Co-op Locations

Faculty Mentors

Rajiv Gupta, M.D., Ph.D.

Assistant Radiologist
Massachusetts General Hospital

Instructor of Radiology
Harvard Medical School


Biography: Dr. Gupta earned his MD at Cornell University and his PhD in Computer Science at the State University of New York at Stony Brook. In addition to serving as the CIMIT Site Miner for MGH, he is the director of the MGH Ultra-high Resolution Volume CT Lab. An Assistant Professor in radiology at Harvard Medical School, Dr. Gupta’s clinical specialties include Cardiovascular and Neuroradiology. Prior to joining MGH, Dr. Gupta was a Computer Scientist at GE Global Research Center in Niskayuna, NY, conducting research in medical imaging, non-destructive evaluation of aircraft engine parts, and computer vision. He also served on the faculty of University of Southern California, Los Angeles, in the Department of Electrical Engineering Systems.

Research and Expertise: My research and clinical interests are in Neuro and Cardiac radiology. My research is focused on development and clinical applications of ultra-high resolution CT. In my lab, we are developing and testing a prototype CT system based on flat-panel X-ray detectors with a resolution of 150 microns. This system, when fully developed, has the potential to open a new window on human anatomy and physiology by enabling, in a single scanner, dynamic imaging of temporally evolving processes, image-guided interventions, and high-resolution computed tomography. This system has already resulted in several significant clinical results. For example, we have demonstrated that the trabecular structure of the bone in Anorexia Nervosa is significantly altered even when the conventional metrics for measuring it (e.g., DEXA) are normal. Furthermore, this change is irreversible and results in fundamentally altered biomechanics of the bone leading to significantly increased fracture risk in adult life, even after the disease has been fully treated. For dynamic CT imaging, we have developed a new technique and demonstrated its feasibility in an animal model of aneurysms. Algorithms for efficient and accurate reconstruction of dynamic data are under development. The paradigm of volumetric CT using a large area detector also enables image-guided interventions in the same scanner. To this end, we are developing a new genre of low-cost, X-ray transparent, disposable robots that can operate inside the bore of the scanner.