Abstract
Objective: Futile recanalization (FR)—a poor functional outcome despite successful reperfusion after mechanical thrombectomy (MT)—remains a significant issue in acute ischemic stroke owing to large vessel occlusion. This study aimed to identify predictors of FR, focusing on CT perfusion (CTP) parameters using our institutional retrospective data.
Methods: Patients who underwent MT at our institution between April 2015 and February 2023 were retrospectively reviewed. FR was defined as a 90-day modified Rankin Scale (mRS) score of 3–6 despite successful reperfusion (modified thrombolysis in cerebral infarction ≥2b). Patients with internal carotid artery (ICA) or M1 segment of the middle cerebral artery occlusion, pre-stroke mRS 0–2, stroke etiology classified as cardioembolic or embolic stroke of undetermined source, and available CTP were included. The ischemic core was defined as cerebral blood volume (CBV) <1.0 mL/100 g on CTP, and the Alberta Stroke Program Early CT Score (ASPECTS) was also evaluated. Clinical, imaging, and procedural variables were compared between the FR group and those with a favorable outcome (mRS 0–2) after successful reperfusion. Multivariable logistic regression was performed, including imaging markers and variables with p <0.1 in univariate analyses as covariates. Receiver-operating characteristic (ROC) analyses determined thresholds for ASPECTS and CBV-defined core volume, followed by sensitivity analyses.
Results: A total of 531 patients underwent MT during the study period, of whom 136 met the inclusion criteria (mean age 78 ± 11 years, 70 women, 46 ICA occlusions, median ASPECTS 9; interquartile range, 7–10). FR was observed in 69 patients (50.8%). Compared with the favorable outcome group, the FR group had significantly older age, higher baseline NIHSS scores, higher prevalence of diabetes mellitus, lower ASPECTS, larger CBV-defined core volumes, and a greater total number of device passes. Multivariable logistic regression identified older age, higher NIHSS, diabetes mellitus, and a greater total number of device passes as consistently independent predictors of FR. ROC analysis identified CBV-defined core volume ≥28.5 mL as an independent predictor of FR (area under the curve [AUC] 0.62, p = 0.013; adjusted odds ratio [aOR] 3.09, 95% confidence interval [CI] 1.23–8.28; p = 0.02); this association remained significant at ≥30 mL (aOR 2.82, 95% CI 1.14–7.33; p = 0.02) but not at ≥40 mL. ASPECTS ≤8 was also associated with FR (AUC 0.64, p = 0.002; aOR 2.92, 95% CI 1.20–7.44; p = 0.02).
Conclusion: Older age, baseline stroke severity, diabetes mellitus, and multiple device passes were major predictors of FR. Among imaging markers, a CBV-defined core volume of approximately 30 mL emerged as a clinically relevant threshold associated with increased FR risk. These findings suggest that integrating clinical, procedural, and imaging factors may help optimize patient selection, although validation in larger, multicenter studies is warranted.
Introduction
Futile recanalization (FR), commonly defined as a poor functional outcome despite technically successful recanalization (modified thrombolysis in cerebral infarction [mTICI] ≥2b), remains a major clinical challenge in the treatment of acute ischemic stroke (AIS) due to large vessel occlusion (LVO).1,2) Although mechanical thrombectomy (MT) is well established as an effective treatment for AIS with LVO, a considerable proportion of patients still experience poor clinical outcomes despite successful reperfusion.3) Previous studies show several clinical predictors of FR, including older age, baseline stroke severity, and comorbidities such as diabetes mellitus.1,2) Furthermore, imaging biomarkers—especially ischemic core volume assessed using CT perfusion (CTP) imaging—have been explored as potential predictors of FR.4) However, the predictive value of ischemic core volume remains unclear, with some studies reporting significant associations while others fail to confirm its independent predictive utility.4–7) At our institution, we previously investigated the correlation between CTP-derived ischemic core volume using GE Healthcare software (GE Healthcare Inc., Milwaukee, WI, USA) and Alberta Stroke Program Early CT Score (ASPECTS), thereby providing technical validation of the software’s core volume estimation.8) This study aimed to retrospectively analyze institutional data to identify clinical and imaging predictors of FR in patients who underwent MT for anterior circulation LVO with CTP evaluation. We specifically assessed the predictive value of ASPECTS and CTP-derived ischemic core volume as defined by cerebral blood volume (CBV) <1.0 mL/100 g.
Materials and Methods
Study design and population
This retrospective observational study, conducted at a single stroke center, included patients treated between April 2015 and February 2023. We retrospectively reviewed the records of patients who underwent MT for anterior circulation LVO.
