Next Article in Journal
Cubebin Attenuates Methamphetamine-Induced Neurotoxicity Through CREB/BDNF/Caspase-3 Signaling: In Vivo and In Silico Study
Previous Article in Journal
Biliary Drainage for the Preoperative Management of Periampullary Neoplasms: A Retrospective Cohort Study
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Medicina
Volume 61
Issue 9
10.3390/medicina61091566
medicina-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Comparative Diagnostic and Prognostic Performance of SWI and T2-Weighted MRI in Cerebral Microbleed Detection Following Acute Ischemic Stroke: A Meta-Analysis and SPOT-CMB Study

by
Rachel Tan
1,2,3,
Kevin J. Spring
1,2,4,5,6,
Murray Killingsworth
1,2,3,4,7 and
Sonu Bhaskar
1,3,4,8,9,*,†
1
Global Health Neurology Lab, Sydney, NSW 2150, Australia
2
UNSW Medicine and Health, South Western Sydney Clinical Campuses, University of New South Wales (UNSW), Sydney, NSW 2052, Australia
3
Clinical Sciences Stream, Ingham Institute for Applied Medical Research, Sydney, NSW 2170, Australia
4
NSW Brain Clot Bank, NSW Health Pathology, Sydney, NSW 2170, Australia
5
Medical Oncology Group, Ingham Institute for Applied Medical Research, Sydney, NSW 2751, Australia
6
School of Medicine, Western Sydney University, Sydney, NSW 2000, Australia
7
Correlative Microscopy Facility, Department of Anatomical Pathology, NSW Health Pathology, and Liverpool Hospital, Liverpool, NSW 2170, Australia
8
Department of Neurology & Neurophysiology, Liverpool Hospital & South Western Sydney Local Health District (SWSLHD), Sydney, NSW 2170, Australia
9
Department of Neurology, Division of Cerebrovascular Medicine & Neurology, National Cerebral and Cardiovascular Center (NCVC), Suita 564-8565, Osaka, Japan
*
Author to whom correspondence should be addressed.
†
Mailing address: National Cerebral and Cardiovascular Center (NCVC), Department of Neurology, 6-1 Kishibeshimmachi, Suita 564-8565, Osaka, Japan.
Medicina 2025, 61(9), 1566; https://doi.org/10.3390/medicina61091566
Submission received: 27 July 2025 / Revised: 19 August 2025 / Accepted: 27 August 2025 / Published: 30 August 2025
(This article belongs to the Section Neurology)
Browse Figures
Versions Notes

Abstract

Background and Objectives: Cerebral microbleeds (CMBs) are increasingly being considered as potential biomarkers of small vessel disease and cerebral vulnerability, particularly in patients with acute ischemic stroke (AIS). Accurate detection is crucial for prognosis and therapeutic decision-making, yet the relative utility of susceptibility-weighted imaging (SWI) versus T2*-weighted imaging (T2*) remains uncertain. Materials and Methods: We conducted a systematic review and meta-analysis (SPOT-CMB, Susceptibility-weighted imaging and Prognostic Outcomes in Acute Stroke—Cerebral Microbleeds study) of 80 studies involving 28,383 AIS patients. Pooled prevalence of CMBs was estimated across imaging modalities (SWI, T2*, and both), and stratified analyses examined variation by demographic, clinical, and imaging parameters. Meta-analytic odds ratios assessed associations between CMB presence and key outcomes: symptomatic intracerebral hemorrhage (sICH), hemorrhagic transformation (HT), and poor functional outcome (modified Rankin Scale score 3–6) at 90 days. Diagnostic performance was assessed using summary receiver operating characteristic curves. Results: Pooled CMB prevalence was higher with SWI (36%; 95% CI 31–41) than T2* (25%; 95% CI 22–28). CMB presence was associated with increased odds of sICH (OR 2.22; 95% CI 1.56–3.16), HT (OR 1.33; 95% CI 1.01–1.75), and poor 90-day outcome (OR 1.61; 95% CI 1.39–1.86). However, prognostic performance was modest, with low sensitivity (e.g., AUC for sICH: 0.29) and low diagnostic odds ratios. SWI outperformed T2* in detection but offered limited prognostic gain. Access to SWI remains limited in many settings, posing challenges for global implementation. Conclusions: SWI detects CMBs more frequently than T2* in AIS patients and shows stronger associations with adverse outcomes, supporting its value for risk stratification. However, prognostic accuracy remains limited, and our GRADE appraisal indicated only moderate certainty for functional outcomes, with lower certainty for diagnostic accuracy due to heterogeneity and imprecision. These findings highlight the clinical utility of SWI but underscore the need for standardized imaging protocols and high-quality prospective studies.

1. Background

Cerebral microbleeds (CMBs) are small, chronic brain hemorrhages that are increasingly recognized as critical markers of cerebrovascular pathology, particularly following acute ischemic stroke (AIS) [1]. Globally, AIS remains a leading cause of mortality and morbidity, accounting for over 63.5 million disability-adjusted life years (DALYs) in 2019 [2]. The prevalence of CMBs rises from 10% in the general population to 34% in AIS patients, with higher rates observed in hypertensive and elderly individuals [3]. Accurate CMB detection is essential for guiding clinical decision-making, particularly for managing risks such as symptomatic intracerebral hemorrhage (sICH), hemorrhagic transformation (HT), stroke recurrence, and functional decline [4,5,6]. For instance, CMB presence can influence the initiation of anticoagulation therapy or other therapeutic interventions, underscoring its prognostic significance [7]. Beyond being static markers of past hemorrhage, emerging evidence suggests that CMBs may represent dynamic sites of ongoing microvascular injury, inflammation, and impaired glymphatic clearance, a concept we refer to as the Living Lesion Paradigm (under review) [8]. This paradigm has important implications for both imaging-based detection and clinical outcomes in AIS patients, warranting a re-evaluation of CMBs as active contributors to disease progression rather than inert remnants [8].
Currently, CMB detection relies on magnetic resonance imaging (MRI) sequences, with hemosiderin-sensitive techniques such as T2*-weighted imaging (T2*) and susceptibility-weighted imaging (SWI) being the most effective [9]. T2* is widely used for its accessibility but has limited sensitivity [10], while SWI offers superior contrast and sensitivity, though it is not yet standard in many acute stroke protocols [11]. In contrast, other imaging modalities, such as fluid-attenuated inversion recovery (FLAIR) and non-contrast computed tomography (NCCT), are less reliable for CMB detection due to poor sensitivity to blood degradation products [12,13], although their effectiveness when combined with SWI or T2* remains unclear. Despite these advancements, there remains a limited understanding of the extent to which SWI offers additional diagnostic benefit over T2*. Additionally, the lack of standardized imaging protocols and variability in reported diagnostic accuracy across modalities pose significant challenges for clinical implementation [14]. Addressing these gaps, our systematic review and meta-analysis (SPOT-CMB, Susceptibility-weighted imaging and Prognostic Outcomes in Acute Stroke—Cerebral Microbleeds study) aimed to systematically evaluate CMB prevalence based on the use of SWI or T2*, quantify their relative detection performance, and explore the clinical implications of CMB burden in AIS patients.
The primary objective of this study was to undertake a comprehensive evaluation of CMBs in the context of AIS. Our objectives were to (1) estimate the pooled prevalence of CMBs in AIS patients using SWI compared with T2*; (2) assess how prevalence varies according to patient demographics, stroke subtype, imaging parameters, and regional differences; and (3) evaluate the association of CMBs with clinically relevant outcomes, including sICH, HT, and functional outcome at 90 days.

2. Materials and Methods

2.1. Literature Search and Study Selection

A comprehensive literature search was conducted using PubMed, EMBASE, the Cochrane Central Register of Controlled Trials, Scopus, and Web of Science to identify relevant studies published between January 2000 and May 2025. The search strategy incorporated a combination of keywords, including “cerebral microbleeds”, “microhemorrhages”, “ischemic stroke”, “cerebral infarction”, “susceptibility-weighted imaging”, “T2-star imaging”, “flair attenuated inversion recovery”, or “non-contrast computed tomography”. The full search strategy is provided in the Online Supplementary Information (Search Strategy). In addition to database searches, the reference lists of relevant studies, systematic reviews, and meta-analyses were screened to capture any additional eligible studies. The study selection process, inclusion criteria, and subgroup analyses conducted as part of this meta-analysis were summarized using a PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) flow diagram (Figure 1). This review was conducted following the PRISMA 2020 guidelines (Supplemental Table S1) and the MOOSE (Meta-analysis of Observational Studies in Epidemiology) reporting standards (Supplemental Table S2), as detailed in the Supplementary Information. This study was registered in Open Science, registration number “fks6z” (https://osf.io/fks6z/, accessed on 19 July 2025).

2.2. Inclusion and Exclusion Criteria

Studies were deemed eligible for inclusion if they met the following criteria: (a) patients diagnosed with AIS; (b) patients aged 18 years or older; (c) reported data on the baseline presence of CMBs; (d) availability of comparative data between CMB-positive and CMB-negative groups for relevant post-stroke outcomes; (e) applied an appropriate study design with a minimum sample size of at least 20 patients.
Studies were excluded if they were: (a) systematic reviews, meta-analyses, case reports, or narrative reviews; (b) involved animal experiments; (c) did not provide access to full-text articles; (d) lacked relevant data on baseline CMB status or post-stroke outcomes; (e) were not published in English; (f) duplicated publications.

2.3. Data Extraction

All article titles and abstracts were initially screened using EndNote (Clarivate Analytics, London, UK) to exclude studies that did not meet the predefined eligibility criteria. Full-text articles for potentially relevant studies were then assessed in detail to determine final eligibility for inclusion in the systematic review and meta-analysis. Data extraction was performed using a standardized data collection form, which captured the following information from each study:
(1)
Study characteristics: author, country, publication year, study name or registry, study design, cohort size;
(2)
Participant characteristics: age, sex, comorbidities, number of patients with CMBs at baseline, stroke subtype, CMB location, and specific characteristics of patients with AIS;
(3)
Imaging parameters: MRI sequence type for CMB detection, field strength, slice thickness.
The ‘SWI and T2*’ subgroup is defined as studies that visualized CMBs in their patients using either SWI or T2* sequences. Slice thickness was extracted as reported and categorized using study-defined thresholds: Thin (≤2 mm), Medium (2.1–4.9 mm), and Thick (≥5 mm), based on radiological conventions commonly applied in neuroimaging studies [15,16];
(4)
Definition and criteria of various parameters: CMBs, sICH, poor functional outcome;
(5)
Clinical outcomes: occurrence of sICH, HT, and mRS score for functional outcome at 90 days, assessed in relation to the presence or absence of CMBs.
Disagreements were resolved through discussions, and if unresolved, adjudicated by a third reviewer.

2.4. Methodological Quality Assessment of Included Studies

The methodological quality of the included studies was assessed using the Modified Jadad Analysis (MJA) [17] (Supplemental Table S3). In addition, potential bias related to funding sources was examined by reviewing each study’s declarations of funding and conflicts of interest (Supplemental Table S4).

2.5. Certainty of Evidence Assessment

We evaluated the certainty of evidence using the Grading of Recommendations, Assessment, Development and Evaluations (GRADE) approach. Each outcome (symptomatic intracerebral hemorrhage, hemorrhagic transformation, and poor functional outcome) was assessed for study limitations (risk of bias), inconsistency, indirectness, imprecision, and publication bias. A Summary of Findings (SoF) table (SPOT-CMB GRADE SoF) was constructed, presenting pooled effect estimates, absolute effects, and certainty ratings. This allowed us to make a transparent comparison of strengths and limitations across outcomes and facilitated interpretation of the prognostic value of CMBs.

2.6. Statistical Analyses

All statistical analyses were conducted using STATA version 13.0 (StataCorp, College Station, TX, USA). Baseline characteristics of the included study populations were extracted, with means and standard deviations (SDs) estimated from medians and interquartile ranges (IQRs) when necessary, following the method proposed by Wan et al. [18].
The pooled prevalence of CMBs among AIS patients across different imaging modalities was calculated using the “metaprop” package, applying a random-effects meta-analysis for proportions derived from individual studies. Exact 95% confidence intervals (CIs) were generated using the “cimethod (exact”) and “ftt” commands. To investigate associations between CMB presence and clinical outcomes, random-effects meta-analyses were performed using the DerSimonian and Laird method. This analysis was restricted to studies reporting baseline CMB data and outcomes related to CMB presence or absence. The random-effects method was consistently applied across all subgroup analyses, which included comparisons based on the imaging modality used (SWI, T2*, or SWI and T2*).
Forest plots were generated to present pooled odds ratios (ORs), 95% CIs, and inter-study heterogeneity. Heterogeneity was assessed using the I2 statistic, with thresholds defined as follows: <30% indicating low heterogeneity, 30–50% moderate, 50–70% substantial, and >75% considerable heterogeneity. Cochran’s Q test p values and Tau-squared (τ2) were also reported to further quantify heterogeneity. Sensitivity analyses were conducted using the “metainf” package, assessing the influence of individual studies on the overall estimates by systematically excluding one study at a time.
Potential publication bias was evaluated using Egger’s regression test, visual inspection of funnel plots, and Deeks’ funnel plot asymmetry test, generated with the “metabias”, “metafunnel”, and “midas” packages, respectively. Publication bias was assessed for all primary outcomes (prevalence, sICH, HT, and functional outcome) using Egger’s test, funnel plots, and Deeks’ test. Asymmetry within the funnel plot, along with significant Egger’s test results, was considered indicative of publication bias. To evaluate diagnostic performance and the association between CMBs and clinical outcomes, the “midas’ package was used. Accuracy plots summarizing pooled estimates of sensitivity, specificity, likelihood ratios, diagnostic odds ratios (DORs), and other test performance metrics were generated using the midas command with the res(all) option. Summary Receiver Operating Characteristic (SROC) curves were constructed with 95% confidence and prediction contours using the “plot sroc(both)” function. Lastly, Fagan’s Nomogram was generated to illustrate the relationships among pre-test probability, likelihood ratios, and post-test probability, using the “midas” package. All statistical tests were two-sided, with significance set at p < 0.05.

3. Results

3.1. Description of Included Studies

A total of 1464 studies were initially identified through electronic database searches. After duplicate records were removed, 848 studies remained for screening. Following a detailed review of titles and abstracts, 696 studies were excluded based on relevance and inability to retrieve full-text reports. Of the 152 full-text articles assessed, 72 were excluded for multiple reasons: 22 studies included transient ischemic stroke (TIA) and/or hemorrhagic stroke in the patient cohort, 16 studies included atrial fibrillation and/or hypertension in the patient cohort, 12 studies had unsuitable age ranges, 4 studies were missing CMB data at baseline, 12 studies looked at a specific stroke subtype, and 6 studies did not report the primary outcomes of interest. Ultimately, 80 studies, comprising 28,383 patients, were included in this meta-analysis. Among these studies, 46 detected CMBs using T2* [19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64], 30 used SWI [65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94], and 4 used T2* and SWI [7,95,96,97].
Table 1, Table 2, and Table 3 present an overview of the clinical characteristics, risk factors, and outcomes of participants across the studies, respectively. Table 4 summarizes findings related to heterogeneity and estimated pooled prevalence of CMBs across different modalities and clinical parameters. Additional insight into the association between CMBs and prognostic outcomes such as sICH, HT, and mRS scores at 90 days is presented in Table 5, while Table 6 presents information on diagnostic and prognostic performance. It is important to note that variations in the definitions of CMBs and sICH existed across the studies.
A comprehensive evaluation of methodological quality and funding bias is presented in Supplemental Table S3 and Supplemental Table S4. Assessment of publication bias using Egger’s test revealed no significant evidence of small-study effects across key outcomes, as illustrated in Supplemental Table S5. Finally, Supplemental Table S6 presents the results from Deeks’ test, providing further assessment of publication bias.

3.2. Prevalence of CMBs Using Different Imaging Modalities

A comprehensive analysis of all 80 studies (n = 28,383) revealed the overall pooled prevalence of CMBs to be 29% (95% CI: 0.26; 0.31) in AIS patients. Subgroup analysis based on the type of modality used revealed that patients undergoing SWI for CMB detection [65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94] had the highest pooled prevalence, at 36% (95% CI: 0.31; 0.41) (Figure 2). Studies that used T2* [19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64] had a prevalence of 25% (95% CI: 0.22; 0.28), while those that used SWI and T2* [7,95,96,97] exhibited a prevalence of 25% (95% CI; 0.18; 0.32) (Figure 2). Notably, significant heterogeneity persisted within these subgroups (I2 = 95.87%, p < 0.001), with a heterogeneity chi2 of 1912.84 (p < 0.001, d.f. 79). The high heterogeneity observed for pooled prevalence estimates (I2 > 90%) reflected methodological and population variability across the 80 included studies. In contrast, lower heterogeneity values in outcome analyses (e.g., sICH I2 = 29.7%, mRS I2 = 0%) arose from smaller and more clinically homogeneous subsets of studies. Sensitivity analyses confirmed that no single study drove the pooled estimates, though heterogeneity persisted in prevalence analyses. While formal meta-regression was limited by data availability, subgroup analyses (by age, hypertension, imaging parameters) partially reduced heterogeneity, supporting these factors as potential contributors. Overall, these results highlight the influence of detection methods on reported CMB burden.

