Advancements and future trends in machine learning for lung cancer: a comprehensive bibliometric analysis

Abstract

Background: In recent years, significant progress has been made in lung cancer screening, diagnosis, and treatment with the continuous development of machine learning (ML).

Methods: To systematically explore the evolution and core driving factors of ML in lung cancer research since 2004, we conducted a comprehensive bibliometric analysis of 1,826 academic papers retrieved from the Web of Science Core Collection.

Results: This study reveals that the USA is at the forefront of applying ML in lung cancer research. The institutional analysis indicates that Harvard University plays a key role as a leading institution in this field. In the author co-occurrence network analysis, Madabhushi Anant stood out as a significant contributor to the application of ML in lung cancer research. Additionally, journal co-occurrence analysis shows that the SCI REP-UK published the highest volume of papers in this area. It is worth noting that several prestigious medical journals, including NEW ENGL J MED, NATURE, and CA-CANCER J CLIN, have shown significant interest in this research field. The burst citation analysis of keywords and references indicates that research hotspots have evolved from early attention to "breast cancer" and "radiotherapy" (2004-2012) to a focus on "computer-aided diagnosis" (2013-2017). Since 2018, "texture analysis", "computer-aided detection", "survival prediction", and "radiomics" have emerged as new research trends.

Conclusion: As ML continues to be applied more extensively and deeply in lung cancer, "computer-aided detection," "survival prediction," and "radiomics" are emerging as vital areas, deserving more attention from researchers.

Keywords: Bibliometrics; Citespace; Lung cancer; Machine learning.