PhD Defence Sneha Mithun

Keywords: Artificial Intelligence, Natural Language Processing, Deep Learning, Electronic medical records, Lung Cancer, Nuclear Medicine
 

"Natural Language Processing for Classification and Clinical Concept Extraction from Imaging Reports in Oncology"

Cancer is the second leading cause of death worldwide. Advancements in cancer management and research involve vast amounts of data analysis. Most of this data exists as free text on the EHR systems. Imaging reports constitute one of the largest free text data in oncology. Different imaging modalities are used for diagnosis as well as prognostication at different stages of cancer management. The information present in the imaging reports may be utilized for the development of clinical decision support systems. The use of artificial intelligence and natural language processing for development of such systems requires large quantities of data. The extraction of such data may even be automated by using natural language processing. This study addresses the challenges of the availability of rich clinical data, technological impediments with making the data available for further research, lack of uniformity in data storage and presentation, and basic flaws in real-world data. This work also demonstrates the automated extraction of clinical concepts from the free text imaging reports and addresses the different methods for tackling imbalanced language datasets.

Click here for the live stream.