on stroke evaluation in the United States.N Engl J Med. 2020; 383:400–401. doi: 10.1056/NEJMc2014816CrossrefMedlineGoogle Scholar
  • 11. Lansberg MG, Thijs VN, Bammer R, Olivot JM, Marks MP, Wechsler LR, Kemp S, Albers GW. The MRA-DWI mismatch identifies patients with stroke who are likely to benefit from reperfusion.Stroke. 2008; 39:2491–2496. doi: 10.1161/STROKEAHA.107.508572LinkGoogle Scholar
  • 12. Bang OY, Saver JL, Kim SJ, Kim GM, Chung CS, Ovbiagele B, Lee KH, Liebeskind DS. Collateral flow predicts response to endovascular therapy for acute ischemic stroke.Stroke. 2011; 42:693–699. doi: 10.1161/STROKEAHA.110.595256LinkGoogle Scholar
  • 13. Cereda CW, Christensen S, Campbell BCV, Mishra NK, Mlynash M, Levi C, Straka M, Wintermark M, Bammer R, Albers GW, et al. A benchmarking tool to evaluate computer tomography perfusion infarct core predictions against a DWI standard.J Cereb Blood Flow Metab. 2016; 36:1780–1789. doi: 10.1177/0271678X15610586CrossrefMedlineGoogle Scholar
  • 14. Menon BK, Xu H, Cox M, Saver JL, Goyal M, Peterson E, Xian Y, Matsuoka R, Jehan R, Yavagal D, et al. Components and trends in door to treatment times for endovascular therapy in get with the guidelines-stroke Hospitals.Circulation. 2019; 139:169–179. doi: 10.1161/CIRCULATIONAHA.118.036701LinkGoogle Scholar
  • 15. Rebello LC, Bouslama M, Haussen DC, Dehkharghani S, Grossberg JA, Belagaje S, Frankel MR, Nogueira RG. Endovascular treatment for patients with acute stroke who have a large ischemic core and large mismatch imaging profile.JAMA Neurol. 2017; 74:34–40. doi: 10.1001/jamaneurol.2016.3954CrossrefMedlineGoogle Scholar
  • 16. Lansberg MG, Straka M, Kemp S, Mlynash M, Wechsler LR, Jovin TG, Wilder MJ, Lutsep HL, Czartoski TJ, Bernstein RA, et al; DEFUSE 2 Study Investigators. MRI profile and response to endovascular reperfusion after stroke (DEFUSE 2): a prospective cohort study.Lancet Neurol. 2012; 11:860–867. doi: 10.1016/S1474-4422(12)70203-XCrossrefMedlineGoogle Scholar

eLetters(0)

eLetters should relate to an article recently published in the journal and are not a forum for providing unpublished data. Comments are reviewed for appropriate use of tone and language. Comments are not peer-reviewed. Acceptable comments are posted to the journal website only. Comments are not published in an issue and are not indexed in PubMed. Comments should be no longer than 500 words and will only be posted online. References are limited to 10. Authors of the article cited in the comment will be invited to reply, as appropriate.

Comments and feedback on AHA/ASA Scientific Statements and Guidelines should be directed to the AHA/ASA Manuscript Oversight Committee via its Correspondence page.