Abstract
Introduction
Collateral circulation is a critical determinant of outcomes in acute ischemic stroke (AIS). This bibliometric analysis evaluates research trends and future directions on this topic.
Methods
A systematic search was conducted in the Web of Science Core Collection to identify publications on AIS and collateral circulation between 1981 and 2025. Bibliometric analysis and visualization were performed using VOSviewer, CiteSpace, and the R package “bibliometric.”
Results
The analysis included 482 publications. China was the most productive countries, leading in publication volume (642). The University of California System emerged as top contributing institutions. Stroke was the most influential journals in terms of H-index (33), TP (53), and TC (3796). David S. Liebeskind and Gregory W. Albers were identified as leading authors. Keyword co-occurrence and cluster analysis revealed four main thematic areas: imaging and diagnostic assessment, treatment strategies, prognosis and outcome prediction, and pathophysiological mechanisms and hemodynamics. In addition, burst keyword analysis showed that terms like “large vessel occlusion”, “score”, “management”, and “endovascular thrombectomy” remained active through 2025.
Conclusion
Research on collateral circulation in AIS has evolved from exploring basic mechanisms to focusing on clinical applications. The current landscape integrates imaging-based collateral assessment with treatment decisions and prognosis to guide precise endovascular therapy. Future research will likely concentrate on collateral-guided precision treatment, imaging scores, and strategies for extending the therapeutic window. This study provides a comprehensive overview and valuable insights for researchers.
Introduction
Acute ischemic stroke (AIS) is a leading cause of death and disability worldwide, presenting substantial challenges to global public health. It typically results from the interruption of cerebral blood flow, leading to ischemia and infarction, which damage neurons in the affected brain regions, thereby exacerbating both mortality and disability risks. 1 The Global Burden of Disease study reports that in 2010, approximately 11.6 million new cases of stroke occurred globally. 2 More recent data from 2020 further emphasize the increasing incidence and burden of stroke, with low-income countries accounting for 77% of stroke-related deaths and 89% of stroke-related disability-adjusted life years.3,4 Projections suggest that by 2030, global mortality due to ischemic stroke will rise from approximately 3.3 million in 2019 to 4.9 million. 5
Collateral circulation plays a critical role in influencing the outcomes of AIS by providing alternative blood flow pathways to ischemic brain tissue. When major cerebral arteries are occluded, the presence of strong collateral circulation can help sustain the perfusion of ischemic areas, reducing neuronal injury and improving patient prognosis. 6 Previous studies have demonstrated that patients with better collateral circulation tend to experience smaller infarct volumes and more favorable clinical outcomes. 7 This protective mechanism is especially crucial in AIS, as it helps mitigate the severity of ischemia and delays the onset of irreversible damage. 8 In some instances, effective collateral circulation can even compensate for the loss of a primary artery, enhancing both functional recovery and survival rates.9,10 Despite its recognized importance, the relationship between collateral circulation and AIS remains inadequately understood, with limited research exploring the underlying mechanisms and its full impact on stroke outcomes.
Bibliometrics is a quantitative approach that systematically analyzes literature within a specific research domain, providing valuable insights into its development and trends.11,12 However, no bibliometric studies focusing on collateral circulation and AIS have been published to date, with prior research primarily concentrating on the clinical applications and imaging evaluation of collateral circulation.12,13 This study seeks to fill this critical gap by conducting an in-depth bibliometric analysis, providing a thorough and insightful overview of the evolving landscape of research on collateral circulation and AIS.