3.2.1. Stratified by Age

Age-related differences were examined in this meta-analysis. For SWI-detected CMBs, 13 studies (n = 1436) [65,70,72,73,78,79,80,81,82,83,85,86,92] assessed pooled CMB prevalence in adults aged 65 and older, while 6 studies (n = 673) [67,68,75,76,89,90] focused on those under 65. The meta-analysis indicated an estimated pooled prevalence of 35% (95% CI: 0.28; 0.43) in the older group, and 36% (95% CI: 0.27; 0.47) in the younger group.
For T2*-weighted MRI, 24 studies (n = 2893) [21,24,28,29,32,33,35,37,38,39,40,43,44,47,48,52,54,55,56,58,59,60,62,63] examined adults 65 and older, and 10 (n = 787) [20,22,25,30,31,46,49,50,51,64] focused on those under 65. The corresponding pooled prevalence was 25% (95% CI: 0.21; 0.30) and 22% (95% CI: 0.18; 0.26), respectively. These results align with established evidence that advancing age is a major determinant of small vessel pathology and higher CMB prevalence, consistent with population-based imaging studies [98]. Significant heterogeneity was observed for both modalities, with a chi2 of 377.01 (p < 0.001, d.f. 18) for SWI and 779.72 (p < 0.001, d.f. 33) for T2*.

3.2.2. Stratified by Hypertension Rates

Given the potential role of vascular risk factors in CMB development, we next examined hypertension’s influence. In studies using SWI, 14 studies (n = 1980) [65,66,68,75,76,77,78,79,81,85,86,89,90,92] assessed CMB prevalence in cohorts with an average hypertension rate of 65% or higher, while 5 studies (n = 486) [70,80,82,83,88] focused on cohorts with rates below 65%. The meta-analysis indicated an estimated pooled prevalence of 36% (95% CI: 0.29; 0.43) in the higher hypertension group and 37% (95% CI: 0.27; 0.48) in the lower hypertension group.
For studies using T2*, 15 (n = 1631) [19,21,22,29,32,33,35,42,46,47,56,57,58,59,64] assessed cohorts with hypertension rates of 65% or higher, and 12 (n = 893) [20,24,25,30,45,48,49,52,55,60,62,63] focused on rates below 65%. The corresponding pooled prevalence was 26% (95% CI: 0.0.23; 0.29) and 21% (95% CI: 0.17; 0.27), respectively. This gradient in prevalence highlights the well-established association between hypertension and the development of microangiopathic changes that underlie CMBs [1]. Significant heterogeneity was observed for both modalities, with a chi2 of 418.37 (p < 0.001, d.f. 18) for SWI and 300.82 (p < 0.001, d.f. 26) for T2*.

3.2.3. Stratified by Regional Variation

Geographic disparities were then analyzed to reflect differences in population-specific risk factors. Twenty-nine studies assessed regional differences in CMB prevalence among AIS patients using SWI. In European cohorts, 5 studies (n = 393) [70,71,78,79,83] had a pooled prevalence of 27% (95% CI: 0.18; 0.37), while in Asian cohorts, 20 studies (n = 2820) [65,66,67,68,72,73,75,77,80,81,82,84,86,87,88,89,90,91,92,93] had a pooled prevalence of 41% (95% CI: 0.37; 0.46). In North America, 2 studies (n = 120) [76,85] had a pooled prevalence of 24% (95% CI: 0.21; 0.28), while in Africa, 2 studies (n = 107) [69,94] had a pooled prevalence of 26% (95% CI: 0.22; 0.30).
Forty-six studies assessed regional differences in CMB prevalence among AIS patients using T2*. In European cohorts, 13 studies (n = 774) [20,25,32,34,36,41,42,45,53,59,60,61,63] had a pooled prevalence of 21% (95% CI: 0.19; 0.24), while in Asian cohorts, 27 studies (n = 2970) [19,21,22,23,26,27,29,30,31,33,35,37,38,39,40,43,44,46,47,48,49,54,55,56,57,58,62] had a pooled prevalence of 28% (95% CI: 0.19; 0.24). In North America, 4 studies (n = 444) [50,51,52,64] had a pooled prevalence of 18% (95% CI: 0.14; 0.22), while 2 multinational studies (n = 97) [24,28] had a pooled prevalence of 15% (95% CI: 0.12; 0.18). Significant heterogeneity was observed between regions, with a chi2 of 559.19 (p < 0.001, d.f. 28) for SWI and 844.41 (p < 0.001, d.f. 45) for T2*. The consistently higher prevalence observed in Asian cohorts may reflect differences in genetic susceptibility, vascular risk profiles, and lifestyle factors. These findings underscore the importance of considering population-level variation when interpreting CMB burden.

3.2.4. Stratified by Use of FLAIR

The addition of FLAIR sequences to SWI protocols were analyzed, where 17 studies (n = 2090) [52,65,66,67,73,75,77,79,81,83,84,86,87,88,89,93,94] assessed the estimated pooled prevalence of CMBs amongst AIS patients when using SWI and FLAIR, while 10 studies (n = 846) [68,69,70,71,72,76,78,80,82,92] estimated pooled prevalence when SWI was used but not FLAIR. The meta-analysis revealed an estimated pooled prevalence of 38% (95% CI: 0.31; 0.44) when FLAIR was added, and 33% (95% CI: 0.25; 0.42) when FLAIR was not added.
Twenty eight (28) studies (n = 2131) [19,24,25,26,27,29,30,32,35,36,38,40,41,43,44,45,46,47,50,51,52,55,57,58,59,61,62,64] assessed the estimated pooled prevalence of CMBs amongst AIS patients when using T2* and FLAIR, while 18 studies (n = 2154) [20,21,22,23,28,31,33,34,37,39,42,48,49,53,54,56,60,63] estimated pooled prevalence when T2* was used but not FLAIR. The meta-analysis revealed an estimated pooled prevalence of 24% (95% CI: 0.21; 0.27) when FLAIR was added, and 26% (95% CI: 0.20; 0.31) when FLAIR was not added. Significant heterogeneity was observed, with a chi2 of 555.50 (p < 0.001, d.f. 26) for SWI and 844.41 (p < 0.001, d.f. 45) for T2*.

3.2.5. Stratified by Use of NCCT

The effect of incorporating NCCT with SWI protocols was analyzed, where 6 studies (n = 509) [67,78,81,88,93,94] examined the estimated pooled prevalence of CMBs amongst AIS patients when using SWI and NCCT, while 21 studies (n = 2427) [65,66,68,69,70,71,72,73,75,76,77,79,80,82,83,84,85,86,87,89,92] estimated pooled prevalence when SWI was used but not NCCT. The meta-analysis revealed an estimated pooled prevalence of 44% (95% CI: 0.21; 0.33) when NCCT was used, and 24% (95% CI: 0.21; 0.26) when NCCT was not used. While intriguing, this discrepancy likely reflects the very small number of studies in the NCCT+SWI subgroup and should not be overinterpreted. NCCT is generally insensitive to CMBs, and further validation is required before firm conclusions can be drawn [98].
Sixteen studies (n = 2194) [20,21,26,28,29,30,33,37,38,39,43,48,52,54,62,64] examined the estimated pooled prevalence of CMBs amongst AIS patients when using T2* and NCCT, while 30 studies (n = 2091) [19,22,23,24,25,27,31,32,34,35,36,40,41,42,44,45,46,47,49,50,51,53,55,56,57,58,59,60,61,63] estimated pooled prevalence when T2* was used but not NCCT. The meta-analysis revealed an estimated pooled prevalence of 27% (95% CI: 0.21; 0.33) when NCCT was added, and 24% (95% CI: 0.21; 0.26) when NCCT was not added. Significant heterogeneity was observed, with a chi2 of 555.50 (p < 0.001, d.f. 26) for SWI and 844.41 (p < 0.001, d.f. 45) for T2*.

3.2.6. Stratified by Use of Slice Thickness

The effect of scanner parameters, such as slice thickness, was explored. Eleven studies (n = 712) [44,66,68,72,78,80,81,84,86,93,94] assessed pooled CMB prevalence when using thin slices, 4 studies (n = 326) [69,71,85,95] looked at medium slices, and 33 studies (n = 3397) [7,20,21,22,23,24,26,27,28,29,30,31,33,34,35,36,37,39,41,42,43,45,46,47,49,52,53,60,62,73,76,77,79] examined thick slices.
A comparison between SWI and T2* was unable to be made due to a lack of studies within the slice thickness subgroups. The meta-analysis revealed an estimated pooled prevalence of 40% (95% CI: 0.32; 0.49) for thin slices, 23% (95% CI: 0.18; 0.28) for medium slices, and 25% (95% CI: 0.22; 0.29) for thick slices. These results reinforced prior imaging studies showing that thinner slices increase lesion detectability and should be adopted as standard where feasible to optimize CMB detection [15]. The chi2 for heterogeneity was 809.02 (p < 0.001, d.f. 47).

3.2.7. Stratified by Field Strength

Field strength was another parameter that influenced CMB prevalence. Fourteen studies (n = 2048) [65,66,68,71,73,75,76,77,78,80,82,84,86,89] examined the estimated pooled prevalence of CMBs amongst AIS patients when using SWI at 3 Tesla, while 8 studies (n = 521) [67,69,72,81,83,87,93,94] estimated pooled prevalence when SWI was used at a lower field strength of 1.5 Tesla. The meta-analysis revealed an estimated pooled prevalence of 37% (95% CI: 0.31; 0.43) at 3 Tesla, and 36% (95% CI: 0.26; 0.47) at 1.5 Tesla.
Nine studies (n = 927) [30,42,44,46,50,51,53,62,64] examined the estimated pooled prevalence of CMBs amongst AIS patients when using T2* at 3 Tesla, while 22 studies (n = 2074) [19,20,21,22,24,25,29,31,32,33,34,35,36,38,39,40,43,45,49,55,59,61] estimated pooled prevalence when T2* was used at a lower field strength of 1.5 Tesla. The meta-analysis revealed an estimated pooled prevalence of 23% (95% CI: 0.18; 0.28) at 3 Tesla and 27% (95% CI: 0.23; 0.31) at 1.5 Tesla. While counterintuitive, these findings may reflect small subgroup sizes and methodological inconsistencies across studies rather than the true superiority of 1.5T. Larger, harmonized datasets are needed to clarify the relationship between field strength and CMB detection. Significant heterogeneity was observed, with a chi2 of 370.43 (p < 0.001, d.f. 21) for SWI and 460.20 (p < 0.001, d.f. 30) for T2*.

3.2.8. Stratified by Stroke Subtype

Stroke subtype was another important factor associated with variation in prevalence estimates. Using SWI, 5 studies (n = 129) [54,66,75,85,89] assessed pooled CMB prevalence in patients with an atherothrombotic stroke subtype, 5 studies (n = 168) [54,66,75,85,89] focused on lacunar stroke, 5 (n = 159) [54,66,75,85,89] on cardio-embolism, and 4 (n = 61) [54,66,85,89] on undetermined stroke subtypes. The meta-analysis indicated an estimated pooled prevalence of 23% (95% CI: 0.08; 0.42) in atherothrombotic stroke, 26% (95% CI: 0.17; 0.37) in lacunar stroke, 25% (95% CI: 0.11; 0.43) in cardio-embolism, and 20% (95% CI: 0.10; 0.32) in undetermined stroke subtype.
Using T2*, 5 studies (n = 218) [19,23,33,43,59] assessed pooled CMB prevalence in patients with an atherothrombotic stroke subtype, 5 studies focused on lacunar stroke (n = 206) [19,23,33,43,59], 5 (n = 67) [19,23,33,43,59] on cardio-embolism, and 3 (n = 66) [33,43,59] on undetermined stroke subtypes. The meta-analysis indicated an estimated pooled prevalence of 25% (95% CI: 0.12; 0.39) in atherothrombotic stroke, 39% (95% CI: 0.25; 0.53) in lacunar stroke, 24% (95% CI: 0.14; 0.35) in cardio-embolism, and 27% (95% CI: 0.20; 0.33) in undetermined stroke subtype. Significant heterogeneity was observed, with a chi2 of 229.98 (p < 0.001, d.f. 18) for SWI and 119.90 (p < 0.001, d.f. 17) for T2*.

3.2.9. Stratified by CMB Location

The anatomical location of CMBs also contributed to variability in prevalence estimates. Using SWI, 11 studies (n = 549) [65,68,70,73,77,79,80,83,87,89,90] assessed pooled CMB prevalence in lobar locations, 9 (n = 172) [65,68,73,77,79,80,83,87,89] studies focused on infratentorial locations, 9 (n = 240) [65,68,73,77,79,80,83,87,89] on deep locations, and 10 (n = 816) [65,68,70,73,77,79,80,83,89,90] on mixed locations. The meta-analysis indicated an estimated pooled prevalence of 29% (95% CI: 0.24; 0.24) in lobar regions, 12% (95% CI: 0.07; 0.19) in infratentorial regions, 18% (95% CI: 0.14; 0.21) in deep regions, and 49% (95% CI: 0.39; 0.60) in patients with mixed CMB locations.
Using T2*, 12 studies (n = 419) [31,35,44,45,46,47,49,50,53,59,60,64] assessed pooled CMB prevalence in lobar locations, 5 (n = 19) [31,35,44,47,49] studies focused on infratentorial locations, 7 (n = 200) [31,35,44,46,47,49,64] on deep locations, and 11 (n = 476) [31,35,44,45,46,47,49,50,53,59,60] on mixed locations. The meta-analysis indicated an estimated pooled prevalence of 37% (95% CI: 0.29; 0.46) in lobar regions, 8% (95% CI: 0.02; 0.19) in infratentorial regions, 33% (95% CI: 0.20; 0.47) in deep regions, and 46% (95% CI: 0.36; 0.55) in patients with mixed CMB locations. Significant heterogeneity was observed, with a chi2 of 1021.49 (p < 0.001, d.f. 38) for SWI and 446 (p < 0.001, d.f. 34) for T2*.

3.3. Association of CMBs with Prognostic Outcomes

Table 5 summarizes the association between various prognostic outcomes and CMB prevalence in patients with AIS, while Table 6 presents information on diagnostic and prognostic performance. For more detailed information on these associations, refer to the supplemental figures (Supplemental Figures S4–S9), which provide information on publication bias and sensitivity analyses.

3.3.1. Symptomatic Intracranial Hemorrhage (sICH)

To explore the prognostic implications of CMBs in AIS, the association with sICH, a critical complication following stroke, was analyzed. The meta-analysis included a total of 14 studies, comprising 6163 patients, sub-grouped as follows: 4 studies [70,73,74,76] using SWI, 9 studies [24,26,28,42,51,54,59,60,63] using T2*, and 1 [95] employing both SWI and T2* for CMB detection, as evident in Figure 3. Various criteria were used to define sICH (Table 3), which introduced some heterogeneity in outcome reporting. CMB presence was associated with an overall increased risk of sICH, with an OR of 2.216 (95% CI: 1.555; 3.159, p < 0.0001). This trend was more pronounced when SWI and T2* were used to detect CMBs, presenting an OR of 2.916 (1.294; 6.574, p = 0.010). However, this relied on a limited dataset from a single study, warranting cautious interpretation. Similarly, within the subgroups of patients who received SWI and those that received T2* for CMB detection, there were increased odds of sICH (SWI: OR 2.687; CI: 0.722; 10.007, p = 0.141; T2*: OR 2.13; CI: 1.435; 3.160 s), but only T2* and studies that used SWI and T2* obtained statistical significance.
The overall heterogeneity of the meta-analysis was low, with an I2 of 29.7%, which was not statistically significant (p = 0.140). Visual inspection of the funnel plot revealed slight asymmetry (Supplemental Figure S5); however, Egger’s regression test showed no statistically significant evidence of small-study effects (p = 0.656) (Supplemental Figure S4). Similarly, Deeks’ test demonstrated no significant evidence of publication bias (p = 0.97) (Supplemental Figure S8).
Further sensitivity analyses included Fagan’s Nomogram, which demonstrated a weakly positive likelihood ratio of 2 (Supplemental Figure S9), and the SROC curve, which showed poor diagnostic performance with an AUC of 0.29 (Supplemental Figure S7). These findings are consistent with the influence analysis (Supplemental Figure S6) and the diagnostic performance summary stratified by imaging modality (Table 6). Specifically, SWI demonstrated a sensitivity of 0.05, specificity of 0.98, and a DOR of 3, while T2* showed a sensitivity of 0.09, specificity of 0.96, and a DOR of 2. These findings confirmed that while CMBs increase the risk of sICH (OR 2.216), their diagnostic performance remains limited, as reflected in the low AUC (0.29). Thus, CMBs should be interpreted as contributory risk markers rather than standalone predictors of post-stroke hemorrhage.

3.3.2. Hemorrhagic Transformation (HT)

HT represents another important complication in AIS patients. The meta-analysis incorporated 21 studies, encompassing a total of 6049 patients, which were divided into subgroups based on imaging modality: 7 studies [69,72,73,76,78,84,85] using SWI, 12 studies [20,24,26,35,38,40,47,48,52,54,62,64] using T2*, and 2 studies [7,95] implementing both SWI and T2* for CMB detection (Figure 3). Definitions of HT were not always specified and varied across studies. The presence of CMBs was linked to a significantly higher risk of HT, with an OR of 1.332 (95% CI: 1.013; 1.750, p = 0.040). This association appeared more pronounced in the subset of studies using both SWI and T2*, which demonstrated an OR of 1.788 (95% CI: 1.033; 3.094, p = 0.038). Within the modality-specific subgroups, increased odds of HT were also observed for both SWI (OR 1.402; 95% CI: 0.910; 2.163, p = 0.125) and T2* (OR 1.229; 95% CI: 0.820; 1.843, p = 0.319), though statistical significance was achieved only for the combined SWI and T2* group.
Heterogeneity across studies was moderate, with an I2 statistic of 53.5% and a significant p-value (p = 0.002). Visual inspection of the funnel plot revealed minor asymmetry, suggestive of possible small-study effects (Supplemental Figure S5); however, Egger’s regression and Deeks’ test did not indicate significant publication bias (Supplemental Tables S5 and S6).
Additional sensitivity analyses were performed, including Fagan’s Nomogram (Supplemental Figure S9) and the SROC plot, with an AUC value of 0.56 (Supplemental Figure S7). These results were consistent with findings from the diagnostic performance summary (Supplemental Table S6). SWI demonstrated a sensitivity of 0.34, specificity of 0.75, and a DOR of 2, while T2* showed a sensitivity of 0.21, specificity of 0.82, and a DOR of 1. The pooled analysis demonstrated a significant association between CMBs and HT (OR 1.332), most evident in the combined SWI/T2* subgroup. However, moderate heterogeneity (I2 = 53.5%) reflected variability in population characteristics and imaging protocols, consistent with prior HT meta-analyses.

