A comprehensive literature search was conducted using the Web of Science Core Collection (WoSCC), Scopus and EMBASE databases for peer-reviewed literature published between the 1^st of January 2005 and 31^st of March 2025 inclusive. The literature search was performed with the assistance of a trained librarian. The search terms used were ((“extracellular vesicle*” OR “exosome*” OR “microvesicle*” OR “cell-derived microparticle*” OR “small vesicle*” OR “ectosome**) AND (“cutaneous melanoma*” OR “melanoma*” OR “cutaneous malignant melanoma*” OR “cutaneous metastatic melanoma*”) AND (“biomarker*” OR “liquid biops*”) NOT (“uveal melanoma*” OR “oral melanoma*” OR “Merkel cell”)). The key words were chosen as they were broad ranging enough to include all relevant articles, while limiting output to the specific research question. The search was limited to the English language and to journal articles only as these represent the most up-to-date scientific repository. A dedicated search of the “grey literature” (unpublished trials, theses, reports and conference notes) was not undertaken as sufficient data was obtained through the conventional literature search. No restrictions on study design were set in the initial search. The abstracts identified were independently reviewed by two reviewers using the Rayyan software (15). Any disagreement between the reviewers was highlighted. The full-text articles were then reviewed by a third reviewer and a final consensus was reached.

All studies that investigated the role of EVs in CM between 2005 and 2025 were included. This included studies: (a) characterising EVs cargo, (b) investigating the role of EVs as biomarkers, or (c) linking EVs with CM progression, metastasis, resistance or response to therapy (Table 1).

Review articles and articles relating to the role of EVs in other biological or pathological processes not specific to CM were not deemed eligible for inclusion. Studies without a clear methodology for EVs isolation and characterisation, studies lacking EVs-specific analysis and studies looking into EVs as transporters for therapeutic agents were also excluded (Table 1). Finally, we excluded animal studies except murine models. This was to ensure data homogeneity and reduce inter-species variability in our data.

Several software tools are available to analyse and visualise bibliometric data. Bibliometrix, a free R statistical package, developed by Aria and Cuccurullo (14), was employed in this study to analyse and map bibliographic data (16). Bibliometrix for RStudio incorporates Biblioshiny, a graphical interface that enables non-coders to perform comprehensive analyses with optimised visual representations. Compared to most free software (e.g. CiteSpace or VOSviewer), Bibliometrix does not only focus on the data visualisation but also on the correctness and statistical completeness of the results (16). It combines bibliometric techniques such as co-word analysis, co-citation network analysis and generation of collaboration networks to analyse a research field’s evolution (17). It was, therefore, felt that this software would most completely assess and depict publication trends, author networks, institutional productivity, international collaborators, keyword co-occurrence and citation metrics in the field of EVs in CM.

Thresholds for bibliometric network inclusion were chosen to optimise interpretability while preserving literature coverage. Minimum occurrence cut-offs limited the influence of sparsely connected nodes, and sensitivity analyses demonstrated that varying the threshold resulted in no substantive change in cluster composition or thematic interpretation.

First, the included articles were imported into the WoSCC database as it is compliant with the Bibliometrix software and allows comprehensive data analysis through its online analysis tool. The WoSCC analysis tool was used to yield bibliometric parameters including the number of annual publications, the output of countries/regions, total citations per publication (CPP) and the Hirsch index (H-index). The H-index is commonly applied to quantify and standardise the scientific impact of researchers, institutions, and countries. Higher H-indices generally correspond to greater academic influence (18). Journal Impact Factors (IF) and quartile rankings were retrieved from the 2025 edition of Journal Citation Reports (JCR).

The latest version 2025.05.0 + 496 of RStudio for mac was downloaded and a bibliometrix package was established within the R environment to analyse and map the bibliographic data. The included studies were exported in BibTex format, before being uploaded to Biblioshiny for analysis. The bibliometrix functions were then used to create descriptive and co-citation networks.

The “Preferred Reporting Items for Systematic Reviews and Meta-Analyses” (PRISMA) guidelines were used for study selection in this bibliometric analysis. The initial search yielded 2800 publications; which was reduced to 2038 after duplicates were removed. 1706 papers were excluded by title, abstract and keywords alone, leaving 332 manuscripts for full-text review. 44 articles were deemed ineligible after full-text review and the remaining 288 articles were deemed suitable for inclusion in the bibliometric analysis. The reasons as to why the articles were excluded are listed in the PRISMA flow diagram (Figure 1).

