The Web of Science (WoS) core database was searched on October 31, 2023, for all the data used here. Following the outlined procedure, the search algorithm used was: TS = (osteosarcoma) AND TS = (exosomes OR EVs) AND TS = (marker OR biomarkers). Moreover, to further investigate the exosome-derived diagnostic biomarkers, we additionally searched for studies on tumor- and exosome-derived biomarkers from the 2013–2023 period using the TS = (tumor OR cancer) AND TS = (exosome OR EVs) AND TS = (marker OR biomarker) algorithm. After a thorough and repeated screening process, we retained only articles and reviews, eliminated the literature that was not relevant to our research question, constrained the language to plain English, extracted the data, and stored them in the TXT format.

The files were in plain-text format and included complete citations for references to improve the bibliometric analysis and the visualization of our findings. Figure 1A (OS) and 1B (cancer throughout the body) present an in-depth overview of the selected information. Furthermore, as a component of the qualitative analysis, we conducted a bibliometric screening of the papers with an average annual citation count of over 10.

This study did not necessitate approval from the Committee on Ethical Medicine.

The final collection of studies on OS from the 1995–2023 period included 61 publications on OS and exosome-derived biomarkers. Moreover, 1,738 documents on tumor- and exosome-derived biomarkers from 2013 to 2023 were found. The extracted data included the journal name, reference type, publication date, author affiliations and names, and abstract.

Bibliometrix is a bibliometric software package that was designed by the University of Federico II in Naples, Italy, and is implemented using the R programming language. According to previous research, R4.0.3 bibliometric procedures have been used to automatically assess the publication data (13, 14). This study incorporated data on annual publication counts, publications categorized by country, authors and institutions, journals, co-cited references, and keywords. To assess the quality of the studies, metrics such as the number of published papers, citation counts, and the H-index of the authors were used (14). The H-index of the authors was used to measure the output of researchers in this research field (15). CiteSpace, a scientometric research tool developed by the School of Computing and Intelligence at Drexel University (14), was used for the co-author analysis of the country/author/institution, the analysis of co-citations for a journal or reference, contribution analysis of the keywords, and visualization using grids and overlays, among others.

On November 1, 2023, a scientific literature search of the WoS core repository was performed to obtain all relevant published papers related to OS and exosomes/EVs, which serve as biomarkers in clinical diagnosis. Based on the search terms mentioned above, after screening, we initially obtained a total of 62 documents, of which 36 articles (58.06%), 25 reviews (40.32%), and one article were retracted over an almost thirty-year period, from 1995 to 2023 ( Figure 2 ). The annual growth rate was 8.16% and the rate of international co-authorship was 16.13%, as shown in Figure 2A . Additionally, between 1995 and 2023, although only one article was published in 1995, 2002, and 2006, the annual literature published experienced a rapid upward trend from 2016 to 2021, followed by a rapid decline from to 2021 to 2023 ( Figure 2B ). The evolution trend of the number of cumulative publications fits the fitting curve y = 7.3545x - 18.945 (R² = 0.8854) ( Figure 2B ), which demonstrates that exosomes/EVs have progressively attracted the attention of researchers and may be a long-standing and promising research hotspot.

The corresponding authors were primarily from China, the USA, Australia, Canada, Brazil, France, and Italy ( Figure 3A ). The average paper citations for the top 10 affiliations and leading countries are listed in Tables 1 and 2 . China and Italy are the main countries in which these studies were published.

The University of Kansas Medical Center and National Cancer Center, Japan, are the primary affiliations of the authors of these studies.

The nation that published the largest number of documents is depicted in Figure 3B . The number of papers associated with China increased quickly and consistently over time, differing from that associated with the USA, which moderately increased ( Figure 3B ). Table 1 lists the top 10 nations with the highest average article citation rates, along with the individual citation rates in global terms. The average number of article citations from the United Kingdom (71.00%), Italy (44.90%), and China (26.4%) ranked the highest, which is indicative of the superior quality of their articles in this research field ( Table 1 ). Additionally, Figure 3C shows that the most cited studies were published in China (739) and Italy (359). The top 10 affiliations and article rankings are presented in Table 2 . The University of Kansas Medical Center, National Cancer Center, Japan, and the University of Kansas are the main institutions associated with research in this field.

