Exploring Watson's Role as an Intelligent Assistant in Radiology (Full Transcript)

Speaker 1: With advances in machine learning and artificial intelligence, a new role is emerging for machines as intelligent assistants to radiologists in their clinical workflows. But what systematic clinical thought process are these machines using? Are they similar enough to those of radiologists to be trusted as assistants? In this live demonstration, clinicians can select a case from various sub-specialties, attempt to make a diagnosis, and see how a work-in-progress Watson technology attempts to assist the same case. Watson uses sophisticated medical imaging, deep learning, and clinical inference technologies to analyze patient cases using a systematic clinical thought process. To experience the eyes of Watson, you can select a case for analysis. You can examine the imaging study and read the associated case description. Select the appropriate conclusion. To see how Watson would attempt this case in real time, select Ask Watson. Watson examines the case description first, analyzes text, and highlights the relevant clinical concepts found. It summarizes the findings and updates the evolving clinical inference. Next, it analyzes the imaging study systematically by successively looking for more and more meaningful features that could potentially be anomalous until it locates a potential lesion. For this, it uses a combination of low-level image processing to highlight suspicious regions that are then classified as potential masses using deep learning networks trained by prior anomalous data labeled by clinical experts. Watson then takes the clinical concepts derived from the imaging exam and case descriptions and begins its reasoning process using clinical knowledge. Paths are explored and scored in the knowledge graph while searching for related concepts and facts that lead to the specified conclusions starting from the chosen clinical concepts. The retained conclusion is the one which has overwhelming evidence from the number of high-scoring paths that lead up to it. The final clinical inference can be seen on the right. If you concur with Watson's inference, you can change your conclusion if you like. You can continue with another case or take our evaluation survey. Eyes of Watson is a joint effort by RSNA and IBM Research to show how machines of the future may be able to assist radiologists. Thank you for watching.