Deterioration of peri-infarct microvascular hemodynamics following successful reperfusion in acute ischemic stroke predicts infarct growth

Abstract

Background:

Microvascular hemodynamic deterioration is a proposed mechanism underlying the no-reflow phenomenon in acute ischemic stroke (AIS). Experimental studies suggested that relative heterogeneity of capillary transit time (RTH) is a potential imaging biomarker of microvascular impairment. We investigated the relationship between peri-infarct RTH and infarct growth, measured on immediate post-thrombectomy perfusion MRI and functional outcomes in AIS patients with successful reperfusion.

Methods:

In this retrospective case-control study from University of California, Los Angeles, AIS patients from 2015 to 2020 were included if they had excellent reperfusion (mTICI≥2C) and diffusion- and perfusion-weighted MRIs at baseline, immediately post-thrombectomy (PT), and 24–48h after thrombectomy. Peri-infarct regions were defined as Tmax≥6s regions on baseline MRI that were not infarcted on immediate PT-MRI. Following image coregistration, quantitative values of perfusion parameters, including RTH, were obtained from peri-infarct regions on immediate PT-MRI. Two primary outcomes were assessed: 1) infarct growth ≥10 ml, and 2) poor functional outcome (90-day mRS>2). Associations between clinical and imaging parameters and outcomes were evaluated using univariate analysis. Age, Sex, and covariates that were significant in univariate analysis were included in logistic regression analyses.

Results:

Among 47 patients included in the analysis (62% female, mean age 72±13 years, median NIHSS 14), 20 (43%) patients had substantial infarct progression and poor functional outcome. Peri-infarct RTH was significantly (p=0.005) higher in patients with infarct growth >10ml than those without and remained as an independent predictor (OR 1.4 per 10% increase, 95% CI 1.0–2.0, p=0.04). Similarly, the peri-infarct RTH was significantly higher (p<0.001) in patients with poor functional outcomes than those with good functional outcomes and remained as a significant predictor of poor functional outcome (OR 1.4 per 10% increase, 95% CI 1.0–1.9, p=0.03).

Conclusions:

Increased peri-infarct RTH on immediate PT-MRI predicts substantial infarct progression and poor functional outcomes in successfully reperfused AIS patients. Results highlight the role of microvascular dysfunction depicted by elevated RTH as a potential marker of no-reflow phenomenon despite excellent macrovascular reperfusion.

Graphical Abstract

Introduction

Timely thrombolytic therapy and thrombectomy are the most effective treatments for acute ischemic stroke (AIS) with large vessel occlusions (LVO). Approximately 20% of LVO patients treated with thrombolytics¹ and 75–85% of those treated with thrombectomy² have achieved successful reperfusion. Despite a high reperfusion rate, over half of patients do not achieve functional independence^2,3, highlighting a critical gap in understanding the pathophysiology and factors contributing to tissue reperfusion. One potential theory is the no-reflow phenomenon resulting from microvascular hemodynamic deterioration^4,5 despite macrovascular reperfusion, leading to downstream tissue injury and poor functional outcome.

The no-reflow phenomenon in brain ischemia has been described in experimental studies^6,7; however, clinical evaluation of this phenomenon remains inconsistent and controversial, with various methodologies and results^5,8–12. Multiple methods have been proposed to evaluate microvascular perfusion^13,14, including transcranial Doppler¹¹, CT or MR perfusion to calculate cerebral blood flow (CBF) and volume (CBV)^5,9,15, mean transit time (MTT)¹⁶, and time to maximum (Tmax)^8,10,17. Previous studies have developed models to calculate intra-voxel transit time standard deviation, named as capillary transit time heterogeneity (CTH)¹⁸, and its ratio to MTT, named as relative transit time heterogeneity (RTH), derived from dynamic susceptibility contrast (DSC) MR perfusion. In a tissue model with critical capillary dysfunction, RTH demonstrated more drastic changes than other perfusion parameters, such as CBF and CTH, making it a promising biomarker for cerebral microvascular hemodynamics¹⁹.