Q.Y. Bao (4, 11.33%), followed by Y.H. Shen (4, 11.33%) and J. Wang (3, 12.00%) were the authors with the highest number of article citations in this field, as shown in Figure 4A . In addition, all of these authors are those of studies highly cited every year (light blue circles in Figure 4A ). The local impact of the authors was assessed using the H-index, which was 4, 4, 3, and 2 for Q.Y. Bao, Y.H. Shen, J. Wang, and the other authors (2), respectively, as shown in Figure 4B .

Then, the collaboration between authors based on clusters was analyzed, and the connecting lines indicate that the authors were co-authors. The authors Asano, Nanfumi, Chikuda, Hirotaka, Ochiya, and Takahiro (centrality = 0.02) were the core authors with a tremendous mediation centrality ( Figure 4C ). In addition, a cluster evaluation of the cooperative institutes was conducted, as shown in Figure 4D . The articles written by Chinese authors were mostly published online in the “Frontiers in Oncology” journal. The authors from the USA mostly published their studies in the journals “Cancer Biomarkers” and “Biomarker Insight,” as exhibited in Figure 4D .

The most important sources were published in “Frontiers in Oncology,” “International Journal of Molecular Sciences,” “Cancers,” and “Cancer Letters,” which are the main journals in the field of exosomes as OS diagnostic biomarkers ( Figure 5A ). With a high rating in the relevant industry, “Frontiers in Oncology” is a distinguished Journal Citation Reports Division 1 journal, with an impact factor of 4.70. Moreover, most sources cited locally were mainly published in “Oncotarget” and “Plos One”, followed by “Journal of Extracellular Vesicles” and “Cancer Research” ( Figure 5B ). Furthermore, according to Figure 5C , “Frontiers in Oncology,” “International Journal of Molecular Sciences,” “Cancer Letters,” and “Cancers” were the main journals in which the main sources that contributed to the local impact according to the H-index were published. Additionally, we found that the related literature began to be published in 2019 ( Figure 5D ).

Choosing an appropriate journal is crucial for scientific researchers, because it allows them to align their research priorities with the journal’s specific aims and scope. This alignment is essential to establish a robust theoretical foundation for OS, particularly when exploring the role of exosomes as biomarkers.

The most globally cited document was “Osteoclast-derived microRNA-containing exosomes selectively inhibit osteoblast activity” (total citation: = 201), which was published in the journal “Cell Discovery” by W.J. Sun, et al., in 2016. The second most globally cited one, “Mesenchymal stroma: Role in osteosarcoma progression” (total citation: = 109) was published in “Cancer Letters” by M. Cortini, et al., in 2017. The third most cited one, “Multimodal transfer of MDR by exosomes in human osteosarcoma” (total citation: = 102) was published in the “International Journal of Oncology” by E. Torreggiani et al., in 2016 ( Figure 6A ). In addition, the most locally cited documents were “Exosomal miR-675 from metastatic osteosarcoma promotes cell migration and invasion by targeting CALN1” (local citation = 17), which was authored by L.Z. Gong and Q.Y. Bao et al. and published in “Biochemical and Biophysical Research Communications” in 2018. This one was followed by “Multimodal transfer of MDR by exosomes in human osteosarcoma” (local citation = 15), which was authored by E. Torreggiani, et al. and published in the “International Journal of Oncology” in 2016. The study titled “Exosomes containing differential expression of microRNA and mRNA in osteosarcoma that can predict response to chemotherapy” (local citation = 13), which was published in 2017 in the journal “Oncotarget” and was authored by J.F. Xu et al., was the third most locally cited document ( Figure 6B ).

The most relevant key words were: extracellular vesicles, prognostic biomarker, cells, biogenesis, cancer, exosomes, expression, mesenchymal stem cells, progression, and osteosarcoma. These are presented in a plot chart, a word cloud, and a tree map, where the size represents their frequency and importance ( Figures 7A, B, D ).

Furthermore, the main topic subjects were proliferation, mechanism, stem cells, gradually converted into cancer, extracellular vesicles, and exosomes ( Figure 7C ) and, in 2022, lung metastasis and potential serum biomarkers were the most studied subjects. This indicates that a rising number of researchers are exploring the potential connection between the exosomes (or EVs) derived from OS and pulmonary metastasis.

