Stem cell therapy for ischemic stroke: neuroimaging approaches and evidence from a systematic review

But why go thru all the trouble of stem cells if exosomes are the reason for the benefits? Which must be why no one seems to be monitoring stem cell survival.

Application of stem cell-derived exosomes in ischemic diseases: opportunity and limitations 

The latest here:

 Stem cell therapy for ischemic stroke: neuroimaging approaches and evidence from a systematic review

Bin Jiang^1*Moss Zhao²Elizabeth Tong¹Yongkai Liu¹Ates Fettahoglu¹Wen-Kai Weng³Michael E. Moseley¹Max Wintermark⁴Gary K. Steinberg²Greg Zaharchuk¹

• ¹Department of Radiology, School of Medicine, Stanford University, Stanford, CA, United States
• ²Department of Neurosurgery, School of Medicine, Stanford University, Stanford, CA, United States
• ³Department of Medicine - Med/Blood and Marrow Transplantation, School of Medicine, Stanford University, Stanford, CA, United States
• ⁴Department of Neuroradiology, MD Anderson Center, The University of Texas, Houston, TX, United States

Purpose: Cell-based therapy is a promising approach for ischemic stroke treatment. This systematic review and meta-analysis aimed to consolidate clinical evidence on the use of neuroimaging to evaluate stem cell therapy across all stages of stroke recovery.

Methods: A systematic search was conducted in 5 databases in July 2025. They were included if neuroimaging analysis, regardless of cell source, route administration or dosage were reported. The level of evidence and risk of bias were assessed using the ROB-2 or ROBINS-I tool. Imaging data from all included articles were extracted, and randomized-effect meta-analyses were performed when two or more outcomes were available for any reported imaging parameter.

Results: Thirty articles were included in the systematic review, of which four were eligible for meta-analysis. Meta-analysis of subacute stroke patients revealed no significant differences in infarct volume reduction at 3 months (SMD = −0.50; 95% CI: −1.15 to 0.51; p = 0.13; I² = 63%) or 1 year (SMD = −1.02; 95% CI: −3.63 to 1.60; p = 0.45; I² = 92%) between treatment and control group. Chronic stroke patients exhibited less overall volume loss. There was a trend toward improved white matter recovery and motor cortex activity, reflected in increased DTI and fMRI parameters. SPIO-labeled autologous stem cells recently proved safe in patients, with T2* imaging showing engraftment and migration.

Conclusion: Advanced neuroimaging offers a valuable non-invasive tool for assessing the effects of stem cell therapy in ischemic stroke. However, substantial heterogeneity in imaging protocols and reporting limits cross-study comparisons. Standardization of neuroimaging methodology is essential to advance future research and clinical translation.

Introduction

Stroke remains one of the leading causes of death and long-term disability worldwide. Among its subtypes, ischemic stroke accounts for the vast majority of cases, arising from obstruction of cerebral blood flow and subsequent brain tissue damage. The growing global burden is striking: from 1990 to 2019, the incidence of ischemic stroke increased by 70% and prevalence by over 100%, with parallel rises in mortality and disability (1). These statistics underscore the urgent need for more effective therapeutic strategies.

Currently, the most effective treatments for ischemic stroke are reperfusion-based interventions, including intravenous thrombolysis and endovascular therapy (EVT). While these approaches can restore blood flow and improve outcomes, their clinical application is limited by narrow therapeutic windows, strict eligibility criteria, and incomplete functional recovery in many patients (2). As a result, a substantial proportion of stroke survivors are left with persistent neurological deficits, highlighting the unmet need for restorative therapies that extend beyond the acute treatment window.

Stem cell therapy has emerged as a promising approach in this context. Building on the 2012 Nobel Prize–winning discovery that somatic cells can be reprogrammed into pluripotent stem cells, preclinical and clinical studies have explored the potential of stem cells to enhance brain repair after ischemic stroke. These studies suggest that stem cells may exert beneficial effects through multiple mechanisms, including promoting neurogenesis, angiogenesis, and synaptic plasticity, as well as modulating neuroinflammation and immune response (3–5). Clinical trials to date have demonstrated safety and feasibility, with early signals of efficacy across different stem cell types, transplantation routes, and stages after stroke onset (6–9).

Despite these advances, critical challenges remain in optimizing stem cell therapy for stroke, particularly in objectively assessing therapeutic effects in the human brain. Neuroimaging offers tools to address this need, providing noninvasive biomarkers of structural and functional brain changes. Techniques such as diffusion tensor imaging (DTI), functional MRI (fMRI), perfusion imaging, and positron emission tomography (PET) can reveal the mechanisms of action, monitor brain repair, and potentially predict clinical outcomes. While prior reviews have summarized stem cell mechanisms, delivery methods, and safety, comparatively little attention has been given to the role of neuroimaging in this field (2, 10–14).

The aim of this systematic review is therefore to synthesize current evidence on structural and functional neuroimaging of stem cell therapy in ischemic stroke. By focusing on imaging biomarkers, we highlight how neuroimaging contributes to understanding treatment mechanisms, evaluating efficacy, and guiding the future development of regenerative therapies for stroke patients.

