Kidney medicine meets computer vision: a bibliometric analysis

Chen Junren, Rui Chen, Liangyin Chen, zhanglei-scu, Wang-wei, Xiaoxi Zeng

TL;DR

This paper provides the first bibliometric analysis of computer vision (CV) applications in kidney medicine, revealing a rapidly growing research landscape dominated by deep learning for image segmentation. The findings highlight key trends, influential players, and future directions, offering valuable insights for researchers and clinicians.

What Problem Does This Paper Address?

Kidney diseases are a major global health concern, leading to significant morbidity and mortality. Early and accurate diagnosis, along with effective treatment, are crucial, especially given the strain on healthcare systems. Computer vision (CV) technology offers a promising avenue to enhance these processes by enabling machines to "see" and interpret medical images. However, the field of CV applied to kidney medicine is evolving rapidly, and a comprehensive understanding of its research landscape, emerging trends, and key contributors has been lacking. This paper aims to fill that gap by providing a bibliometric analysis to map out the current state and future trajectory of this interdisciplinary field.

Methods

The researchers conducted a comprehensive bibliometric analysis using data from the Web of Science Core Collection database, covering publications from January 1, 1900, to December 31, 2022. They employed a specific search strategy focusing on "kidney" and related terms combined with various CV techniques. The analysis involved:

  • Publication Trend Analysis: Tracking the annual number of publications to identify growth patterns.
  • Influence Analysis: Using metrics like Local Citation Score (LCS) and Global Citation Score (GCS), and their per-year equivalents (LCS/t, GCS/t), to assess the impact of publications, authors, and journals.
  • Network Analysis: Utilizing tools like VOSviewer and Python libraries (Matplotlib, Seaborn) to visualize:
  • Co-authorship Networks: To understand collaboration patterns among researchers.
  • Co-occurrence Networks: To identify key research themes and emerging hotspots based on keyword frequency.
  • Citation Maps: To visualize the relationships and evolution of influential research within the field.
  • Categorization of Methodologies: Classifying CV techniques into manual feature-based, machine learning-based, and deep learning-based approaches.

Datasets

The study did not introduce new datasets. Instead, it analyzed existing publications from the Web of Science Core Collection. The "datasets" in this context refer to the vast collection of research papers themselves, which were used as the source material for the bibliometric analysis. The analysis focused on the metadata and content of these publications to understand research trends and applications.

Key Insights and Ideas

  • Explosive Growth: The field of CV in kidney medicine has seen a dramatic increase in publications, particularly since 2016, indicating a surge in research interest and activity.
  • Deep Learning Dominance: Deep learning, especially for image segmentation tasks, is the leading CV technique in this domain, outperforming traditional machine learning in many complex applications.
  • Shift Towards AI: The research is increasingly driven by artificial intelligence (AI) principles, moving beyond basic image processing to more sophisticated analysis and diagnosis.
  • Expanding Research Scope: The focus is broadening from general kidney imaging to specific kidney-related objects (e.g., renal cells, glomeruli) and moving towards 3D imaging and more diverse data modalities.
  • US Leads, China Rising: The United States is the leading country in terms of publications and international collaborations, with China showing significant growth and collaboration.
  • Interdisciplinary Nature: The research thrives at the intersection of computer science, biomedical engineering, and medicine, with journals like IEEE Transactions on Medical Imaging and Medical Image Analysis being highly influential.
  • Collaboration is Key: While individual countries often collaborate internally, there's a strong trend towards international collaboration, particularly between the US and China, which is crucial for advancing the field.

