In the current study, a comprehensive literature search was conducted in the WOSCC database on the topics of the CS. A total of 4181 publications in English were retrieved, which could erase the potential of ungrasping the authors’ intention due to language misinterpretation. Our statistical and quantitative analysis found a significant improvement in research on the field of CS and an increasing number of orthopedic and oncology experts focused their insights on this region. Although those studies have been highly extensive, they are relatively chaotic and lack hotspots analysis. Bibliometric analysis is an interdisciplinary science emerging in recent years, which could use mathematical and statistical methods to make analysis on a specific field, and thus offer the possibility of visualizing the present state from multiple dimensions and making reasonable prediction of future directions base on the published record. Therefore, we summarized in this study the contributions and research collaborations of the countries, institutions and authors in the field of CS, as well as the research hotspots and new trends of investigations. Researchers new to this field can easily gain useful and relevant information from our bibliometric analysis, and avoid repeated research.
From Figure 2, we could see that the annual growth trend of publication outputs related to CS is generally on the rise. It can be divided into four stages, including a stagnation period (before 1970), a slow growth period (1970 to 2000), a plateau period (2000 to 2010) and a rapid growth period (after 2010). The studies on CS were sparse before 1970s with publications less than 5 per year. Since then, there was an increment in the number of publications till 2000s. After a short plateau period from 2000 to 2010, the researches made a robust rise. We found that the studies during the first peak of publication were mainly morphology, histology, pathology, immunochemistry, and molecular analysis of CS, which reveals the primary focus of researchers in the early stage was to identify the histological and pathological features of CS, as well as mutations that may be associated with CS in the molecular level. Although the number of studies published from 2000 to 2010 was still relatively large, we can speculate from the plateau stage that it may be a symbol demonstrating the oversaturation of previous studies and the determination of new research directions being made afterwards. Since then, the number of publications began to show a consistent upward trend again.
The United States had the highest absolute productivity ranking, which contributed more than half of the publications, followed by Japan, China, Germany and Italy. We could learn from the Figure 4 that the institutions or universities outstanding in each cluster were mainly from the United States, especially the Harvard University, which was a prominent contributor of 3 clusters. Among the top 5 countries, only China is a developing country, showing that the research capacity of developing countries in this filed is weak, which could also be reflected by the low average citations (11.27 times). This situation might be influenced by the amount of financial support available for conduction of basic experiments on CS. Therefore, developing countries should come up with strategies to promote research in this field and actively learn from developed countries. The developed countries should make more collaborations with developing countries to promote the medical progression. Although the United States published the most records worldwide, Netherlands (which ranks in the 8th) had the highest average citations (36.48 times). Citations differed from the number of documents, as did the cooperation of countries and regions: United States, Italy, England, Netherlands, Germany, France and Canada showed the strongest cooperation, with a high total link strength among the countries. This was mainly because the studies on CS were published earlier in these countries, which was indicated by the dynamics and trends of countries over the years in the Figure 3B.
Similarly, the analysis of institutions and authors almost matched the distribution by countries and districts. The institutes from United States (Harvard University, Mayo Clinic, Memory Sloan Kettering Cancer Center and University of Texas MD Anderson Cancer Center), European Universities (Leiden University, University of Bologna and Heidelberg University), Japan Universities (Okayama University and Niigata University) and Chinese Universities (China Medical University), have taken leading places in CS papers, with steady collaborations in global groups. Moreover, among the authors who contributed to publications, the top 50 authors were situated in the collaboration networks that account for large proportions of research, mainly from the above-mentioned countries and institutes. However, the network map analysis for authors could generate almost no clusters and lines representing link strength, which indicates that less active collaborations had been implemented during the past few decades among the prolific authors. We call for strengthening the exchanges and cooperation between researchers to better promote the development of his field.
The dynamics of a field can be characterized, in part, by articles with citation burst. References with citation burst refer to the sudden increase of citations of certain documents in a certain period, which can help us find emerging topics and research topics that have attracted much attention in a certain field. We selected the top 25 references with intense citation bursts. The first 3 references with citation burst were studies concentrating on the histological and biological features of CS. Two of them investigated the biosynthesis of proteoglycan aggregates from CS cells and both published on Journal of Biological Chemistry[20, 24]. And then, the four references with citation burst from 1990 to 2000 shifted the research emphasis to myxoid CS, a rare subtype of CS[17, 25-27]. During the 2000 to 2010, the researchers started to put their insight to the genetic level and related pathways in CS. While in the second decade of 21 century, the references burst were most clinical studies which investigated novel therapeutic approaches and prognosis of patients with CS. The Leiden University ranked the first for contributing 6 papers among the top 25 references with citation burst, followed by Harvard University, which were also illustrated in the Figure 4 that these universities were at leading position in this field. The Cancer, Annals of Oncology, Journal of Biological Chemistry and Journal of Joint and Bone Surgery were the journals published the most references with citation burst (3, 2, 2 and 2 records, respectively). New researchers could read articles from these journals preferentially to get the hotspots on this field, and give priority to these journals above others for new manuscript submission.