Patients were initially screened using the following inclusion criteria:
- 1) acute occlusion of the internal carotid artery (ICA) or the M1 segment of the middle cerebral artery (MCA);
- 2) pre-stroke modified Rankin Scale (mRS) score of 0–2;
- 3) stroke etiology classified as cardioembolic or embolic stroke of undetermined source (ESUS);
- 4) successful reperfusion, defined as mTICI grade ≥2b;
- 5) availability of pretreatment CTP imaging performed using standardized acquisition protocols at our institution; and
- 6) initiation of thrombectomy within 24 hours from symptom onset or last known well (LKW).
Patients with in-hospital onset of stroke were excluded.
Stroke etiology classification
Stroke etiology was classified according to the TOAST criteria for cardioembolic stroke and the definition of ESUS proposed by Hart et al.9,10) To avoid terminological confusion, the term cryptogenic stroke was not used in this study.
Cardioembolic stroke included cases due to atrial fibrillation (either previously documented or newly detected paroxysmal AF during continuous electrocardiographic monitoring), valvular heart disease, cardiomyopathy, and intracardiac thrombus.
ESUS was defined as non-lacunar infarction without major risk cardioembolic sources and without ≥50% stenosis in the corresponding artery. Importantly, branch atheromatous disease was explicitly excluded from ESUS in this study.
Patent foramen ovale–related embolism was classified as cardioembolic stroke when judged clinically evident; otherwise, if no major-risk cardioembolic source was identified and ESUS criteria were met, cases were classified as ESUS, in line with prior reports.11)
Tandem occlusions involving both the extracranial ICA and the intracranial M1 segment were excluded.
Treatment eligibility criteria
At our institution, patients were eligible for MT based on the following clinical criteria:
- 1) initiation of treatment within 24 hours of symptom onset or LKW;
- 2) ASPECTS of 6 or greater; and
- 3) National Institutes of Health Stroke Scale (NIHSS) score of 6 or greater.
Eligibility for treatment was determined through case-by-case discussions among stroke neurologists and neurointerventional neurosurgeons for cases with uncertain treatment indications, such as those with relatively extensive early ischemic changes (ASPECTS ≤5) or with mild clinical symptoms (NIHSS <5) despite sizable ischemic lesions on imaging.
Intravenous thrombolysis with alteplase was administered according to Japanese stroke treatment guidelines.12,13) Specifically, alteplase (0.6 mg/kg) was given to patients who arrived within 4.5 hours of symptom onset or LKW, or to those arriving beyond 4.5 hours if diffusion–FLAIR mismatch was present on MRI and no contraindications existed.
All eligible patients received intravenous alteplase, and we did not adopt a strategy of intentionally skipping thrombolysis.
Imaging protocol and ischemic core assessment
In this study, non-contrast CT (NCCT) was used to determine the ASPECTS. CTP imaging was performed upon admission and processed using Advantage Workstation CT Perfusion 4D software (GE Healthcare Inc., Milwaukee, WI, USA).8) Ischemic core was defined as CBV <1.0 mL/100 g, following previously published studies. This threshold was adopted based on prior validation studies demonstrating high concordance between CBV-defined core volumes using GE software and those from RAPID software (iSchemaView, Menlo Park, CA, USA), which is widely used in clinical trials.14,15)
Data collection and outcome measures
Patients were categorized into the FR group, defined as those who achieved successful reperfusion (mTICI ≥2b) but had an unfavorable outcome at 90 days (mRS 3–6), or the effective recanalization (ER) group, defined as those who achieved successful reperfusion with a favorable outcome (mRS 0–2). Demographic variables (age, sex), clinical severity (NIHSS), vascular risk factors (atrial fibrillation, hypertension, hyperlipidemia, diabetes mellitus), smoking status, and pre-stroke mRS were all collected.
Imaging parameters included the ASPECTS, ischemic core volume defined by CBV <1.0 mL/100 g, and occlusion site (ICA or M1).
Time metrics (LKW-to-door, door-to-puncture, puncture-to-reperfusion), reperfusion status (mTICI 2b/3 and first-pass mTICI 3 recanalization), and procedures (number of device passes, first-line strategy categorized as stent retriever [SR] alone, aspiration catheter [AC] alone, or combined [SR + AC] treatment) were recorded.