The definition of the no-reflow phenomenon in AIS has evolved as more treatment options and diagnostic imaging methods are available¹⁴. To date, there is no standard criterion for defining the presence of the no-reflow phenomenon in humans, as the intermediate pathophysiology immediately after reperfusion remains largely unknown, given clinical feasibility and limited detection methods. In previous animal studies^6,20,21, the no-reflow phenomenon occurred within the area of the initial perfusion deficit. While many human studies used perfusion status within the infarcted area alone as a marker for the no-reflow phenomenon^9,15, leaving the reperfused peri-infarct areas unexplored. Recent studies have suggested that hemodynamic changes in peri-infarct tissue precede subsequent infarct progression and functional outcome²².

In this study, the investigators aimed to take advantage of a unique design of including AIS patients who had three sets of MRIs including one immediately following mechanical thrombectomy. Having access to immediate post-thrombectomy MR perfusion provided an opportunity to investigate microvascular heterogeneity defined by RTH as a measure of no-reflow phenomenon and assess its role in the prediction of infarct progression and functional outcome immediately following successful reperfusion. We explored the relationship between peri-infarct RTH and infarct growth, measured on immediate post-thrombectomy perfusion MRI and functional outcomes in AIS patients with successful reperfusion.

Method and Material

The data that support the findings of this study are available from the corresponding author upon reasonable request. The corresponding author had full access to all the data in the study and takes responsibility for its integrity and the data analysis. This retrospective case-control study follows the STROBE reporting guideline²³.

Ethical Statement

This retrospective study was conducted under an approved institutional review board protocol (Pr-20–000724). Written informed consent was waived by the institutional review board due to its retrospective nature. This study adheres to the Strengthening the Reporting of Observational Studies in Epidemiology reporting guideline.

Patients

In this retrospective single institutional study at University of California, Los Angeles, AIS patients with anterior circulation LVO who underwent thrombectomy at University of California, Los Angeles affiliated hospitals between 2015 and 2020 were included if they had 1) excellent reperfusion (defined as modified Thrombolysis in Cerebral Infarction [mTICI] score 2c or 3); 2) baseline pre-treatment MRP and DWI, 3) immediate post-thrombectomy MRP and DWI (within 2 hours after thrombectomy); and 3) Follow-up MRI DWI within 24–48 hours after thrombectomy. Note that the post-thrombectomy MRI was optional and performed at the neurologists’ discretion when the patient’s condition permitted.

Patients were excluded if the perfusion study at baseline or post-thrombectomy had poor quality, rendering the calculation of perfusion maps unreliable. Patients were also excluded if final MRI was of non-diagnostic quality for defining final infarct volume due to motion or parenchymal hematoma.

Patients’ baseline clinical characteristics were collected, including age, sex, NIHSS at presentation, last known well to imaging time, imaging to endovascular thrombectomy time, occlusion site, IV thrombolysis, mTICI score, and stroke sitology.

DSC post-processing

Clinically used DSC perfusion parametric maps including Tmax, MTT, CBF, and CBV were available from automated post-processing performed using Olea Sphere 3.0 (SP 36, Olea Medical, France). In order to generate CTH maps, the perfusion data were reprocessed by Cercare software (Version 16.1, Cercare Medical, Denmark).

All MR images including DWI and perfusion parametric maps were then registered and normalized to Montreal Neurological Institute (MNI) space, using MATLAB (MathWorks Inc, Natick, MA, 2024a) and SPM 12 (Wellcome Trust Centre for Neuroimaging, UCL, UK).

The peri-infarct area was defined as areas of Tmax ≥ 6s on baseline MR subtracted by infarction on PT-MR (Figure 1), to encompass the entire reperfused, salvaged ischemic tissue while mitigating the effect of noise outside the original hypoperfused area. From the immediate PT-MR, mean values of Tmax, CBF, CBV, MTT, CTH, and RTH defined as CTH/MTT, were extracted within the peri-infarct area for each patient.

