Medical Physics Seminar – Monday, March 3, 2014
Towards Fully-Automated Abdominal CT Image Interpretation
Ronald M. Summers, MD, PhD. (guest of Dr. Robert Jeraj)
Senior Investigator, Clinical Image Processing Service, Imaging Biomarkers & Computer-Aided Diagnosis Laboratory, Bethesda, MD
The need for decision support systems in radiology is growing given the dramatic increase in imaging utilization, intensity and workload. My laboratory at NIH focuses on the application of advanced image processing and machine learning techniques to provide decision support for radiology image interpretation. As an abdominal radiologist and CT subspecialist, I have chosen to focus my research on the development of decision support for abdominal CT image interpretation.
In this talk I will discuss our approach to achieving full automation of abdominal CT interpretation. Topics will include fully automated detection and segmentation of the major abdominal organs and their lesions, the spine and spine lesions and lymphadenopathy. I will show how fully automated quantitation of bone mineral density and visceral fat can be obtained from routine abdominal CT scans without additional radiation exposure to detect osteoporosis and patients at risk for metabolic syndrome. I will also present selected recent research in CT colonography computer-aided detection (CAD) including our work on using Internet-based crowdsourcing to gain insight into observer performance studies of CAD.