Imaging Informatics

Richard Robb, Ph.D., is the leader of the team that developed “Analyze,” a scientific image analysis and visualization program that is widely is regarded as the tool for choice for research-related processing of 3-dimensional image data sets. He has collaborated with clinical investigators and basic researchers on many projects that require manipulation of 3-D data sets.

Dr. Robb is currently working to develop a seamless, user-transparent interface between Analyze and the “Insight Toolkit,” which an open-source software system developed by the National Library of Medicine to provide leading-edge segmentation and registration algorithms in two, three and more dimensions for imaging research applications.

Some other collaborative initiatives of Dr. Robb’s include:

• Enhanced tracked colonoscopy with Gianrico Farrugia, M.D., and William Sandborn, M.D., - seeking to develop a significant adjunctive feature in routine colonoscopy in which the gastro-endoscope is tracked in real-time in 6 DOF space
• Spine surgery planning with Paul Huddleston, M.D., and Michael Yaszemski, M.D., – seeking to develop improved image-guided interventional capabilities for facilitating optimal spine surgery planning for patients who have spinal abnormalities, including metastatic lesions in the vertebra
• MRI diffusion image analysis of tumor response to treatment with Tim Kaufmann, M.D., – quantitative characterization of these diffusion maps may be an early marker for graded tumor response to treatment, suggesting intra-treatment modification of dose, different pharmaceuticals or stopping altogether
• Intensity modulated radiation therapy with Janelle Molloy, Ph.D., – developing 4D intensity-modulated radiation therapy for cancer treatment
• Analysis of microscopic tumor growth patterns with Andrei Ougolkov, M.D., Ph.D., – developing quantitative image analysis of tumor growth patterns as may be revealed in 3D construction of serial sections of fixed specimens

Imaging-based Assessment of Tumor Progression
Bradley Erickson, M.D., Ph.D., is studying the specific question of how to present images for maximum sensitivity for brain tumor progression. He is also one of the funded subject matter experts in the caBIG in vivo imaging workspace, being a co-lead of the software special interest group; and a member of the team that was awarded one of the recent caBIG contracts for algorithm validation toolkit (AVT Phase 1).

Imaging Psychophysics
Steve Langer, Ph.D., is investigating optimum display options for soft-copy image interpretation. Trade-off of resolution and image contrast is being assessed. He is collaborating with Dr. Erickson to investigate image compression using a metric that aims to predict the response of the Human Visual System to lossy compression. The results of that work will be to show if the prediction accuracy of the JND HVS model is invariant across display types and anatomy. This would permit automatic, intelligent image processing that may not only compress images, but enhance conspicuity as well.