The Impact of AI on Radiology Reporting

For decades, the radiological reporting process has been roughly the same. Of course, the introduction of PACS and digitization changed many things in radiology, but the radiological report—the final product of the examination—did not change. Last decades structured reporting and standardization are being introduced to improve the value of the radiological report. The introduction of artificial intelligence (AI) in radiology enhanced the research and implementation of all kinds of imaging tools in clinical practice. Again, the radiological report did not change. However, since the introduction of Natural Language Processing (NLP) models, it has become possible to enhance the radiological report process by using AI. In particular, a subset of NLP models, the Large Language Models (LLMs) are viewed with great interest since these models are capable of text understanding as well as text generation and can be used in the reporting process. Such tools can improve the radiological report by for instance by providing text or contextual suggestions, checking report completeness or adding useful information. Furthermore, it can also help with reusing (clinical) textual data. This chapter will focus on the radiological reporting process and how natural language processing can enhance the radiologic report.