Evaluation of Large Ischemic Cores to Predict Outcomes of Thrombectomy: A Proposal of a Novel Treatment Phase

Abstract

Background

Endovascular treatment of large ischemic cores is challenging. The severity of ischemic stress is assessed using the apparent diffusion coefficient (ADC). We aimed to evaluate the ADC in patients with a low Alberta Stroke Program Early CT [Computed Tomography] Score using diffusion‐weighted imaging and whether it correlates with clinical outcomes.

Methods

This study included consecutive patients with acute large ischemic stroke (Alberta Stroke Program Early CT Score‐diffusion‐weighted imaging ≤5) who underwent endovascular treatment with successful recanalization between April 2014 and March 2023. The most frequent ADC (peak ADC) and diffusion‐weighted imaging lesion volumes were assessed. The primary outcome was the 3‐month modified Rankin Scale (mRS) score. Good (mRS score, 0–3) and poor clinical outcomes (mRS score, 4–6) were compared to confirm whether ADC was associated with clinical outcomes.

Results

In total, 78 patients were enrolled in this study; 30 had an mRS score of 0 to 3 at 3 months. The peak ADC in these patients was significantly higher than that in patients with mRS scores of 4 to 6 (P = 0.0002). In multivariate analysis, peak ADC was strongly associated with good clinical outcomes (odds ratio, 1.231; P = 0.0135) rather than onset‐to‐recanalization time and ischemic core volume. The optimal peak ADC threshold for discriminating between the mRS groups was 520×10⁻⁶ mm²/s with a sensitivity of 75% and a specificity of 73%. Good clinical outcomes were more frequently observed in patients with peak ADC ≥520×10⁻⁶ mm²/s (P<0.0001).

Conclusion

In large ischemic cores, diffusion‐weighted imaging lesions with peak ADCs ≥520×10⁻⁶ mm²/s are associated with favorable outcomes. Evaluation of the ischemic core is necessary to confirm endovascular treatment.

Nonstandard Abbreviations and Acronyms

ADC

apparent diffusion coefficient

ASPECTS

Alberta Stroke Program Early CT [Computed Tomography] Score

DWI

diffusion‐weighted imaging

EVT

endovascular treatment

mRS

modified Rankin Scale

Clinical Perspective

What Is New?

• The most frequent apparent diffusion coefficient (peak apparent diffusion coefficient) was associated with clinical outcomes after endovascular treatment in patients with a large ischemic core, with peak apparent diffusion coefficient ≥520×10⁻⁶ mm²/s associated with favorable outcomes.

• This study evaluated whether ischemic cores are important in ascertaining endovascular treatment in patients with large ischemic strokes.

What Are the Clinical Implications?

• Endovascular treatment can be considered even in cases with a large ischemic core that does not show a mismatch between clinical symptoms and diffusion‐weighted imaging lesions (“clinical‐diffusion equal”), and this is the next target following the phase of “clinical‐diffusion mismatch.”

Endovascular treatment (EVT) is effective in some patients with acute ischemic stroke attributable to large‐vessel occlusion in the anterior circulation. ¹ , ² , ³ , ⁴ , ⁵ These studies evaluated salvageable regions using computed tomography or magnetic resonance imaging (MRI), the Alberta Stroke Program Early CT [Computed Tomography] Score (ASPECTS), ⁶ and computed tomography perfusion; patients with large ischemic cores were excluded. ¹ , ² , ³ , ⁴ , ⁵

