Assessment of Computed Tomography Perfusion Research Landscape: A Topic Modeling Study

Area of Science:

• Medical Imaging
• Radiology
• Data Science in Medicine

Background:

• The exponential growth in scholarly articles presents challenges for researchers in synthesizing information.
• Computed tomography perfusion (CTP) has seen rapid advancements, necessitating a structured overview of its research landscape.
• Topic modeling offers a novel approach to analyze and understand large volumes of scientific literature.

Purpose of the Study:

• To apply topic modeling to the scientific literature of computed tomography perfusion (CTP).
• To identify and analyze the prevalent and trending research topics within CTP.
• To provide insights into the evolving focus of CTP research over time.

Main Methods:

• A comprehensive literature search was performed on the Scopus database (2000-2023).
• The BERTopic model was employed to derive and group research topics from identified CTP articles.
• Linear regression models were used to analyze topic trends, particularly for the 2020s.

Main Results:

• A total of 12 distinct topics were identified, including 'Tumor Vascularity', 'Stroke Assessment', 'Myocardial Perfusion', and 'Artificial Intelligence'.
• The most prevalent topics were 'Tumor Vascularity' and 'Stroke Assessment'.
• Emerging trends in the 2020s highlight 'Postprocessing' and 'Artificial Intelligence' as the most significant trending topics.

Conclusions:

• Topic modeling is an effective method for navigating and understanding the expanding CTP research field.
• The identified topics provide a structured overview of CTP applications and research focus areas.
• The increasing prominence of 'Postprocessing' and 'Artificial Intelligence' signals key future directions in CTP research.