Figure 1.

An example of a patient with increased peri-infarct relative heterogeneity of transit time (RTH, defined as CTH/MTT ratio) and significant interval infarction progression since thrombectomy despite TICI 2c reperfusion. The patient had an unfavorable outcome and died at 90 days. Note that the RTH elevation we observed in this cohort was often patchy and focal, in contrast to the non-increased RTH demonstrated in Figure 3.

PT=Post-thrombectomy, CTH=Capillary transit heterogeneity, MTT=Mean transit time, DWI=Diffusion-weighted imaging, ROI=Region of interest.

Due to rise of ADC value following reperfusion²⁴, post-treatment infarct on post-thrombectomy and 24–48 hour follow-up MRI (24–48 hour) was segmented on DWI B1000 images using a voxel-based signal intensity method with neuroradiologist review and manual correction when necessary²⁵.

The volume of infarct growth was calculated by subtracting the volume of infarct on PT MRI from the volume of infarct on 24–48 hour MR (Figure 1). Infarct growth greater than 10 ml was defined as substantial infarct progression, as this threshold demonstrated meaningful deterioration in functional outcome in previous literature^26,27.

The primary exposure of interest in this study was elevated peri-infarct RTH.

Statistical analysis

Two main outcomes were considered for this study: 1) Substantial infarction growth ≥10 mL and 2) poor functional outcome defined as a modified Rankin Scale (90-day mRS) score greater than 2. Stata Statistical Software 19 (StataCorp LLC, TX) was used for statistical analysis. All data were expressed as mean ± standard deviation or Median (interquartile range) as appropriate. T-test or Mann-Whitney U test was used for univariate analysis.

We performed outcome-specific backward logistic regression analyses using baseline clinical and imaging variables. We first fit univariate models for each continuous predictor and χ² or Fisher’s exact test for categorical predictors, then multivariable logistic regression models including a maximum of five candidate predictors, which remained significant in univariate analysis. Age and sex were included a priori in all models as potential confounders regardless of statistical significance. We performed linear regression analysis between infarct growth and clinical and imaging variables. Given that the distribution of infarct growth volume was highly skewed, the log of infarct growth volume was used in linear regression analysis. A small constant (0.1) was added to cases with infarct growth of 0 ml before log transformation. Receiver operating characteristic (ROC) analysis was performed to diagnose substantial infarct progression and poor functional outcome. P<0.05 was considered statistically significant in demographic and clinical characteristics. The Benjamini-Hochberg method was used to correct for p-value when comparing 6 perfusion parameter characteristics for two outcomes because of multiple comparisons.

Results

A total of 280 patients with anterior circulation AIS underwent thrombectomy between 2015–2020. Among 71 AIS patients who had MR at baseline, immediate post-thrombectomy, and at 24–48 hours, fifty patients had mTICI ≥2c. Three were excluded due to poor perfusion image quality. The flow diagram is shown in Figure S1. Among 47 patients included in the analysis, 29 were females (62%) with an average age (mean± SD) of 72±13 years. The clinical characteristics of the overall cohort are summarized in Table 1.

Table 1.

Baseline characteristics and comparative analysis of clinical and imaging variables for outcome 1 (substantial infarct growth > 10mL)

