Dr. Gunvant Chaudhari presented on natural language processing at the RSNA Annual Meeting

By Dr. Gunvant Chaudhari, Diagnostic Radiology Resident

With gracious support from the Charles A. Rohrmann Endowment for Radiology Resident Educational Excellence, I had the privilege of co-presenting an educational session at the 2024 Radiological Society of North America (RSNA) Annual Meeting, in Chicago Illinois. Our presentation was titled, “Basics of Natural Language Processing in Radiology.” This session was a collaborative effort with Dr. Jae Ho Sohn, assistant professor of thoracic radiology at UCSF, and Sohn Research Lab members Dr. Tim Chen and Dr. Cody Savage.

Our presentation introduced attendees to foundational concepts in natural language processing (NLP) and its applications to radiology. We developed an interactive online Python notebook with theory and practical examples about NLP concepts relevant to radiology, such as preprocessing techniques and medical ontologies. The notebook also provided a hands-on demonstration of an example pipeline for identifying radiology reports with abnormal findings using basic NLP techniques.

One of the key takeaways from this session was the importance of translating complex technical topics into accessible insights for a primarily clinician audience. Engaging with attendees during the Q&A offered valuable perspectives on the specific needs and concerns of radiologists. Beyond the presentation, attending RSNA also allowed me to gain knowledge from various educational sessions and build connections with peers and mentors. Thank you to the Rohrmann Endowment for making this opportunity possible.