Materials and Methods
Search Strategies and Data Collection
A literature search was conducted in the Web of Science Core Collection (WoSCC), specifically using the Science Citation Index Expanded (SCIE) and the Social Sciences Citation Index (SSCI). 14 The search strategy was as follows: (TS = (“Bypass circulation” OR “collateral circulation” OR “Arterial collaterogenesis” OR “Collateral blood flow”)) AND TS = (“Acute Cerebral Infarction” OR “Acute Ischemic Stroke”).15,16 Publications from January 1, 1981, to July 10, 2025, were included in this study. Only articles published in English were considered. Data was collected in text format, with studies screened to exclude literature irrelevant to this research. Two investigators independently performed the initial screening phase by examining the titles and abstracts of all identified articles. Publications considered relevant to the research topic were included for subsequent full-text evaluation. Disagreements regarding inclusion were resolved by consensus between the researchers. The dataset included publication and citation counts, article titles, author information, institutional affiliations, countries/regions, journals, and keywords, all of which were used for subsequent bibliometric analysis.
Statistical Analysis
The bibliometric analysis was conducted using Microsoft Excel (Version 2402), VOSviewer (version 1.6.20), CiteSpace (version 6.3.R1), and the R package “bibliometric” (version 4.3.3).
VOSviewer is a powerful tool for mapping institutional collaborations, co-authorship, citations, and co-citations. It also facilitates co-occurrence analysis to explore associations among keywords. 17 In its visualizations, node size represents the number of publications, line thickness indicates link strength, and nodes of the same color denote clusters of related items. Link strength is determined by the number of connections an item has with others, such as the degree of collaboration between authors. 17
To identify emerging trends and research hotspots, CiteSpace was used to analyze burst keywords, with time slicing set from January 1994 to July 2025. The time slice was set to 1 year, and the node type was specified as keywords. For keyword analysis, the threshold was set to the top 5 items in each slice, and pruning was performed using the pathfinder algorithm combined with merged networks. Based on these settings, a visual analysis was conducted to generate a keyword timeline map for the research on collateral circulation in AIS.
Additionally, the R package “bibliometric” is a crucial R tool for comprehensive bibliometric analysis, developed in the R programming language. 18 It supports the entire process of data import, transformation, analysis, and scientific visualization, making it well-suited to meet the needs of bibliometric research.
To assess journal influence, Journal Citation Reports (JCR) metrics were utilized, including the impact factor (IF) for 2024 and quartile ranking (Q1-Q4) for 2023. To quantify the academic impact of individuals and journals, we employed the H-index, M-index, and G-index. The H-index, the most widely used bibliometric metric, is defined as the number of papers an author has published that have each been cited at least h times.19,20 The G-index, an improvement on the H-index, measures the global citation performance of a set of articles, 21 while the M-index represents the average number of citations received by the papers contributing to the H-index. 22 In this study, the H-index, M-index, and G-index for each author were retrieved from WoSCC.
Results
An Overview of Publications
A total of 482 articles were included in the final analysis (Figure 1). These publications involved 3262 authors from 841 institutions across 42 countries/regions, and appeared in 144 journals, collectively citing 8509 references. Publication output on collateral circulation in AIS increased steadily from 2000 to 2025, peaking in 2022 with 51 articles (Figure 2).
Flowchart of the literature screening process.
Growth trend of the publications worldwide.
Country Insights and Trends
The top 10 most productive countries accounted for 95.6% (461/482) of all publications. China led in total publications (TP = 626), followed by the United States (TP = 424) and Italy (209). In terms of total citations (TC), Canada ranked first (TC = 5279), ahead of the United States (TC = 4933) and China (1763) (Table 1
Analysis of countries.
Publication and Citation Profiles of Leading Countries.
Note(s): Articles: Publications of Corresponding Authors only. Freq: Frequence of Total Publications. MCP ratio: Proportion of Multiple Country Publications. TP: Total Publications. TP rank: Rank of Total Publications. TC: Total Citations. TC rank: Rank of Total Citations. Average Citations: The average number of citations per publication.
Institutions Insights and Trends
As shown in Figure 4a, the University of California System led in publication output with 54 articles, followed by Stanford University (48) and Institut National de la Santé et de la Recherche Médicale (INSERM; 46). Among the 101 institutions engaged in international collaborations (with at least three collaborative publications), the University of California, Los Angeles exhibited the highest collaboration link strength (32), followed by Stanford University and Capital Medical University (both 31) (Figure 4b).