Keywords co-occurrence, together with the reference citation burst, could both indicate the research trends and hotspots. A lot of keywords were used in the studies published on research in the field of CS, while the majority of the keywords appeared once and only a little of them had a frequency of greater than five, revealing that only few keywords are frequently used. In the current bibliometric study, cluster analysis of co-occurrence keywords showed that there were 9 clusters in this filed. Of the 9 clusters in visualization maps to explore key topics and future directions, cluster 0 represented the studies on human CS cell and the timeline view demonstrated that this cluster last the longest period of time with the investigation on this topic still continuing. The keywords in the cluster #0 mainly consist antigen, monoclonal antibody, inhibitor, pathway and isocitrate dehydrogenase, which describe the new hotspots of CS in addition to surgical resection. As a chemo- and radioresistant tumor, conventional adjuvant therapies show little benefit and the surgical resection is the mainstay of treatment. While the patients with unresectable or recurrent disease still have a poor prognosis. Much research interest should be focused on the other novel therapies, such as molecularly-target agents, inhibiting abnormal MicroRNAs expression[28-31], inactivation of glucose metabolism[32, 33], downregulating the tumor cell activity[34-36], suppressing the matrix metalloproteinase (MMP)[37-39], targeting IDH mutations[40-43] and Indian hedgehog (IHH) signaling pathway[44-47] and immunotherapy[48-50], all of which apparently requires substantial basic experiments on CS cells to make some breakthrough of these aspects. Cluster #1, #3 and #7 were the studies investigating the treatments of CS on different anatomical locations. We could learn from the 3 clusters labels that CS is commonly originate from these regions[51-55]. The common keywords in these 3 clusters were chemotherapy, radiotherapy, surgery, resection and follow-up, meaning that these studies were concentrating on the clinical efficacy and safety of preexisting treatments, and the prognosis of patients with CS. While the reconstruction in the long bone and pelvis is an intractable issue after resection, which is not often encountered in the skull base CS. Modified methods for reconstruction based on current techniques and a better design of prosthesis are the hot topics of these investigations[56-60], and that explains why the cluster #3 and #7 are much enduring. The gene p53 (cluster #2) could suppress tumor cell, playing the role of monitoring aberrant cell division and inducing cell apoptosis if repair process could not be initiated. Although the tumor suppressor p53 gene has been extensively studied in the field of cancer, the mutations of p53 are thought to be relatively low in CS compared to other alterations in the genetic level. And a higher incidence of p53 mutation is found in the late stage of CS or other atypical CS, which only accounts a small fraction of CS. It suggested that the p53 may be correlated with the progression instead of initiation of CS[55, 61-63]. It is why the cluster #2 had an early stagnation of investigation before 2007. While combined with other treatment regimen currently used for certain altered pathways, appropriate p53 target agents can be potential therapeutic alternatives[3, 64]. There is a same situation for the cluster #4, which elucidated the substances of extracellular cartilage matrix (ECM) of CS cells to explore the tumor microenvironment (TME). The studies on this topic were used to attract much attention before 2000, while since then, researchers shifted their sights into other regions. Enhancing endoplasmic reticulum (ER) stress (cluster #6) has shown great effect as a tumor suppressor of CS. In both vivo and vitro studies, treatment with enhancing ER stress could induce apoptosis of CS cells via mitochondrial dysfunction, which could be a promising approach in CS treatment[65-67]. However, no updating researches have been forthcoming in recent years, which could suggest the clinical efficacy of this modality needs to be further ascertained. Previous studies have revealed that connective tissue growth factor (CTGF) could be induced by hypoxia and therefore promotes tumor angiogenesis in human CS cells[68, 69]. However, the studies to date failed to demonstrate meaningful clinical activity of target the gene of CTGF on the prevention of CS.
Limitations and strengths
Nevertheless, we were aware of several potential limitations of this current study. First, we only searched records from the WOSCC and only included papers in English and in form of article and review, which might result in selection bias. Second, the number of related papers on CS may increase rapidly, that the constant updating of the database may lead to a difference between bibliographic analysis data and the actual research progress. Third, the error of the database itself may also lead to a deviation of the results, such as mislabeling document types and less inclusive of the documents. Forth, although the data had been manually standardized, bias might exist due to the authors with the same name or the keywords of various expressions. Finally, by using the software, we might have overlooked some information, which may result in errors in data analysis, even if there are some reasonable and unavoidable differences on the same outcomes between CiteSpace and VOSviewer. Acknowledging the limitations above-mentioned, however, to be the best of our knowledge, this is the first study to perform bibliometric analysis on research in the topic of CS. The current study aimed to identify collaboration networks among authors, countries, and institutions and illustrate the development trends, current hotspots and future directions in this field.