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

INTRODUCTION: Parathyroid tumor, in particular carcinoma, is fairly rare among neoplasms of the endocrine system, unlike its benign counterpart. However, there is no bibliometric analysis in the field of parathyroid tumors comprehensively summarizing and discussing a large number of publications by a machine learning-based method. MATERIALS AND METHODS: Parathyroid tumor-related publications in PubMed from January 2001 to December 2020 were searched using the MeSH term “parathyroid neoplasms”. Latent Dirichlet allocation was adopted to identify the research topics from the abstract of each publication using Python. RESULTS: A total of 3,301 parathyroid tumor-associated publications were identified from the past 20 years, and included in further analyses. Research articles and case reports occupied the most proportion of publications, while the number of clinical studies and clinical trials decreased, especially in recent years. Technetium Tc 99m sestamibi was most studied among the diagnosis-related MeSH terms, while parathyroidectomy was among the treatment-related MeSH terms. The Latent Dirichlet allocation analyses showed that the top topics were (99m)Tc-MIBI imaging, parathyroidectomy, gene expression in the cluster of diagnosis research, treatment research, and basic research. Notably, scarce connections were shown between the basic research cluster and the other two clusters, indicating the requirements of translational study turning basic biological knowledge into clinical practice. CONCLUSION: The annual scientific publications on parathyroid tumors have scarcely changed during the last two decades. (99m)Tc-MIBI imaging, parathyroidectomy, and gene expression are the most concerned topics in parathyroid tumor research.