Symptomatic intracranial hemorrhage (sICH) was defined according to the ECASS III criteria as any ICH associated with neurological deterioration, defined as an increase of ≥4 points in the NIHSS score from baseline (at admission) or death, when the deterioration was judged to be causally related to the hemorrhage.16) In addition to sICH, procedure-related subarachnoid hemorrhage (SAH) was defined as any SAH detected on post-procedural imaging related to the endovascular procedure.
Statistical analysis
Continuous and categorical variables were compared using the Mann–Whitney U test and chi-square test, respectively. Variables with a p-value <0.1 in univariate analysis were entered into a multivariable logistic regression model to identify independent predictors of FR.
The predictive value of ischemic imaging markers was assessed using 3 separate multivariable models:
- Model 1 included both ASPECTS and core volume defined by CBV <1.0 mL/100 g;
- Model 2 included ASPECTS alone;
- Model 3 included CBV-defined core volume alone.
In our previous study, the correlation coefficient between ASPECTS and core volume defined by CBV <1.0 mL/100 g was ρ = −0.40, indicating only a moderate correlation.8) To mitigate the effects of multicollinearity, separate models were constructed rather than including multiple ischemic core indicators within a single model. Adjusted odds ratio (aOR) and 95% confidence intervals (CIs) were then calculated for each variable.
Additionally, receiver-operating characteristic (ROC) curve analyses were performed to determine the optimal cutoff values of ASPECTS and CBV-defined core volume using the maximum Youden index. Each variable was dichotomized at the identified threshold and evaluated separately in multivariable logistic regression models. Sensitivity analyses using exploratory thresholds of CBV-defined core volume were also conducted. The adjustment set for these models was based on variables with p <0.1 in univariate analysis, consistent with the main analysis.
All statistical analyses were performed using JMP Pro, version 12.0 (SAS Institute Inc., Cary, NC, USA). A p-value <0.05 was considered statistically significant.
Ethical considerations
This study was approved by the institutional review board of Ise Red Cross Hospital (approval number: ER2020-74). Patient data were analyzed in anonymized form in accordance with the Declaration of Helsinki.
Results
During the study period, 531 patients underwent MT for acute anterior circulation LVO. After excluding patients with in-hospital stroke onset (n = 59), 472 patients with out-of-hospital stroke onset were assessed for eligibility. Figure 1 illustrates the patient inclusion flow chart. Reasons for exclusion included posterior circulation stroke, M2 segment of the MCA occlusion, anterior cerebral artery occlusion, pre-stroke mRS score of 3–5, non-cardioembolic and non-ESUS stroke, unsuccessful reperfusion, or absence of pretreatment CTP imaging. Patients meeting multiple exclusion criteria were counted only once in the final exclusion tally.
A total of 136 patients met all inclusion criteria and were included in the final analysis. The median age was 80 years (interquartile range [IQR], 72–86), and 70 patients (51%) were women. Forty-six patients (34%) had ICA occlusion, and the median ASPECTS was 9 (IQR, 7–10). Eight patients (5.9%) had ASPECTS ≤5. FR was observed in 69 patients (50.8%). Compared to the ER group, the FR group exhibited significantly older age (median 84 vs. 76 years, p <0.01), higher baseline NIHSS score (median 25 vs. 19, p <0.01), and a higher prevalence of diabetes mellitus (35% vs. 18%, p = 0.03). In imaging, the FR group had lower ASPECTS (median 8 vs. 10, p <0.01) and larger CBV-defined core volume (40.7 vs. 29.8 mL, p = 0.02). Procedural variables also differed: the FR group had longer puncture-to-reperfusion times (median 47 vs. 38 min, p = 0.03) and required more device passes (mean 2.0 vs. 1.5, p = 0.03) (Table 1).