Key Results

  • Publication Volume: The annual publication count exceeded 200 in 2021 and 2022, showcasing recent rapid growth.
  • Leading Institutions: The National Institutes of Health (NIH), Harvard University, and the University of California System were among the top institutions.
  • Top Countries: The US (470 publications), China (296 publications), and Germany (127 publications) led in publication output.
  • Influential Journals: Proceedings of SPIE, Lecture Notes in Computer Science (LNCS), and IEEE Transactions on Medical Imaging (TMI) published the most papers.
  • Dominant Techniques: Deep learning-based segmentation and machine learning-based texture analysis are the most frequently used techniques.
  • Keyword Trends: "Segmentation," "CT," "deep learning," and "kidney" were the most frequent keywords, highlighting the focus on deep learning for segmentation from CT images.
  • Methodology Evolution: CV algorithms are shifting from traditional feature engineering to AI, with deep learning showing greater influence than machine learning in citation analysis.

Limitations and Future Work

The authors acknowledge several limitations:

  • Language Bias: The study was limited to English-language publications, potentially missing important research from other languages.
  • Database Scope: Reliance solely on the Web of Science Core Collection might have excluded publications from other major databases like Scopus and PubMed.
  • Citation Metric Nuances: Citation counts, especially per year (GCS/t, LCS/t), might not fully capture the influence of older, foundational papers or newer, rapidly gaining traction works. Self-citation could also introduce bias.

For future work, the authors suggest:

  • Differentiating Publication Eras: Future studies should differentiate between early movers, mid movers, and late movers to provide a more equitable comparison of influence.
  • Strengthening Collaboration: Encouraging greater transnational collaboration, particularly with leading countries like the US and China, is essential for mutual benefit and advancing the field.
  • Addressing Data Challenges: Continued research is needed to overcome data scarcity and quality issues, especially for complex tasks like medical image segmentation, potentially through robust models and weakly supervised learning.
  • Exploring New Algorithms: The paper highlights the potential of Vision Transformers (ViT) and Vision Mamba for improved global information awareness and efficient modeling, suggesting these as avenues for future research.
  • Real-time Applications: Future work should focus on developing robust models for embedded surgical devices and improving real-time performance in medical image tasks.

Summary updated

Abstract

Background and objective: Rapid advances in computer vision (CV) have the potential to facilitate the examination, diagnosis, and treatment of diseases of the kidney. The bibliometric study aims to explore the research landscape and evolving research focus of the application of CV in kidney medicine research. Methods: The Web of Science Core Collection was utilized to identify publications related to the research or applications of CV technology in the field of kidney medicine from January 1, 1900, to December 31, 2022. We analyzed emerging research trends, highly influential publications and journals, prolific researchers, countries/regions, research institutions, co-authorship networks, and co-occurrence networks. Bibliographic information was analyzed and visualized using Python, Matplotlib, Seaborn, HistCite, and Vosviewer. Results: There was an increasing trend in the number of publications on CV-based kidney medicine research. These publications mainly focused on medical image processing, surgical procedures, medical image analysis/diagnosis, as well as the application and innovation of CV technology in medical imaging. The United States is currently the leading country in terms of the quantities of published articles and international collaborations, followed by China. Deep learning-based segmentation and machine learning-based texture analysis are the most commonly used techniques in this field. Regarding research hotspot trends, CV algorithms are shifting toward artificial intelligence, and research objects are expanding to encompass a wider range of kidney-related objects, with data dimensions transitioning from 2D to 3D while simultaneously incorporating more diverse data modalities. Conclusion: The present study provides a scientometric overview of the current progress in the research and application of CV technology in kidney medicine research. Through the use of bibliometric analysis and network visualization, we elucidate emerging trends, key sources, leading institutions, and popular topics. Our findings will offer valuable insights for future clinical practice and research directions on the use of CV in kidney medicine research.
BibTeX
@inproceedings{junren2024kidney,
  author = {Chen Junren and Rui Chen and Liangyin Chen and zhanglei-scu and Wang-wei and Xiaoxi Zeng},
  title = {{Kidney medicine meets computer vision: a bibliometric analysis}},
  booktitle = {International Urology and Nephrology},
  year = {2024},
  doi = {10.1007/s11255-024-04082-w},
}
Figure