3.3.3. mRS 3-6 at 90 Days

The presence of CMBs was also evaluated for its impact on long-term functional outcomes, specifically, disability at 90 days post-stroke, as measured by the mRS. The meta-analysis included a total of 11 studies involving 5499 patients, which were divided into subgroups based on imaging modality: 4 studies [70,73,76,85] using SWI, 6 studies [42,54,59,60,62,63] using T2*, and 1 study [95] implementing both SWI and T2* for CMB detection (Figure 3). All studies consistently defined poor functional outcome at 90 days as an mRS score between 3 to 6. Overall, the presence of CMBs was associated with significantly increased odds of poor functional outcome at 90 days (OR 1.606; 95% CI: 1.387; 1.858, p < 0.0001). This association appeared more pronounced in the subset of studies using SWI, which demonstrated an OR of 1.727 (95% CI: 1.303; 2.289, p < 0.0001). This was followed by the SWI and T2* combined subgroup (OR 1.727; 95% CI: 0.976; 2.555, p = 0.063), although caution is warranted as it relied on a limited data set of 1 study. Increased odds were also observed for the T2* subgroup (OR 1.572; 95% CI: 1.282; 1.927, p < 0.0001), though statistical significance was achieved only for studies that used SWI or T2*.
No heterogeneity existed across the included studies, with an I2 statistic of 0% and a p-value of 0.524. Visual inspection of the funnel plot revealed reasonable symmetry (Supplemental Figure S5), with Egger’s regression and Deeks’ test not indicating significant publication bias (Supplemental Tables S5 and S6).
Additional sensitivity analyses were performed, including Fagan’s Nomogram (Supplemental Figure S9) and the SROC plot, with an AUC value of 0.58 (Supplemental Figure S7). These results were consistent with findings from the diagnostic performance summary (Supplemental Table S6), which reported an overall sensitivity of 0.49, specificity of 0.62, and a DOR of 2. Subgroup analysis by imaging modality using the “midas” model could not be performed due to the small number of included studies and high variability. Nonetheless, the association with poor 90-day outcome (OR 1.606) was strikingly consistent, with no heterogeneity (I2 = 0%). This underscores CMBs as reliable prognostic indicators of post-stroke disability, representing the most robust outcome signal observed in this meta-analysis.

3.4. Methodological Quality

The Modified Jadad Scores (Supplemental Table S3) indicated moderate methodological quality across the included studies, with most scoring between 3.5 and 5.5. Variability in study design, blinding, and reporting likely contributed to the heterogeneity observed in the pooled estimates. These limitations highlight the need for higher-quality, standardized research to improve the reliability of future meta-analyses.

4. Discussion

This meta-analysis is distinct in terms of providing pooled prevalence estimates of CMBs in AIS patients based on different imaging modalities, with additional analyses exploring how prevalence varies by patient demographics, stroke subtypes, and clinical settings. Moreover, the meta-analysis also identifies that CMBs are associated with significantly increased odds of sICH, HT, and mRS scores, highlighting the clinical relevance of CMB detection in AIS. To our knowledge, no previous meta-analyses have directly compared SWI and T2* for detecting CMBs in AIS patients.
To further contextualize these findings, we summarized the certainty of evidence for each outcome using the GRADE approach (see Table 7). The findings suggest that while CMBs are consistently associated with increased risk of sICH, HT, and poor functional outcomes, the certainty of evidence is limited by observational design, heterogeneity in definitions, and imprecision. The GRADE SoF table provides a transparent appraisal of where evidence is more robust (e.g., poor functional outcome with moderate certainty) versus where conclusions should be interpreted cautiously (e.g., diagnostic accuracy of CMBs, very low certainty). Incorporating GRADE facilitates balanced interpretation, highlights current gaps, and underscores the need for high-quality prospective studies.
Among AIS patients, the overall pooled prevalence of CMBs was 36% when detected using SWI, which is approximately 1.7 times higher than the prevalence observed with T2* (25%). This difference was statistically significant (p < 0.0001) and is consistent with earlier observational studies of AIS patients, which similarly report that SWI detects between 1.2 to 1.7 times more CMBs compared to T2* [11,99]. The improved detection rate with SWI is likely attributable to its incorporation of phase imaging, which enhances magnetic susceptibility contrast, improves spatial resolution, and reduces artifacts [84]. Interestingly, when SWI and T2* were used in combination, the pooled prevalence remained at 25%. This lower-than-expected estimate likely reflects the small number of studies in this subgroup (4 studies vs. >30 studies in the individual SWI and T2* groups), which reduces representativeness and increases susceptibility to the influence of outliers. In addition, differences in T2* acquisition parameters in the combined subgroup may have attenuated the incremental benefit of SWI, effectively lowering the pooled prevalence estimate despite SWI’s known higher sensitivity. Notably, Kidwell et al. [7] reported a low CMB prevalence rate of 12% (95% CI: 0.04; 0.26), which may have disproportionately affected the combined result. Understanding these modality-dependent differences is crucial, as higher detection rates with SWI may influence clinical decision-making and risk stratification in AIS populations. The notable heterogeneity observed across studies assessing CMB prevalence likely reflects considerable methodological and clinical variability. Differences in imaging protocols, patient characteristics, CMB definitions, and interpretation differences among radiologists may further exacerbate heterogeneity across studies.
Building on these findings, the present meta-analysis has identified a range of clinical and methodological factors that influence CMB prevalence among AIS patients. A consistently higher prevalence of CMBs was observed with SWI compared to T2* across most subgroups, including those defined by age, hypertension status, geographic region, and imaging parameters. Within each imaging modality, distinct trends emerged with patient demographics. Specifically, older age and higher rates of hypertension were associated with increased CMB prevalence when using T2*. In contrast, among patients assessed with SWI, CMB prevalence appeared lower in these same subgroups. The trends observed using SWI contradict several studies [99,100,101] which consistently reported that the prevalence of CMBs rises substantially with both age and hypertension, irrespective of imaging modality. A possible explanation is that SWI, being more sensitive [11], detects a broader spectrum of CMBs even in younger or lower-risk individuals, thereby diluting the relative differences observed across age and hypertension subgroups.
Analysis of imaging parameters showed that CMB prevalence increased when FLAIR or NCCT were combined with SWI. For T2*, CMB prevalence increased with the addition of NCCT but not with FLAIR. Although previous studies have established that SWI and T2* are more sensitive than FLAIR or NCCT for detecting CMBs [11,13,69,102], few have evaluated the added diagnostic value of including FLAIR or NCCT alongside these sequences. In this meta-analysis, the pooled prevalence of CMBs detected with SWI increased by 15% with the addition of FLAIR and by 33% with NCCT. This may reflect FLAIR’s utility in distinguishing true CMBs from common mimics, such as enlarged perivascular spaces, which may influence specificity in CMB detection [16]. This meta-analysis also supports existing evidence that thinner MRI slices improve CMB detection. Our analysis found a 15% higher pooled prevalence of CMBs in studies using thin slices compared to thick slices, which is consistent with prior imaging studies, such as Nandigam et al. [15], which reported that thick-section gradient echo (GRE) detected only 33% of the CMBs identified with thin-section SWI. In addition, our analysis found that a higher magnetic field strength of 3T was correlated with increased CMB prevalence, particularly with SWI [15,103]. Conversely, the expected benefit of a higher field strength was not observed with T2*; however, this difference was not statistically significant (p = 0.22) and was likely influenced by outlier results from Kato et al. [19] and Wang et al. [43], who reported high CMB prevalences of 47% and 46%, respectively, when 1.5T was used. Further research is therefore necessary. These findings should be interpreted cautiously, as both the NCCT+SWI and 1.5T T2* subgroups were small, making them highly sensitive to single-study effects. Further, variations in scanner calibration, sequence optimization, and patient selection may have contributed to the observed patterns. Additional prospective studies are needed to confirm or refute these observations. Further, when stratified by continent, the highest CMB prevalence was observed in Asian populations, followed by European populations, which was evident across both modality groups. While genetic differences have been proposed [104], variations in age and comorbidities are likely major contributors, as data comparing CMB prevalence across ethnicities remain limited.
When subgrouping by stroke subtype and CMB location, the highest CMB prevalence was observed among patients with lacunar strokes and those with lobar CMBs, regardless of imaging modality. This finding aligns with the known association between lacunar strokes and small vessel disease [105], while the predominance of lobar CMBs may reflect underlying CAA, particularly in older populations [106]. Hence, this analysis builds on previous research by offering a more comprehensive evaluation of the factors linked to CMB prevalence, incorporating a wider spectrum of clinical, demographic, and imaging-related variables.
The meta-analysis also identified differences in how imaging modalities influenced the association between CMBs and prognostic outcomes. Of the studies included for prognostic outcomes, 14 underwent IVT, 5 underwent EVT, 2 had bridging therapy, and 8 had no reperfusion therapy. Overall, SWI was associated with higher odds ratios compared to T2* for sICH, HT, and poor functional outcomes, with respective odds ratios of 1.26, 1.14, and 1.10. Further, the highest odds ratios for sICH and HT outcomes were observed in the subgroups using SWI and T2*, though this subgroup was limited to only one study. While no meta-analysis directly compared SWI to T2* for CMB stroke outcomes, the association with CMBs and increased odds of sICH, HT, and 90-day poor functional outcome (mRS 3-6) is consistent with findings reported in previous studies and meta-analyses [51,107,108,109].
The diagnostic performance of CMB presence in terms of predicting clinical outcomes was modest, as reflected by the AUC values derived from the SROC analyses (Table 6). For sICH, the AUC was notably low at 0.11 when using SWI and 0.30 for T2*, indicating poor discriminatory ability of CMBs as a standalone prognostic marker. Similarly, for HT and poor 90-day functional outcomes, AUCs remained below 0.65, underscoring the limited predictive utility of CMBs and reinforcing the need for multimodal risk stratification approaches. In addition, the presence of CMBs demonstrated limited predictive value for sICH, with both SWI and T2* showing poor sensitivity (SWI: 0.05; T2: 0.09) but high specificity (SWI: 0.98; T2: 0.96). While SWI showed marginally better discriminatory ability (diagnostic odds ratio [DOR] = 3) compared to T2* (DOR = 2), both modalities exhibited overall limited predictive performance for sICH risk stratification.
For HT, SWI demonstrated higher sensitivity (0.34) than T2* (0.21) but lower specificity (0.75 vs. 0.82). DORs were low for both modalities (2 for SWI, 1 for T2*), suggesting limited predictive value. Cross-comparison data were not available for mRS scores across modalities. While SWI detects more CMBs than T2*, demonstrating greater sensitivity, the clinical prognostic value of CMB detection remains limited. This was evident from the low DORs and AUC values below 0.6 for key outcomes such as sICH, HT, and functional decline, indicating poor predictive accuracy. These findings caution against overreliance on CMB presence alone for prognostication and highlight the need for integrated, multimodal risk assessment strategies in acute stroke management.
The results align with the Living Lesion Paradigm, which reframes CMBs not as incidental remnants but as active, evolving markers of neurovascular dysfunction [8]. The higher prevalence of CMBs observed with SWI supports this perspective, suggesting that SWI captures a broader spectrum of dynamic vascular injury [110]. Further studies are warranted to assess whether SWI can detect biological processes such as angiogenesis and extracellular matrix deposition, which are known to occur during the resolution phase of other hemorrhagic pathologies [1]. This paradigm also helps explain the modest prognostic performance observed across imaging modalities, as conventional metrics may not fully capture the ongoing pathophysiological processes underlying CMBs [105,111]. In addition, CMBs appear to reflect a state of continued vulnerability rather than serving as a definitive predictor of outcome, consistent with their low sensitivity and modest DORs observed in the meta-analyses [111]. The increased odds of sICH and HT in patients with CMBs across both SWI and T2* support this paradigm by suggesting that these lesions reflect ongoing pathological processes rather than resolved hemorrhagic events. This aligns with mechanistic data indicating persistent microvascular instability in CMB regions [1]. Viewing CMBs as dynamic lesions highlights the need for sophisticated, temporally informed imaging strategies and integrated biomarkers for longitudinal monitoring and risk stratification, rather than relying on CMB detection for acute prognostic decision-making alone.
Clinically, this evolving understanding has significant implications for optimizing both CMB detection and AIS treatment strategies, particularly in managing the risks of sICH, HT, and poor functional outcomes. Identifying CMBs can prompt stricter vascular risk control and influence the use of antithrombotic therapies, especially in patients with a high CMB burden [7]. Given SWI’s improved ability to detect CMBs over T2* [11], many centers should consider adopting it if feasible. However, given that SWI is not universally available, particularly in low-resource settings, these findings must be interpreted with caution, as access disparities may limit the global applicability of SWI-based diagnostic strategies. In addition, combining SWI with complementary modalities, such as FLAIR or NCCT, can further improve specificity by distinguishing true CMBs from artifacts [16]. Despite its clinical importance, CMB detection faces challenges. Variability in MRI parameters leads to inconsistent detection, and a lack of longitudinal studies limits understanding of long-term outcomes [112]. Furthermore, current risk models for HT and sICH rarely incorporate CMB burden, and limited access to advanced MRI techniques restricts widespread use [14]. Standardized protocols, multicenter studies, and improved access are essential to enhance the clinical utility of CMB detection.

5. Limitations

There are several important limitations to this study that should be acknowledged. First, high levels of heterogeneity were observed across most analyses, particularly regarding CMB prevalence estimates. This variability likely reflects differences in study populations, as well as in MRI parameters, such as field strength, slice thickness, and echo time, which significantly impact CMB detection rates [112]. Standardized imaging protocols are essential to minimize inter-study heterogeneity and improve comparability. Second, varying definitions of CMBs, HT, and sICH were evident across studies, which may have introduced classification bias, influencing both prevalence and prognostic associations. These definitional discrepancies could lead to misclassification bias, either overestimating or underestimating true associations. For example, stricter sICH definitions may exclude clinically relevant hemorrhages, while broader definitions could artificially inflate prevalence and odds ratios. To ensure uniformity in research findings, establishing consistent definitions across centers is essential.
Third, several subgroup analyses, particularly those involving stroke subtypes, CMB locations, or combined SWI and T2* modalities, were based on a small number of studies. This restricts the statistical power and generalizability of these specific findings. Future research should prioritize larger studies, focusing on these subgroups, to clarify observed trends. Fourth, despite the association between CMB presence and adverse outcomes, diagnostic performance analyses revealed low sensitivity and modest DORs, indicating that CMB detection alone provides limited predictive value for sICH, HT, or poor functional outcomes. Fifth, the lack of histopathological validation poses a limitation, as all studies relied on imaging markers rather than direct tissue analysis, raising the risk of false positives from mimics [113]. Future studies should incorporate post-mortem correlation studies to clarify the true accuracy of SWI and T2* for detecting CMBs [114,115]. The identified limitations should be acknowledged when assessing the study’s results and overall implications.

6. Conclusions

In summary, this study found a pooled CMB prevalence of 36% with SWI and 25% with T2* among AIS patients, confirming the superior detection capability of SWI. Increased CMB prevalence was associated with older age, hypertension, lacunar stroke, lobar CMB location, thinner slice thickness, and higher field strength, though the strength and consistency of these associations varied slightly by imaging modality used. Furthermore, CMB presence was associated with adverse prognostic outcomes, including sICH, HT, and poor functional outcomes at 90 days. Although both SWI and T2* detected these associations, SWI demonstrated higher odds ratios across all outcomes, suggesting a stronger link between SWI-detected CMBs and prognosis. Despite this, the prognostic accuracy of CMB detection for predicting adverse outcomes remains limited, with low sensitivity (0.05 to 0.49) and modest DORs (2 to 3) across outcomes. To further qualify these results, our GRADE appraisal indicated that while associations with poor functional outcome are supported by moderate certainty of evidence, other findings—such as diagnostic accuracy—remain of very low certainty due to observational design, heterogeneity, and imprecision. This highlights the clinical value and the limitations of current evidence. Together, these findings underscore the utility of SWI in CMB detection and risk stratification among AIS patients, while also emphasizing the need for cautious interpretation and integration of CMB findings within a broader clinical context. Future research should standardize imaging protocols, reduce heterogeneity, and strengthen the certainty of evidence through high-quality prospective studies, ultimately supporting better identification of high-risk patients and guiding targeted management to improve outcomes.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/medicina61091566/s1, Search strategy; Supplemental Table S1: PRISMA-2020 Checklist; Supplemental Table S2: MOOSE Checklist; Supplemental Table S3: Modified Jadad Analysis for Methodological Quality; Supplemental Table S4: Funding Bias Scores for Studies; Supplemental Table S5: Outputs from Egger’s Test for Publication Bias for Association Variables; Supplemental Table S6: Outputs from Deeks’ Test for Small Study Effects and Publication Bias for Association Variables; Supplemental Figure S1: Forest Plots of CMB Prevalence, Stratified by Age, Hypertension and Regional Variation; Supplemental Figure S2: Forest Plots CMB Prevalence, Stratified by Use of FLAIR, NCCT and Slice Thickness; Supplemental Figure S3: Forest Plots of CMB Prevalence, Stratified by Field Strength, Stroke Subtype and CMB Location; Supplemental Figure S4: Graphs of Egger’s Regression Test for Meta-Analysis on the Association between CMBs and Prognostic Outcomes; Supplemental Figure S5: Graphs of Funnel Plot for Meta-Analysis on the Association between CMBs and Prognostic Outcomes; Supplemental Figure S6: Sensitivity Analysis on Association between CMBs and Prognostic Outcomes; Supplemental Figure S7: Graphs of ROC Plot for Meta-Analysis on the Association between CMBs and Prognostic Outcomes; Supplemental Figure S8: Graphs of Deeks’ Funnel Plot for Meta-Analysis on the Association between CMBs and Prognostic Outcomes; Supplemental Figure S9: Graphs on Fagan’s Plot for Meta-Analysis on the Association between CMBs and Prognostic Outcome.

