A Bibliometric Analysis of 8,276 Publications During the Past 25 Years on Cholangiocarcinoma by Machine Learning

Abstract

IntroductionCholangiocarcinoma (CCA) is the second most common hepatic malignancy. Progress and developments have also been made in the field of CCA management along with increasing scientific publications during the past decades, which reflect topics of general interest and suggest the future direction of studies. The purpose of this bibliometric study is to summarize scientific publications during the past 25 years in the field of CCA using a machine learning method.Material and MethodsScientific publications focusing on CCA from 1995 to 2019 were searched in PubMed using the MeSH term “cholangiocarcinoma.” Full associated data were downloaded in the format of PubMed and extracted in the R platform. Latent Dirichlet allocation (LDA) was adopted to identify the research topics from the abstract of each publication using Python.ResultsA total of 8,276 publications related to CCA from the last 25 years were found and included in this study. The most type of publications remained little changed, while the proportion of clinical trials remained relatively low (7.24% as the highest) and, more significantly, with a further downward trend during the recent years (1.42% in 2019). Neoplasm staging, hepatectomy, and survival rate were the most concerning terms among those who are diagnosis-related, treatment-related, and prognosis-related. The LDA analyses showed chemotherapy, hepatectomy, and stent as the highly concerned research topics of CCA treatment. Meanwhile, conversions from basic studies to clinical therapies were suggested by a poor connection between clusters of treatment management and basic research.ConclusionThe number of publications of CCA has increased rapidly during the past 25 years. Survival analysis, differential diagnosis, and microRNA expression are the most concerned topics in CCA studies. Besides, there is an urgent need for high-quality clinical trials and conversions from basic studies to clinical therapies.

Research topics and trends of lumbar spondylolisthesis: a bibliometric study with machine-learning based text mining

“…Natural language processing (NLP) is a popular research field where human language could be decoded by machine learning, which has been adopted to analyze academic publications recently(29). LDA is one of the most widely used machine learning algorithms in NLP(18, 25, 26), as it is scalable, computationally fast and close to what the human mind assigns while decoding text words. In the study, publications of Topic 2 remained as the dominant topic from 1990 to 2010, when all kinds of radiographic equipment spread over the world gradually.…”

“…For example, bibliometrics analysis can quantify the contribution of a research field (including different countries, institutions, journals, or authors), and to identify the research trends or topics in a particular field(14). Nowadays, bibliometric analysis has been widely adopted in summarizing medical fields, like emerging techniques(15-17) and common disease(18-20). In spine field, bibliometric studies about cervical myelopathy(21), spinal stenosis(22), spinal cord injury(23), scoliosis(24) are thriving.…”

Research Trends and Hotspots on Connectomes from 2005 to 2021: A Bibliometric and Latent Dirichlet Allocation Application Study

“…The main topics of the connectome were classi ed using Latent Dirichlet allocation (LDA) analysis. LDA is an algorithm that can generate a glossary of terms and then classify the articles included therein by topics [10]. The R package "lda" was utilized to conduct an LDA analysis of the abstract of all publications included.…”

“…Investigating the quantitative relationship of the literature reveals the overview and research hotspots in a speci c eld [8]. Latent Dirichlet allocation (LDA) is a machine-learningbased algorithm that has been widely utilized as a bibliometric tool for identifying research topics and generating clusters of topic terms [9,10].…”

A Comprehensive Overview of the Parathyroid Tumor From the Past Two Decades: Machine Learning-Based Bibliometric Analysis

“…Besides the regular bibliometric analysis methods, machine learning is also developed to analyze human language, such as natural language processing. Among these methods, latent Dirichlet allocation (LDA) is most frequently applied in the scientific publication analysis by identifying specific themes and dividing documents into these themes ( 4 ).…”