Abstract

Introduction:

Vascular malformations are a group of congenital vascular developmental anomalies. According to the International Society for the Study of Vascular Anomalies classification, they are mainly divided into Slow-Flow and Fast-Flow lesions. Slow-Flow lesions are represented by venous malformations, lymphatic malformations, and capillary malformations; Fast-Flow lesions primarily include arteriovenous malformations and arteriovenous fistulae. As a minimally invasive interventional approach, sclerotherapy has become a first-line treatment for Slow-Flow vascular malformations; significant advances in techniques, agents, and combination therapies have been made in recent years. However, a systematic bibliometric analysis of overall research trends, knowledge structure, and frontier hotspots in this field is lacking.

Methods:

Using the Web of Science Core Collection and PubMed databases, we retrieved literature on sclerotherapy for vascular malformations from 2005 to 2025, including 2,040 and 1,855 articles, respectively. VOSviewer, CiteSpace and Bibliometrix were employed to analyze publication trends, collaboration networks, journal distribution, highly cited literature, keyword clustering, and burst detection. Cross-validation across the two databases was performed to enhance robustness.

Results:

Annual publications accelerated from 2015 and showed an explosive increase after 2020. China, the United States, and Japan were the top three publishing countries, with 535, 451, and 106 publications, respectively; the United States led in total citations. Shanghai Jiao Tong University, Shandong University, and Sungkyunkwan University were the most active institutions. Journal distribution followed Bradford’s law, with Zone 1 comprising 30 journals. Three-field (three-map) analysis revealed that core authors such as Lee BB and Do YS formed dense networks with key terms and journals. LDA topic modeling identified 15 topics; evolutionary analysis indicated a field shift from traditional sclerosants toward targeted therapies, electrochemical treatments, and artificial intelligence. Burst-term analysis confirmed sirolimus and intralesional bleomycin as sustained frontiers. PubMed and WoS results were highly consistent.

Discussion:

This study provides the first systematic 20-year knowledge map of the field of sclerotherapy for vascular malformations. China is the most productive country, though average per-article impact remains to be improved. The field is transitioning toward targeted therapies, physicochemical synergistic approaches, and multidisciplinary comprehensive management. Future work should strengthen combined targeted-drug and sclerotherapy approaches, promote AI-assisted precision treatment, and conduct high-quality prospective clinical studies.

1 Introduction

Vascular malformations are congenital anomalies of blood or lymphatic vessels. According to the International Society for the Study of Vascular Anomalies (ISSVA) classification, they are divided into Slow−Flow and Fast−Flow malformations. Slow−Flow malformations primarily include venous malformations, lymphatic malformations, capillary malformations, and mixed malformations composed of these types; fast−flow malformations mainly comprise arteriovenous malformations and arteriovenous fistulas (Wassef et al., 2015; International Society for the Study of Vascular Anomalies, 2025). Unlike infantile hemangiomas, vascular malformations do not involute spontaneously; they typically grow proportionally with body development and may exhibit rapid progression or symptom exacerbation during puberty, pregnancy, or after trauma, persisting lifelong (Mulliken and Glowacki, 1982). Clinical manifestations vary with subtype, location, and lesion size: mild cases may be asymptomatic, whereas severe cases can cause pain, functional impairment, or be life−threatening (Liu et al., 2025). Conventional treatments include surgical excision, sclerotherapy, laser therapy, and symptomatic pharmacotherapy (Raja et al., 2024). However, extensive and complex lesions remain difficult to cure. Surgery is applicable only to focal lesions or for debulking; for diffuse lesions or those in anatomically challenging regions, complete resection is often impossible and recurrence rates are high (Markovic et al., 2020). Sclerotherapy, as a minimally invasive and repeatable interventional approach, has become the first−line treatment for Slow−Flow vascular malformations (particularly venous malformations and macro−cystic lymphatic malformations) (Leal et al., 2023). Recent advances in sclerotherapy include development of composite sclerosant formulations that enhance safety while maintaining efficacy (Liu et al., 2024). Nevertheless, for diffuse or multifocal disease the primary goals of sclerotherapy are symptom control, functional improvement, and enhanced quality of life rather than radiologic cure; thus recurrence and multiple treatment sessions are common, often requiring multidisciplinary management combining surgery or targeted medications (Shiraishi et al., 2026).

Bibliometrics applies mathematical and statistical methods to quantitatively and qualitatively analyze scientific literature, objectively revealing a field’s knowledge structure, developmental trajectory, and research frontiers (Hassan and Duarte, 2024). Although bibliometric analyses have been conducted on venous/lymphatic malformations overall, systematic studies specifically focused on sclerotherapy—the core therapeutic modality—are lacking. This study retrieved literature on sclerotherapy for vascular malformations from the Web of Science Core Collection and PubMed for 2005–2025 and performed bibliometric and visualization analyses using VOSviewer, CiteSpace and Bibliometrix, aiming to systematically map publication trends, key contributors, knowledge bases, research hotspots, and emerging directions to provide evidence for clinical practice and future research.

