OBJECTIVE
To identify research trends in glioma radiomics using topic modeling, to analyze development of this field and to determine areas with potential clinical and scientific significance.
MATERIAL AND METHODS
A comprehensive bibliometric and thematic analysis of literature data on glioma radiomics published between 2014 and 2024 was conducted. We screened the PubMed database using appropriate keywords. After preliminary screening of publications, we used topic modeling and bibliometric analysis to identify key research areas and development trends.
RESULTS
The number of publications devoted to glioma radiomics increased between 2014 and 2024. The key research areas included segmentation and extraction of features of glioma radiographic images, validation and reproducibility of machine learning models, molecular classification of gliomas, and survival prediction. These studies highlight the importance of model reproducibility and application of radiomics together with genomic data analysis and artificial intelligence methods.
DISCUSSION
Glioma radiomics demonstrates significant potential for clinical practice, particularly when integrated with molecular and genomic data. Despite rapid development of this field, there are challenges with model reproducibility and protocol standardization limiting their widespread adoption. Available studies demonstrate the need to use multiple-center data and improve interpretability of radiomics features. Future studies should focus on significant clinically applicable models and better integration of radiomics with artificial intelligence for personalized medicine.
CONCLUSION
Glioma radiomics is advancing towards creation of more significant and clinically applicable models. Validation of models on diverse datasets and integration of radiomics with molecular and genomic methods are key areas of future research. Topic modeling enabled identification of key research topics reflecting current state and promising areas of radiomics application in glial tumor research.