A total of 62 documents were included in the analyses of the collaborations between authors, institutions, and countries using CiteSpace visual software. As shown in Figures 8A, B , T. Ochiya, Y. Araki, and S. Miwa worked closely with one another. The authors Q.Y. Bao and J. Wang also collaborated closely.

Furthermore, as Figure 8C shows, the Shanghai Jiao Tong University, Soochow University, and Fudan University cooperated closely. In addition, the group of the National Cancer Center, Japan, Tokyo Medical University, Okayama University, and the group of Consiglio Nazionale delle Ricerche, University of Bologna, also worked in close collaboration. However, interestingly, the University of Kansas Medical Center reported that most of their articles did not derive from collaborations, which deserves our deep consideration.

The collaboration of groups mainly occurred for groups from China, Japan, the USA, and Italy, which suggests that global teamwork needs to be strengthened ( Figure 8D ).

The frequency and number of keyword occurrences over a given period are valuable for assessing the ongoing and anticipated developments in a particular area of study.

Next, we conducted a co-occurrence network analysis of the keywords using CiteSpace software, and the results were visualized using a visualization diagram, cluster plot, timeline chart, and burstiness graph.

Figure 9A shows the most commonly occurring keywords, which are consistent with the findings presented in Figure 7 . Figures 9B , 10B are the cluster plot and timeline chart, respectively, which show that nine subject phrases were grouped using the CiteSpace software: microRNA, cancer therapy, personalized medicine, sarcoma, canine leishmaniasis, osteoblast, noncoding RNA, IL-6, cartilage cells in vitro, and flow cytometry. The research hotspot gradually shifted toward miRNAs and noncoding RNA, as illustrated in Figure 10B .

Figure 10A shows the tree map, which illustrates the connection between the keywords and their hierarchical arrangement through hierarchical cluster analysis, which revealed that every subject is composed of a succession of keywords related to “osteosarcoma” and “exosomes.”

After applying the Bibliometrix filter with an average citation of ≥ 10 per year, 9 literature pieces (6 articles and 3 reviews) were included ( Table 3 ). Next, taken together with the nine studies and 62 documents retrieved using the bibliometric filter, qualitative analysis was conducted not only to clarify the relative research progress on the topic but also to deeply comprehend the intimate relationship between exosomes and the cancer molecular biological mechanisms.

Liquid biopsy has gained traction in cancer research and is currently undergoing a surge in clinical applications. Liquid biopsies can generally be classified according to the cancer cells they target (17): (i) CTCs, (ii) circulating tumor DNA (ctDNA), and (iii) tumor-derived exosomes and other EVs. Exosomes and other EVs have evolved into versatile platforms with broad and complementary applications. Exosomes contain RNAs, DNAs, proteins, sugar structures, lipids, and metabolites. Moreover, surface molecules on the cytomembrane of exosomes can provide information about the tissue origin, making it possible to enrich the signatures of site-specific cancers.

Furthermore, tumor cells often utilize exosomes, robust components of intercellular communication, to improve their growth (16, 18). Hence, exosome-derived biomarkers show tremendous promise and have been adopted in scientific research and clinics (16, 18–20).

We conducted a thorough analysis of the applications of exosome-derived diagnostic biomarkers for OS using bibliometric estimation to present the research status and hotspots and predict future research trends in this field for the first time. Moreover, R-bibliometric, VOS, and CiteSpace viewer analyses were conducted to ensure the high objectivity and reliability of the data. We estimated the main tumor and exosomal biomarkers via bibliometric analysis and CiteSpace visualization to deeply and comprehensively understand the developmental trend of exosome-derived diagnostic biomarkers in OS disorders. However, this study has some inherent limitations. First, we obtained these documents only from the WoS database, not selecting data from the Embase, Scopus, PubMed, among others, which may have led to biased data. Second, since the use of exosomal biomarkers for the diagnosis of tumor tissues is an emerging research field, related literature, especially on OS, is scarce. Moreover, the latest online literature published in high-quality journals may be ignored owing to the lower citation counts, as per the citation analysis performed. Third, no relevant keywords with the strongest citation bursts were found based on the CiteSpace results, which demonstrates that the literature on this topic is scarce. Fourth, we did not use uniform criteria when using the CiteSpace software, thus affecting the output data, which may swing slightly under different parameters.