Author Contributions

S.B. conceptualized and led the SPOT-CMB study and the associated meta-analysis, providing intellectual leadership, developing the overarching study framework, and supervising the Global Health Neurology Lab team. He oversaw all aspects of study design, data interpretation, and manuscript development. R.T. performed the literature review, data extraction, statistical analysis, and contributed substantially to data visualization, drafting, and revision of the manuscript. K.J.S. and M.K. provided domain-specific insights, contributed to the discussion of imaging parameters, and offered critical feedback throughout the drafting process. All authors have read and agreed to the published version of the manuscript.

Funding

This study received no direct funding. S.B. received separate financial support through the Grant-in-Aid for Scientific Research (KAKENHI) funded by the Japan Society for the Promotion of Science (JSPS), Japanese Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan (Grant ID: 23KF0126). S.B. was also awarded the JSPS International Fellowship supported by MEXT and the Australian Academy of Science for the period 2023–2025 (Grant ID: P23712).

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The original contributions presented in this study are included in the article. Further inquiries can be directed to the corresponding author.

Acknowledgments

We gratefully acknowledge the support of the JSPS International Fellowship (Grant ID: P23712) and the Grant-in-Aid for Scientific Research (KAKENHI) (Grant ID: 23KF0126). We also thank the teams at the Global Health Neurology Lab and NSW Brain Clot Bank for their valuable input throughout the development of the SPOT-CMB study.

Conflicts of Interest

S.B. reports leadership or fiduciary roles with the following organizations: National Cerebral and Cardiovascular Center (Osaka, Japan) as Visiting Director (2023–2025); Rotary District 9675 (Sydney, Australia) as District Chair for Diversity, Equity, and Inclusion; the Global Health and Migration Hub Community, Global Health Hub Germany (Berlin, Germany) as Chair and Founding Member; and editorial board memberships at PLOS One, BMC Neurology, Frontiers in Neurology, Frontiers in Stroke, Frontiers in Public Health, Journal of Aging Research, Neurology International, Diagnostics, and BMC Medical Research Methodology. He also serves as a Member of the College of Reviewers for the Canadian Institutes of Health Research (CIHR), Government of Canada; Director of Research for the World Headache Society (Bengaluru, India); Scientific Review Committee Member at Cardiff University Biobank (UK); Chair of the Rotary Reconciliation Action Plan (RAP), Rotary District 9675 (NSW, Australia); Healthcare and Medical Adviser for Japan Connect (Osaka, Japan); and Expert Adviser/Reviewer for the Cariplo Foundation (Milan, Italy). These roles are unrelated to the submitted work. Other authors (R.T., K.J.S., M.K.) declare no conflicts of interest.