Analysis of institutions.
Journals Insights and Trends
The top 10 journals ranked by H-index in the field of collateral circulation in AIS are summarized in Table 2. Stroke ranked as the most influential journal, leading in H-index (33), number of publications (TP = 53), and total citations (TC = 3796), followed by the Journal of Cerebral Blood Flow and Metabolism (H-index = 13). Frontiers in Neurology ranked second in publications (TP = 36), while the American Journal of Neuroradiology ranked second in total citations (TC = 1001). Most of the top journals were in JCR Q1.
Bibliometric Indicators of High-Impact Journals.
Note(s): IF: Impact Factor in 2024, indicating the average number of citations to recent articles published in the journal. JCR: The quartile ranking of the journal in the Journal Citation Reports in 2024, indicating the journal's ranking relative to others in the same field (Q1: top 25%, Q2: 25%-50%, Q3: 50%-75%, Q4: bottom 25%). TP: Total Publications. TP rank: Rank of Total Publications. TC: Total Citations. TC rank: Rank of Total Citations. Average Citations: The average number of citations per publication. PY start: Publication Year Start, indicating the year the journal started publication.
In the journal co-occurrence analysis (Figure 5a), Stroke had the highest link strength (467), followed by the Journal of Cerebral Blood Flow and Metabolism (122) and Frontiers in Neurology (102). In the journal coupling analysis (Figure 5b), Stroke also led with the highest coupling strength (16,835), followed by Frontiers in Neurology (10,558) and the Journal of Stroke & Cerebrovascular Diseases (6872).
Analysis of journals.
Authors Insights and Trends
The top 10 most influential authors are listed in Table 3. Based on H-index, David S. Liebeskind (H-index=18, TP = 27) was the most influential, followed by Max Wintermark (H-index=14, TP = 19) and Gregory W. Albers (H-index=12, TP = 17). A. Shuaib stood out with the highest total citations (TC = 5066) despite a moderate H-index of 8. Co-authorship analysis revealed nine major research clusters (Figure 6). Collaboration network analysis among 133 authors with international publications showed Gregory W. Albers had the strongest collaborative links (123), followed by Jeremy J. Heit (121) and Maarten G. Landsberg (110).
Visualization map depicting the collaboration among different authors.
Publication and Citation Profiles of High-Impact Authors.
Note(s): H-index: The h-index of the journal, which measures both the productivity and citation impact of the publications. G-index: The g-index of the journal, which gives more weight to highly-cited articles. M-index: The m-index of the journal, which is the h-index divided by the number of years since the first published paper. TP: Total Publications. TP rank: Rank of Total Publications. TC: Total Citations. TC rank: Rank of Total Citations. Average Citations: The average number of citations per publication. PY start: Publication Year Start, indicating the year the journal started publication.
Analysis of the Keywords
Keyword co-Occurrence Network Analysis
Four keyword clusters were identified based on co-occurrence analysis (Figure 7a and Table 4). Cluster 1 contained terms primarily linked to therapeutic processes and clinical decision-making, including “management”, “guidelines”, “interventional management”, “mechanical thrombectomy”, “recanalization”, “reperfusion”, “revascularization”, “safety”, “selection”, and “trial”. Cluster 2 was characterized by terminology associated with imaging modalities and pathological features, such as “acute ischemic-stroke”, “angiography”, “collateral circulation”, “ct”, “diffusion”, “mri”, “perfusion”, “penumbra”, “infarction”, “occlusion”, “stenosis”, and “thrombolysis”. Cluster 3 comprised keywords referring to specific imaging techniques and revascularization approaches, including “computed-tomography”, “ct-angiography”, “endovascular therapy”, “intraarterial thrombolysis”, “intravenous thrombolysis”, “middle cerebral-artery”, “thrombolytic therapy”, and “tissue-plasminogen activator”. Cluster 4 encompassed terms related to collateral vessel status and vascular characteristics, such as “circulation”, “ct angiography”, “leptomeningeal collaterals”, “severity”, “vessels”, “predictor”, and “outcomes”.