Methods

Ethics approval by an Institutional Review Board (IRB) was not required for this systematic review.

Protocol and registration

The search strategy and written methodology were developed with the assistance of the medical librarians, following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) (15), the Peer-Review of Electronic Search Strategies (PRESS) (16), the National Academies (IOM) Standards for Systematic Reviews (17), and the Cochrane guidelines (18).

Inclusion and exclusion criteria

Studies were included if they were published in a peer-reviewed journal, evaluated ischemic stroke in an adult population, administered a specific type of stem cells or stimulating factors for stroke, and included results based on one or more evaluations of functional imaging parameters, including MRI or PET (Table 1). Exclusion criteria were studies that were not full research articles, animal studies, dissertations, or review articles. Additionally, studies focusing on populations with hemorrhagic stroke or strokes from sickle cell or moyamoya diseases were excluded. This selection was chosen to enhance the focus on imaging related to arterial ischemic infarcts.

Table 1

Table 1. Inclusion and exclusion criteria for studies identified in the search (PICOS).

Information sources and search strategy

MEDLINE, EMBASE, Scopus, Cochrane Central Register of Controlled Trials and Web of Science were searched by a medical librarian (EW). To identify any missing reports, we also hand-searched the references lists from included studies and identified articles meeting the inclusion criteria, contacting authors and experts, and examining related articles in PubMed and Google Scholar. An updated search was conducted in July 2025. Complete, reproducible search strategies for each database are provided in the Supplementary materials.

Study selection and data extraction

All titles, abstracts, and full manuscripts underwent review by two unique individuals within the authorship team, utilizing Covidence systematic review software. In cases of discrepancies, two authors (BJ, MZ) conducted a joint review to reach a final decision regarding study inclusion. To enhance transparency, the predefined PICOS-based inclusion and exclusion criteria (Table 1) were applied consistently throughout title/abstract and full-text review. After full-text screening, we collected the data using a data extraction form. Data collection included: stage of ischemic stroke (acute [<7 days], subacute [1 ~ 12 weeks] or chronic [>3 months]), route, dosage, and type of stem cells introduced, baseline NIHSS, the number of subjects in the intervention and the control group (if present), infarct size and how it was reported (volume, percentage, and etc.), DTI parameters (FA, MD, and etc.), functional MRI, PET changes, and other imaging parameters (such as SPIO labeled T2*).

Risk of bias assessment and level of evidence

Two authors (BJ, YL) evaluated the risk of bias using the ROB-2 (Cochrane risk-of-bias tool) for the randomized control studies (19), and ROBINS-I (Risk of Bias in Non-Randomized Studies of Interventions) tool for non- randomized control studies (20). The following domains were assessed: confounding, selection of participants, classification of interventions, deviations from intended interventions, missing data, measurement of outcomes, selection of reported results, and overall risk of bias. The risk of bias was then classified as high, moderate, or low according to the ROB-2 and ROBINS-I tool. The corresponding author was contacted to retrieve missing data. To grade the overall strength of evidence from the included studies, the level of evidence of each included study was assessed by The Oxford Centre for Evidence-Based Medicine Levels of Evidence and scored accordingly (CEBM) (21).

Reporting bias assessment

If at least 10 studies were available for an outcome, we planned to assess reporting bias using funnel plots.

Statistical analysis

We employed a random-effects meta-analysis to combine the study results. Meta-analyses were conducted using Review Manager version 5.4 software (RevMan 2020). For continuous variables, we computed the standard mean difference (SMD) and 95% confidence interval (CI). In cases where numerical outcome data were not directly reported in the original publications, graphical data were extracted. We obtained these values either by contacting the corresponding authors or, when necessary, by digitizing published figures using Web Plot Digitizer¹ (22). This ensured inclusion of studies that otherwise lacked extractable numerical data. Heterogeneity was quantified with the I² statistic, with values of 25, 50, and 75% considered low, moderate, and high, respectively. A two-tailed p value < 0.05 was considered statistically significant.

Results

Search results

A total of 3,838 articles from 5 databases were initially identified. After removing duplicates, 2,685 articles remained. Full-text review was completed on 106 studies, and 30 studies were included in this review (Figure 1). Studies were excluded for the following reasons: (1) not peer-reviewed (n = 45); (2) no imaging to outcome analysis or insufficient imaging data (n = 15); (3) ineligible population (n = 11); (4) non-English (n = 3); (5) wrong intervention (n = 1); (6) wrong indication (n = 1).

Figure 1

Figure 1. PRISMA flowchart illustrating the process of selecting studies included.

Characteristics and quality of evidence of included studies

Of the 30 included studies, 22 were clinical trials, with 8 studies involving sub-analyses of the same study. These studies were conducted in various regions: 4 in North America, 5 in Europe, 12 in Asia, and 1 in Africa (Table 2). All studies incorporated MRI or PET. Among them, 16 studies compared infarct sizes between baseline and the latest available follow-up based on T1-weighted images, 6 included the identification of abnormal high signals on T2 FLAIR images immediately after treatment, 9 included analysis of DTI images, 4 involved functional MRI analysis, 2 studies analyzed perfusion MRI changes, and 5 included PET analysis.