Research Trends in the Application of Artificial Intelligence in Oncology: A Bibliometric and Network Visualization Study
Open Access
- 31 August 2022
- journal article
- research article
- Published by IMR Press in Frontiers in Bioscience-Landmark
- Vol. 27 (9) , 254
Abstract
Background: The past decade has seen major advances in the use of artificial intelligence (AI) to solve various biomedical problems, including cancer. This has resulted in more than 6000 scientific papers focusing on AI in oncology alone. The expansiveness of this research area presents a challenge to those seeking to understand how it has developed. A scientific analysis of AI in the oncology literature is therefore crucial for understanding its overall structure and development. This may be addressed through bibliometric analysis, which employs computational and visual tools to identify research activity, relationships, and expertise within large collections of bibliographic data. There is already a large volume of research data regarding the development of AI applications in cancer research. However, there is no published bibliometric analysis of this topic that offers comprehensive insights into publication growth, co-citation networks, research collaboration, and keyword co-occurrence analysis for technological trends involving AI across the entire spectrum of oncology research. The purpose of this study is to investigate documents published during the last decade using bibliometric indicators and network visualization. This will provide a detailed assessment of global research activities, key themes, and AI trends over the entire breadth of the oncology field. It will also specifically highlight top-performing authors, organizations, and nations that have made major contributions to this research domain, as well as their interactions via network collaboration maps and betweenness centrality metric. This study represents the first global investigation of AI covering the entire cancer field and using several validated bibliometric techniques. It should provide valuable reference material for reorienting this field and for identifying research trajectories, topics, major publications, and influential entities including scholars, institutions, and countries. It will also identify international collaborations at three levels: micro (that of an individual researcher), meso (that of an institution), and macro (that of a country), in order to inform future lines of research. Methods: The Science Citation Index Expanded from the Web of Science Core Collection was searched for articles and reviews pertaining exclusively to AI in cancer from 2012 through 2022. Annual publication trends were plotted using Microsoft Excel 2019. CiteSpace and VOSViewer were used to investigate the most productive countries, researchers, journals, as well as the sharing of resources, intellectual property, and knowledge base in this field, along with the co-citation analysis of references and keywords. Results: A total of 6757 documents were retrieved. China produced the most publications of any country (2087, 30.89%), and Sun Yat Sen University the highest number (167, 2.47%) of any institute. WEI WANG was the most prolific author (33, 0.49%). RUI ZHANG ranked first for highest betweenness centrality (0.21) and collaboration criteria. Scientific Reports was found to be the most prolific journal (208, 3.18%), while PloS one had the most co-citations (2121, 1.55%). Strong and ongoing citation bursts were found for keywords such as “tissue microarray”, “tissue segmentation”, and “artificial neural network”. Conclusions: Deep learning currently represents one of the most cutting-edge and applicable branches of AI in oncology. The literature to date has dealt extensively with radiomics, genomics, pathology, risk stratification, lesion detection, and therapy response. Current hot topics identified by our analysis highlight the potential application of AI in radiomics and precision oncology.Keywords
This publication has 123 references indexed in Scilit:
- Lung cancer probability in patients with CT-detected pulmonary nodules: a prespecified analysis of data from the NELSON trial of low-dose CT screeningThe Lancet Oncology, 2014
- The Increasing Incidence of Young-Onset Colorectal Cancer: A Call to ActionMayo Clinic Proceedings, 2014
- Quantitative Image Analysis of Cellular Heterogeneity in Breast Tumors Complements Genomic ProfilingScience Translational Medicine, 2012
- Performance of computer-aided detection of pulmonary nodules in low-dose CT: comparison with double reading by nodule volumeEuropean Radiology, 2012
- Radiomics: Extracting more information from medical images using advanced feature analysisEuropean Journal of Cancer, 2012
- Reduced Lung-Cancer Mortality with Low-Dose Computed Tomographic ScreeningNew England Journal of Medicine, 2011
- The Lung Image Database Consortium (LIDC) and Image Database Resource Initiative (IDRI): A Completed Reference Database of Lung Nodules on CT ScansMedical Physics, 2011
- Software survey: VOSviewer, a computer program for bibliometric mappingScientometrics, 2009
- Intraoperative Evaluation of Breast Tumor Margins with Optical Coherence TomographyCancer Research, 2009
- CiteSpace II: Detecting and visualizing emerging trends and transient patterns in scientific literatureJournal of the American Society for Information Science and Technology, 2005