Exploring Radiology Reports of Brain Cancer Patients Using the Advanced Large Language Model ChatGPT

In this study, radiology reports concerning brain cancer have been examined. These reports have been structured more effectively using the ChatGPT language model. The processes of topic modeling with NMF, summarization, and automatic classification of the textual reports have been conducted based on prompts given to the language model. The NMF topic modeling algorithm identified the most significant topics within the reports, text summarization simplified the reports for better comprehension by patients, and the automatic summarization contributed to decision support systems in the diagnostic and treatment processes of healthcare professionals. This paper shares the results obtained from these three crucial stages applied to the reports.