Between January 1^st 2005 and March 31^st 2025 a total of 288 publications met our inclusion criteria. The growth of the annual publication output and annual citations are shown in Figure 2. Compared to 2005 (n=1), the growth rate of publications in the field was 4100%, with 42 publications in 2020. There was subsequently a slight decrease in publication output between 2020 and 2023 possibly stemming from the overall reduction in original research output during the COVID-19 pandemic (19). Notably, the highest surge in citations was in 2021 (n=33, citations = 3620), and remained high up to and including 2024, reflecting the significance of this rapidly expanding research area. The top 10 most cited publications are shown in Table 2. The most frequently cited study was published in Nature Medicine in 2012 by Peinado et al. entitled “Melanoma exosomes educate bone marrow progenitor cells toward a pro-metastatic phenotype through MET”, with 3,238 citations (20). The annual citation rate (ACR), which was calculated as the ratio between the number of citations and the number of years since its publication, is 231.29 citations per year for this article. The second most cited article (citations = 2,326; ACR = 290.75) was published in Nature by Chen and colleagues in 2018 and is entitled “Exosomal PD-L1 contributes to immunosuppression and is associated with the anti-PD-1 response” (21).

The distribution of all 288 articles covered 40 countries. The 10 countries with the highest publication productivity are illustrated in Table 3. The United States ranked first in the number of publications (27.8% of 288 articles; n=82), followed by China (16.61%, n=49), Italy (13.9%, n=41), Germany (11.53%, n=34) and Japan (6.78%, n=20). Notably, collaborations between countries were included in the counts, resulting in a higher number of counts than included studies i.e. some studies had more than one country/region of origin. The collaboration between countries is further reflected in Figure 3a. Based on the distribution of single-country (SCP) versus multiple -country publications (MCP) shown, international collaboration in the field of EVs in CM appears limited, with relatively few publications involving cross-country collaboration. Interestingly, China which is the second most productive country, has the lowest collaboration rate (2.2%), while the United States has a 15.5% collaboration rate (Figure 3a). This shows that there is still great scope for collaboration in the field of EVs in CM.

International cooperation network analysis among different countries/regions in EVs melanoma research is shown in Figure 3b. Based on this collaboration map, link strength represents the degree of international research collaboration, measured by the number of co-authored publications between countries. Countries depicted in a darker shade of blue have demonstrated higher productivity in the field. Further analysis revealed that the closest collaborative relationship with the United States was Germany (n=13), followed by Switzerland (n=6), Spain (n=6), Italy (n=5) and France (n=4). Within Europe, collaboration was observed primarily among Western and Southern European research hubs, including Germany, France, Spain and Italy, while broader inter-European collaboration was limited (Figure 3b).

The annual publication output of the six most productive countries was identified. As demonstrated in Figure 3c, the most prolific country for annual publications was the United States. This leading position was maintained by the United States between 2012 and 2025 and demonstrates the impact this country has had on the current research landscape in the field. Interestingly, while China only began to stand out in the field of EVs research in melanoma in 2020, the country has demonstrated a very rapid increase in scientific production between 2020 and 2025 (Figure 3c).

Among the top 10 countries with regards to citations, the United States was the most cited (12,898), followed by Italy (6,047) and Spain (4,234). Spain however had the highest average citations per paper (222.84), largely attributed to the landmark paper by Peinado H. et al. (20) (Table 3).

The distribution of institutional affiliations consists of 612 entries. The 10 most relevant affiliations are shown in Table 4, with Chongqing Medical University having the highest output (n=45), followed by the University of Pittsburgh (n=41) and University of Louisville (n=36). Interestingly, Chongqing Medical University entered the field of EV research in CM in 2017, 9 years after the University of Pittsburgh, yet has shown a remarkable surge in research output in the past 5 years (Figure 4).

The articles included in this study were published in 151 journals. The top 10 most productive journals are shown in Table 5. The CiteScore represents the number of citations a journal receives relative to the number of articles it publishes.

The Journal of Extracellular Vesicles was the most productive journal for publishing research related to the role of EVs in CM diagnosis, pathogenesis, progression and monitoring, with a total of 15 publications (IF 15.5), followed by the International Journal of Molecular Sciences (n=14), Scientific Reports (n=10), Cancers (n=9) and PLOS ONE (n=8), respectively. The Journal of Extracellular Vesicles had the highest overall CiteScore and impact factor, while Cancer Research has the highest H-index (Table 5).

The authors of the 288 publications were analysed. When considering publication outputs, Chen Y, was the most productive author with 14 publications, followed by Wang J (n=12), Li X (n=11), Hood JL (n=10) and Zhang Y (n=10). Based on H-index, Hood JL had the highest local impact with an H-index of 10, followed by Chen Y and Wang J, both with an H index of 9 (Table 6). Figure 5 illustrates the author collaboration network of EVs research in CM. To reflect the relatively small size of the field while maintaining network interpretability, authors were included if they had contributed to at least two publications; inclusion of single-publication authors resulted in a highly fragmented network, whereas higher thresholds excluded a substantial proportion of active contributors. Collaborations were defined by co-authorship within the same publication. Node size is proportional to the number of publications per author, while links represent co-authorship relationships, with link thickness corresponding to collaboration strength (number of shared publications). Colours indicate clusters of authors identified algorithmically based on the density of collaborative links.