References

  1. Agarwal, A.; Ajmera, P.; Sharma, P.; Kanekar, S. Cerebral microbleeds: Causes, clinical relevance, and imaging approach—A narrative review. J. Neurosci. Rural Pract. 2024, 15, 169–181. [Google Scholar] [CrossRef]
  2. Pu, L.; Wang, L.; Zhang, R.; Zhao, T.; Jiang, Y.; Han, L. Projected Global Trends in Ischemic Stroke Incidence, Deaths and Disability-Adjusted Life Years From 2020 to 2030. Stroke 2023, 54, 1330–1339. [Google Scholar] [CrossRef]
  3. Kilinc, D.; Vernooij, M.W.; Bron, E.E.; Biessels, G.J.; Vinke, E.J. Normative Population–Derived Data for MRI Manifestations of Cerebral Small Vessel Disease. Stroke 2024, 55, 2863–2871. [Google Scholar] [CrossRef]
  4. Cho, A.H.; Kwon, H.S.; Lee, M.H.; Park, J.H.; Heo, S.H.; Yu, S.; Kwon, S.U. Asia Pacific Stroke Conference 2020. Abstracts of the Annual Conference of the Asia Pacific Stroke Organization (APSO), Virtual Conference, Korea, December 4–6, 2020. Cerebrovasc. Dis. 2020, 49, 1–149. [Google Scholar] [CrossRef]
  5. Hori, S.; Okamoto, S.; Kubo, M.; Horie, Y.; Kuroda, S. Cerebral microbleeds is a predictor of recurrent small vessel cerebrovascular disease: Evaluation based on the recurrent stroke pattern. J. Stroke Cerebrovasc. Dis. 2024, 33, 107812. [Google Scholar] [CrossRef]
  6. Ferro, D.A.; van den Brink, H.; Exalto, L.G.; Boomsma, J.M.F.; Barkhof, F.; Prins, N.D.; van der Flier, W.M.; Biessels, G.J.; Scheltens, P.; Teunissen, C.E.; et al. Clinical relevance of acute cerebral microinfarcts in vascular cognitive impairment. Neurology 2019, 92, e1558–e1566. [Google Scholar] [CrossRef]
  7. Kidwell, C.S.; Saver, J.L.; Villablanca, J.P.; Duckwiler, G.; Fredieu, A.; Gough, K.; Leary, M.C.; Starkman, S.; Gobin, Y.P.; Jahan, R.; et al. Magnetic Resonance Imaging Detection of Microbleeds Before Thrombolysis. Stroke 2002, 33, 95–98. [Google Scholar] [CrossRef] [PubMed]
  8. Tan, R.; Bhaskar, S. Cerebral Microbleeds as Living Lesions: Rethinking Their Role in Stroke Pathophysiology and Prognosis. 2025; Under review. [Google Scholar]
  9. Haller, S.; Vernooij, M.W.; Kuijer, J.P.A.; Larsson, E.-M.; Jäger, H.R.; Barkhof, F. Cerebral Microbleeds: Imaging and Clinical Significance. Radiology 2018, 287, 11–28. [Google Scholar] [CrossRef]
  10. Shams, S.; Martola, J.; Cavallin, L.; Granberg, T.; Shams, M.; Aspelin, P.; Wahlund, L.O.; Kristoffersen-Wiberg, M. SWI or T2*: Which MRI Sequence to Use in the Detection of Cerebral Microbleeds? The Karolinska Imaging Dementia Study. Am. J. Neuroradiol. 2015, 36, 1089–1095. [Google Scholar] [CrossRef] [PubMed]
  11. Shao, L.; Wang, M.; Ge, X.H.; Huang, H.D.; Gao, L.; Qin, J.C. The use of susceptibility-weighted imaging to detect cerebral microbleeds after lacunar infarction. Eur. Rev. Med. Pharmacol. Sci. 2017, 21, 3105–3112. [Google Scholar]
  12. Kim, Y.; Luby, M.; Burkett, N.-S.; Norato, G.; Leigh, R.; Wright, C.B.; Kern, K.C.; Hsia, A.W.; Lynch, J.K.; Adil, M.M.; et al. Fluid-Attenuated Inversion Recovery Hyperintense Ischemic Stroke Predicts Less Favorable 90-Day Outcome after Intravenous Thrombolysis. Cerebrovasc. Dis. 2021, 50, 738–745. [Google Scholar] [CrossRef] [PubMed]
  13. El-Serougy, L.G.; El-Rakhawy, M.M.; Ashamallah, G.A.; Mustafa, W.F. Reliability of magnetic susceptibility weighted imaging in detection of cerebral microbleeds in stroke patients. Egypt. J. Radiol. Nucl. Med. 2017, 48, 225–229. [Google Scholar] [CrossRef][Green Version]
  14. Puy, L.; Pasi, M.; Rodrigues, M.; van Veluw, S.J.; Tsivgoulis, G.; Shoamanesh, A.; Cordonnier, C. Cerebral microbleeds: From depiction to interpretation. J. Neurol. Neurosurg. Psychiatry 2021, 92, 598–607. [Google Scholar] [CrossRef]
  15. Nandigam, R.N.K.; Viswanathan, A.; Delgado, P.; Skehan, M.E.; Smith, E.E.; Rosand, J.; Greenberg, S.M.; Dickerson, B.C. MR Imaging Detection of Cerebral Microbleeds: Effect of Susceptibility-Weighted Imaging, Section Thickness, and Field Strength. Am. J. Neuroradiol. 2009, 30, 338–343. [Google Scholar] [CrossRef]
  16. Charidimou, A.; Krishnan, A.; Werring, D.J.; Rolf Jäger, H. Cerebral microbleeds: A guide to detection and clinical relevance in different disease settings. Neuroradiology 2013, 55, 655–674. [Google Scholar] [CrossRef]
  17. Jadad, A.R.; Moore, R.A.; Carroll, D.; Jenkinson, C.; Reynolds, D.J.M.; Gavaghan, D.J.; McQuay, H.J. Assessing the quality of reports of randomized clinical trials: Is blinding necessary? Control. Clin. Trials 1996, 17, 1–12. [Google Scholar] [CrossRef] [PubMed]
  18. Wan, X.; Wang, W.; Liu, J.; Tong, T. Estimating the sample mean and standard deviation from the sample size, median, range and/or interquartile range. BMC Med. Res. Methodol. 2014, 14, 135. [Google Scholar] [CrossRef] [PubMed]
  19. Kato, H.; Izumiyama, M.; Izumiyama, K.; Takahashi, A.; Itoyama, Y. Silent Cerebral Microbleeds on T2*-Weighted MRI. Stroke 2002, 33, 1536–1540. [Google Scholar] [CrossRef]
  20. Nighoghossian, N.; Hermier, M.; Adeleine, P.; Blanc-Lasserre, K.; Derex, L.; Honnorat, J.; Philippeau, F.; Dugor, J.F.; Froment, J.C.; Trouillas, P. Old Microbleeds Are a Potential Risk Factor for Cerebral Bleeding After Ischemic Stroke. Stroke 2002, 33, 735–742. [Google Scholar] [CrossRef]
  21. Fan, Y.H.; Zhang, L.; Lam, W.W.M.; Mok, V.C.T.; Wong, K.S. Cerebral Microbleeds as a Risk Factor for Subsequent Intracerebral Hemorrhages Among Patients With Acute Ischemic Stroke. Stroke 2003, 34, 2459–2462. [Google Scholar] [CrossRef]
  22. Lee, S.H.; Bae, H.J.; Kwon, S.J.; Kim, H.; Kim, Y.H.; Yoon, B.W.; Roh, J.K. Cerebral microbleeds are regionally associated with intracerebral hemorrhage. Neurology 2004, 62, 72–76. [Google Scholar] [CrossRef]
  23. Naka, H.; Nomura, E.; Wakabayashi, S.; Kajikawa, H.; Kohriyama, T.; Mimori, Y.; Nakamura, S.; Matsumoto, M. Frequency of asymptomatic microbleeds on T2*-weighted MR images of patients with recurrent stroke: Association with combination of stroke subtypes and leukoaraiosis. Am. J. Neuroradiol. 2004, 25, 714–719. [Google Scholar] [PubMed]
  24. Kakuda, W.; Thijs, V.N.; Lansberg, M.G.; Bammer, R.; Wechsler, L.; Kemp, S.; Moseley, M.E.; Marks, M.P.; Albers, G.W. Clinical importance of microbleeds in patients receiving IV thrombolysis. Neurology 2005, 65, 1175–1178. [Google Scholar] [CrossRef]
  25. Werring, D.J.; Coward, L.J.; Losseff, N.A.; Jager, H.R.; Brown, M.M. Cerebral microbleeds are common in ischemic stroke but rare in TIA. Neurology 2005, 65, 1914–1918. [Google Scholar] [CrossRef]
  26. Kim, H.S.; Lee, D.H.; Ryu, C.W.; Lee, J.H.; Choi, C.G.; Kim, S.J.; Suh, D.C. Multiple Cerebral Microbleeds in Hyperacute Ischemic Stroke: Impact on Prevalence and Severity of Early Hemorrhagic Transformation After Thrombolytic Treatment. Am. J. Roentgenol. 2006, 186, 1443–1449. [Google Scholar] [CrossRef]
  27. Naka, H.; Nomura, E.; Takahashi, T.; Wakabayashi, S.; Mimori, Y.; Kajikawa, H.; Kohriyama, T.; Matsumoto, M. Combinations of the presence or absence of cerebral microbleeds and advanced white matter hyperintensity as predictors of subsequent stroke types. Am. J. Neuroradiol. 2006, 27, 830–835. [Google Scholar]
  28. Fiehler, J.; Albers, G.W.; Boulanger, J.-M.; Derex, L.; Gass, A.; Hjort, N.; Kim, J.S.; Liebeskind, D.S.; Neumann-Haefelin, T.; Pedraza, S.; et al. Bleeding Risk Analysis in Stroke Imaging Before ThromboLysis (BRASIL). Stroke 2007, 38, 2738–2744. [Google Scholar] [CrossRef] [PubMed]
  29. Soo, Y.O.; Yang, S.R.; Lam, W.W.; Wong, A.; Fan, Y.H.; Leung, H.H.; Chan, A.Y.; Leung, C.; Leung, T.W.; Wong, L.K. Risk vs benefit of anti-thrombotic therapy in ischaemic stroke patients with cerebral microbleeds. J. Neurol. 2008, 255, 1679–1686. [Google Scholar] [CrossRef]
  30. Han, J.; Gao, P.; Lin, Y.; Zhang, J.; Xu, L.; Xue, J. Three-tesla magnetic resonance imaging study of cerebral microbleeds in patients with ischemic stroke. Neurol. Res. 2013, 31, 900–903. [Google Scholar] [CrossRef] [PubMed]
  31. Jeon, S.B.; Kwon, S.U.; Cho, A.H.; Yun, S.C.; Kim, J.S.; Kang, D.W. Rapid appearance of new cerebral microbleeds after acute ischemic stroke. Neurology 2009, 73, 1638–1644. [Google Scholar] [CrossRef]
  32. Orken, D.N.; Kenangil, G.; Uysal, E.; Forta, H. Cerebral Microbleeds in Ischemic Stroke Patients on Warfarin Treatment. Stroke 2009, 40, 3638–3640. [Google Scholar] [CrossRef][Green Version]
  33. Sun, J.; Soo, Y.O.; Lam, W.W.; Wong, K.S.; Zeng, J.S.; Fan, Y.H. Different distribution patterns of cerebral microbleeds in acute ischemic stroke patients with and without hypertension. Eur. Neurol. 2009, 62, 298–303. [Google Scholar] [CrossRef]
  34. Dassan, P.; Brown, M.M.; Gregoire, S.; Keir, G.; Werring, D.J. UK Stroke Forum 2011 Abstracts, Glasgow, 29 November–1 December 2011. Int. J. Stroke 2011, 6, 1–65. [Google Scholar] [CrossRef]
  35. Soo, Y.O.; Siu, D.Y.; Abrigo, J.; Yu, S.; Ng, N.; Ahuja, A.T.; Wong, L.K.; Leung, T.W. Risk of intracerebral hemorrhage in patients with cerebral microbleeds undergoing endovascular intervention. Stroke 2012, 43, 1532–1536. [Google Scholar] [CrossRef]
  36. Gregoire, S.M.; Scheffler, G.; Jäger, H.R.; Yousry, T.A.; Brown, M.M.; Kallis, C.; Cipolotti, L.; Werring, D.J. Strictly Lobar Microbleeds Are Associated with Executive Impairment in Patients With Ischemic Stroke or Transient Ischemic Attack. Stroke 2013, 44, 1267–1272. [Google Scholar] [CrossRef] [PubMed]
  37. Kimura, K.; Aoki, J.; Shibazaki, K.; Saji, N.; Uemura, J.; Sakamoto, Y. New Appearance of Extraischemic Microbleeds on T2*-Weighted Magnetic Resonance Imaging 24 Hours After Tissue-type Plasminogen Activator Administration. Stroke 2013, 44, 2776–2781. [Google Scholar] [CrossRef] [PubMed]
  38. Moriya, Y.; Takahashi, W.; Kijima, C.; Yutani, S.; Iijima, E.; Mizuma, A.; Honma, K.; Uesugi, T.; Ohnuki, Y.; Nagata, E.; et al. Predictors for hemorrhagic transformation with intravenous tissue plasminogen activator in acute ischemic stroke. Tokai J. Exp. Clin. Med. 2013, 38, 24–27. [Google Scholar]
  39. Naka, H.; Nomura, E.; Kitamura, J.; Imamura, E.; Wakabayashi, S.; Matsumoto, M. Antiplatelet Therapy as a Risk Factor for Microbleeds in Intracerebral Hemorrhage Patients: Analysis Using Specific Antiplatelet Agents. J. Stroke Cerebrovasc. Dis. 2013, 22, 834–840. [Google Scholar] [CrossRef]
  40. Takahashi, W.; Moriya, Y.; Mizuma, A.; Uesugi, T.; Ohnuki, Y.; Takizawa, S. Cerebral microbleeds on T2*-weighted images and hemorrhagic transformation after antithrombotic therapies for ischemic stroke. J. Stroke Cerebrovasc. Dis. 2013, 22, e528–e532. [Google Scholar] [CrossRef] [PubMed]
  41. Brundel, M.; Kwa, V.I.H.; Bouvy, W.H.; Algra, A.; Kappelle, L.J.; Biessels, G.J. Cerebral Microbleeds Are Not Associated with Long-Term Cognitive Outcome in Patients with Transient Ischemic Attack or Minor Stroke. Cerebrovasc. Dis. 2014, 37, 195–202. [Google Scholar] [CrossRef]
  42. Dannenberg, S.; Scheitz, J.F.; Rozanski, M.; Erdur, H.; Brunecker, P.; Werring, D.J.; Fiebach, J.B.; Nolte, C.H. Number of Cerebral Microbleeds and Risk of Intracerebral Hemorrhage After Intravenous Thrombolysis. Stroke 2014, 45, 2900–2905. [Google Scholar] [CrossRef]
  43. Wang, B.G.; Yang, N.; Lin, M.; Lu, B. Analysis of risk factors of hemorrhagic transformation after acute ischemic stroke: Cerebral microbleeds do not correlate with hemorrhagic transformation. Cell Biochem. Biophys. 2014, 70, 135–142. [Google Scholar] [CrossRef]
  44. Yan, S.; Chen, Y.; Zhang, X.; Liebeskind, D.S.; Lou, M. New microbleeds after thrombolysis: Contiguous thin-slice 3T MRI. Medicine 2014, 93, e99. [Google Scholar] [CrossRef]
  45. Turc, G.; Sallem, A.; Moulin, S.; Tisserand, M.; Machet, A.; Edjlali, M.; Baron, J.C.; Leclerc, X.; Leys, D.; Mas, J.L.; et al. Microbleed Status and 3-Month Outcome After Intravenous Thrombolysis in 717 Patients With Acute Ischemic Stroke. Stroke 2015, 46, 2458–2463. [Google Scholar] [CrossRef] [PubMed][Green Version]
  46. Zhang, C.; Li, Z.; Wang, Y.; Zhao, X.; Wang, C.; Liu, L.; Pu, Y.; Zou, X.; Pan, Y.; Du, W.; et al. Risk factors of cerebral microbleeds in strictly deep or lobar brain regions differed. J. Stroke Cerebrovasc. Dis. 2015, 24, 24–30. [Google Scholar] [CrossRef]
  47. Shi, Z.S.; Duckwiler, G.R.; Jahan, R.; Tateshima, S.; Gonzalez, N.R.; Szeder, V.; Saver, J.L.; Kim, D.; Ali, L.K.; Starkman, S.; et al. Mechanical thrombectomy for acute ischemic stroke with cerebral microbleeds. J. Neurointerv. Surg. 2016, 8, 563–567. [Google Scholar] [CrossRef] [PubMed]
  48. Yang, N.; Lin, M.; Wang, B.G.; Zeng, W.Y.; He, Y.F.; Peng, H.Y.; Zeng, J.; Wu, Z.Y.; Zhong, Y. Low level of low-density lipoprotein cholesterol is related with increased hemorrhagic transformation after acute ischemic cerebral infarction. Eur. Rev. Med. Pharmacol. Sci. 2016, 20, 673–678. [Google Scholar]
  49. Potigumjon, A.; Watcharakorn, A.; Dharmasaroja, P.A. Prevalence of Cerebral Microbleeds in Thai Patients with Ischemic Stroke. J. Neurosci. Rural Pract. 2019, 08, 216–220. [Google Scholar] [CrossRef] [PubMed][Green Version]
  50. Shahjouei, S.; Tsivgoulis, G.; Singh, M.; McCormack, M.; Noorbakhsh-Sabet, N.; Goyal, N.; Alexandrov, A.W.; Alexandrov, A.V.; Zand, R. Racial Difference in Cerebral Microbleed Burden among Ischemic Stroke Patients. J. Stroke Cerebrovasc. Dis. 2017, 26, 2680–2685. [Google Scholar] [CrossRef]
  51. Zand, R.; Tsivgoulis, G.; Singh, M.; McCormack, M.; Goyal, N.; Ishfaq, M.F.; Shahripour, R.B.; Nearing, K.; Elijovich, L.; Alexandrov, A.W.; et al. Cerebral Microbleeds and Risk of Intracerebral Hemorrhage Post Intravenous Thrombolysis. J. Stroke Cerebrovasc. Dis. 2017, 26, 538–544. [Google Scholar] [CrossRef]
  52. Nagaraja, N.; Tasneem, N.; Shaban, A.; Dandapat, S.; Ahmed, U.; Policeni, B.; Olalde, H.; Shim, H.; Samaniego, E.A.; Pieper, C.; et al. Cerebral Microbleeds are an Independent Predictor of Hemorrhagic Transformation Following Intravenous Alteplase Administration in Acute Ischemic Stroke. J. Stroke Cerebrovasc. Dis. 2018, 27, 1403–1411. [Google Scholar] [CrossRef]
  53. Braemswig, T.B.; Villringer, K.; Turc, G.; Erdur, H.; Fiebach, J.B.; Audebert, H.J.; Endres, M.; Nolte, C.H.; Scheitz, J.F. Predictors of new remote cerebral microbleeds after IV thrombolysis for ischemic stroke. Neurology 2019, 92, e630–e638. [Google Scholar] [CrossRef]
  54. Choi, K.-H.; Kim, J.-H.; Kang, K.-W.; Kim, J.-T.; Choi, S.-M.; Lee, S.-H.; Park, M.-S.; Kim, B.-C.; Kim, M.-K.; Cho, K.-H. Impact of Microbleeds on Outcome Following Recanalization in Patients With Acute Ischemic Stroke. Stroke 2019, 50, 127–134. [Google Scholar] [CrossRef]
  55. Liang, Y.; Chen, Y.-K.; Liu, Y.-L.; Mok, V.C.T.; Ungvari, G.S.; Chu, W.C.W.; Tang, W.-K.; Kim, J.S.; Kim, J.-M. Exploring causal pathways linking cerebral small vessel diseases burden to poststroke depressive symptoms with structural equation model analysis. J. Affect. Disord. 2019, 253, 218–223. [Google Scholar] [CrossRef] [PubMed]
  56. Nam, K.W.; Kwon, H.M.; Kim, H.L.; Lee, Y.S. Left ventricular ejection fraction is associated with small vessel disease in ischaemic stroke patients. Eur. J. Neurol. 2019, 26, 747–753. [Google Scholar] [CrossRef] [PubMed]
  57. Takahashi, Y.; Saito, S.; Yamamoto, Y.; Uehara, T.; Yokota, C.; Sakai, G.; Nishida, N.; Takahashi, R.; Kalaria, R.N.; Toyoda, K.; et al. Visually-Rated Medial Temporal Lobe Atrophy with Lower Educational History as a Quick Indicator of Amnestic Cognitive Impairment after Stroke. J. Alzheimers Dis. 2019, 67, 621–629. [Google Scholar] [CrossRef] [PubMed]