Keyword Clustering Analysis.
Analysis of Burst Keywords
As shown in Figure 7b, the strength of the top 20 keywords with the strongest bursts ranged from 3.42 to 6.41. The term “large vessel occlusion” had the highest burst strength (6.41). Notably, since 2022 the keywords “association”, “management”, “risk”, and “endovascular thrombectomy” have been more prominently concentrated. Furthermore, the keywords “large vessel occlusion”, “score”, “management”, and “endovascular thrombectomy” continued to exhibit bursts through 2025.
Discussion
General Information
This bibliometric analysis indicated a continued growth in research on collateral circulation in AIS. In terms of publication volume and international collaboration, the United States and China are the leading countries. The University of California system and Stanford University are recognized as top institutions in this field, while Stroke and Frontiers in Neurology were identified as the most influential journals. David S. Liebeskind and Gregory W. Albers are regarded as the most prominent authors in terms of publication volume and citation impact.
The United States and China are the leading contributors to global research on collateral circulation in AIS. This is likely attributable to the substantial national gross domestic product in both countries, which provides ample resources for clinical studies, as well as a growing focus on holistic treatment strategies, particularly in conservative management for teenage patients. 23 Additionally, the “Healthy China 2030” initiative has significantly facilitated research advancements in this area. 24 Leading institutions such as Capital Medical University have expanded their research capacities, further bolstering position of China in this field. Notably, collaborations with Western research centers, particularly in stroke-related fields, have enhanced China's visibility in high-impact journals, amplifying its citation influence. In the United States, funding from the National Institutes of Health has supported prominent institutions like Stanford University and the University of California, enabling them to increase both the output and citation impact of their research by leveraging advanced infrastructure and extensive collaborative networks. 25
Research Hotspots and Trends
Our study provides a clear overview of the development process in research on collateral circulation in AIS and identifies its latest hotspots through keyword cluster analysis and burst keyword analysis.
Cluster 1: Reperfusion Treatment and Clinical Management
This cluster primarily reflects how clinical decision-making in reperfusion workflows increasingly depends on imaging-based characterization of collateral circulation. CTA and CTP have become essential for defining perfusion mismatch, which represents the difference between salvageable penumbra and infarct core. Foundational studies established that robust collaterals are linked to smaller core volumes and higher mismatch ratios, 26 supporting earlier treatment activation and triage. Subsequent comparisons of imaging modalities demonstrated that dynamic CTA (dCTA) achieves higher concordance with CTP than optimized multiphase CTA (omCTA) when selecting patients for reperfusion therapy. 27 More recent evidence further showed that the circle of Willis and leptomeningeal collaterals act synergistically to augment perfusion; patients with favorable combined collateral profiles exhibit smaller infarcts and larger mismatch ratios. 28 Importantly, CTP-derived mismatch remains an independent predictor of clinical outcomes across a spectrum of collateral grades, 29 integrating functional perfusion information into therapeutic planning. Overall, the work grouped in this cluster illustrates how collateral-related imaging markers directly inform time-sensitive management pathways in AIS, linking diagnostic evaluation with practical clinical workflow optimization.
Cluster 2: Imaging Evaluation and Pathological Features
Studies in this cluster examine how ischemic pathology and imaging-defined collateral characteristics shape treatment responses, particularly in endovascular therapy. Mechanical thrombectomy has become the standard of care for large-vessel occlusion, with stent retrievers achieving reperfusion rates above 80%. 30 However, these high technical success rates do not guarantee good outcomes, prompting deeper evaluation of pathological and collateral determinants. A 2025 review emphasized that inadequate collateral flow or prolonged hypoperfusion may lead to poor outcomes despite recanalization, a pattern termed “futile recanalization”. 31 Additional investigations have identified specific pathological factors influencing thrombectomy performance: erythrocyte-rich thrombi are more easily extracted, 32 and procedural efficiency is crucial, as most successful recanalizations occur within two passes. 33 Collectively, this cluster underscores how imaging and pathological assessment of collaterals, clot composition, and perfusion status guide the refinement of treatment strategies and shape expectations regarding therapeutic efficacy.