	Overall cohort (n=47)	Substantial infarct progression (n=20)	No substantial infarct progression (n=27)	p-value (unadjusted)
Female; n (%)	29 (62)	12 (60)	17 (63)	0.8‡
Age, years	72 (64 – 83)	67 (61 – 76)	78 (67 – 86)	0.03 †
NIHSS at presentation	14 (9 – 19)	15 (10 – 20)	12 (8 – 18)	0.4†
Baseline penumbra (Tmax> 6 sec), mL	133 (94 – 172)	140 (98 – 179)	111 (85 – 170)	0.4†
Baseline core, mL	12 (6 – 31)	31 (16 – 52)	7 (2 – 12)	<0.001†
Core-to-penumbra ratio, %	10 (5 – 31)	29 (12 – 44)	7 (1 – 12)	<0.001†
Last known well to imaging, min	117 (65 – 295)	102.5 (65 – 273)	148 (65 – 321)	0.5†
Imaging to EVT, min	70 (60 – 90)	66.5 (60 – 82.5)	78 (55 – 90)	0.4†
IV thrombolysis, n(%)	15 (32)	8 (40)	7 (26)	0.3‡
Occlusion site, n(%)				0.9‡
ICA	5 (11)	2 (10)	3 (11)
M1	39 (83)	17 (85)	22 (82)
M2	3 (6)	1 (5)	2 (7)
Stroke etiology, n(%)				0.7‡
Cardioembolic	30 (64)	12 (60)	18 (67)
Large artery atherosclerosis	7 (15)	4 (20)	3 (11)
Cryptogenic	6 (13)	3 (15)	3 (11)
Other	4 (9)	1 (5)	3 (11)
mTICI; n (%)				0.03 ‡
2c	29 (62)	16 (80)	13 (48)
3	18 (38)	4 (20)	14 (52)
Peri-infarct perfusion values obtained from immediate post-thrombectomy MRIs
mean Tmax, sec	3.8 (2.6 – 5.3)	4.2 (3.2 – 5.9)	3.5 (2.5 – 4.3)	0.08†
mean MTT, sec	5.8 (5.2 – 7.3)	5.4 (4.9 – 7.8)	6.0 (5.2 – 7.2)	0.7†
mean CTH, sec	3.0 (2.5 – 3.9)	3.4 (2.6 – 4.8)	2.8 (2.4 – 3.6)	0.047 † *
mean rCBF	1.2 (1.0 – 1.4)	1.2 (0.9 – 1.4)	1.2 (1.1 – 1.4)	0.3†
mean rCBV	1.2 (1.1 – 1.5)	1.2 (0.9 – 1.5)	1.2 (1.1 – 1.4)	0.4†
mean RTH	0.7 (0.5 – 0.9)	0.9 (0.6 – 1.0)	0.6 (0.5 – 0.8)	0.005 † **

Data are presented as median (interquartile range), unless otherwise specified

EVT= Endovascular thrombectomy, IV= intravenous

TICI= Thrombolysis in cerebral infarction, PT= post-thrombectomy

RTH= Relative Transit time Heterogeneity, CTH= capillary transit time heterogeneity

CBF= cerebral blood flow, CBV=cerebral blood volume

MTT=Mean transit time, Tmax= time to maximum

^*Not statistically significant after corrected by Benjamini-Hochberg procedure

^**Remain statistically significant after corrected by Benjamini-Hochberg procedure

^†by Mann-Whitney U test

^‡by χ² or fisher’s exact test

Outcome-1: substantial infarct growth ≥ 10 mL

Twenty (43%) patients had infarct growth of ≥ 10 mL. The comparative univariate analysis of the baseline demographic and imaging variables among patients with and without substantial infarct progression are summarized in Table 1. Univariate analysis in patients who had substantial infarct progression (≥10 mL), showed significantly younger age (p=0.03), higher percentages of mTICI 2c (p=0.03), higher baseline core infarct volume (p<0.001), and higher peri-infarct RTH (p=0.005) values (see Table 1). The peri-infarct CTH was not significant between the two groups after correction for multiple comparisons. Notably, the peri-infarct tissue had a benign perfusion profile in patients with substantial infarct progression, with a median Tmax of 4s, and median rCBF and rCBV greater than 120% (Table 1). RTH also demonstrated a positive association with log infarct growth volume with a coefficient of 0.52 (p<0.001, Figure 2A).

Figure 2.

Scatter plot of (A) log infarct growth volume and peri-infarct relative heterogeneity of transit time (RTH) and (B) 90-day mRS and peri-infarct RTH.