2 Methodology2.1 Data sources and search strategies

This study retrieves literature from two independent databases: the Web of Science Core Collection (WoSCC) and PubMed. The Science Citation Index Expanded (SCI-Expanded) within WoSCC serves as the primary data source for bibliometric mapping and visual analysis, as it covers high-quality journals and its export format is compatible with VOSviewer and CiteSpace. PubMed is used as a supplementary data source for cross-database consistency analysis of publication trends, as well as supplementary analysis of journals, authors, countries, and institutions. The retrieval date is April 7, 2026, covering the time span from January 1, 2005, to December 31, 2025. The search strategy is constructed around keywords related to the treatment of vascular malformation sclerosis, using a combination of subject headings and free-text terms. The specific search formula in WoSCC is shown in Figure 1.The literature types are limited to Articles and Reviews, and the language is restricted to English. After deduplication, title and abstract screening, and full-text screening, a total of 2,040 WoSCC literature items were included. The parallel analysis in PubMed employs the same core search formula, with language restricted to English, yielding a total of 1,855 literatures after the same screening process. The complete retrieval strategy for WoSCC and PubMed, along with detailed inclusion and exclusion criteria, can be found in the Supplementary Materials.

Workflow of the data search strategy.

2.2 Bibliometric analysis

We employed three bibliometric software tools for data analysis and visualization. The analysis process included data cleaning, descriptive statistics, collaboration network analysis, co-occurrence and co-citation analysis, emergence detection, and visualization presentation. Before network analysis, the author and institution names were standardized; the thesaurus function of VOSviewer was used to merge different name variants of the same institution into a standard form, and author disambiguation was completed by cross-checking the full author names, affiliated institutions, and accessible ORCID identifiers. The process for handling authors with the same name is detailed in the Supplementary Materials. VOSviewer (version 1.6.20) was used to construct collaboration networks of countries, institutions, and authors, as well as keyword co-occurrence networks and co-citation networks of journals and references. Association strength was used as the normalization method. Node size represents publication volume or citation frequency, line thickness represents association strength, and color represents different groups identified by clustering algorithms. The network layout was optimized using the Force Atlas2 algorithm. Complete parameter settings can be found in the Supplementary Materials. CiteSpace (version 6.4.R1) was used for emergence detection of keywords and references, generating keyword cluster timelines, and overlaying journal dual maps. The emergence detection algorithm was utilized to identify emerging research frontiers, while the time zone view was used to display the trajectory of thematic evolution. Complete parameter settings can be found in the Supplementary Materials. Topic modeling using latent Dirichlet allocation (LDA) was conducted using the topicmodels package in R. The number of topics was determined through grid search, selecting optimal parameters based on a combination of perplexity, topic coherence, stability, and interpretability. Model stability was verified by running with different random seeds, and topic interpretation was independently completed by two authors who discussed to reach consensus. The complete results of the grid search, model verification parameters, and interpretation processes can be found in the Supplementary Materials. Bibliometrix (R package, version 4.3.3) and its web interface Biblioshiny were used for descriptive statistics, Bradford’s law verification, three-field plot drawing, and thematic evolution analysis. The annual growth rate calculation employed an exponential growth model.

3 Results3.1 Publication trends

According to the retrieval strategy described above, this study included 2,040 publications published between 2005 and 2025, comprising 1,748 Articles and 219 Reviews. The total number of references was 42,122, and the average citations per paper were 21.87. Basic information on research into sclerotherapy for vascular malformations is presented in Table 1.

DescriptionResultsMain informationTimespan2005:2025Sources (Journals, Books, etc)735Documents2040Average citations per document21.87References42122KeywordKeyword Plus(ID)3718AuthorAuthors8597Co-authors per document6.04International co-authorships%13.91Document typesArticle1748article; early access10article; proceedings paper62Article; retracted publication3Review219Review; early access1

Summary of Basic Bibliographic Information on Sclerotherapy for Vascular Malformations.

3.2 Contributions of countries to global publications

The annual publication trend for research on sclerotherapy for vascular malformations is shown in Figure 2A. From 2005 to 2014, the number of publications increased slowly (40 to 104 papers per year); from 2015 to 2018, the field entered an accelerated growth phase (rising from 87 to 117 papers); after a brief decline in 2019, publication numbers have grown explosively since 2020, reaching a peak of 160 papers in 2025. Based on the fitted model of cumulative publication counts in Figure 2B, research output in this field is expected to continue steady growth in the coming years.

