Mapping Trends in Neurosurgery: A Natural Language Processing and Machine Learning Analysis of Spine, Functional, and Skull Base Surgery

Introduction: Neurosurgery is a rapidly advancing field with a growing rate of publications that reflect both emerging and declining research trends. To analyze these developments, we developed a topic modeling algorithm to elucidate trends in key topics within spine, functional, and skull base neurosurgery.

Methods: In April 2024, we queried the Web of Science database using the keyword “neurosurgery”, yielding a dataset of over 185,000 articles from five major journals. Textual data were pre-processed and analyzed using BERTopic, identifying frequently occurring topics with at least 50 articles. Topics were manually by neurosurgical subdiscipline, and linear regressions assessed trends in topic prevalence over time. We then compared prevalent topics by article count to scientifically relevant topics, normalized by citation count divided by average publication year.

Results: Analysis of 185,000 articles identified a total of 226 topics relating to spine, functional, neural engineering, and skull base surgery. Spinal neurosurgery demonstrated the strongest positive trend, with peripheral nerve topics surging since 2015 and minimally invasive techniques peaking in 2018. Brain-related topics, such as ICV drugs and epilepsy, and skull base topics, such as meningioma and vestibular schwannoma, also grew significantly after 2015. Overall, trend analysis revealed increased focus on neurosurgical residency, training programs, and peripheral nerve topics. Journals with a specific focus, like Journal of Neurosurgery: Spine, featured more subdiscipline-specific topics, while general neurosurgery journals reflected broader topics.

Conclusion : Trends across spine, functional, and skull base neurosurgery reveal evolving research priorities and advancements. Understanding these trends offers insight into current progress and future directions in neurosurgery, highlighting the need for natural language processing techniques to keep pace with the expanding literature and support the field’s advancement.