Logistic regression analysis demonstrated that the peri-infarct RTH (OR 1.4 per 10% increase, 95% CI 1.0–2.0, p=0.04) and baseline core infarct volume (OR 1.1 per 1 ml increase, 95%CI 1.0–1.2, p<0.001) are independent predictors of substantial infarct progression after correcting for age and sex (Table S1). mTICI was dropped from the final model. The ROC analysis for RTH performance to predict substantial infarct progression resulted in the AUC 0.74 with CI 95% (0.60–0.89), sensitivity of 60%, and specificity of 81% at the Youden threshold of 0.85 (Figure S2).

In multivariate linear regression analysis, the peri-infarct RTH (β 0.17 per 10% increase, 95% CI 0.06–0.28, p=0.004) and baseline core infarct volume (β 0.04, 95% CI 0.02–0.06, p<0.001) again demonstrated significant association with the log infarct growth volume (R² = 0.58), after correcting for age and sex (Table S2).

Outcome-2: poor functional outcome (90D-mRS>2)

Twenty (43%) patients had poor functional outcome (90-day-mRS >2). The comparative univariate analysis of the baseline demographic and imaging variables among patients with and without poor functional outcomes is summarized in Table 2. Univariate analysis in patients with poor functional outcome showed significantly higher baseline NIHSS (p=0.016), larger infarct growth (p=0.049), shorter imaging-to-thrombectomy time (p=0.04), different proportions of stroke etiology (p=0.01), and significantly higher peri-infarct RTH values (p=0.006) (Table 2). The peri-infarct MTT difference between the two groups was not significant after correction for multiple comparisons. RTH also demonstrated a positive association with the numeric 90-day mRS with a coefficient of 0.41 (p=0.006, Figure 2B)

Table 2.

Comparative analysis of clinical and imaging variables for outcome 2 (90-day mRS scores)

	90-day-mRS >2 (n=20)	90-day-mRS ≤2 (n=27)	p-value (unadjusted)
Female; n (%)	15 (75)	14 (52)	0.1‡
Age, years	77 (68 – 86)	68 (61 – 80)	0.2†
NIHSS at presentation	16.5 (11– 21)	12 (7 – 18)	0.019 †
Infarct growth, mL	11.2 (3.9 – 30.2)	7.4 (1.0 – 10.9)	0.049 †
Baseline penumbra (Tmax> 6 sec), mL	129 (77 – 181)	133 (94 – 169)	0.9†
Baseline core, mL	15 (9 – 31)	12 (2 – 31)	0.3†
Core-to-penumbra ratio, %	16 (8 – 36)	7 (2 – 29)	0.1†
Last known well to imaging, min	176 (65 – 395)	110 (65 – 273)	0.4†
Imaging to EVT, min	67.5 (52 – 76.5)	80 (65 – 90)	0.04†
IV thrombolysis, n(%)	8 (40)	7 (26)	0.3‡
Occlusion site, n(%)
ICA	3 (15)	2 (7)	0.2‡
M1	17 (85)	22 (82)
M2	0 (0)	3 (11)
Stroke etiology, n(%)
Cardioembolic	12 (60)	18 (67)	0.01‡
Large artery atherosclerosis	6 (30)	1 (4)
Cryptogenic	0 (0)	6 (22)
Other	2 (10)	2 (7)
mTICI; n(%)			0.3‡
2c	14 (70)	15 (55)
3	6 (30)	12(45)
Peri-infarct perfusion values obtained from immediate post-thrombectomy MRIs
mean Tmax, sec	4.0 (3.2 – 5.8)	3.5 (2.4 – 5.3)	0.3†
mean MTT, sec	5.3 (5.0 – 6.3)	6.5 (5.2 – 7.3)	0.048 † *
mean CTH, sec	3.1 (2.7 – 3.8)	2.8 (2.4 – 3.9)	0.3†
mean rCBF	1.2 (1.0 – 1.5)	1.2(1.1 – 1.3)	0.5†
mean rCBV	1.2 (1.0 – 1.9)	1.2 (1.1– 1.4)	0.6†
mean RTH	0.9 (0.7–1.1)	0.6 (0.5 – 0.8)	<0.001 † **