Publication trends and international cooperation in vascular malformations treatment research. (A) Global annual publication trend; (B) Cumulative publication volume and polynomial fitting curve (y = 0.016x² + 4.984x + 39.75, R² = 0.8734); (C) Distribution of publication volume by major countries; (D) Distribution of citation influence by country (total citation frequency vs average number of citations); (E) Cumulative publication trend in China; (F) Cumulative publication trend in the United States; (G) Cumulative publication trend in Japan; (H) National scientific research cooperation network map.

By corresponding author country (Figure 2C), China produced the most publications over the past two decades (535 papers, 26.2%), followed by the United States (451 papers, 22.1%), Japan (106 papers, 5.2%), South Korea (87 papers, 4.3%), and India (83 papers, 4.1%). In terms of total citations, the United States ranked first, reflecting higher average impact per paper; China, Japan, Germany, and the United Kingdom followed (Figure 2D). Notably, the number of new publications from the United States, China, and Japan increased throughout the study period, with particularly marked growth after 2020 (Figures 2E–G). The international collaboration network is shown in Figure 2H. Five clusters were identified, with prominent Western European–American clusters centered on the United States and Germany and an Asia–Pacific cluster centered on China and Japan, although collaboration intensity between China and the United States was limited.

3.3 Distribution of publishing institutions

Among the top 15 institutions by publication count, Shanghai Jiao Tong University ranks first, followed by Shandong University. Five are from China, seven from the United States, and the remainder from South Korea, Belgium, Canada, and others. Shanghai Jiao Tong University in China exhibited the most rapid growth, reaching 213 papers in 2025 to take first place; Shandong University in China significantly accelerated after 2015, accumulating 125 papers; Sungkyunkwan University in South Korea led early but slowed later, accumulating 124 papers; Harvard University in the United States started publishing early but has shown weak growth in recent years, accumulating 89 papers; the remaining institutions each accumulated between 40 and 80 papers (Figure 3A). Figure 3B presents the cumulative publication trends of the top 15 institutions: Chinese institutions experienced strong growth after 2015 and have established a leading advantage; although there are many U.S. institutions, most have slowed in later growth; European and Canadian institutions have also maintained steady output.

Distribution and cooperation network of major research institutions. (A) Top 15 institutions by publishing papers; (B) Time evolution trend of the top 15 institutions publishing papers; (C) Institutional cooperation network map.

In the analysis of institutional collaboration networks, 14 clusters were identified (Figure 3C). Institutions such as Harvard University, Boston Children’s Hospital, and Harvard Medical School form a core U.S. research cluster, collaborating closely with the University of Pennsylvania, Johns Hopkins University, the University of Washington, and others. KU Leuven, University College London, the University of Amsterdam, and Ludwig Maximilian University of Munich form a European research cluster. Shanghai Jiao Tong University, Shandong University, Fudan University, the Chinese Academy of Medical Sciences, Sungkyunkwan University, and Seoul National University comprise an East Asian research cluster, within which Shanghai Jiao Tong University maintains relatively close collaborative ties with domestic Chinese institutions and Sungkyunkwan University in South Korea.

3.4 Distribution of authors’ publications

Consistent with Lotka’s law, authors publishing six or more papers account for less than 0.1%. As shown in Table 2, 20 authors published 20 or more papers. WANG Y from Guang’anmen Hospital, China, is the most prolific author in the field, with 40 publications. WANG H from Guilin University, China, has the highest h-index, indicating strong citation impact; closely following is DO YS from Sungkyunkwan University, South Korea. Notably, among the top 20 most productive authors, 15 are from China, three from South Korea, and two from Germany, reflecting the dominant role of Chinese researchers in this field.

AuthorArticlesCountryH-indexAffilliationsPublication year startedWANG Y40CHINA17Guang’anmen Hospital Jinan Hospital2007YANG X36CHINA11Shanghai Jiao Tong University2014LI J33CHINA15Shandong University, Jinan2009CHEN H30CHINA12Shanghai Jiaotong University2009LI X30CHINA14Chongqing Medical University2013ZHANG J30CHINA15Chongqing Medical University2008WANG L26CHINA11Sichuan University2007DO YS25KOREA18Sungkyunkwan University School of Medicine2005LIN X25CHINA10Shanghai Jiao Tong University2008LIU Y25CHINA13Guizhou University2008ZHANG X25CHINA11Children’s Hospital Affiliated to Shandong University2015WANG24CHINA11Qingdao University2010WANG H24CHINA35Guilin University2010WOHLGEMUTH WA24GERMANY10University Hospital Halle2015CHEN Y21CHINA9Southern Medical University2010FAN X20CHINA12China-Japan Friendship Hospital2010KIM DI20KOREA17Yonsei University2005KIM YW20KOREA17Hanyang University2005WILDGRUBER M20GERMANY9Regensburg University Medical Center2017ZHANG L20CHINA14Weifang No. 2 People’s Hospital,2007

Ranking of prolific authors.