Table 1 Comparison of patient characteristics and procedural variables between FR and ER groups
| Variables |
Total (n = 136) |
FR (n = 69) |
ER (n = 67) |
p-value |
| Age (years), median (IQR) |
80 (72–86) |
84 (75–89) |
76 (69–82) |
<0.01 |
| Male sex, n (%) |
66 (49) |
30 (43) |
36 (54) |
0.23 |
| NIHSS, median (IQR) |
21 (16–28) |
25 (20–29) |
19 (12–25) |
<0.01 |
| Left-sided lesion, n (%) |
72 (53) |
40 (56) |
32 (44) |
0.23 |
| Intravenous alteplase treatment, n (%) |
54 (40) |
25 (36) |
29 (43) |
0.40 |
| Cerebrovascular risk factors and comorbidities |
|
|
|
|
| Atrial fibrillation, n (%) |
71 (52) |
34 (49) |
37 (55) |
0.49 |
| Hypertension, n (%) |
96 (71) |
47 (68) |
49 (73) |
0.52 |
| Hyperlipidemia, n (%) |
50 (37) |
26 (38) |
24 (36) |
0.82 |
| Diabetes mellitus, n (%) |
36 (26) |
24 (35) |
12 (18) |
0.03 |
| Ischemic stroke, n (%) |
25 (18) |
12 (17) |
13 (19) |
0.76 |
| Current smoking, n (%) |
37 (27) |
19 (28) |
18 (27) |
0.93 |
| Stroke etiology |
|
|
|
|
| Cardioembolic stroke, n (%) |
117 (86) |
58 (84) |
59 (88) |
0.50 |
| ESUS, n (%) |
19 (14) |
11 (16) |
8 (12) |
|
| Premorbid antithrombotic agents |
|
|
|
|
| Antiplatelet use, n (%) |
28 (21) |
18 (26) |
10 (15) |
0.11 |
| Anticoagulant use, n (%) |
31 (23) |
14 (20) |
17 (25) |
0.48 |
| Pre-stroke mRS |
|
|
|
|
| mRS 0, n (%) |
85 (63) |
36 (52) |
49 (73) |
0.01 |
| mRS 1, n (%) |
29 (21) |
16 (23) |
13 (19) |
|
| mRS 2, n (%) |
22 (16) |
17 (25) |
5 (7) |
|
| Imaging |
|
|
|
|
| ASPECTS, median (IQR) |
9 (7–10) |
8 (6–10) |
10 (8–10) |
<0.01 |
| CBV-defined core volume (mL), median (IQR) |
35.2 (17.2–63.6) |
40.7 (22.8–78.5) |
29.8 (12.4–53.8) |
0.02 |
| Occluded vessels |
|
|
|
|
| ICA, n (%) |
46 (34) |
22 (48) |
24 (52) |
0.63 |
| MCA-M1, n (%) |
90 (66) |
47 (52) |
43 (48) |
| Time metrics |
|
|
|
|
| LKW to door (min), median (IQR) |
155 (90–345) |
140 (81–471) |
160 (91–324) |
0.71 |
| Door to puncture (min), median (IQR) |
53 (43–68) |
51 (11–100) |
56 (31–170) |
0.43 |
| Puncture to reperfusion (min), median (IQR) |
43 (31–66) |
47 (20–219) |
38 (12–154) |
0.03 |
| Door to reperfusion (min), median (IQR) |
97 (80–130) |
103 (49–273) |
93 (59–218) |
0.65 |
| LKW to reperfusion (min), median (IQR) |
267 (196–474) |
263 (112–1677) |
269 (126–1662) |
0.80 |
| Reperfusion status |
|
|
|
|
| mTICI 2b, n (%) |
73 (54) |
42 (61) |
31 (46) |
0.09 |
| mTICI 3, n (%) |
63 (46) |
27 (39) |
36 (54) |
|
| First-pass mTICI 3 recanalization, n (%) |
42 (31) |
19 (28) |
23 (34) |
0.39 |
| Procedures |
|
|
|
|
| Total device passes, mean ± SD |
1.8 ± 1.1 |
2.0 ± 1.2 |
1.5 ± 0.8 |
0.03 |
| First-line thrombectomy procedures |
|
|
|
|
| SR alone, n (%) |
15 (11) |
7 (10) |
8 (12) |
0.37 |
| AC alone, n (%) |
19 (14) |
7 (10) |
12 (18) |
|
| Combined (SR + AC), n (%) |
102 (75) |
55 (80) |
47 (70) |
|
| Symptomatic intracranial hemorrhage, n (%) |
3 (2) |
2 (2.9) |
1 (1.5) |
0.58 |
| Procedure-related subarachnoid hemorrhage, n (%) |
9 (7) |
5 (7) |
4 (6) |
0.76 |
Ischemic core was defined as regions with CBV <1.0 mL/100 g on CT perfusion. AC, aspiration catheter; ASPECTS, Alberta Stroke Program Early CT Score; CBV, cerebral blood volume; ER, effective recanalization; ESUS, embolic stroke of undetermined source; FR, futile recanalization; ICA, internal carotid artery; IQR, interquartile range; LKW, last known well; MCA-M1, M1 segment of the middle cerebral artery; mRS, modified Rankin Scale; mTICI, modified thrombolysis in cerebral infarction; NIHSS, National Institutes of Health Stroke Scale; SD, standard deviation; SR, stent retriever
Three separate multivariable logistic regression models were conducted. In Model 1, which included both ASPECTS and CBV-defined core volume, older age (aOR 1.11, 95% CI 1.05–1.18, p <0.01), higher baseline NIHSS score (aOR 1.10, 95% CI 1.04–1.17, p <0.01), diabetes mellitus (aOR 5.46, 95% CI 1.79–18.9, p <0.01), and a greater number of device passes (aOR 2.01, 95% CI 1.16–3.71, p = 0.03) were independently associated with FR. In Model 2 (ASPECTS only), ASPECTS was also found to be a significant predictor (aOR 0.70, 95% CI 0.52–0.92, p = 0.04). In Model 3 (CBV-defined core volume only), failure to achieve mTICI 3 (aOR 2.60, 95% CI 1.02–6.95, p = 0.045) and a CBV-defined core volume (aOR 1.01, 95% CI 1.01–1.03, p = 0.03) were also significant predictors (Table 2).