Cluster 3: Imaging-Guided Revascularization Strategies
This cluster concentrates on how imaging-derived collateral metrics intersect with revascularization approaches to inform prognostic evaluation. Clinical outcome scales such as the modified Rankin Scale (mRS) and NIH Stroke Scale (NIHSS) are increasingly complemented by imaging biomarkers. The regional leptomeningeal collateral (rLMC) score derived from CTA shows strong associations with mRS outcomes, with higher collateral grades corresponding to smaller infarcts and improved functional recovery. 34 Recent data further indicate that the circle of Willis and leptomeningeal pathways each contribute independently to outcome variation, while their combined effect yields the most favorable mismatch ratios and 3-month mRS scores. 28 Emerging CTP markers, such as collateral time, also correlate with infarct growth and functional outcomes, 35 extending prognostic insight beyond static anatomical scoring. In summary, this cluster highlights the growing emphasis on integrating imaging-defined collateral characteristics with revascularization strategies to refine individualized outcome prediction.
Cluster 4: Collateral Pathways and Prognostic Vascular Features
The final cluster addresses the anatomical, physiological, and hemodynamic mechanisms that underpin collateral circulation, forming the basis for interpreting imaging patterns and clinical outcomes. Computational modeling studies have shown that cerebral autoregulation can sustain near-normal perfusion even in the presence of anatomical variations, 36 whereas impaired autoregulation may trigger maladaptive redistribution of flow. 37 Anatomical variants, such as a fetal-type posterior cerebral artery, may constrain collateral compensation between anterior and posterior circulations, thereby increasing mismatch risk. 38 At the vascular level, arteriogenesis—driven by endothelial and immune-cell signaling—plays a critical role in adaptive collateral remodeling during ischemia. 39 Together, studies in this cluster emphasize that collateral function arises from the interaction of vascular structure, dynamic hemodynamics, and remodeling processes. These mechanistic insights provide a biological and physiological substrate that helps interpret imaging phenotypes, understand variability in revascularization benefit, and contextualize outcome-related findings in AIS.
Burst Keyword Analysis
Burst keyword analysis further revealed the continued prominence of terms such as “large vessel occlusion”, “score”, “management”, and “endovascular thrombectomy” into 2025.
The term “large vessel occlusion” (LVO) shows a strong burst trend, highlighting its growing significance in AIS research. LVO significantly influences stroke severity and prognosis, largely mediated by cerebral collateral circulation. Nevertheless, specific clinical or demographic determinants of outcomes in LVO remain poorly defined and warrant further investigation. 40 In recent years, interventions targeting LVO—especially mechanical thrombectomy—have become a major therapeutic focus. 41 There is also increasing emphasis on the use of advanced imaging to characterize LVO and tailor treatments according to occlusion subtype. Evidence suggests that LVO patients derive substantial benefit from mechanical thrombectomy, with better collateral circulation associated with higher rates of successful recanalization and improved reperfusion outcomes. 42
The term “score” also displays research momentum. Scoring systems integrating collateral grading, infarct volume, and clinical severity are increasingly used to guide thrombectomy decisions. For example, collateral status derived from CT angiography has been associated with better recanalization outcomes and reduced infarct growth, though its role in trial selection criteria remains limited. 43 The integration of collateral scores into existing prognostic models has improved outcome prediction, especially in populations with borderline clinical indications.