Figure 4A shows the citation network of authors in the field of sclerotherapy for vascular malformations. We identified 161 authors with at least five publications; these authors form five clusters (distinguished by color). Lee BB (green cluster), Lin Xiaoxi (center of the blue cluster), and Wohlgemuth WA (yellow cluster) occupy key nodes in the network, indicating their work is widely cited by subsequent researchers. Authors such as Fishman SJ, Vikkula M, and Richter GT show tight citation linkages within the red cluster. Do YS, as one of the core nodes of the green cluster, together with Lee BB, constitutes an important research direction in the field, reflecting the growing influence of Chinese and Korean researchers.

Author collaboration and co-citation network. (A) Author co-occurrence network (authors with ≥ 5 articles); (B) Author co-citation network (authors with ≥ 40 citations).

Figure 4B presents the co-citation network of authors in the field of sclerotherapy for vascular malformations. We identified 150 co-cited authors with at least 40 citations; these authors form three clusters (distinguished by color). Lee BB and Boon LM serve as core nodes of the red cluster and have active co-citation relationships with authors such as Dompmartin A and Richter GT. Adams DM, Ozeki M, Wohlgemuth WA, and Seront E exhibit close co-citation ties in the blue cluster, forming another group focused on sclerotherapy techniques and clinical outcome studies. Wiegand S, Balakrishnan K, and others form an independent research group within the green cluster.

3.5 Journal distribution

The journal distribution analysis shows that literature in this field follows a typical concentration–dispersion pattern, consistent with Bradford’s law describing core, related, and peripheral zones (Supplementary Figure 1). The core Zone 1 comprises 30 journals, which published about one-third of the papers(Supplementary Table 1). The top 20 journals by publication volume are listed in Table 3; JOURNAL OF PEDIATRIC SURGERY and PHLEBOLOGY are tied for highest (57 articles each), followed by CARDIOVASCULAR AND INTERVENTIONAL RADIOLOGY (49 articles) and JOURNAL OF VASCULAR AND INTERVENTIONAL RADIOLOGY (47 articles). High-impact journals such as CHEMICAL ENGINEERING JOURNAL (IF = 13.2, Q1) and JOURNAL OF MEMBRANE SCIENCE (IF = 9.0, Q1) also contain related publications, though with lower article counts. A three-field analysis linking authors, keywords, and journals shows that core authors, high-frequency terms, and principal publishing journals form a tightly integrated knowledge-production network (Supplementary Figure 2), reflecting the main structure and disciplinary boundaries of knowledge production in this field.

Journal nameH-indexTCNPJournal citation reports quartileImpact factorCARDIOVASCULAR AND INTERVENTIONAL RADIOLOGY1989449Q22.9JOURNAL OF PEDIATRIC SURGERY19124857Q12.4JOURNAL OF VASCULAR AND INTERVENTIONAL RADIOLOGY1897647Q22.6PHLEBOLOGY18123857Q31.6JOURNAL OF MEMBRANE SCIENCE17116418Q19.0INTERNATIONAL JOURNAL OF PEDIATRIC OTORHINOLARYNGOLOGY1669138Q41.3JOURNAL OF VASCULAR SURGERY1580318Q23.6JOURNAL OF VASCULAR SURGERY-VENOUS AND LYMPHATIC DISORDERS1447837Q33.0DERMATOLOGIC SURGERY1351732Q32.2JOURNAL OF ORAL AND MAXILLOFACIAL SURGERY1352921Q40.4JOURNAL OF CRANIOFACIAL SURGERY1246051Q31.0LARYNGOSCOPE1249218Q32.0OTOLARYNGOLOGY-HEAD AND NECK SURGERY1268014Q32.5CHEMICAL ENGINEERING JOURNAL11104411Q113.2EUROPEAN JOURNAL OF VASCULAR AND ENDOVASCULAR SURGERY11104615Q16.8AMERICAN JOURNAL OF ROENTGENOLOGY1026112Q16.1BRITISH JOURNAL OF ORAL & MAXILLOFACIAL SURGERY1032014Q31.9EUROPEAN RADIOLOGY1034612Q24.7PEDIATRIC SURGERY INTERNATIONAL1028918Q31.6PLOS ONE1041814Q32.6

Ranking of major publishing journals.

Regarding citation influence, the United States has the highest total citations (11,715), followed by China (10,666) and South Korea (2,118). However, Singapore has the highest citations per article (60.2), followed by Belgium (53.5) and Austria (42.4), indicating that small but highly specialized research teams may produce work of outstanding impact. Canada (30.3) and the Netherlands (31.7) also have relatively high citations per article, reflecting the quality of their research (Table 4).

CountryArticlesTCAverage article citationsUSA15361171526CHINA13141066619.9KOREA242211824.3GERMANY295162921.4JAPAN368160815.2