  58. Jiang, Q.; Ryu, W.-S.; Jeong, S.-W.; Kim, D.-E. Total small vessel disease burden and functional outcome in patients with ischemic stroke. PLoS ONE 2020, 15, e0242319. [Google Scholar] [CrossRef]
  59. Capuana, M.L.; Lorenzano, S.; Caselli, M.C.; Paciaroni, M.; Toni, D. Hemorrhagic risk after intravenous thrombolysis for ischemic stroke in patients with cerebral microbleeds and white matter disease. Neurol. Sci. 2020, 42, 1969–1976. [Google Scholar] [CrossRef]
  60. Derraz, I.; Cagnazzo, F.; Gaillard, N.; Morganti, R.; Dargazanli, C.; Ahmed, R.; Lefevre, P.-H.; Riquelme, C.; Mourand, I.; Gascou, G.; et al. Microbleeds, Cerebral Hemorrhage, and Functional Outcome After Endovascular Thrombectomy. Neurology 2021, 96, E1724–E1731. [Google Scholar] [CrossRef]
  61. Jabłoński, B.; Gójska-Grymajło, A.; Ossowska, D.; Szurowska, E.; Wyszomirski, A.; Rojek, B.; Karaszewski, B. New Remote Cerebral Microbleeds on T2*-Weighted Echo Planar MRI After Intravenous Thrombolysis for Acute Ischemic Stroke. Front. Neurol. 2022, 12, 744701. [Google Scholar] [CrossRef]
  62. Lee, S.-J.; Hwang, Y.-H.; Hong, J.M.; Choi, J.W.; Park, J.H.; Park, B.; Kang, D.-H.; Kim, Y.-W.; Kim, Y.-S.; Hong, J.-H.; et al. Influence of cerebral microbleeds on mechanical thrombectomy outcomes. Sci. Rep. 2022, 12, 3637. [Google Scholar] [CrossRef]
  63. Agbonon, R.; Forestier, G.; Bricout, N.; Benhassen, W.; Turc, G.; Bretzner, M.; Pasi, M.; Benzakoun, J.; Seners, P.; Derraz, I.; et al. Cerebral microbleeds and risk of symptomatic hemorrhagic transformation following mechanical thrombectomy for large vessel ischemic stroke. J. Neurol. 2024, 271, 2631–2638. [Google Scholar] [CrossRef]
  64. Zand, R.; Shahjouei, S.; Tsivgoulis, G.; Singh, M.; McCormack, M.; Noorbakhsh-Sabet, N.; Goyal, N.; Alexandrov, A.V. Cerebral Microbleeds are Associated with Higher Mortality Among Ischemic Stroke Patients. J. Stroke Cerebrovasc. Dis. 2018, 27, 3036–3042. [Google Scholar] [CrossRef] [PubMed]
  65. Lau, K.K.; Wong, Y.K.; Teo, K.C.; Chang, R.S.K.; Tse, M.Y.; Hoi, C.P.; Chan, C.Y.; Chan, O.L.; Cheung, R.H.K.; Wong, E.K.M.; et al. Long-Term Prognostic Implications of Cerebral Microbleeds in Chinese Patients With Ischemic Stroke. J. Am. Heart Assoc. 2017, 6, e007360. [Google Scholar] [CrossRef]
  66. Gao, T.; Wang, Y.; Zhang, Z. Silent cerebral microbleeds on susceptibility-weighted imaging of patients with ischemic stroke and leukoaraiosis. Neurol. Res. 2013, 30, 272–276. [Google Scholar] [CrossRef] [PubMed]
  67. Bai, Q.; Zhao, Z.; Sui, H.; Xie, X.; Chen, J.; Yang, J.; Zhang, L. Susceptibility-weighted imaging for cerebral microbleed detection in super-acute ischemic stroke patients treated with intravenous thrombolysis. Neurol. Res. 2013, 35, 586–593. [Google Scholar] [CrossRef] [PubMed]
  68. Huang, Z.; Yin, Q.; Sun, W.; Zhu, W.; Li, Y.; Liu, W.; Xiao, L.; Duan, Z.; Cai, Q.; Liu, D.; et al. Microbleeds in ischemic stroke are associated with lower serum adiponectin and higher soluble E-selectin levels. J. Neurol. Sci. 2013, 334, 83–87. [Google Scholar] [CrossRef]
  69. Elnekeidy, A.E.; Yehia, A.; Elfatatry, A. Importance of susceptibility weighted imaging (SWI) in management of cerebro-vascular strokes (CVS). Alex. J. Med. 2019, 50, 83–91. [Google Scholar] [CrossRef][Green Version]
  70. Gratz, P.P.; El-Koussy, M.; Hsieh, K.; von Arx, S.; Mono, M.-L.; Heldner, M.R.; Fischer, U.; Mattle, H.P.; Zubler, C.; Schroth, G.; et al. Preexisting Cerebral Microbleeds on Susceptibility-Weighted Magnetic Resonance Imaging and Post-Thrombolysis Bleeding Risk in 392 Patients. Stroke 2014, 45, 1684–1688. [Google Scholar] [CrossRef]
  71. Horstmann, S.; Möhlenbruch, M.; Wegele, C.; Rizos, T.; Laible, M.; Rauch, G.; Veltkamp, R. Prevalence of atrial fibrillation and association of previous antithrombotic treatment in patients with cerebral microbleeds. Eur. J. Neurol. 2014, 22, 1355–1362. [Google Scholar] [CrossRef]
  72. Liu, C.; Dong, Z.; Xu, L.; Khursheed, A.; Dong, L.; Liu, Z.; Yang, J.; Liu, J. MR image features predicting hemorrhagic transformation in acute cerebral infarction: A multimodal study. Neuroradiology 2015, 57, 1145–1152. [Google Scholar] [CrossRef] [PubMed]
  73. Yan, S.; Jin, X.; Zhang, X.; Zhang, S.; Liebeskind, D.S.; Lou, M. Extensive cerebral microbleeds predict parenchymal haemorrhage and poor outcome after intravenous thrombolysis. J. Neurol. Neurosurg. Psychiatry 2015, 86, 1267–1272. [Google Scholar] [CrossRef] [PubMed]
  74. Zhao, G.; Sun, L.; Wang, Z.; Wang, L.; Cheng, Z.; Lei, H.; Yang, D.; Cui, Y.; Zhang, S. Evaluation of the role of susceptibility-weighted imaging in thrombolytic therapy for acute ischemic stroke. J. Clin. Neurosci. 2017, 40, 175–179. [Google Scholar] [CrossRef]
  75. Akhtar, N.; Salam, A.; Kamran, S.; D’Souza, A.; Imam, Y.; Own, A.; ElSotouhy, A.; Vattoth, S.; Bourke, P.; Bhutta, Z.; et al. Pre-existing Small Vessel Disease in Patients with Acute Stroke from the Middle East, Southeast Asia, and Philippines. Transl. Stroke Res. 2017, 9, 274–282. [Google Scholar] [CrossRef]
  76. Chacon-Portillo, M.A.; Llinas, R.H.; Marsh, E.B. Cerebral microbleeds shouldn’t dictate treatment of acute stroke: A retrospective cohort study evaluating risk of intracerebral hemorrhage. BMC Neurol. 2018, 18, 33. [Google Scholar] [CrossRef]
  77. Lau, K.K.; Lovelock, C.E.; Li, L.; Simoni, M.; Gutnikov, S.; Küker, W.; Mak, H.K.F.; Rothwell, P.M. Antiplatelet Treatment After Transient Ischemic Attack and Ischemic Stroke in Patients With Cerebral Microbleeds in 2 Large Cohorts and an Updated Systematic Review. Stroke 2018, 49, 1434–1442. [Google Scholar] [CrossRef]
  78. Ozbek, D.; Ozturk Tan, O.; Ekinci, G.; Midi, I. Risk of hemorrhage in ischemic stroke and its relationship with cerebral microbleeds. Clin. Neurol. Neurosurg. 2018, 168, 112–117. [Google Scholar] [CrossRef] [PubMed]
  79. Purrucker, J.C.; Wolf, M.; Haas, K.; Siedler, T.; Rizos, T.; Khan, S.; Heuschmann, P.U.; Veltkamp, R. Microbleeds in ischemic vs hemorrhagic strokes on novel oral anticoagulants. Acta Neurol. Scand. 2018, 138, 163–169. [Google Scholar] [CrossRef]
  80. Zhao, F.F.; Gao, H.Y.; Gao, Y.; Zhao, Z.; Li, J.; Ning, F.B.; Zhang, X.N.; Wang, Z.G.; Yu, A.L.; Guo, Y.Y.; et al. A Correlational Study on Cerebral Microbleeds and Carotid Atherosclerosis in Patients with Ischemic Stroke. J. Stroke Cerebrovasc. Dis. 2018, 27, 2228–2234. [Google Scholar] [CrossRef]
  81. Li, G.-F.; Wu, Y.-L.; Wang, S.; Shi, Y.-H.; Zhao, R.; Liu, F.-D.; Liu, Y.-S.; Zhuang, M.-T.; Zhao, Y.; Sun, Q.; et al. Previous chronic symptomatic and asymptomatic cerebral hemorrhage in patients with acute ischemic stroke. Neuroradiology 2018, 61, 103–107. [Google Scholar] [CrossRef]
  82. Xu, C.X.; Xu, H.; Yi, T.; Yi, X.Y.; Ma, J.P. Cerebral Microbleed Burden in Ischemic Stroke Patients on Aspirin: Prospective Cohort of Intracranial Hemorrhage. Front. Neurol. 2021, 12, 742899. [Google Scholar] [CrossRef] [PubMed]
  83. Diker, S.; Gelener, P.; Eker, A.; Kaymakamzade, B.; Mut, S.; Erem, A.; Balyemez, U. Association between cerebral microbleeds and inflammatory biomarkers in patients with ischemic stroke. Egypt. J. Neurol. Psychiatr. Neurosurg. 2022, 58, 43. [Google Scholar] [CrossRef]
  84. Khaladkar, S.M.; Chanabasanavar, V.; Dhirawani, S.; Thakker, V.; Dilip, D.; Parripati, V.K. Susceptibility Weighted Imaging: An Effective Auxiliary Sequence That Enhances Insight Into the Imaging of Stroke. Cureus 2022, 14, e24918. [Google Scholar] [CrossRef] [PubMed]
  85. Nagaraja, N.; Farooqui, A.; Bin Zahid, A.; Kaur, S. Factors associated with the presence of cerebral microbleeds and its influence on outcomes of stroke not treated with alteplase. Clin. Neurol. Neurosurg. 2021, 207, 106798. [Google Scholar] [CrossRef]
  86. Zhang, X.; Liu, L.; Jiang, N.; Liu, Y.; Wang, Q.; Tang, X.; Zhai, Q.; Zhao, L. Correlation of lipoprotein-associated phospholipase A2 and cerebral microbleeds in patients with acute ischaemic stroke. BMC Neurol. 2022, 22, 482. [Google Scholar] [CrossRef]
  87. Bao, Y.; Gu, J.; Lv, T.; Chen, M.; Zhao, K.; Yang, Y.; Gu, D. Correlation between blood pressure variability and deep cerebral microbleeds in patients with acute ischemic stroke. Folia Neuropathol. 2023, 61, 309–316. [Google Scholar] [CrossRef]
  88. Chen, L.; Liu, F.; Tian, X.; Zhang, T.; Zhang, J.; Ran, F. Impact of cerebral microbleeds on cognitive functions and its risk factors in acute cerebral infarction patients. Neurol. Res. 2023, 45, 564–571. [Google Scholar] [CrossRef]
  89. Guo, X.; Xing, Y.; Teng, Z.; Shen, Z.; Guo, X.; Lv, P.; Tian, S. Gender heterogeneity in the influencing factors for cerebral microbleeds in acute ischemic stroke patients. Curr. Med. Res. Opin. 2023, 39, 1045–1054. [Google Scholar] [CrossRef]
  90. Wang, M.; Yang, Y.; Wang, Y.; Luan, M.; Xu, L.; Zhong, M.; Zheng, X. Dual antiplatelet therapy in acute ischaemic stroke with or without cerebral microbleeds. Eur. J. Neurosci. 2023, 57, 1197–1207. [Google Scholar] [CrossRef]
  91. Wang, M.; Yang, Y.; Wang, Y.; Luan, M.; Zhong, M.; Xu, L.; Zheng, X. Comparative effectiveness of dual to single antiplatelet therapy after one year versus seven years in patients with acute ischemic stroke combined with cerebral microbleeds. J. Clin. Neurosci. 2023, 112, 73–79. [Google Scholar] [CrossRef]
  92. Hou, Y.; Xiao, Q.; Chen, Z.; Cheng, J.; Shi, Z.; Zhu, Q.; Liu, C. Risk factors and prognosis of acute ischemic stroke complicated with cerebral microbleeds. Neurol. Asia 2024, 29, 591–596. [Google Scholar] [CrossRef]
  93. Luo, Y.; Gao, K.; Fawaz, M.; Wu, B.; Zhong, Y.; Zhou, Y.; Haacke, E.M.; Dai, Y.; Liu, S. Automatic detection of cerebral microbleeds using susceptibility weighted imaging and artificial intelligence. Quant. Imaging Med. Surg. 2024, 14, 2640–2654. [Google Scholar] [CrossRef] [PubMed]
  94. Nasreldein, A.; Shoamnesh, A.; Foli, N.; Makboul, M.; Salah, S.; Faßbender, K.; Walter, S. Prevalence and Risk Factors of Cerebral Microbleeds among Egyptian Patients with Acute Ischemic Stroke. Neuroepidemiology 2024, 59, 1–9. [Google Scholar] [CrossRef]
  95. Schlemm, L.; Braemswig, T.B.; Boutitie, F.; Vynckier, J.; Jensen, M.; Galinovic, I.; Simonsen, C.Z.; Cheng, B.; Cho, T.-H.; Fiehler, J.; et al. Cerebral Microbleeds and Treatment Effect of Intravenous Thrombolysis in Acute Stroke. Neurology 2022, 98, E302–E314. [Google Scholar] [CrossRef]
  96. Zhao, D.X.; Gootee, E.; Johansen, M.C. Atrial cardiopathy is associated with cerebral microbleeds in ischemic stroke patients. Front. Neurol. 2022, 13, 982926. [Google Scholar] [CrossRef]
  97. Brauner, R.; Gory, B.; Lapergue, B.; Sibon, I.; Richard, S.; Kyheng, M.; Labreuche, J.; Desilles, J.P.; Blanc, R.; Piotin, M.; et al. Effect of small vessel disease severity on blood pressure management after endovascular therapy in theBP TARGETtrial. Eur. J. Neurol. 2023, 30, 1676–1685. [Google Scholar] [CrossRef]
  98. De Simone, M.; Fontanella, M.M.; Choucha, A.; Schaller, K.; Machi, P.; Lanzino, G.; Bijlenga, P.; Kurz, F.T.; Lövblad, K.-O.; De Maria, L. Current and Future Applications of Arterial Spin Labeling MRI in Cerebral Arteriovenous Malformations. Biomedicines 2024, 12, 753. [Google Scholar] [CrossRef] [PubMed]
  99. Fandler-Höfler, S.; Eppinger, S.; Ambler, G.; Nash, P.; Kneihsl, M.; Lee, K.-J.; Lim, J.-S.; Shiozawa, M.; Koga, M.; Li, L.; et al. Sex Differences in Frequency, Severity, and Distribution of Cerebral Microbleeds. JAMA Netw. Open 2024, 7, e2439571. [Google Scholar] [CrossRef]
  100. Liang, C.; Wang, J.; Feng, M.; Zhang, N.; Guo, L. White matter changes, duration of hypertension, and age are associated with cerebral microbleeds in patients with different stages of hypertension. Quant. Imaging Med. Surg. 2022, 12, 119–130. [Google Scholar] [CrossRef]
  101. Lyu, L.; Shen, J.; Zeng, C.; Ji, J.; Hu, W.; Wei, T.; Mao, W. Cerebral microbleeds are associated with blood pressure levels in individuals with hypertension. Clin. Exp. Hypertens. 2019, 42, 328–334. [Google Scholar] [CrossRef] [PubMed]
  102. Alemany, M.; Stenborg, A.; Terent, A.; Sonninen, P.; Raininko, R. Coexistence of Microhemorrhages and Acute Spontaneous Brain Hemorrhage: Correlation with Signs of Microangiopathy and Clinical Data. Radiology 2006, 238, 240–247. [Google Scholar] [CrossRef]
  103. Stehling, C.; Wersching, H.; Kloska, S.P.; Kirchhof, P.; Ring, J.; Nassenstein, I.; Allkemper, T.; Knecht, S.; Bachmann, R.; Heindel, W. Detection of asymptomatic cerebral microbleeds: A comparative study at 1.5 and 3.0 T. Acad. Radiol. 2008, 15, 895–900. [Google Scholar] [CrossRef] [PubMed]
  104. De Kort, A.M.; Verbeek, M.M.; Schreuder, F.H.B.M.; Klijn, C.J.M.; Jäkel, L. Prevalence of Cerebral Amyloid Angiopathy Pathology and Strictly Lobar Microbleeds in East-Asian Versus Western Populations: A Systematic Review and Meta-Analysis. J. Stroke 2024, 26, 179–189. [Google Scholar] [CrossRef] [PubMed]
  105. Martinez-Ramirez, S.; Greenberg, S.M.; Viswanathan, A. Cerebral microbleeds: Overview and implications in cognitive impairment. Alzheimer’s Res. Ther. 2014, 6, 33. [Google Scholar] [CrossRef]
  106. Yilmaz, P.; Ikram, M.A.; Ikram, M.K.; Niessen, W.J.; Viswanathan, A.; Charidimou, A.; Vernooij, M.W. Application of an Imaging-Based Sum Score for Cerebral Amyloid Angiopathy to the General Population: Risk of Major Neurological Diseases and Mortality. Front. Neurol. 2019, 10, 1276. [Google Scholar] [CrossRef]
  107. Tipirneni, S.; Stanwell, P.; Weissert, R.; Bhaskar, S.M.M. Prevalence and Impact of Cerebral Microbleeds on Clinical and Safety Outcomes in Acute Ischaemic Stroke Patients Receiving Reperfusion Therapy: A Systematic Review and Meta-Analysis. Biomedicines 2023, 11, 2865. [Google Scholar] [CrossRef]
  108. Tsivgoulis, G.; Zand, R.; Katsanos, A.H.; Turc, G.; Nolte, C.H.; Jung, S.; Cordonnier, C.; Fiebach, J.B.; Scheitz, J.F.; Klinger-Gratz, P.P.; et al. Risk of Symptomatic Intracerebral Hemorrhage After Intravenous Thrombolysis in Patients With Acute Ischemic Stroke and High Cerebral Microbleed Burden: A Meta-analysis. JAMA Neurol. 2016, 73, 675–683. [Google Scholar] [CrossRef]
  109. Wilson, D.; Charidimou, A.; Ambler, G.; Fox, Z.V.; Gregoire, S.; Rayson, P.; Imaizumi, T.; Fluri, F.; Naka, H.; Horstmann, S.; et al. Recurrent stroke risk and cerebral microbleed burden in ischemic stroke and TIA: A meta-analysis. Neurology 2016, 87, 1501–1510. [Google Scholar] [CrossRef] [PubMed]
  110. Chojdak-Łukasiewicz, J.; Dziadkowiak, E.; Zimny, A.; Paradowski, B. Cerebral small vessel disease: A review. Adv. Clin. Exp. Med. 2021, 30, 349–356. [Google Scholar] [CrossRef]
  111. Daugherty, A.M.; Raz, N. Incident risk and progression of cerebral microbleeds in healthy adults: A multi-occasion longitudinal study. Neurobiol. Aging 2017, 59, 22–29. [Google Scholar] [CrossRef]
  112. Schrag, M.; Greer, D.M. Clinical Associations of Cerebral Microbleeds on Magnetic Resonance Neuroimaging. J. Stroke Cerebrovasc. Dis. 2014, 23, 2489–2497. [Google Scholar] [CrossRef] [PubMed]
  113. Yates, P.A.; Villemagne, V.L.; Ellis, K.A.; Desmond, P.M.; Masters, C.L.; Rowe, C.C. Cerebral microbleeds: A review of clinical, genetic, and neuroimaging associations. Front. Neurol. 2014, 4, 205. [Google Scholar] [CrossRef] [PubMed]
  114. De Reuck, J.; Auger, F.; Cordonnier, C.; Deramecourt, V.; Durieux, N.; Pasquier, F.; Bordet, R.; Maurage, C.A.; Leys, D. Comparison of 7.0-T T2*-magnetic resonance imaging of cerebral bleeds in post-mortem brain sections of Alzheimer patients with their neuropathological correlates. Cerebrovasc. Dis. 2011, 31, 511–517. [Google Scholar] [CrossRef] [PubMed]