Table 2 Multivariable logistic regression models for predictors of FR
| Predictor |
Model 1 |
Model 2 |
Model 3 |
| aOR |
95% CI |
p-value |
aOR |
95% CI |
p-value |
aOR |
95% CI |
p-value |
| Age |
1.11 |
1.05–1.18 |
<0.01 |
1.11 |
1.05–1.18 |
<0.01 |
1.11 |
1.05–1.17 |
<0.01 |
| NIHSS |
1.10 |
1.04–1.17 |
<0.01 |
1.12 |
1.06–1.20 |
<0.01 |
1.10 |
1.04–1.17 |
<0.01 |
| Diabetes mellitus |
5.46 |
1.79–18.9 |
<0.01 |
5.11 |
1.71–17.4 |
<0.01 |
5.02 |
1.71–16.5 |
<0.01 |
| Pre-stroke mRS 0–1 vs. 2 |
2.22 |
0.65–8.25 |
0.22 |
2.01 |
0.60–7.68 |
0.28 |
2.82 |
0.85–10.6 |
0.10 |
| ASPECTS |
0.77 |
0.56–1.03 |
0.09 |
0.70 |
0.52–0.92 |
0.01 |
— |
— |
— |
| CBV-defined core volume |
1.01 |
0.99–1.02 |
0.13 |
— |
— |
— |
1.01 |
1.00–1.03 |
0.03 |
| Puncture to reperfusion |
0.99 |
0.98–1.01 |
0.49 |
0.99 |
0.98–1.01 |
0.64 |
0.99 |
0.98–1.01 |
0.37 |
| mTICI 2b vs. 3 |
2.55 |
0.98–7.00 |
0.06 |
2.34 |
0.90–6.28 |
0.08 |
2.60 |
1.02–6.95 |
0.05 |
| Total device passes |
2.01 |
1.16–3.71 |
0.02 |
1.86 |
1.09–3.38 |
0.03 |
1.99 |
1.17–3.61 |
0.02 |
Ischemic core was defined as regions with CBV <1.0 mL/100 g on CT perfusion. aOR, adjusted odds ratio; ASPECTS, Alberta Stroke Program Early CT Score; CBV, cerebral blood volume; CI, confidence interval; FR, futile recanalization; mRS, modified Rankin Scale; mTICI, modified thrombolysis in cerebral infarction; NIHSS, National Institutes of Health Stroke Scale
The median CBV-defined core volume in our cohort was 35.2 mL (IQR 17.2–63.6). ROC curve analysis identified 28.5 mL as the optimal cutoff of CBV-defined core volume for predicting FR (area under the curve [AUC] 0.62, p = 0.013; Fig. 2). CBV ≥28.5 mL was independently associated with FR (aOR 3.09, 95% CI 1.23–8.28, p = 0.02). For practical interpretability, exploratory sensitivity analyses were also performed using thresholds of CBV-defined core volume at 30 and 40 mL. Consistent results were observed for CBV ≥30 mL (aOR 2.82, 95% CI 1.14–7.33, p = 0.02), whereas CBV ≥40 mL did not reach significance (aOR 1.57, 95% CI 0.65–3.90, p = 0.32). For ASPECTS, the optimal cutoff was 8 (AUC 0.64, p = 0.002; Fig. 2), and ASPECTS ≤8 was independently associated with FR (aOR 2.92, 95% CI 1.20–7.44, p = 0.02). Detailed results are presented in Table 3.