On MRI, the infarct core is defined as a diffusion‐weighted imaging (DWI) lesion with an apparent diffusion coefficient (ADC) ≤620×10⁻⁶ mm²/s. ⁷ A recent report stated that, in acute ischemic stroke, DWI lesions with mean ADC ≥520×10⁻⁶ mm²/s are still viable and salvageable using EVT. ⁸ This study demonstrated that ADC is the only factor associated with DWI reversal; consequently, neurologic function can be restored in lesions associated with DWI reversal. ⁸ An approximate ADC of 520×10⁻⁶ to 620×10⁻⁶ mm²/s may include the possibility of DWI reversal after successful reperfusion, and ADC <520×10⁻⁶ mm²/s may result in persistent infarction. ⁹ Gwak et al reported that lesion volume at 3 ADC thresholds (≤620×10⁻⁶ mm²/s, ≤540×10⁻⁶ mm²/s, and ≤520×10⁻⁶ mm²/s), and a low ADC ≤540×10⁻⁶ mm²/s/ADC ≤620×10⁻⁶ mm²/s ratio were reliable predictors of good outcomes following EVT in large ischemic stroke. ¹⁰ These studies focused on evaluating the infarct core attributable to large‐vessel occlusion and reported that mild ADC reduction may result in good outcomes. ⁸ , ¹⁰ However, the evaluation of large ischemic cores using ADC is not well established, and only a few studies have focused on ADC in patients with ischemic stroke. Yoshimura et al recently reported that patients with large ischemic cores had better functional outcomes after endovascular thrombectomy than those who received medical treatment. However, in the endovascular therapy group, a modified Rankin Scale (mRS) score of 0 to 3 at 90 days was noted in only 31% of patients, and the rate of intracranial hemorrhage within 48 hours was 58%. ¹¹ Therefore, examining factors associated with clinical outcomes in patients with large ischemic cores who underwent EVT is necessary.

We hypothesized that a decrease in ADC would reflect the clinical outcomes in patients with large ischemic cores who underwent EVT. To test this hypothesis, the ADC of all voxels in the initial DWI lesion was measured using a software program (BD score, Pixspace, Japan). Factors associated with good clinical outcomes were examined by comparing patients with good outcomes (mRS score, 0–3) and those with poor outcomes (mRS score, 4–6) at 3 months among patients who underwent EVT for a large ischemic core.

Methods

Patient Selection

All the data generated in this study are available from the corresponding author on reasonable request. This retrospective cohort study included patients with acute ischemic stroke who underwent endovascular mechanical thrombectomy between April 2014 and March 2023 at 3 hospitals (University of Occupational and Environmental Health, Kitakyushu General Hospital, and Moji Medical Center) in Japan. The study protocol was approved by the Institutional Review Board of the University of Occupational and Environmental Health. The requirement for written informed consent was waived because of the retrospective nature of the study. This study was conducted in accordance with the principles of the Declaration of Helsinki. The study results are reported according to the Strengthening the Reporting of Observational Studies in Epidemiology statement. ¹²

The inclusion criteria were as follows: (1) a diagnosis of symptomatic ischemic stroke; (2) anterior circulation occlusion (internal carotid artery, M1, or M2); (3) initial MRI data available; (4) successful recanalization using EVT; and (5) ASPECTS‐DWI ¹³ ≤5. The degree of vessel occlusion at presentation and after treatment was defined using the modified Thrombolysis in Cerebral Infarction classification. ¹⁴ This study defined successful recanalization as Thrombolysis in Cerebral Infarction grade ≥2b. Those who did not undergo MRI were excluded.

Assessments of ADC and Volumes of Ischemic Core

An ischemic core was defined as a DWI lesion caused by large‐vessel occlusion on the initial MRI scan. These areas were automatically extracted using the BD score (PixSpace software, Kitakyushu, Japan) (Figure 1). This software was programmed to identify and extract ischemic cores as high‐intensity DWI areas with ADC <800×10⁻⁶ mm²/s. The ADC was calculated from each voxel (0.9×0.9×5 mm³) of the DWI lesion. The ischemic core volume was calculated by summing the voxels. The most frequent ADC was evaluated as the peak ADC, using a voxel‐based volume histogram (Figure 1). The bin width was set to 20×10⁻⁶ mm²/s, and each grade of ADC on the histogram was obtained using a moving average per 20×10⁻⁶ mm²/s.

Figure 1.

Method of measuring peak apparent diffusion coefficient (ADC) and ischemic core volume. A, Initial diffusion‐weighted imaging (DWI) lesion attributable to left middle cerebral artery occlusion. B, The ischemic core (DWI lesion) was automatically extracted by the software (yellow region), and the volume was 82.2 mL. C, The peak ADC value (the most frequent ADC value) was 530×10⁻⁶ mm²/s.