  115. Yilmazer-Hanke, D.; Mayer, T.; Muller, H.P.; Neugebauer, H.; Abaei, A.; Scheuerle, A.; Weis, J.; Forsberg, K.M.E.; Althaus, K.; Meier, J.; et al. Histological correlates of postmortem ultra-high-resolution single-section MRI in cortical cerebral microinfarcts. Acta Neuropathol. Commun. 2020, 8, 33. [Google Scholar] [CrossRef]
Medicina 61 01566 g001
Figure 1. PRISMA Flowchart: Inclusion of Studies in the Meta-Analysis. Illustration depicting the flow of study selection according to the PRISMA guidelines, leading to the inclusion of studies in the meta-analysis. Abbreviations: CMB = cerebral microbleed, N = number of studies, n = cohort size, TIA = transient ischemic attack, T2* = T2 Gradient Echo Imaging, SWI = Susceptibility Weighted Imaging, sICH = symptomatic intracranial hemorrhage, HT = hemorrhagic transformation, mRS = Modified Rankin Scale.
Figure 1. PRISMA Flowchart: Inclusion of Studies in the Meta-Analysis. Illustration depicting the flow of study selection according to the PRISMA guidelines, leading to the inclusion of studies in the meta-analysis. Abbreviations: CMB = cerebral microbleed, N = number of studies, n = cohort size, TIA = transient ischemic attack, T2* = T2 Gradient Echo Imaging, SWI = Susceptibility Weighted Imaging, sICH = symptomatic intracranial hemorrhage, HT = hemorrhagic transformation, mRS = Modified Rankin Scale.
Medicina 61 01566 g001
Medicina 61 01566 g002
Figure 2. Forest Plots: Pooled Prevalence of CMBs based on Different Imaging Modalities [7,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97]. (a) Prevalence of CMBs in AIS patients assess using SWI. (b) Prevalence of CMBs in AIS patients assessed using T2*. (c) Prevalence of CMBs in AIS patients assesed using both SWI and T2*. The red diamonds indicate the pooled prevalence estimates with 95% confidence intervals, and the red dashed vertical lines represent the overall pooled prevalence across studies. Abbreviations: CMB = cerebral microbleed, CI = confidence interval, T2* = T2 Gradient Echo Imaging, SWI = Susceptibility Weighted Imaging, N = number of patients, C = cohort size, P = prevalence, AIS = acute ischemic stroke, ES = effect size.
Figure 2. Forest Plots: Pooled Prevalence of CMBs based on Different Imaging Modalities [7,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97]. (a) Prevalence of CMBs in AIS patients assess using SWI. (b) Prevalence of CMBs in AIS patients assessed using T2*. (c) Prevalence of CMBs in AIS patients assesed using both SWI and T2*. The red diamonds indicate the pooled prevalence estimates with 95% confidence intervals, and the red dashed vertical lines represent the overall pooled prevalence across studies. Abbreviations: CMB = cerebral microbleed, CI = confidence interval, T2* = T2 Gradient Echo Imaging, SWI = Susceptibility Weighted Imaging, N = number of patients, C = cohort size, P = prevalence, AIS = acute ischemic stroke, ES = effect size.
Medicina 61 01566 g002
Medicina 61 01566 g003
Figure 3. Forest Plots: Prognostic Outcome Analysis of CMBs Stratified by Imaging Modality [7,20,24,26,28,35,38,40,42,47,48,51,52,54,59,60,62,63,64,69,70,72,73,74,76,78,84,85,95]. (a) Association between CMBs and sICH among AIS patients. (b) Association between CMBs and HT among AIS patients. (c) Association between CMBs and poor functional outcome, defined as mRS scores of 3-6, among AIS patients. The blue diamonds represent the pooled odds ratios with 95% confidence intervals within each subgroup and for the overall analysis. Abbreviations: CMB = cerebral microbleed, sICH = symptomatic intracerebral hemorrhage, HT = hemorrhagic transformation, mRS = modified Rankin Score, AIS = acute ischemic stroke, CI = confidence interval, DL = DerSimonian and Laird method.
Figure 3. Forest Plots: Prognostic Outcome Analysis of CMBs Stratified by Imaging Modality [7,20,24,26,28,35,38,40,42,47,48,51,52,54,59,60,62,63,64,69,70,72,73,74,76,78,84,85,95]. (a) Association between CMBs and sICH among AIS patients. (b) Association between CMBs and HT among AIS patients. (c) Association between CMBs and poor functional outcome, defined as mRS scores of 3-6, among AIS patients. The blue diamonds represent the pooled odds ratios with 95% confidence intervals within each subgroup and for the overall analysis. Abbreviations: CMB = cerebral microbleed, sICH = symptomatic intracerebral hemorrhage, HT = hemorrhagic transformation, mRS = modified Rankin Score, AIS = acute ischemic stroke, CI = confidence interval, DL = DerSimonian and Laird method.
Medicina 61 01566 g003
Table 1. Clinical Characteristics of Studies Included in Meta-Analysis for Acute Ischemic Stroke Patients.
Table 1. Clinical Characteristics of Studies Included in Meta-Analysis for Acute Ischemic Stroke Patients.
AuthorYearContinentStudy DesignCohortAge Mean (±Standard Deviation (SD))Male, n (n%)Number of CMBsCMB DefinitionCMB Imaging
OverallCerebral Microbleed (CMB)No CMB
Agbonon et al. [63]2024EuropeRetrospective44568.3 (±15.2)71.7 (±13)-229 (51)70-T2 Gradient Echo Imaging (T2*GRE)
Akhtar et al. [75]2018AsiaRetrospective71854.7 (±14)--594 (83)166<5 mmSusceptibility Weighted Imaging (SWI)
Bai et al. [67]2013AsiaProspective11361.6 (±10.8)---46-SWI
Bao et al. [87]2023AsiaRetrospective199----922–10 mmSWI
Braemswig et al. [53]2019EuropeProspective396---103 (26)115<10 mmT2*GRE
Brauner et al. [97]2023EuropeProspective24673.6 (±13.3)--117 (48)72-T2*GRE, SWI
Brundel et al. [41]2014EuropeProspective155----19-T2*GRE
Capuana et al. [59]2021EuropeProspective43468.3 (±13.3)69 (±12.6)68.1 (±13.8)264 (61)101<10 mmT2*GRE
Chacon-Portillo et al. [76]2018North AmericaRetrospective29263 (±15)--240 (82)622–10 mmSWI
Chen et al. [88]2023AsiaRetrospective190---104 (55)82<10 mmSWI
Choi et al. [54]2019AsiaProspective153269.4 (±11.8)72 (±11.2)68.9 (±11.9)855 (56)165-T2*GRE
Dannenberg et al. [42]2014EuropeProspective326---159 (49)81≤10 mmT2*GRE
Dassan et al. [34]2011EuropeRetrospective20----5-T2*GRE
Derraz et al. [60]2021EuropeProspective51369.4 (±25.9)80.8 (±15.7)67.3 (±25.4)243 (47)89≤10 mmT2*GRE
Diker et al. [83]2022EuropeRetrospective12766.6 (±14.4)68.5 (±12.9)63.6 (±15.6)74 (58)47<10 mmSWI
Elnekeidy et al. [69]2014AfricaProspective46----5-SWI
Fan et al. [21]2003AsiaProspective12168 (±11)69.5 (±11)67.1 (±10.9)82 (68)43-T2*GRE
Fiehler et al. [28]2007MultinationalRetrospective57068.3 (±13.3)--341 (60)86<5 mmT2*GRE
Gao et al. [66]2008AsiaRetrospective114----20<10 mmSWI
Gratz et al. [70]2014EuropeProspective39268.1 (±13.7)--223 (57)79<5 mmSWI
Gregoire et al. [36]2013EuropeProspective254----59-T2*GRE
Guo et al. [89]2023AsiaRetrospective23063.8 (±11)66.5 (±10.8)61.3 (±11.1)160 (70)1112–10 mmSWI
Han et al. [30]2009AsiaRetrospective24761.3 (±11.4)64.6 (±11)60 (±11.6)176 (71)72≤5 mmT2*GRE
Horstmann et al. [71]2015EuropeProspective645----165≤10 mmSWI
Hou et al. [92]2024AsiaRetrospective20068.3 (±9.5)70.7 (±8.6)65.3 (±10.5)144 (72)112-SWI
Huang et al. [68]2013AsiaProspective12663.8 (±13)64.6 (±12.7)63.2 (±13.3)83 (66)632–10 mmSWI
Jablonski et al. [61]2021EuropeProspective49---23 (47)14-T2*GRE
Jeon et al. [31]2009AreaRetrospective23764 (±12.8)--142 (60)75≤5 mmT2*GRE
Kakuda et al. [24]2005MultinationalProspective7070.8 (±29.2)70 (±32)71 (±29)31 (44)11<5 mmT2*GRE
Kato et al. [19]2002AsiaRetrospective113---65 (58)53-T2*GRE
Khaladkar et al. [84]2022AsiaProspective20----13-SWI
Kidwell et al. [7]2002North AmericaRetrospective41----5<5 mmT2*GRE, SWI
Ho et al. [26]2006AsiaRetrospective65---37 (57)25<5 mmT2*GRE
Kimura et al. [37]2013AsiaProspective22476.2 (±10.6)--121 (54)72-T2*GRE
Lau et al. [65]2017AsiaProspective100369 (±12)--601 (60)450<10 mmSWI
Lau et al. [77]2018AsiaProspective1003---601 (60)450-SWI
Lee et al. [22]2004AsiaRetrospective14464.6 (±9.1)--75 (52)50≤5 mmT2*GRE
Lee et al. [62]2022AsiaRetrospective57767 (±13)70.8 (±10.4)66.7 (±12.8)322 (56)91<10 mmT2*GRE
Li et al. [81]2019AsiaRetrospective18071.5 (±12.4)--100 (56)902–10 mmSWI
Liang et al. [55]2019AsiaProspective56367 (±10.2)--333 (59)76-T2*GRE
Liu et al. [72]2015AsiaProspective8767.3 (±12.5)--49 (56)162–5 mmSWI
Luo et al. [93]2024AsiaRetrospective206----123≤10 mmSWI
Moriya et al. [38]2013AsiaRetrospective7173 (±10)--50 (70)14-T2*GRE
Nagaraja et al. [85]2021North AmericaRetrospective19666.1 (±14)72 (±13)63.6 (±14.4)98 (50)582–10 mmSWI
Nagaraja et al. [52]2018North AmericaRetrospective36667 (±15)74.1 (±12.5)64.9 (±15.2)198 (54)95<10 mmT2*GRE
Naka et al. [23]2004AsiaProspective66----12-T2*GRE
Naka et al. [39]2013AsiaProspective150272.6 (±12)--881 (59)542<10 mmT2*GRE
Naka et al. [27]2006AsiaProspective183----53-T2*GRE
Nam et al. [56]2019AsiaProspective84168 --516 (61)257<10 mmT2*GRE
Nasreldein et al. [94]2024AfricaProspective364----102-SWI
Nighoghossian et al. [20]2002EuropeProspective10060 (±13)--58 (58)202–5 mmT2*GRE
Orken et al. [32]2009EuropeProspective14165.8 (±12.2)69.6 (±10.7)64.7 (±12.4)82 (58)31<5 mmT2*GRE
Ozbek et al. [78]2018EuropeProspective14868 (±14.8)--84 (57)662–10 mmSWI
Potigumjon et al. [49]2017AsiaRetrospective200616660126 (63)39<10 mmT2*GRE
Purrucker et al. [79]2018EuropeProspective29078.6--150 (52)362–10 mmSWI
Ryu et al. [58]2020AsiaProspective47766 (±14)--294 (62)125≤10 mmT2*GRE
Schlemm et al. [95]2022EuropeProspective4596871.767289 (63)98≤10 mmT2*GRE, SWI
Shahjouei et al. [50]2017North AmericaRetrospective76062.1 (±13.9)--391 (51)122≤10 mmT2*GRE
Shi et al. [47]2016AsiaProspective20666.8 (±17.6)77 (±14)65 (±18)87 (42)37<10 mmT2*GRE
Soo et al. [35]2012AsiaProspective13367.36767.4-232–10 mmT2*GRE
Soo et al. [29]2008AsiaProspective90868.4 (±11.9)71.2 (±10)67.3 (±11.8)524 (58)252-T2*GRE
Sun et al. [33]2009AsiaRetrospective99868.3 (±11.7)71.4 (±10)67.2 (±12)588 (59)2732–10 mmT2*GRE
Takahashi et al. [57]2019AsiaProspective69---45 (65)19-T2*GRE
Takahashi et al. [40]2013AsiaRetrospective18774 (±11)--112 (60)63-T2*GRE
Turc et al. [45]2015EuropeProspective717---351 (49)150≤10 mmT2*GRE
Wang et al. [43]2014AsiaProspective34865.2 (±13.1)--207 (59)1602–5 mmT2*GRE
Wang et al. [90]2023AsiaRetrospective58164.365.663.5388 (67)225<10 mmSWI
Wang et al. [91]2023AsiaRetrospective732----279<10 mmSWI
Werring et al. [25]2005EuropeProspective8662.1 (±16.1)--57 (66)20<10 mmT2*GRE
Xu et al. [82]2021AsiaProspective45967.3 (±11.7)69 (±11.3)66.1 (±12)314 (68)1872–10 mmSWI
Yan et al. [73]2015AsiaRetrospective33366.2 (±13)--223 (67)133≤10 mmSWI
Yan et al. [44]2014AsiaProspective12167.3 (±12.5)72.2 (±13)-77 (64)57≤10 mmT2*GRE
Yang et al. [48]2016AsiaProspective34865.2 (±13.1)--207 (59)1602–5 mmT2*GRE
Zand et al. [64]2018North AmericaRetrospective77261.9 (±14.2)64.9 (±13.2)61.3 (±14.3)398 (52)124≤10 mmT2*GRE
Zand et al. [51]2017North AmericaProspective67262 (±14)64.8 (±14.1)61
(±14)		350 (52)103≤10 mmT2*GRE
Zhang et al. [86]2022AsiaProspective24267.5 (±9.5)69.5 (±9.9)66.7
(±9.2)		158 (65)71≤10 mmSWI
Zhang et al. [46]2015AsiaRetrospective696606659516 (74)162≤10 mmT2*GRE
Zhao et al. [74]2017AsiaProspective6062.3 (±12.5)--38 (63)142–5 mmSWI
Zhao et al. [80]2018AsiaProspective19868.1 (±8.7)--109 (55)91<10 mmSWI
Zhao et al. [96]2022North AmericaProspective12059.6--65 (54)39<10 mmT2*GRE, SWI
Abbreviations: CMB = cerebral microbleed, n = number, T2*GRE = T2 Gradient Echo Imaging, SWI = Susceptibility Weighted Imaging, SD = Standard Deviation.
Table 2. Rates of Clinical Risk Factors in Acute Ischemic Stroke Patients with Cerebral Microbleeds included in the Meta-Analysis.
Table 2. Rates of Clinical Risk Factors in Acute Ischemic Stroke Patients with Cerebral Microbleeds included in the Meta-Analysis.
Clinical Risk Factors, n (n%)
AuthorYearAtrial Fibrillation Hyper-lipidaemia Hypertension Coronary Artery Disease Prior Stroke/Transient Ischemic StrokeSmokingDiabetes Mellitus
Agbonon et al. [63]2024-25 (14)46 (18)---6 (10)
Akhtar et al. [75]2018-------
Bai et al. [67]2013-------
Bao et al. [87]2023-------
Braemswig et al. [53]2019-------
Brauner et al. [97]2023-------
Brundel et al. [41]2014-------
Capuana et al. [59]202117 (24)-77 (27)--18 (22)15 (20)
Chacon-Portillo et al. [76]2018-------
Chen et al. [88]2023--35 (61)17 (47)-37 (44)26 (42)
Choi et al. [54]2019-------
Dannenberg et al. [42]2014-------
Dassan et al. [34]2011-------
Derraz et al. [60]202138 (25)36 (23)65 (21)21 (25)19 (31)28 (14)14 (21)
Diker et al. [83]202221 (42)17 (38)34 (41)11 (44)7 (27)-15 (35)
Elnekeidy et al. [69]2014-------
Fan et al. [21]20033 (50)11 (41)32 (38)--17 (34)11 (28)
Fiehler et al. [28]2007-------
Gao et al. [66]2008-------
Gratz et al. [70]2014-------
Gregoire et al. [36]2013-------
Guo et al. [89]2023-32 (38)88 (53)--30 (43)34 (45)
Han et al. [30]2009--63 (40)-26 (40)34 (26)17 (24)
Horstmann et al. [71]2015-------
Hou et al. [92]20249 (56)-83 (58)11 (61)-50 (55)55 (63)
Huang et al. [68]2013-14 (38)53 (56)--18 (49)10 (59)
Jablonski et al. [61]2021-------
Jeon et al. [31]2009-------
Kakuda et al. [24]2005-2 (12)8 (19)--6 (20)4 (21)
Kato et al. [19]2002-------
Khaladkar et al. [84]2022-------
Kidwell et al. [7]2002-------
Ho et al. [26]2006-------
Kimura et al. [37]2013-------
Lau et al. [65]2017-------
Lau et al. [77]2018-------
Lee et al. [22]2004-------
Lee et al. [62]202242 (15)-71 (20)-24 (24)19 (14)27 (17)
Li et al. [81]2019-------
Liang et al. [55]2019-------
Liu et al. [72]2015-------
Luo et al. [93]2024-------
Moriya et al. [38]2013-------
Nagaraja et al. [85]202115 (58)27 (38)52 (34)15 (34)33 (48)-20 (29)
Nagaraja et al. [52]201814 (24)48 (33)67 (31)22 (39)25 (49)19 (19)23 (28)
Naka et al. [23]2004-------
Naka et al. [39]2013-------
Naka et al. [27]2006-------
Nam et al. [56]2019-------
Nasreldein et al. [94]2024-------
Nighoghossian et al. [20]2002-------
Orken et al. [32]2009--27 (24)-7 (27)5 (13)6 (22)
Ozbek et al. [78]2018-------
Potigumjon et al. [49]20176 (15)21 (18)33 (27)1 (10)9 (25)10 (20)10 (18)
Purrucker et al. [79]2018-------
Ryu et al. [58]2020-------
Schlemm et al. [95]202216 (32)-64 (26)-14 (24)-22 (30)
Shahjouei et al. [50]2017-------
Shi et al. [47]201616 (20)10 (16)26 (19)11 (26)5 (15)-13 (30)
Soo et al. [35]2012-20 (18)20 (20)-12 (22)12 (21)8 (20)
Soo et al. [29]200819 (28)138 (25)200 (32)19 (25)83 (46)64 (34)76 (26)
Sun et al. [33]200919 (28)148 (25)211 (32)---81 (25)
Takahashi et al. [57]2019-------
Takahashi et al. [40]2013-------
Turc et al. [45]2015-------
Wang et al. [43]2014-------
Wang et al. [90]2023-82 (36)174 (44)---81 (42)
Wang et al. [91]2023-------
Werring et al. [25]2005-------
Xu et al. [82]202110 (43)4 (31)120 (45)--99 (44)44 (39)
Yan et al. [73]2015-------
Yan et al. [44]2014-------
Yang et al. [48]2016-------
Zand et al. [64]201813 (17)51 (20)110 (18)-43 (22)45 (16)44 (17)
Zand et al. [51]2017-------
Zhang et al. [86]20229 (30)-54 (34)16 (27)-26 (30)27 (39)
Zhang et al. [46]2015-124 (22)149 (27)--68 (21)53 (19)
Zhao et al. [74]2017-------
Zhao et al. [80]2018--25 (52)--44 (46)13 (54)
Zhao et al. [96]2022-------
Abbreviations: CMB = cerebral microbleed, n = number., AF = atrial fibrillation, HL = hyperlipidaemia, HTN = hypertension, CAD = coronary artery disease, PS = prior stroke, TIA = transient ischemic attack, DM = diabetes mellitus.
Table 3. Prognostic Outcomes of Studies Selected for Meta-Analysis.
Table 3. Prognostic Outcomes of Studies Selected for Meta-Analysis.
AuthorYearReperfusion TherapySymptomatic Intracranial Hemorrhage (sICH) DefinitionsICH, n (n%)Hemorrhagic Transformation (HT), n (n%)Modified Ranking Scale (mRS) 3–6 at 90 Days, n (n%)
