The development trend of research about uveal melanoma was shown by the quantity variation of academic papers. By analyzing the literature over the last 20 years, we found a steady growth of studies on UM in the last 5 years. As a malignancy, such growth rates reflect a more comprehensive understanding of the disease. And it is noteworthy that there are several time points when the discovery of new related genes might lead to a new wave of research. Although the mechanism of the disease is not fully understood, we do have a better understanding of the disease.
The most productive institutes are mainly located in the United States, the United Kingdom, the Netherlands, France, and Germany. Six of the top 10 research institutions are from Europe. Considering the high-risk population for UM, it is not difficult to infer that Europe and the United States are UM research centers. We are very grateful to these core authors for their contributions to the field, but the significance of the visual analysis results about the authors is not very obvious. These authors are studying this field for many years, so their findings have a longer exposure time. However, the authors' results can be used as a reference indicator when researchers choose their collaborators.
Through the bibliographic coupling analysis of source journals, we can determine the core journals in studies of UM. In Figure2, the apparent clustering of journals is determined. The results show a clear stratification of clinical and basic research and uneven distribution. This implicates the difficulties that basic research encounters in translating results into clinical applications. In the top 3 journals, Ophthalmology and British Journal of Ophthalmology belong to the same cluster, while Investigative ophthalmology & visual science belongs to the first cluster. Articles published in these journals are more likely to be cited. In other words, articles in these journals are more likely to disseminate their findings.
For the analysis of keywords, we used co-occurrence analysis. Through this analysis, we can see the hotspots that authors in this field are most concerned about and draw a knowledge background map. Therefore, we conducted a cluster analysis to explore the main topics of UM research. UM, keywords formed six main clusters and clustered together keywords with similar research topics (Figure 3). Combining the characteristics of UM and the current status of UM research, the 6 clusters are analyzed as follows:
Cluster 1 (Red): in this cluster, keywords are mainly related to traditional therapies. When treating uveal melanoma, the most important thing is to reduce mortality. However, patients and ophthalmologists are also committed to maintaining the visual function, cosmetic appearance, and quality of living[6,7]. Treating small to medium melanomas with Ru106 was a success.[8-10] According to different studies, regular treatments like enucleation, brachytherapy, charged particle irradiation, and local resection have similar survival outcomes[11-13]. It is noteworthy that studies on traditional treatment methods still accounted for the majority, but the studies were generally older.
Cluster 2 (Green): genetically related prognosis of melanoma was focused in this cluster. GNAQ, a stimulatory αq subunit of heterotrimeric G-proteins, was found to be mutated in 40% of UMs[14]. Some studies showed that 83% of UM had somatic mutations in GNAQ or GNA11. Constitutive activation of the pathway involving these two genes appears to be a major contributor to the development of uveal melanoma[15]. These two genes caused a wave of research around 2010. Prior to this, around 2003, research in this area was focused on BRAF mutations. Although Mutations in the BRAF gene enhance the kinase activity have been described in >60% of cutaneous melanomas and premalignant melanocytic lesions, it is not common in primary uveal melanoma[16,17]. Within this cluster, several genetic research hotspots are beginning to emerge, they are strongly clustered in time, and the research is generally close to us.
Cluster 3 (Blue): metastasis of UM is most mentioned in this cluster. In this cluster, the research approach is more oriented towards oncology. Of greatest concern were UM metastasis, apoptosis, and invasion. Monosomy 3 as a significant predictor of both relapse-free and overall survival of UM[18] is in this cluster's leading position. In this cluster, however, there is no trend regarding specific genes or key nodes. This indicates that there are still many unclear mechanisms waiting to be studied in this area. Researchers can look for new research directions in this cluster.
Cluster 4 (Orange): comparing with cluster 3，the research approach is more oriented towards ophthalmology. Within this cluster, the hotspots of research are the choroid, uvea, ciliary body. Immunohistochemistry, as a critical histopathological examination[19], appeals in this cluster. Also appearing is plaque radiotherapy, a common treatment method[20].
Cluster 5 (Purple): most keywords in this cluster mutation of BAP1 were noticed. Some specific genes such as BAP1, EIF1AX, and SF3B1 are related to uveal melanoma metastasis[21] and have prognostic value in UM. Many studies found loss of BAP1 in uveal melanoma metastasis may be mainly involved in the progression of uveal melanoma to an aggressive, metastatic phenotype[19,21]. A high prevalence of liver metastases, as a character of UM, is clustered here. Unlike clusters 2 and 3, there is an exact object of study, and the direction of study is essentially the same. Research on this gene is worth continuing and is expected to translate into clinical results.
Cluster 6 (Brown)：Since immunotherapy has dramatically changed the treatment approach to cutaneous melanoma[22], it's no surprise that immunotherapy has become the hotspot in the latest cluster. The drugs currently in the spotlight are ipilimumab, pembrolizumab, and nivolumab[23-25]. The biomarker, which is closely related to immunotherapy, appears in this cluster at the same time. In this cluster, the literature's average publication date is relatively new, and the study objects are relatively specific.
By comparing the WoSCC and PubMed databases, PubMed has fewer metadata to analyze, the aggregation effect is not apparent, and the connection is not strong enough to reveal the deep connection between individual studies in the field. This study's limitation is that the PubMed database tags are not uniform, resulting in lower quality of metadata acquisition compared to WoSCC. Therefore, the amount of literature covered in this study is not comprehensive.
In conclusion, the analysis of the UM literature for the past 20 years by using scientific data visualization tools permits researchers to find references for their potential research directions, to determine potential research collaborators, and to find proper journals for publishing their articles regarding UM. More importantly, it identifies current hotspots including immunotherapy, BAP1, and GNAQ, in UM research for the last two decades.
Since medical research's speed now exceeds the learning speed of general physicians and researchers, it is more important to use scientific tools to organize the unknown knowledge. Databases for standardized management of literature tags are more critical in the knowledge explosion era than knowledge itself. In addition, these results help clinicians to find the newest relevant literature and clinical information for state-of-the-art treatment concepts for UM.