MRI Protocol

The initial MRI was performed using a 3.0‐T scanner (Signa Premier GE Healthcare, Chicago, IL), with protocols including DWI/ADC, T2*, fluid‐attenuated inversion recovery, and intracranial magnetic resonance angiography. DWI was performed with single‐shot spin‐echo diffusion echo planar imaging using the following parameters: 220‐mm field of view, 27 5‐mm axial sections with a 1.0‐mm gap, and a b‐value of 1000 s/mm² along the 3 orthogonal directions. The axial image slices were 256×256 with voxel dimensions of 0.9 mm along the x and y axes, and 5 mm along the z axis.

Study Design

We collected the following information from the medical records: age, sex, prestroke mRS score, initial National Institutes of Health Stroke Scale score, risk factors (hypertension, diabetes, hyperlipidemia, current smoking status, and prior stroke), occluded artery, emergency treatment (EVT or tPA [tissue‐type plasminogen activator]+EVT), ASPECTS‐DWI, stroke onset‐to‐imaging time, imaging‐to‐recanalization time, stroke onset‐to‐recanalization time, peak ADC values, and volume of DWI lesions attributable to large‐vessel occlusion. Outcomes included parenchymal hemorrhage, National Institutes of Health Stroke Scale score at 7 days, and mRS score at 3 months. The study end point was the mRS score 3 months after stroke.

A good clinical outcome was defined as a mRS score of 0 to 3. Patients with good (mRS score of 0–3 at 3 months) and poor outcomes (mRS score of 4–6 at 3 months) were compared to examine whether the ADC values of whole DWI lesions were associated with clinical outcomes.

Statistical Analysis

Statistical analyses were performed using GraphPad Prism 7 software (GraphPad Software, La Jolla, CA). Categorical variables are expressed as frequencies and percentages, whereas continuous variables are expressed as mean±SD or median (interquartile range). Baseline characteristics were compared between patients with good and poor outcomes using the χ2 test for categorical variables or the Mann‐Whitney U test for continuous variables. Logistic regression analysis was performed using JMP software (SAS Institute, Cary, NC). Statistical significance was set at P<0.05.

Results

This study included 346 consecutive patients who underwent EVT for acute anterior circulation occlusion at 3 hospitals between April 2014 and March 2023. Of them, recanalization (Thrombolysis in Cerebral Infarction ≥2b) was achieved in 283 (82%) patients. Of these 283 patients, 71 who did not undergo MRI were excluded. Among the remaining patients, 78 with large ischemic stroke (ASPECTS‐DWI ≤5) were enrolled in this study (Figure 2). Total 30 patients had good clinical outcomes after 3 months.

Figure 2.

Study flowchart. ASPECTS indicates Alberta Stroke Program Early CT [Computed Tomography] Score; DWI, diffusion‐weighted imaging; EVT, endovascular treatment; mRS, modified Rankin Scale; and TICI, Thrombolysis in Cerebral Infarction.

Baseline clinical and imaging characteristics and patient outcomes are summarized in the Table . There were no significant differences in the proportions of male sex, risk factors, occluded artery, tPA, imaging‐to‐recanalization time, and recanalization state between patients with good and poor outcomes; however, age, prestroke mRS score, initial National Institutes of Health Stroke Scale score, ASPECTS‐DWI, onset‐to‐imaging time, onset‐to‐recanalization time, peak ADC, and DWI high lesion volume were significantly different between the groups.

Table 1.