The network demonstrates a fragmented collaboration structure, characterised by several small, discrete clusters with limited inter-cluster connectivity. A dominant collaborative cluster is centred around a small number of highly productive authors, including Chen Y and Li X, who show multiple co-authorship links within this group. An important publication by Chen Y and colleagues published in Gene Cancer Therapy in 2024 demonstrates that transfer of exosomal lncRNA from highly metastatic melanoma stem cells to low-metastatic non-stem cells leads to activation of the Wnt signalling pathway and contributes to low-metastatic non-stem cells acquiring metastatic competency (22). In contrast, most other authors form smaller, isolated clusters or dyads, with minimal connections to the wider network. Overall, the limited number of links between clusters suggests that author-level collaboration in this field remains largely confined to individual research groups, with relatively few cross-group or cross-network partnerships.

The most cited author was Peinado H, for the paper published in Nature Medicine entitled “Melanoma exosomes educate bone marrow progenitor cells toward a pro-metastatic phenotype through MET” (20). An author co-citation network, illustrated through multidimensional scaling is shown in Figure 6. Each node represents an individual publication, with node size proportional to citation frequency within the dataset. Links between nodes indicate citation relationships, with thicker links reflecting stronger citation connections. Colours denote clusters of publications identified algorithmically based on citation patterns, representing groups of studies that are more closely related in terms of referenced literature. The network demonstrates a structured citation landscape with several interconnected clusters spanning different publication periods. A highly central and influential node is represented by the publication by Peinado et al. (2012), which occupies a prominent position within the network and shows extensive citation links across multiple clusters, indicating its role as a foundational reference in the field. Earlier studies form more peripheral clusters, while later publications appear more densely interconnected.

A total of 251 keywords were identified, divided into 16 clusters as depicted in Figure 7. Each node represents an individual keyword, with node size proportional to its frequency of occurrence across the included publications. Links between nodes indicate co-occurrence of keywords within the same articles, with thicker links reflecting stronger associations. Colours denote clusters of keywords identified algorithmically based on co-occurrence patterns, representing related thematic areas within the literature. The map is dominated by the central keywords “melanoma” and “extracellular vesicles,” which exhibit the highest frequencies and the strongest associations with multiple surrounding terms. Closely linked keywords include “exosomes,” “tumour microenvironment,” “metastasis,” “biomarkers,” and “immune response,” reflecting key biological and translational themes explored in EV research in CM.

Figure 8 presents a thematic map derived from the Keywords Plus analysis, illustrating the structure and development of research themes in extracellular vesicle studies in cutaneous melanoma. Themes were generated based on co-occurrence patterns of Keywords Plus terms and positioned according to their relevance (centrality) on the x-axis and degree of development (density) on the y-axis. Bubble size reflects the frequency of Keywords Plus within each theme. Basic themes, located in the lower right quadrant, are characterised by high relevance but lower internal development and include broadly defined topics such as “extracellular vesicles,” “cells,” and “membrane vesicles,” reflecting foundational concepts that underpin much of the literature. Motor themes, situated in the upper right quadrant, demonstrate both high relevance and high development, with Keywords Plus such as “DNA,” “liquid biopsy,” and immune checkpoint–related terms (e.g. “nivolumab”), indicating well-established and actively investigated areas.

Niche themes in the upper left quadrant, including terms related to cancer-associated fibroblasts, circular RNA, and matrix interactions, represent more specialised and internally cohesive research areas with limited connectivity to the broader field. Emerging or declining themes, located in the lower left quadrant, include Keywords Plus such as “in vivo,” and “caveolin-1,” suggesting areas of evolving research focus.

To determine recent trend topics, we assessed the change in keyword frequency over time. Earlier research activity was characterised by a focus on general vesicle-related terms and methodological concepts, including “microparticles,” “release,” and “proteomic analysis.” From approximately 2016 onwards, there is a marked increase in the frequency of biologically and clinically oriented keywords such as “extracellular vesicles,” “metastasis,” “immune response,” “diagnosis,” and “biomarkers.” More recent years show the emergence of translational and immunologically focused topics, including “PD-1,” “migration,” and “immunity,” reflecting a shift towards clinical relevance and therapeutic application within the field. (Figure 9).

The most cited reference is the 2012 paper published in Nature Medicine by Peinado H (n=111) (20), followed by the paper entitled “Exosome-mediated transfer of mRNAs and microRNAs is a novel mechanism of genetic exchange between cells” by Valadi H published in 2007 in Nature Cell Biology (n=55) (23). Figure 6 demonstrates a strong link between the two nodes, representative of a high co-citation frequency. Table 7 provides a detailed breakdown of the top 10 most cited references in the field of EVs in CM. The most frequently cited references across the included studies were from the Journal of Extracellular Vesicles (n = 414) and Cancer Research (n = 506).