OverallCerebral Microbleed (CMB)No-CMBOverallCMBNo-CMBOverallCMBNo-CMB
Agbonon et al. [63]2024Endovascular Thrombolysis (EVT)ECASS-II34 (7.6)6 (1.4)28 (6.3)---194 (43.6)31 (7.0)163 (36.6)
Capuana et al. [59]2021Intravenous Thrombolysis (IVT)SITS-MOST13 (3.0)7 (1.6)6 (1.4)---130 (30.0)39 (9.0)91 (21.0)
Chacon-Portillo et al. [76]2018IVTNINDS6 (2.0)3 (1.0)3 (1.0)46 (15.8)12 (4.1)34 (11.6)63 (21.6)16 (6.2)42 (14.4)
Choi et al. [54]2019IVT/EVTECASS-I69 (4.5)17 (1.1)52 (3.4)420 (27.4)66 (4.3)354 (23.1)865 (56.4)103 (6.7)763 (49.8)
Dannenberg et al. [42]2014IVTECASS-III10 (3.1)7 (2.1)3 (0.9)---158 (48.4)50 (15.3)108 (33.1)
Derraz et al. [60]2021EVTECASS-II66 (12.9)15 (2.9)51 (9.9)---281 (54.8)59
(11.5)		222 (43.3)
Elnekeidy et al. [69]2014-----10 (21.7)1 (2.2)9 (19.6)---
Fiehler et al. [28]2007IVTECASS-I18 (3.2)5 (0.9)13 (2.3)------
Gratz et al. [70]2014IVT/EVTPROACT-II21 (5.4)3 (0.8)18 (4.6)---193 (49.2)52 (13.3)141 (36.0)
Kakuda et al. [24]2005IVTECASS-II7 (10.0)0 (0)7 (10.0)32 (45.7)3 (4.3)29 (41.4)---
Khaladkar et al. [84]2022-----18 (90)13 (65)5 (25)---
Kidwell et al. [7]2002IVT----15 (36.6)2 (4.9)13 (31.7)---
Ho et al. [26]2006IVT-5 (12.2)3 (7.3)2 (4.9)17 (41.5)8 (19.5)9 (22.0)---
Lee et al. [62]2022EVT----170 (29.5)32 (55.5)138 (21.9)288 (49.9)59 (10.2)229 (39.7)
Liu et al. [72]2015-----17 (19.5)5 (5.7)12 (13.8)---
Moriya et al. [38]2013IVT----26 (36.6)6 (8.5)20 (28.2)---
Nagaraja et al. [52]2018-----87 (23.8)32 (8.7)55 (15.0)---
Nagaraja et al. [85]2021-----22 (11.2)6 (3.1)16 (8.2)36 (18.4)12 (6.1)24 (12.2)
Nighoghossian et al. [20]2002IVT----26 (26.0)10 (10.0)16 (16.0)---
Ozbek et al. [78]2018-----41 (27.7)18 (12.2)23 (15.5)---
Schlemm et al. [95]2022IVTSITS-MOST, ECASS-II, ECASS-III, NINDS26 (5.7)11 (2.4)15 (3.3)102 (22.2)21 (4.6)46 (10.0)125 (27.2)34 (7.4)91 (19.8)
Shi et al. [47]2016EVT----91 (44.2)14 (6.8)77 (37.4)---
Soo et al. [35]2012EVT----7 (5.3)1 (0.8)6 (4.5)---
Takahashi et al. [40]2013-----27 (14.4)5 (2.7)22 (11.8)---
Yan et al. [73]2015IVTECASS-II8 (2.4)6 (1.8)2 (0.6)102 (30.6)48 (14.4)54 (16.2)206 (61.9)140 (42.0)66 (19.8)
Yang et al. [48]2016-----35 (10.0)10 (2.9)25 (7.2)---
Zand et al. [51]2017IVTECASS-II25 (3.7)5 (0.7)20 (3.0)------
Zand et al. [64]2018IVT----6 (0.8)3 (0.4)3 (0.4)---
Zhao et al. [74]2017IVTECASS-II2 (3.3)2 (3.3)0 (0)------
Abbreviations: CMB = cerebral microbleed, n = number, IVT = intravenous thrombolysis, EVT = endovascular thrombolysis, sICH = symptomatic intracranial hemorrhage, HT = hemorrhagic transformation, mRS = Modified Rankin Scale.
Table 4. Meta-Analysis Results for Prevalence of Cerebral Microbleeds: Summary Effects and Heterogeneity.
Table 4. Meta-Analysis Results for Prevalence of Cerebral Microbleeds: Summary Effects and Heterogeneity.
ModalitySubgroupPooled Prevalence (Effect Size)95% Confidence IntervalWeight (%)Heterogeneity χ2 (Degrees of Freedom)p-ValueI2 (%)z-Scorep-Value (z-Test)
T2 Gradient Echo Imaging (T2*)-0.250.22–0.2857.74844.41 (45)094.6728.820
Susceptibility Weighted Imaging (SWI)-0.360.31–0.4137.44563.55 (29)094.8525.610
Both-0.250.18–0.324.8212.99 (3)076.9011.670
Overall-0.290.26–0.311001912.84 (79)095.8735.040
Age
T2*<65 years0.220.18–0.2629.2075.77 (9)088.1218.580
≥65 years0.250.21–0.3070.80674.41 (23)096.5919.140
Overall0.240.21–0.28100779.72 (33)095.7724.200
SWI<65 years0.360.27–0.4731.66103.61 (5)095.1711.570
≥65 years0.350.28–0.4368.34261.71 (12)095.4115.560
Overall0.360.30–0.42100377.01 (18)095.2319.920
Hypertension
T2*<65% HTN0.210.17–0.2744.34161.27 (11)093.1814.730
≥65% HTN0.260.23–0.2955.66107.71 (14)087.0027.150
Overall0.240.21–0.27100300.82 (26)091.3827.940
SWI<65% HTN0.370.27–0.4826.2265.62 (4)093.9010.960
≥65% HTN0.360.29–0.4373.78351.70 (13)096.3016.520
Overall0.360.30–0.42100418.37 (18)095.7020.240
Fluid Attenuated Inversion Recovery (FLAIR)
T2*FLAIR0.240.21–0.2760.69333.22 (27)091.9025.270
No FLAIR0.260.20–0.3139.31499.28 (17)096.6015.950
Overall0.250.22–0.28100844.41 (45)094.6728.820
SWIFLAIR0.380.31–0.4463.04348.57 (16)095.4118.860
No FLAIR0.330.25–0.4236.96185.34 (9)095.1412.300
Overall0.360.31–0.41100555.50 (26)095.3222.700
Non-contrast Computed Tomography (NCCT)
T2*NCCT0.270.21–0.3335.74573.23 (15)097.3814.890
No NCCT0.240.21–0.2664.26260.83 (29)088.8828.430
Overall0.250.22–0.28100844.41 (45)094.6728.820
SWINCCT0.440.34–0.5422.3962.59 (5)092.0113.270
No NCCT0.330.28–0.3977.61467.38 (20)095.7219.050
Overall0.360.31–0.41100555.50 (26)095.3222.700
Field Strength in Tesla (T)
T2*1.50.270.23–0.3168.57252.74 (21)091.7622.770
3T0.230.18–0.2831.43112.65 (8)092.9016.680
Overall0.250.22–0.29100460.20 (30)093.4826.040
SWI1.5T0.360.26–0.4735.48106.63 (7)093.4410.850
3T0.370.31–0.4364.52261.40 (13)095.0319.390
Overall0.370.32–0.42100370.43 (21)094.3323.040
Slice Thickness
OverallThin ≤ 2 mm0.400.32–0.4913.36139.05 (10)092.8114.100
Medium 2.1–4.9 mm0.230.18–0.28510.84 (3)0.0172.3315.560
Thick ≥ 5 mm0.250.22–0.2941.78545.62 (32)094.1425.020
Overall 0.280.25–0.31100809.02 (47)094.1929.720
Region
T2*Asia0.280.24–0.3359.14645.90 (26)095.9721.400
Europe0.210.19–0.2427.2541.79 (12)071.2925.630
North America0.180.14–0.229.419.97 (3)084.9716.170
Multinational0.150.12–0.184.21---18.060
Overall0.250.22–0.28100844.41 (45)094.6728.820
SWIAfrica0.260.22–0.306.4---19.450
Asia0.410.37–0.4668.91260.13 (19)092.7028.140
Europe0.270.18–0.3717.6369.37 (4)094.239.340
North America0.240.21–0.287.06---21.500
Overall0.360.32–0.41100559.19 (28)094.9925.440
Stroke Subtype
T2*Atherothrombotic0.250.12–0.3928.0346.29 (4)091.365.740
Lacunar0.390.25–0.5329.7335.05 (4)088.598.240
Cardioembolic0.240.14–0.356.5911.31 (4)0.0264.657.090
Undetermined0.270.20–0.3317.11---12.710
Overall0.290.23–0.36100119.90 (17)085.8214.310
SWIAtherothrombotic0.230.08–0.4227.31104.67 (4)096,184.190
Lacunar0.260.17–0.3726.6219.57 (4)079.568.450
Cardioembolic0.250.11–0.4326.3761.15 (4)093.464.960
Undetermined0.200.10–0.3219.711.40 (3)0.0173.695.600
Overall0.240.18–0.30100229.98 (18)092.1711.940
Cerebral Microbleed Location
T2*Deep0.330.20–0.4719.7660.02 (6)090.007.470
Infratentorial0.080.02–0.1913.6919.12 (4)079.083.150
Lobar0.370.29–0.4634.7893.38 (11)088.2213.210
Mixed0.460.36–0.5531.7684.44 (10)088.1614.150
Overall0.340.28–0.41100446 (34)092.3816.160
SWIDeep0.180.14–0.2123.0419.94 (8)0.0159.8716.520
Infratentorial0.120.07–0.1923.0486.75 (8)090.786.550
Lobar0.290.24–0.3428.2545.14 (10)077.8518.790
Mixed0.490.39–0.6025.68155.02 (9)094.1913.460
Overall0.270.21–0.331001021.49 (38)096.2814.860
Abbreviations: T2* = T2 Gradient Echo Imaging, SWI = Susceptibility Weighted Imaging, FLAIR = fluid attenuated inversion recovery, NCCT = non-contrast computed tomography, T = tesla.
Table 5. Meta-Analysis Results for Association of Cerebral Microbleeds with Prognostic Outcomes: Summary Effects and Heterogeneity.
Table 5. Meta-Analysis Results for Association of Cerebral Microbleeds with Prognostic Outcomes: Summary Effects and Heterogeneity.
OutcomeModalityEffect MeasureSummary EffectsHeterogeneity ⍺Heterogeneity Variance Estimates
DerSimonian and Laird Random-Effects Method (REDL)Tests of Overall EffectCochran’s QHI2 ≤ *p-Valueτ2 ≤ †
Odds Ratio (OR) (95% Confidence Interval)
Symptomatic intracranial hemorrhage (sICH)T2 Gradient Echo Imaging (T2*)OR2.13 [1.435; 3.160]p = 0.000, z = 3.75411.081.1827.8%0.1970.0949
Susceptibility Weighted Imaging (SWI)OR2.687 [0.722; 10.007]p = 0.141, z = 1.4746.861.5156.3%0.0760.972
BothOR2.916 [1.294; 6.574]p = 0.010, z = 2.5810.00---0
OverallOR2.216 [1.555; 3.159]p = 0.000, z = 4.40218.491.1929.7%0.1400.122
Hemorrhagic transformation (HT)T2*OR1.229 [0.820; 1.843]p = 0.319, z = 0.99732.951.7366.6%0.0010.282
SWIOR1.402 [0.910; 2.163]p = 0.125, z = 1.5358.641.2030.6%0.1950.0956
BothOR1.788 [1.033; 3.094]p = 0.038, z = 2.0760.700.840.0%0.4010
OverallOR1.332 [1.013; 1.750]p = 0.040, z = 2.0541.161.4753.5%0.0020.174
Modified Ranking Scale (mRS) 3–6 at 90 DaysT2*OR1.572 [1.282; 1.927]p = 0.000, z = 4.3466.061.1017.5%0.3000.0114
SWIOR1.727 [1.303; 2.289]p = 0.000, z = 3.7982.680.950.0%0.4440
BothOR1.579 [0.976; 2.555]p = 0.063, z = 1.8590.00---0
OverallOR1.606 [1.387; 1.858]p = 0.000, z = 6.3449.090.950.0%0.5240
Abbreviations: T2* = T2 Gradient Echo Imaging, SWI = Susceptibility Weighted Imaging, sICH = symptomatic intracranial hemorrhage, HT = hemorrhagic transformation, mRS = Modified Rankin Scale, OR = odds ratio, CI = confidence interval, REDL = DerSimonian and Laird random-effects method, Q = heterogeneity measures were calculated from data with 95% confidence intervals (95% CI), based on non-central χ2 (common effect) distribution for Cochran’s Q test, H = relative excess in Cochran’s Q over degrees of freedom, I2 = proportion of total variation in effect estimate due to between study heterogeneity (based on Cochran’s Q test), τ2 = between-study variance to test comparisons of heterogeneity among subgroups, * = values of I2 ≤ are percentages, α = heterogeneity measures were calculated from the data with 95% Cis based on gamma (random effects) distribution for Q, † = heterogeneity variance estimates (tau≤) were derived from the DerSimonian and Laird method.
Table 6. Summary of Diagnostic Performance and Heterogeneity from MIDAS Meta-Analysis.
Table 6. Summary of Diagnostic Performance and Heterogeneity from MIDAS Meta-Analysis.
OutcomeModalityParameterEstimate95% Confidence Interval (CI)
Symptomatic Intracranial Hemorrhage (sICH)Susceptibility Weighted Imaging (SWI)Sensitivity0.05[0.03; 0.08]
Specificity0.98[0.95; 0.99]
Positive Likelihood Ratio2.8[0.7; 11.2
Negative Likelihood Ratio0.97[0.93; 0.1.01]
Diagnostic Odds Ratio3[1; 12]
Pretest Probability of Disease0.04-
Area under ROC Curve (AUROC)0.11[0.08; 0.14]
Interstudy Variation in Sensitivity (ICC_SEN)0.01[0.00; 0.07]
Interstudy Variation in Specificity (ICC_SPE)0.17[0.00; 0.50]
Heterogeneity (Chi-square)2.333, degrees of freedom (df) = 2, p = 0.156
Inconsistency (I2)14[0; 100]
T2 Gradient Echo Imaging (T2*)Sensitivity0.09[0.07; 0.12]
Specificity0.96[0.93; 0.97]
Positive Likelihood Ratio2.1[1.4; 3.1]
Negative Likelihood Ratio0.95[0.93; 0.97]
Diagnostic Odds Ratio2[1; 3]
Pretest Probability of Disease0.16-
AUROC0.30[ 0.26; 0.34]
ICC_SEN0.02[0.00; 0.07]
ICC_SPE0.11[0.00; 0.22]
Heterogeneity (Chi-square)29.382, df = 2, p < 0.0001
I293[87; 99]
Hemorrhagic Transformation (HT)SWISensitivity0.34[0.15; 0.61]
Specificity0.75[0.62; 0.85]
Positive Likelihood Ratio1.4[1.0; 2.0]
Negative Likelihood Ratio0.87[0.69, 1.11]
Diagnostic Odds Ratio2[1, 3]
Pretest Probability of Disease0.23-
AUROC0.65[0.61; 0.69]
ICC_SEN0.37[0.03; 0.72]
ICC_SPE0.16[0.00; 0.37]
Heterogeneity (Chi-square)44.168, df = 2, p < 0.001-
I295[92; 99]
T2*Sensitivity0.21[0.12; 0.35]
Specificity0.82[0.69; 0.90]
Positive Likelihood Ratio1.2[0.8; 1.7]
Negative Likelihood Ratio0.96[0.88; 1.05]
Diagnostic Odds Ratio1[1; 2]
Pretest Probability of Disease0.21-
AUROC0.52[0.48; 0.56]
ICC_SEN0.30[0.10; 0.50]
ICC_SPE0.32[0.13; 0.52]
Heterogeneity (Chi-square)334.234, df = 2, p < 0.001-
I299[99; 100]
Modified Rankin Scale (mRS) 3-6 at 90 daysOverallSensitivity0.49[0.41; 0.58]
Specificity0.62[0.54; 0.69]
Positive Likelihood Ratio1.3[1.2; 1.4]
Negative Likelihood Ratio0.82[0.75; 0.89]
Diagnostic Odds Ratio2[1; 2]
Pretest Probability of Disease0.46-
AUROC0.58[0.54; 0.62]
ICC_SEN0.09[0.05; −0.12]
ICC_SPE0.08[0.05; 0.10]
Heterogeneity (Chi-square)170.018, df = 2, p < 0.0001-
I299[98; 99]
Abbreviations: sICH = symptomatic intracranial hemorrhage, HT = hemorrhagic transformation, mRS = Modified Rankin Scale, CI = confidence interval, AUROC = area under ROC curve, ICC_SEN = interstudy variation in sensitivity, ICC_SPE = interstudy variation in specificity, I2 = inconsistency, df = degrees of freedom, p = p-value, T2* = T2 Gradient Echo Imaging, SWI = Susceptibility Weighted Imaging.
Table 7. GRADE Summary of Findings: Cerebral Microbleeds (CMBs) in Acute Ischemic Stroke (AIS)—SPOT-CMB Study.
Table 7. GRADE Summary of Findings: Cerebral Microbleeds (CMBs) in Acute Ischemic Stroke (AIS)—SPOT-CMB Study.
OutcomeNo. of Studies (Participants)Study DesignRelative Effect (95% CI)Assumed Risk (control)Risk with CMBsAbsolute EffectCertainty of EvidenceReasons
Symptomatic intracerebral hemorrhage (sICH)14 (~6163)Observational (meta-analysis, random-effects)OR 2.22 (1.56–3.16)40 per 100088 per 100048 more per 1000⊕⊕◯◯ Low to Moderate−1 risk of bias (variable definitions), −1 imprecision (subgroup variability), +1 consistent association
Hemorrhagic transformation (HT)21 (~6049)Observational (meta-analysis, random-effects)OR 1.33 (1.01–1.75)150 per 1000190 per 100040 more per 1000⊕⊕◯◯ Low−1 risk of bias, −1 inconsistency (I2 = 53.5%), −1 indirectness (definitions variable)
Poor functional outcome (mRS 3–6 at 90 days)11 (~5499)Observational (meta-analysis, random-effects)OR 1.61 (1.39–1.86)350 per 1000470 per 1000120 more per 1000⊕⊕⊕◯ Moderate−1 risk of bias, +1 consistency (I2 = 0%)
CMB prevalence by imaging modality (SWI vs. T2*)80 (~28,383)Observational (meta-analysis)SWI 36% (95% CI: 31–41); T2* 25% (22–28)——11% higher detection with SWI⊕⊕◯◯ Low−1 inconsistency (high heterogeneity), −1 indirectness, +1 strong magnitude of effect
Diagnostic accuracy for sICH prediction14 (~6163)Observational (diagnostic meta-analysis)AUC 0.29; DOR 2–3——Poor sensitivity (<10%) but high specificity (>95%)⊕◯◯◯ Very low−1 risk of bias, −1 indirectness, −1 imprecision
GRADE Working Group grades of evidence. ⊕⊕⊕⊕ High: Very confident that the true effect lies close to the estimate. ⊕⊕⊕◯ Moderate: Moderately confident; true effect likely close but may differ. ⊕⊕◯◯ Low: Limited confidence; true effect may differ substantially. ⊕◯◯◯ Very low: Very little confidence; true effect likely substantially different. Abbreviations: AIS = acute ischemic stroke; CMBs = cerebral microbleeds; SPOT-CMB = Susceptibility-weighted imaging and Prognostic Outcomes in Acute Stroke—Cerebral Microbleeds study; OR = odds ratio; DOR = diagnostic odds ratio; AUC = area under the ROC curve; SWI = susceptibility-weighted imaging; T2* = T2*-weighted imaging.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Tan, R.; Spring, K.J.; Killingsworth, M.; Bhaskar, S. Comparative Diagnostic and Prognostic Performance of SWI and T2-Weighted MRI in Cerebral Microbleed Detection Following Acute Ischemic Stroke: A Meta-Analysis and SPOT-CMB Study. Medicina 2025, 61, 1566. https://doi.org/10.3390/medicina61091566

AMA Style

Tan R, Spring KJ, Killingsworth M, Bhaskar S. Comparative Diagnostic and Prognostic Performance of SWI and T2-Weighted MRI in Cerebral Microbleed Detection Following Acute Ischemic Stroke: A Meta-Analysis and SPOT-CMB Study. Medicina. 2025; 61(9):1566. https://doi.org/10.3390/medicina61091566

Chicago/Turabian Style

Tan, Rachel, Kevin J. Spring, Murray Killingsworth, and Sonu Bhaskar. 2025. "Comparative Diagnostic and Prognostic Performance of SWI and T2-Weighted MRI in Cerebral Microbleed Detection Following Acute Ischemic Stroke: A Meta-Analysis and SPOT-CMB Study" Medicina 61, no. 9: 1566. https://doi.org/10.3390/medicina61091566

APA Style

Tan, R., Spring, K. J., Killingsworth, M., & Bhaskar, S. (2025). Comparative Diagnostic and Prognostic Performance of SWI and T2-Weighted MRI in Cerebral Microbleed Detection Following Acute Ischemic Stroke: A Meta-Analysis and SPOT-CMB Study. Medicina, 61(9), 1566. https://doi.org/10.3390/medicina61091566

Article Metrics

Back to TopTop