Comparison of Patients With mRS Score 0 to 3 and mRS Score of 4 to 6 at 3 Months

Characteristics	Total (n = 78)	mRS score 0–3 at 3 mo (n = 30)	mRS score 4–6 at 3 mo (n = 48)	P value
Age, y	78±11	74±12	81±9	0.0074 *
Male sex	38 (49)	15 (50)	23 (48)	0.8579
Prestroke mRS score
0	45 (58)	26 (33)	7 (9)	0.0248 *
1	23 (77)	5 (17)	2 (6)
2	22 (46)	21 (44)	5 (10)
Initial NIHSS score	19 (16–25)	16 (14–19)	21 (17–28)	0.0002 *
Risk factors
Hypertension	48 (62)	15 (50)	33 (69)	0.0977
Diabetes	17 (22)	6 (20)	11 (23)	0.7615
Hyperlipidemia	14 (18)	3 (10)	11 (23)	0.1481
Current smoker	19 (24)	9 (30)	10 (21)	0.3588
Prior stroke	21 (27)	5 (17)	16 (33)	0.1064
Occluded artery
Internal carotid artery	25 (32)	50 (64)	3 (4)	0.3991
M1	7 (23)	22 (73)	1 (3)
M2	18 (38)	28 (58)	2 (4)
Emergency treatment
Intravenous tPA+EVT	42 (54)	19 (63)	23 (48)	0.1839
EVT	36 (46)	11 (37)	25 (52)
ASPECTS‐DWI	4 (3–5)	5 (4–5)	4 (3–5)	0.0058 *
Onset‐to‐imaging time, min	111 (84–207)	98 (71–146)	120 (100–303)	0.0282 *
Imaging‐to‐recanalization time, min	97 (73–124)	95 (77–122)	109 (60–132)	0.8283
Onset‐to‐recanalization time, min	230 (175–334)	202 (158–249)	242 (199–368)	0.0237 *
Peak ADC, ×10−6 mm2/s	489 (455–573)	557 (517–600)	469 (444–519)	0.0002 *
DWI high lesion volume, mL	65 (45–82)	48 (28–68)	73 (50–91)	0.0027 *
Recanalization state
TICI 2b	44 (56)		27 (56)	0.9712
TICI 3	34 (44)	13 (44)	21 (44)
Outcomes
mRS score at 3 mo	4 (2–5)	2 (1–3)	5 (5–6)	<0.0001 *

Values are expressed as mean±SD, number (percentage), or median (interquartile range). ADC indicates apparent diffusion coefficient; ASPECTS, Alberta Stroke Program Early CT [Computed Tomography] Score; DWI, diffusion‐weighted imaging; EVT, endovascular treatment; mRS, modified Rankin Scale; NIHSS, National Institutes of Health Stroke Scale; TICI, Thrombolysis in Cerebral Infarction; and tPA, tissue‐type plasminogen activator.

^* P<0.05.

Logistic regression analysis was performed to identify factors associated with good clinical outcomes (age, onset‐to‐recanalization time, DWI lesion volume, and peak ADC value). In multivariate analysis, peak ADC (odds ratio, 1.231 [95% CI, 1.029–1.473]; P = 0.0135) was strongly associated with good clinical outcomes, whereas onset‐to‐recanalization time and DWI lesion volume were not (Table S1). Furthermore, 73% (22/30) of patients with good clinical outcomes had peak ADC ≥520×10⁻⁶ mm²/s, whereas 75% (36/48) of patients with poor outcomes had peak ADC <520×10⁻⁶ mm²/s (Figure 3). To estimate the ADC threshold for predicting good clinical outcomes using EVT, receiver operating characteristic analysis was performed between the ADC values of patients with good and poor clinical outcomes (Figure 3). The optimal peak ADC threshold for discriminating good clinical outcomes was 520×10⁻⁶ mm²/s (sensitivity, 75%; specificity, 73%; and area under the curve, 0.75). According to our previous report ⁸ and the result of this analysis, the ratio of the volume of the lesion with ADC ≥520×10⁻⁶ mm²/s and the DWI lesion volume (DWI) could reflect the proportion of viable regions within the ischemic core. Bivariate analysis confirmed the relationship between the peak ADC and the ratio of ADC ≥520×10⁻⁶ mm²/s/DWI. A scatterplot visualizing the results of this analysis revealed a linear relationship (P<0.0001) (Figure S1).

Figure 3.

Comparison of peak apparent diffusion coefficient (ADC) between patients with modified Rankin Scale (mRS) score of 0 to 3 and those with mRS score of 4 to 6. Overall, 73% (22/30) of patients with mRS score of 0 to 3 had peak ADC ≥520×10⁻⁶ mm²/s, and 75% (36/48) of patients with mRS score of 4 to 6 had peak ADC <520×10⁻⁶ mm²/s. Receiver‐operating characteristic analysis was performed using peak ADC of mRS scores of 0 to 3 and 4 to 6. The optimal peak ADC threshold for good clinical outcomes was 520×10⁻⁶ mm²/s, with a sensitivity of 75% and a specificity of 73%. The area under the curve is 0.75.