Table 3 Multivariable logistic regression analysis using threshold-based cut-offs of ischemic imaging markers for predictors of FR
| Predictor |
aOR |
95% CI |
p-value |
| CBV-defined core volume ≥28.5 mL |
3.09 |
1.23–8.28 |
0.02 |
| CBV-defined core volume ≥30 mL |
2.82 |
1.14–7.33 |
0.02 |
| CBV-defined core volume ≥40 mL |
1.57 |
0.65–3.90 |
0.32 |
| ASPECTS ≤8 |
2.92 |
1.20–7.44 |
0.02 |
Multivariable models were adjusted for age, baseline NIHSS, diabetes mellitus, pre-stroke mRS (0–1 vs. 2), puncture-to-reperfusion time, mTICI (2b vs. 3), and total device passes. Ischemic core was defined as regions with CBV <1.0 mL/100 g on CT perfusion. aOR, adjusted odds ratio; ASPECTS, Alberta Stroke Program Early CT Score; CBV, cerebral blood volume; CI, confidence interval; FR, futile recanalization; mRS, modified Rankin Scale; mTICI, modified thrombolysis in cerebral infarction; NIHSS, National Institutes of Health Stroke Scale
Discussion
This study aimed to identify factors that predict FR after MT for LVO in the anterior circulation. In our analysis, FR was defined as a 90-day mRS score of 3–6, despite successful reperfusion (mTICI ≥2b), consistent with definitions widely adopted in previous literature.1,2) Importantly, our inclusion criteria limited patients to those with a pre-stroke mRS score of 0–2, indicating baseline functional independence. Consequently, an mRS score of 3 or higher at 90 days indicates a substantial decline in functional status and a clinically significant loss of independence. For patients with a pre-stroke mRS of 2, progression to mRS 3 indicates a newly required level of assistance with daily activities, thus justifying its inclusion in the FR definition.
This study confirms several previously reported predictors of FR after MT, including older age, higher baseline NIHSS, diabetes mellitus, and a higher number of device passes.1,2) Furthermore, failure to achieve mTICI 3 was marginally associated with FR in Model 3 (p = 0.045); however, this association did not consistently achieve statistical significance across all models. Older patients and those with diabetes mellitus tend to have reduced neurological recovery potential, even with successful recanalization, attributed to poorer collateral status and impaired neuroplasticity.1,2) Notably, diabetes mellitus showed the strongest association with FR among all predictors in our models. This may reflect broader metabolic vulnerability or the influence of unmeasured confounders such as glycemic control or comorbid metabolic conditions. Therefore, patient selection remains a vital component of endovascular therapy for AIS.
Additionally, this study underscores the importance of procedural performance in reducing FR rates. A greater number of device passes consistently emerged as an independent predictor of FR across all models, which is consistent with prior findings that emphasize the value of rapid and complete reperfusion.1–3) Conversely, while failure to achieve mTICI 3 showed only marginal significance in Model 3, its inconsistent association across models suggests a limited predictive contribution in this cohort. Despite ongoing technical advances, prolonged procedure times and multiple device passes continue to negatively affect outcomes by increasing the risk of endothelial injury, distal embolization, and hemorrhagic complications.17) We also evaluated the relationship between first-line device strategy (SR alone, AC alone, or combined [SR + AC]) and FR; however, no significant association was found. In our cohort, most patients underwent combined (SR + AC) treatment as the first-line approach, which resulted in a potential imbalance across groups. This imbalance may have limited the power to detect differences among device strategies; therefore, further studies using a more evenly distributed device selection are warranted to clarify the impact of thrombectomy technique on FR.
Imaging findings also played a significant role, as both lower ASPECTS and larger ischemic core volumes on CTP were associated with FR in the univariate analysis. This finding aligns with previous studies reporting that lower ASPECTS scores are associated with poor functional outcomes in patients who underwent MT.1,2) Our multivariable analysis identified ASPECTS as an independent predictor of FR when included alone (Model 2), suggesting its utility as a qualitative marker of early ischemic change.
In addition, core volume, defined by CBV <1.0 mL/100 g, was also an independent predictor of FR in a separate model (Model 3). This finding supports the role of perfusion-derived quantitative metrics in outcome prediction, consistent with literature emphasizing ischemic burden, as assessed by CTP, as a key determinant.4–6) Notably, the predictive value of ischemic core volume was not evident when included in the same model as ASPECTS, suggesting potential redundancy between these two indicators.