Data are presented as median (interquartile range), unless otherwise specified

EVT= Endovascular thrombectomy, IV= intravenous

TICI= Thrombolysis in cerebral infarction, PT= post-thrombectomy

RTH= Relative Transit time Heterogeneity, CTH= capillary transit time heterogeneity

CBF= cerebral blood flow, CBV=cerebral blood volume

MTT=Mean transit time, Tmax= time to maximum

^*Not statistically significant after corrected by Benjamini-Hochberg procedure

^**Remain statistically significant after corrected by Benjamini-Hochberg procedure

^†by Mann-Whitney U test

^‡by χ² or fisher’s exact test

In logistic regression analysis, peri-infarct RTH (OR 1.4 per 10% increase, 95% CI 1.0–1.9, p=0.03), NIHSS at presentation (OR 1.2, 95% CI 1.0–1.4, p=0.04), and imaging to thrombectomy time (OR 0.94, 95% CI 0.9–1.0, p=0.02) remained as significant predictors of poor 90-day functional outcome after correcting for age and sex (Table S3). Infarct growth volume and mean peri-infarct MTT were dropped from the final model. ROC analysis of RTH on immediate PT-MRIs for predicting poor functional outcome showed an AUC of 0.79 (95% CI 0.65–0.92, Figure S3). Using Youden index analysis to define an optimal cutoff point, RTH ≥ 0.65 yielded a sensitivity of 85% and specificity of 67% for predicting poor functional outcome. The median 90-day mRS was 3 (IQR 2–5) in patients with mean peri-infarct RTH ≥ 0.65, and 1 (IQR 0–1) in those with mean peri-infarct RTH < 0.65.

Example cases with increased and non-increased peri-infarct RTH are demonstrated in Figure 1 and 3.

Figure 3.

An example of a patient with non-increased peri-infarct relative heterogeneity of transit time (RTH) and minimal interval infarction progression since thrombectomy with TICI 2c reperfusion. Patient had a favorable outcome with mRS of 0 at 90 days. Note that there is a subtle decrease in RTH compared to the contralateral region.

PT=Post-thrombectomy, CTH=Capillary transit heterogeneity, MTT=Mean transit time, DWI=Diffusion-weighted imaging, ROI=Region of interest.

Discussion

The present study provides novel insights into the value of DSC perfusion MR-based microvascular hemodynamic imaging markers for the assessment of the no-reflow phenomenon. We found that increased peri-infarct mean RTH immediately after thrombectomy is an independent predictor of substantial infarct progression and poor 90-day functional outcome after successful reperfusion. Importantly, the association between peri-infarct RTH and infarct progression remained significant when infarct growth was analyzed as a continuous variable, supporting the robustness of the observed association.

The no-reflow phenomenon is manifested as poor functional outcomes despite successful thrombectomy in over half of the LVO patients^2,3,28. Previous animal studies^6,20,29 have suggested that the no-reflow phenomenon after macrovascular recanalization is due to alterations in capillary flow at the microvascular level. Proposed mechanisms include microvascular occlusion, constriction and compression, and neutrophil plugging^4,6,13,29, all of which ultimately impede blood flow in the capillary bed. Despite its clinical relevance and experimental animal studies, the no-reflow phenomenon is poorly characterized in humans. Several factors have hindered clinical investigation of the no-reflow phenomenon. First, there remains no universally accepted imaging definition or diagnostic criterion as reperfusion treatment advances^12,14; Second, conventional imaging techniques lack sensitivity to detect microvascular-level pathophysiological changes^19,30; Third, no-reflow can occur shortly after thrombectomy²², when functional imaging is not routinely performed due to patients’ clinical vulnerability and MR scanner availability, limiting the observation window for the no-reflow phenomenon.

Because the treatment options have advanced since the initial animal study, the description of the no-reflow phenomenon has evolved and warrants an up-to-date definition. Before the emergence of thrombolytic therapy, the no-reflow phenomenon was described as a lack of tissue reperfusion at the microvascular level after macrovascular recanalization in the animal models^20,31. Later, the mismatch theory³² and angiographic reperfusion^16,33, such as the mTICI score³⁴, were derived from the concept of the no-reflow phenomenon in animal models after the development of reperfusion treatment⁴, pushing the boundaries of the no-reflow definition. Despite successful reperfusion therapy (mTICI≥2c), the infarct can still progress in the salvaged and reperfused penumbral tissue, reflecting a gap in the no-reflow phenomenon that remains unaddressed in stroke management. Therefore, it is crucial to first understand both macrovascular and microvascular hemodynamic changes during the initial reperfused stage.

Traditional perfusion parameters such as CBF, CBV, or Tmax reflect perfusion status within a relatively larger vascular bed, and have been used as surrogate markers of microvascular compromise^8,9,15. In our study, the peri-infarct areas in patients with substantial infarct progression or worse outcome demonstrated a benign perfusion profile in Tmax, MTT, rCBF, and rCBV, indicating that conventional perfusion parameters were not sensitive to microvascular hemodynamic changes, as suggested by Engedal et al¹⁹. These findings emphasize the need for imaging markers tailored to microvascular level dysfunction to evaluate the no-reflow phenomenon.

CTH, a measure of the standard deviation of MTT, increases when capillary blood flow becomes more heterogeneous and oxygen extraction declines^18,30. However, because CTH is highly correlated with MTT, it often elevates passively with delayed MTT, reflecting a mixture of both capillary bed heterogeneity and delayed transit time¹⁹. In line with this, we observed that CTH was associated with substantial infarct progression but did not predict 3-month functional outcomes.

RTH, defined as CTH/MTT, was developed to mitigate the coupling effect between MTT and CTH, thereby providing a more accurate measure of true microvascular flow heterogeneity. Engedal et al¹⁹ demonstrated through computational modeling of three phases of microvascular dysfunction after ischemia, that RTH remains stable during decreased perfusion pressure phase, increases during the heterogeneous capillary constriction phase, and decreases during the capillary occlusion phase, as capillary beds fail homogeneously¹⁹. This characteristic makes RTH a promising biomarker for detecting the critical intermediate phase of microvascular dysfunction. Engedal et al¹⁹ suggested that decreased RTH predicted infarct if recanalization was not achieved, aligning with capillary occlusion and failure. Our findings added important in-vivo evidence of the predictive value of RTH in recanalized and reperfused tissue, aligning with the experimental data¹⁹. Elevated mean peri-infarct RTH immediately after thrombectomy may reflect an area of heterogeneous capillary constriction in the salvaged tissue-at-risk, despite excellent reperfusion with mTICI ≥2c. Without additional intervention, this deterioration in microvascular function may ultimately lead to capillary occlusion, infarct growth, and a subsequent impact on functional recovery.

The no-reflow phenomenon occurs within or outside the area of the initial perfusion deficit or vascular deficit, as observed in previous animal studies^6,20,21. While many human studies evaluated perfusion status only within the infarcted area as a marker for the no-reflow phenomenon^9,15, leaving the reperfused peri-infarct areas unexplored. The microclot burden significantly increased in the post-recanalized hyperperfused area, as well as in tissue outside the occluded vascular territory²¹, indicating that the ischemic insult is causing a much more extensive downstream brain pathophysiologic change. Literature also suggests that microvascular pericytes in peri-infarct areas are critical in neurogenesis and promote post-stroke recovery^35,36. Hence, hypoperfusion status in the peri-infarct area could contribute to infarct progression³⁷ and worse clinical outcome^5,38. The definition of peri-infarct areas in this study specifically targeted the “salvaged” reperfused tissue-at-risk after successful thrombectomy that demonstrated benign conventional perfusion parameters, whereas some inevitably progress to infarction and lead to worse outcomes. Our study quantitatively addressed the impact of peri-infarct microvascular hemodynamic changes in AIS patients after excellent reperfusion.

Studies suggested that the no-reflow phenomenon occurs in approximately 20–40% acute ischemic stroke patients who achieve reperfusion^9,14,22. In our cohort, patients with elevated peri-infarct RTH had a 2-point increase in 3-month mRS compared to those with low peri-infarct RTH, leaving considerable room for treatment development. We showed that RTH is a promising imaging marker for evaluating microvascular failure, which can be calculated from a conventional DSC perfusion MR sequence.

Our study has several unique strengths that help elucidate the pathophysiology of the no-reflow phenomenon, in addition to the application of peri-infarct RTH. First, the immediate post-thrombectomy DSC MR mitigated confounding factors. When using follow-up perfusion scans obtained 24 hours later to study the no-reflow phenomenon^5,15, the re-occlusion or delayed reperfusion after thrombectomy and before 24 hours follow-up may cause false-positive or false-negative results. Second, the immediate post-thrombectomy DWI data accurately defined areas of core infarct; therefore, the perfusion changes in the core infarct areas won’t be accounted for the peri-infarct no-reflow area. Therefore, there were fewer false-positive peri-infarct no-reflow areas. This is particularly important compared to CT perfusion obtained immediately post-thrombectomy⁸, as the core estimation from NCCT and CT perfusion is less accurate, and decreased CBF or CBV in the core area may be misclassified into no-reflow areas. Third, our methodology is rigorous in defining successful reperfusion, which is defined as mTICI greater than or equal to 2c. Although previous literatures have included 2b in evaluation of no-reflow phenomenon^8,17, mTICI 2b category is heterogeneous and suffers from underlying macrovascular hypoperfusion. Our approach mitigated the confounding factor of perfusion deficit related to macrovascular no-reperfusion in mTICI 2b group and provides a more accurate analysis for the true no-reflow phenomenon.

Our study has several limitations. First, the perfusion imaging post-processing is often modality- and software-dependent, and sometimes rater-dependent. Application of the RTH in other modalities, reconstruction software, and data processing workflows requires validation. Second, our sample size is relatively small, and potential selection bias may be present, given the nature of acquiring immediate post-thrombectomy DSC MR, which is not a routine clinical practice and requires additional institutional resources and efforts. Therefore, the study is not powered to detect differences in all possible clinical or imaging markers. Nevertheless, the data shed light on the crucial stage of reperfusion immediately after thrombectomy, which could provide important evidence for developing targeted treatment. Third, some clinical data are not available in the analysis, such as the number of thrombectomy passes, which limits our ability to further explore the factors associated with the stroke outcomes. Fourth, 90-day mRS is a complex outcome measurement that is affected by many confounding factors. Therefore, results should be interpreted with caution. Fifth, although the present study focused on peri-infarct tissue, no-reflow likely represents a spectrum of microvascular dysfunction involving the infarct core, peri-infarct, and potentially remote tissue. Future studies incorporating all tissue compartments may provide a more comprehensive understanding of post-reperfusion microvascular pathophysiology.

In conclusion, we showed that microvascular hemodynamic changes, as evaluated by RTH, occurred in the peri-infarct area immediately after successful thrombectomy. Elevated RTH in the reperfused peri-infarct tissue immediately after thrombectomy is an independent predictor of substantial infarct progression and poor functional outcomes at 3 months. These novel findings provide in-human evidence of microvascular dysfunction in the no-reflow peri-infarct areas after successful thrombectomy, and may serve as a selection tool for future trials targeting the microvascular dysfunction.

Supplementary Material

Supplemental material checklist

Table S1–3

Figure S1–3

Abbreviations:

AIS

Acute Ischemic Stroke

LVO

large vessel occlusions

RTH

Relative Transit Time Heterogeneity

CBF

cerebral blood flow

CBV

cerebral blood volume

MTT

Mean transit time

Tmax

time to maximum

CTH

capillary transit time heterogeneity

DSC

dynamic susceptibility contrast

mTICI

modified Thrombolysis in Cerebral Infarction

MNI

Montreal Neurological Institute

ROC

Receiver operating characteristic