To compare the outcomes between patients with peak ADC ≥520×10⁻⁶ mm²/s and those with peak ADC <520×10⁻⁶ mm²/s, the National Institutes of Health Stroke Scale score at 7 days of onset of stroke, mRS at 3 months, good clinical outcomes, and rates of parenchymal hemorrhage and mortality are summarized (Table S2). These factors were significantly better in patients with peak ADC ≥520×10⁻⁶ mm²/s than in those with peak ADC <520×10⁻⁶ mm²/s. Good functional outcomes (mRS scores ≤0–3) after 3 months were more frequently observed in patients with peak ADC ≥520×10⁻⁶ mm²/s (P<0.0001). The distribution of the mRS scores at 3 months is shown in Figure 4.

Figure 4.

Comparison of clinical outcomes at 3 months between patients with peak apparent diffusion coefficient (ADC) ≥520×10⁻⁶ mm²/s and those with peak ADC <520×10⁻⁶ mm²/s. Good functional outcomes (modified Rankin Scale [mRS] scores ≤0–3) after 3 months were more frequently observed in patients with peak ADC ≥520×10⁻⁶ mm²/s (P<0.0001).

Discussion

The results of this study demonstrate that in patients with large ischemic cores (ASPECTS‐DWI ≤5), the peak ADC was associated with clinical outcomes following EVT, and a threshold of peak ADC of 520×10⁻⁶ mm²/s was optimal in predicting the clinical outcomes. Clinical outcomes, parenchymal hemorrhage rate, and mortality rate in patients with peak ADC ≥520×10⁻⁶ mm²/s were better than those in patients with peak ADC <520×10⁻⁶ mm²/s.

According to the literature, patients with large infarctions do not receive reperfusion treatment because the ischemic core is already irreversible and associated with the risk of bleeding. ¹⁵ , ¹⁶ , ¹⁷ EVT is ineffective in patients with large ischemic cores. However, Yoshimura et al ¹¹ reported that patients with large ischemic regions had better functional outcomes with endovascular thrombectomy than with medical treatment. Therefore, understanding the efficacy of EVT in patients with large ischemic cores is crucial. We believe that the evaluation of the ischemic core is necessary to treat these patients. Our findings revealed that the peak ADC of the ischemic core, rather than the DWI lesion volume and onset‐to‐recanalization time, was strongly associated with good clinical outcomes. Furthermore, peak ADC ≥520×10⁻⁶ mm²/s could be used as a predictor of effectivity and safety for large ischemic cores.

For the ADC threshold and DWI lesion volume, Gwak et al reported that a low ratio of lesion volume (ADC <540×10⁻⁶ mm²/s) in the ischemic core indicated a small proportion of severe ischemic stress and that this low ratio was associated with good clinical outcomes with EVT. ¹⁰ This study evaluated the volume ratio of DWI lesions with ADC value <540×10⁻⁶ mm²/s and ADC value <620×10⁻⁶ mm²/s, and suggested that DWI lesions with ADC ≥540×10⁻⁶ mm²/s are viable. Umemura et al reported that DWI lesions with mean ADC ≥520×10⁻⁶ mm²/s could be treated using EVT and the neurologic functions restored, and this mean ADC value was qualitatively studied for manually setting the region of interest. ⁸ These studies demonstrated that ADC ≥520×10⁻⁶ mm²/s to 540×10⁻⁶ mm²/s is a suitable threshold and can indicate salvageable brain cells. However, both evaluations were uncertain because they did not consider the distribution of ADC in the ischemic cores. In this study, the peak ADC, which is the most frequent ADC in DWI lesions, was used to represent the ischemic core. Peak ADC is not only a qualitative, but also a quantitative, evaluation because it is calculated from the volume histogram (Figure 1). Furthermore, bivariate analysis revealed a linear relationship between peak ADC and the ratio of ADC ≥520×10⁻⁶ mm²/s/DWI lesion volumes (Figure S1). Therefore, the peak ADC reflects the volume of the viable region in the ischemic core, and could be an indicator of the ischemic core.