However, our additional cutoff analyses demonstrated that even among patients selected as small-core (ASPECTS ≥6), those with a CBV-defined core volume ≥28.5 mL were at an approximately 3-fold higher risk of FR. Importantly, this association remained significant when applying a pragmatic threshold of 30 mL, whereas CBV ≥40 mL did not reach significance, likely reflecting the limited number of larger-core patients in our cohort. The 30-mL cutoff was selected as a clinically interpretable approximation of the ROC-derived value (28.5 mL), while the 40-mL cutoff was tested with reference to the concept of the maximal admission core lesion compatible with favorable outcome proposed by Ribo et al., who reported an upper limit of approximately 42 mL when using CTP-CBV.6) Although their definition of ischemic core (CBV <1.5 mL/100 mL) and perfusion software (Siemens Syngo VPCT, Erlangen, Germany) differed from ours (CBV <1.0 mL/100 g), this concept is consistent with the relevance of exploring thresholds in this range. For ASPECTS, the optimal cutoff was 8, and ASPECTS ≤8 was independently associated with FR, although interpretation should be cautious given the restriction to ASPECTS ≥6 cases. Taken together, these findings suggest that threshold-based imaging markers—particularly a CBV-defined core volume of approximately 30 mL—may enhance the clinical applicability of perfusion imaging for risk stratification, beyond per-unit increases in ischemic core size.
sICH was infrequent in our cohort (n = 3), yet its clinical significance should not be underestimated. A recent meta-analysis by Shen et al. demonstrated a strong association between sICH and FR (aOR 7.37, 95% CI 4.89–11.12),1) highlighting the importance of minimizing hemorrhagic complications to improve the overall effectiveness of MT. SAH related to the procedure was evaluated and showed no significant difference between the FR and ER groups in our cohort. However, its potential impact on study outcomes warrants further investigation in larger-scale studies.
Emerging approaches, including machine learning models that integrate clinical, imaging, and procedural data, show potential for improving FR prediction.18) However, their applicability in routine clinical practice still requires validation.18,19) Until such tools are widely adopted, clinicians should maintain a holistic approach when deciding on MT candidacy, particularly for patients with high-risk profiles.
The predictors identified in this study may support the development of risk stratification tools and facilitate transparent communication during informed consent. Specifically, a large CBV-defined ischemic core volume—shown to be an independent predictor of FR in our analysis—provides a clinically actionable threshold for risk assessment. A lower ASPECTS score was also associated with FR, although its discriminative value appears more limited within this ASPECTS ≥6 cohort. Integrating these predictors into clinical workflows could enhance individualized decision-making and help patients understand that even with successful recanalization, favorable functional outcomes are not guaranteed, particularly in the presence of such FR risk factors.
This study has some limitations. First, its retrospective, single-center design may limit the generalizability of the findings. Moreover, potential selection bias in patient recruitment and variations in institutional treatment protocols, including indications for thrombectomy and post-procedural care, may have affected the study results. Second, although CTP-derived ischemic core volumes were evaluated, the analysis was limited to CBV <1.0 mL/100 g with GE Healthcare software. This vendor-specific approach, while based on available evidence, may have influenced the estimation of ischemic burden and limited the reproducibility of our thresholds. Variability in imaging protocols and software thresholds across institutions may also influence core volume estimation. Furthermore, patient inclusion was restricted to those with ASPECTS ≥6, so the predictive value of ASPECTS ≤8 identified in our analyses should be interpreted cautiously within this limited range. Third, potential factors such as collateral status or frailty markers were not assessed. Larger multicenter studies are needed to validate these findings and refine FR prediction models.
In conclusion, reducing FR requires a balanced approach that accounts for patient characteristics and procedural factors. Efforts should prioritize achieving rapid and complete reperfusion with minimal device passes, alongside careful patient selection. In addition, threshold-based imaging markers may provide greater clinical applicability: patients with a CBV-defined core volume around 30 mL were consistently at increased risk of FR, whereas a lower ASPECTS score (≤8) showed a more limited but supportive predictive value. Such cutoff values may enhance the use of perfusion and NCCT imaging for risk stratification in MT candidates. Future research should aim to validate integrated predictive models and explore additional clinical and imaging indicators to support individualized treatment strategies.
Conclusion
This study identifies older age, higher baseline stroke severity, diabetes mellitus, and multiple device passes, together with imaging markers such as a larger ischemic core volume defined by CBV <1.0 mL/100 g and, more modestly, lower ASPECTS, as independent predictors of FR following MT for anterior circulation LVO. These findings highlight the importance of considering patient characteristics, procedural efficiency, and imaging markers—with particular emphasis on CBV-defined core volume around 30 mL as a clinically actionable threshold—in optimizing strategies for FR risk reduction. Validation in larger, prospective, multicenter studies is warranted, and these insights may contribute to the development of individualized treatment strategies.