In this study, we performed a bibliometric systematic review of research on EVs in CM from 2005 to 2025. Our objective was to present a comprehensive overview of the current research landscape and emerging trends in the field, offering scholars a structured reference framework to understand key findings and identify cutting-edge research frontiers. Our findings indicate a slow, yet steady, rise in citations and publications in the field between 2005 to 2017 with an average of 5 articles published per year. The number of published articles increased significantly between 2018 and 2024 to an average of 32.4 articles published per year. Notably, the United States was the main driving force, with the highest number of annual publication outputs and highest H-index (Table 3). It is important to note the geographical imbalance in the field, with China being the only developing country in the top 10 prolific countries in the past 20 years. China also has the lowest number of collaborative studies when compared to the top 5 most productive countries (Figure 3a). Greater international collaboration may facilitate closer alignment with activities and consensus recommendations of the International Society for Extracellular Vesicles (ISEV), including Minimal Information for Studies of Extracellular Vesicles (MISEV), thereby supporting methodological consistency and reproducibility within the field. Overall, the predominance of relatively limited link strengths on the map in Figure 3b highlights that EV research in cutaneous melanoma remains geographically concentrated and collaborative within select hubs, rather than being supported by broad, global research networks. This pattern may have implications for reproducibility, standardisation, and the translational progression of EV-based biomarkers in melanoma.

When addressing the most prolific institutional affiliations, Chongqing Medical University takes the lead, followed closely by the University of Pittsburgh and the University of Louisville (Table 4). Interestingly, when assessing institutional outputs over time, production from Chongqing Medical University increased dramatically between 2021 and 2024, while other prolific institutions showed a more gradual rise in production (Figure 4). Notably, the most prolific author group led by Chen Y, is affiliated with Chongqing Medical University, China (Figure 5). The fragmented author collaboration structure observed in Figure 5 suggests that EVs research in CM is frequently conducted within individual research groups, with relatively limited cross-network collaboration. While this pattern does not preclude methodological rigour, broader collaboration may support efforts promoted by the ISEV. Increased multi-institutional engagement could help facilitate consistency in methodology, validation across independent cohorts, and the translational development of EVs-based biomarkers in melanoma.

By analysing co-cited references, it is possible to identify foundational literature underpinning research in the field. As outlined in Table 7, the most frequently cited paper was the landmark original research study by Peinado et al., published in Nature Medicine, which described the role of the MET receptor in establishing a pro-metastatic niche through education of bone marrow progenitor cells (20). In keeping with this finding, the citation network analysis in Figure 6 demonstrates that this publication occupies a central position within the network, acting as a key node linking multiple citation clusters. As review articles were excluded from the analysis, the co-citation patterns shown in Table 7 and Figure 6 predominantly reflect the influence of original research studies, highlighting the core experimental work that has shaped the development of EVs research in CM.

Keyword analysis is a powerful tool for identifying research hotspots within a field. Analysis of keyword associations and temporal trend topics demonstrates a progressive evolution in extracellular vesicle research in cutaneous melanoma, from early emphasis on vesicle characterisation and methodological concepts towards biologically and clinically oriented themes. Core keywords such as “extracellular vesicles,” “melanoma,” and “exosomes” underpin the literature, while more recent trend topics highlight increasing focus on metastasis, immune modulation, biomarkers, and therapeutic relevance. In this context, Chen Y and colleagues, ranked as the most prolific author group in the field have contributed extensively to work demonstrating how EVs act as vehicles through which melanoma cells acquire metastatic competency (22) The Keywords Plus–based thematic map further supports this progression, identifying foundational EV concepts as basic themes, alongside more developed motor themes related to liquid biopsy, DNA, and immunotherapy. Collectively, these patterns suggest a maturing research landscape increasingly oriented towards translational application, while also highlighting areas where further methodological harmonisation and validation may be required to support clinical integration. These include variability in extracellular vesicle isolation and characterisation practices, reflected by overlapping use of terms such as “exosomes,” “microvesicles,” and “extracellular vesicles,” as well as heterogeneity in analytical approaches applied to biomarker discovery, including proteomic and nucleic acid–based methods. In addition, the emergence of clinically oriented themes such as “liquid biopsy”, “immune modulation”, and immunotherapy-related markers emphasises the importance of validation across independent cohorts and consistent reporting of experimental and clinical outcomes to support reliable translational application in cutaneous melanoma (Figures 7–9). Integration of the keyword co-occurrence network, thematic map, and temporal trend analyses (Figures 8, 9) indicates that current research frontiers and emerging hotspots in EVs research in CM are concentrated within these interconnected biological and translational domains: