Image Interpretation

Abstract

Image interpretation is the core process of radiological workflow. Current visualization environments contain a set of tools to help in the annotation of relevant imaging findings. However, there still exist important challenges for interoperability between different platforms when working with annotated images. How the annotations and findings are reported is also evolving, moving from traditional descriptive texts towards item-based structured reports. Finally, thanks to the recent advances in the artificial intelligence science, specifically in machine learning algorithms it has been possible to implement a growing number of computer-aided detection solutions to assist radiologists in the image interpretation process. Image interpretation is under a process of paradigm shift, from traditional image reading through observation and free text reporting of the findings, towards the inclusion of new technologies in the loop such as computer-aided detection and diagnosis (CAD), imaging biomarker extraction, and structured reporting. The advance in interoperability between systems to standardize image annotation formats, together with the growing use of structured reporting and AI-assisted image reading, will shape radiology as one of the most relevant data sciences in the era of precision medicine.