To evaluate the ischemic core, the distribution of ADC in DWI lesions must be considered. In this study, a histogram of the ADC values extracted from all DWI lesions facilitated visual comprehension. Figure 5 shows a comparison of the histograms in Figure 2 between the ischemic core and normal brain tissue (contralateral unaffected region) from the raw ADC data. The histogram of normal brain tissue revealed a normal distribution centered at a peak ADC of 800×10⁻⁶ mm²/s. This normal distribution collapsed to the left because of the ischemic stress. The degree of collapse is important for understanding the severity of ischemic stress, and the peak ADC of the collapsed histogram reflects the severity of the ischemic stress. Therefore, we believe that the peak ADC is a useful tool for evaluating the ischemic core.

Figure 5.

Comparison of histograms between ischemic cores and normal brain tissues. Histogram of the normal brain tissue revealed a normal distribution centered at a peak apparent diffusion coefficient (ADC) of 800×10⁻⁶ mm²/s (gray region). The black region represents the histogram of the ischemic core. This normal distribution collapsed to the left because of the ischemic stress.

In acute ischemic stroke, good clinical outcomes following reperfusion treatment are generally associated with time to treatment. ¹⁸ However, this study revealed that onset‐to‐recanalization time was not associated with good clinical outcomes in multivariate analysis (Table S1). Recently, EVT was performed for late‐window stroke in the DAWN (Diffusion‐Weighted Imaging or Computed Tomography Perfusion Assessment With Clinical Mismatch in the Triage of Wake‐Up and Late Presenting Strokes Undergoing Neurointervention With Trevo) and DEFUSE 3 (Endovascular Therapy Following Imaging Evaluation for Ischemic Stroke 3) clinical trials. ¹⁹ , ²⁰ Baek et al reported that a low hypoperfusion intensity ratio obtained from perfusion imaging was associated with favorable outcomes, even in cases with unfavorable ASPECTS and onset‐to‐recanalization time. ²¹ In these studies, the candidates were selected on the basis of imaging data. Puig et al reported that imaging data should be the center of treatment decisions rather than rigid time windows. ²² Our study also revealed that imaging analysis was more important to clinical outcomes than the onset‐to‐recanalization time.

Generally, EVT is performed via volumetric studies using perfusion imaging. ² , ⁵ , ¹⁹ , ²⁰ The ischemic core is defined as the DWI lesion volume (ADC <620×10⁻⁶ mm²/s). However, recent studies on ADC and the present study have revealed the presence of a salvageable region in the ischemic core. ⁸ , ¹⁰ In a large ischemic core, this approach could play an important role in identifying patients who are likely to benefit from EVT and in excluding those who are unlikely to benefit or are at risk of adverse effects. Severe ADC reduction is associated with an increased risk of intracranial hemorrhage. ⁸ Therefore, we believe that peak ADC could provide a novel treatment modality for large ischemic strokes. In this phase, DWI shows large high‐intensity areas corresponding to the occluded artery, indicating that there is no mismatch between clinical symptoms and DWI lesions. Therefore, this novel treatment phase can be called “clinical‐diffusion equal.” In this phase, the evaluation of the ischemic cores is necessary, and we believe that the peak ADC value can play an important role. This novel phase is the next target following the “clinical‐diffusion mismatch” phase. ²³

This evaluation method requires software that accesses the peak ADC values, defined as most frequent value of ischemic cores. For clinical use of this method without specific software, it is necessary to confirm whether manually measured ADC values can be approximated as the peak ADC value. A previous report referred to a manual measurement of the mean ADC value that was associated with DWI reversal. ⁸ Therefore, it is possible that the manual ADC study can be an alternative to calculating peak ADC values using the software, and further studies are necessary.

This study has several limitations. First, it was a retrospective study. Therefore, the design poses a risk of selection bias. Second, only patients who underwent EVT were enrolled. Therefore, it did not reflect all populations with a large infarct core. Third, advances in EVT devices during the study period may have affected our findings.

Conclusions

In stroke lesions with a large ischemic core (ASPECTS‐DWI ≤5, DWI lesions with peak ADC ≥520×10⁻⁶ mm²/s are associated with favorable outcomes. Peak ADC is a new perspective for predicting the effectiveness of EVT in large ischemic cores.

Sources of Funding

None.

Disclosures

None.

Supporting information

Table S1, S2

Figure S1