Since 2022, research on the intersection of “management” has grown rapidly, marking a clear keyword burst trend in bibliometric analysis. Recent studies have shown that robust collateral circulation not only extends the effective time window for reperfusion therapies but also significantly improves clinical outcomes by reducing mortality and disability rates. 9 Correspondingly, management strategies have increasingly shifted toward individualized assessment based on imaging. AI-assisted automated tools, such as e-CTA, are now employed to rapidly identify favorable collateral status, predict functional outcomes, and provide quantifiable metrics to guide clinical decision-making. 44 From a management perspective, future directions include: (1) integrating multimodal imaging to optimize individualized reperfusion treatment strategies 45 ; (2) incorporating collateral status as a key prognostic and treatment-response indicator in clinical workflows. 46
“Endovascular thrombectomy” remains central to AIS treatment research, especially for LVO. Studies have expanded beyond standard indications, exploring novel access strategies and complex anatomical scenarios. For instance, Cai et al demonstrated the feasibility of performing mechanical thrombectomy through chronically occluded proximal arteries when collateral circulation was well developed, achieving successful recanalization and favorable outcomes. 47 Despite successful recanalization, a substantial number of patients fail to achieve favorable long-term outcomes, a phenomenon known as futile reperfusion. Recent reviews further emphasize that despite improved technology, up to half of LVO patients do not achieve functional independence post-thrombectomy, often due to ischemic core expansion or inadequate collateral supply prior to recanalization. 42 In this context, augmenting collateral flow or developing neuroprotective strategies is an emerging area of investigation. Furthermore, the high rate of futile reperfusion suggests that current thrombectomy practices may include patients with minimal salvageable brain tissue, who are unlikely to benefit from recanalization. In this context, collateral status has emerged as a direct predictor of this outcome, with poor collaterals strongly linked to futile reperfusion. 48 However, clinical guidelines currently lack specific criteria for integrating collateral status into thrombectomy decisions, creating uncertainty. 10 Therefore, a key future challenge is to establish definitive thresholds for collateral assessment. Defining such criteria would help identify patients ineligible for thrombectomy, thereby reducing the rate of futile procedures.
Implications for Future Research
Future research in AIS is likely to move toward more individualized, collateral-guided treatment strategies. There will be greater emphasis on refining patient selection using advanced imaging-based scoring systems that integrate collateral quality, infarct size, and timing. Expanded treatment windows-especially in patients with favorable collateral flow-will likely become standard. Management strategies will increasingly focus on integrating collateral circulation into prognostic assessment and treatment planning. Novel interventional techniques and adjunct therapies aimed at enhancing or preserving collateral circulation may emerge. Overall, the field is expected to evolve toward precision endovascular therapy driven by real-time physiological and imaging data.
Limitations
The present bibliometric analysis is not devoid of limitations. Firstly, the study's focus on English-language articles may exclude relevant research published in other languages, thereby limiting the comprehensiveness of the dataset. This linguistic bias could potentially lead to an incomplete picture of the global research landscape. Furthermore, the study's reliance on bibliometric data alone may overlook other important aspects of research, such as its practical applications, patient outcomes, and policy implications. A more holistic approach that integrates qualitative and quantitative data could provide a more comprehensive understanding of the field's impact and relevance.
Conclusion
This study provided a comprehensive overview of collateral circulation and AIS research. This study identified four main clusters: imaging assessment, treatment strategies, prognosis prediction, and pathophysiological mechanisms of collateral circulation in AIS. Future AIS research and treatment are expected to center on individualized, collateral-informed strategies that expand therapeutic windows, optimize patient selection, and support precision endovascular care.
Footnotes
Ethics Approval and Consent to Participate
Not applicable.
Consent for Publication
Not applicable.
Author Contributions
Yangmei Chen carried out the studies, participated in collecting data, and drafted the manuscript. Pian Wang performed the statistical analysis and participated in its design. Pian Wang and Yangmei Chen participated in acquisition, analysis, or interpretation of data and draft the manuscript. All authors read and approved the final manuscript.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability and Materials
All data generated or analysed during this study are included in this published article.