Disclosure Statement
The authors declare that they have no conflicts of interest.
References
- 1) Shen H, Killingsworth MC, Bhaskar SMM. Comprehensive meta-analysis of futile recanalization in acute ischemic stroke patients undergoing endovascular thrombectomy: prevalence, factors, and clinical outcomes. Life (Basel) 2023; 13: 1965.
- 2) Ni H, Liu X, Hang Y, et al. Predictors of futile recanalization in patients with acute ischemic stroke undergoing mechanical thrombectomy in late time windows. Front Neurol 2022; 13: 958236.
- 3) Kitano T, Todo K, Yoshimura S, et al. Futile complete recanalization: patients characteristics and its time course. Sci Rep 2020; 10: 4973.
- 4) Fu M, Yang J, Dong X, et al. Association of critical hypoperfusion biomarkers on CT with futile recanalization and poor outcome after mechanical thrombectomy in acute ischemic stroke. BMC Neurol 2024; 24: 406.
- 5) Dhillon PS, Butt W, Podlasek A, et al. Perfusion imaging for endovascular thrombectomy in acute ischemic stroke is associated with improved functional outcomes in the early and late time windows. Stroke 2022; 53: 2770–2778.
- 6) Ribo M, Tomasello A, Lemus M, et al. Maximal admission core lesion compatible with favorable outcome in acute stroke patients undergoing endovascular procedures. Stroke 2015; 46: 2849–2852.
- 7) Heit JJ, Mlynash M, Christensen S, et al. What predicts poor outcome after successful thrombectomy in late time windows? J Neurointerv Surg 2021; 13: 421–425.
- 8) Nakanishi K, Sano T, Iwaki K, et al. Quantitative evaluation of ischemic core volume in GE’s CT perfusion imaging analysis software and its relationship to Alberta Stroke Program Early CT Score. J Neuroendovasc Ther 2025; 19: 2024–0090.
- 9) Adams HP Jr., Bendixen BH, Kappelle LJ, et al. Classification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. TOAST. Trial of Org 10172 in Acute Stroke Treatment. Stroke 1993; 24: 35–41.
- 10) Hart RG, Diener HC, Coutts SB, et al. Embolic strokes of undetermined source: the case for a new clinical construct. Lancet Neurol 2014; 13: 429–438.
- 11) Elgendy AY, Saver JL, Amin Z, et al. Proposal for updated nomenclature and classification of potential causative mechanism in patent foramen ovale-associated stroke. JAMA Neurol 2020; 77: 878–886.
- 12) Miyamoto S, Ogasawara K, Kuroda S, et al. Japan stroke society guideline 2021 for the treatment of stroke. Int J Stroke 2022; 17: 1039–1049.
- 13) Toyoda K, Koga M, Iguchi Y, et al. Guidelines for intravenous thrombolysis (recombinant tissue-type plasminogen activator), the third edition, March 2019: a guideline from the Japan Stroke Society. Neurol Med Chir (Tokyo) 2019; 59: 449–491.
- 14) Liu QC, Jia ZY, Zhao LB, et al. Agreement and accuracy of ischemic core volume evaluated by three CT perfusion software packages in acute ischemic stroke. J Stroke Cerebrovasc Dis 2021; 30: 105872.
- 15) Li X, Zeng C, Li Y, et al. Comparison of 3 CT perfusion software packages in estimation of ischemic lesions in acute ischemic stroke patients. J Comput Assist Tomogr 2023; 47: 500–506.
- 16) Hacke W, Kaste M, Bluhmki E, et al. Thrombolysis with alteplase 3 to 4.5 hours after acute ischemic stroke. N Engl J Med 2008; 359: 1317–1329.
- 17) Bruggeman AAE, Kappelhof M, den Hartog SJ, et al. Successful reperfusion in relation to the number of passes: comparing outcomes of first pass expanded Treatment in Cerebral Ischemia (eTICI) 2B with multiple-pass eTICI 3. J Neurointerv Surg 2023; 15: 120–126.
- 18) Lin X, Zheng X, Zhang J, et al. Machine learning to predict futile recanalization of large vessel occlusion before and after endovascular thrombectomy. Front Neurol 2022; 13: 909403.
- 19) Meinel TR, Lerch C, Fischer U, et al. Multivariable prediction model for futile recanalization therapies in patients with acute ischemic stroke. Neurology 2022; 99: e1009–e1018.
