Visualization of Brain Structure for Surgical Planning

The 2009 IEEE Visualization Contest is designed to foster innovation in visualization and analysis techniques, and to encourage new applications of visualization research to real-world imaging data. With this year's emphasis on medical imaging, we hope to attract medical image analysis researchers who haven't previously participated in the Visualization conference, and to highlight the utility and relevance of visualization practices to effective medical image analysis.

The focus of the IEEE Visualization Contest is visualization of brain structure for surgical planning of tumor resection. Imaging data from two patient cases have generously been made available by Alexandra J. Golby, M.D, Department of Neurosurgery, Brigham and Women's Hospital, Harvard Medical School. The brain surgery to remove their tumors (resection) was planned in part according to MRI scans that measured different tissue properties and structures.

In addition to standard (scalar-valued) structural scan, patients were also scanned with Diffusion MRI. Diffusion MRI (dMRI) is able to detect the directional organization of white matter tissue in the central nervous system. A popular method of dMRI analysis and visualization is tractography, which follows directions of highest diffusivity to trace out coarse models of axonal pathways in the white matter. One particularly promising application of dMRI is in neurosurgical planning. Neurosurgeons face the task of understanding the white matter structure around the tumor, how it connects elsewhere in the brain, how to access the tumor during surgery, and how to remove as much of the tumor as possible without damaging brain function by impinging on healthy tissue.

A standard way to locate particularly important areas of the brain is by Functional MRI (fMRI), which detects slight changes in the brain's blood supply to infer which cortical areas are activating. We will supply fMRI scans for both patients, as well as registration information to connect the structural, dMRI, and fMRI data.

This is an application domain at the cutting edge of scientific visualization research:

Neurosurgical visualization broadly intersects with clinical medical research and its detailed neuroanatomical terminology. dMRI analysis (especially tractography) is its own active research area, and there is no ground truth for brain connectivity from vivo imaging, so ideally the methods can reveal something about the uncertainty inherent in the results.

Neurosurgical planning exemplifies the use of visualization for open- ended data exploration and complex decision making. The purpose is not to find a hidden object or answer a single question, but to design a method (hopefully interactive) for detecting and depicting neuroanatomical structures.

Effective methods will arise from the careful synthesis of algorithmic analysis and visual representation; there is little value in trying to pack all the image information into a single dense visualization, because this would not illuminate the essential anatomic and structural relationships.

Diffusion MRI visualization has already drawn from the rich vocabulary of scientific visualization techniques, and much great work on dMRI visualization and analysis has already been published at IEEE Visualization in the past. This challenge is an exciting opportunity to demonstrate how further innovations, refinements, and applications of visualization and interaction techniques can benefit a compelling medical application, as demonstrated on real-world clinical data.

Please see additional pages about the data sets and formats here and goals, submission, and learning more here. Those planning to submit are strongly encouraged to subscribe to the mailing list here so that you will receive updates and clarifications as they are sent out.

The exact nature and number of prizes awarded to the best participants is being determined.

Recognizing the fundamental importance of publications for academic careers, we are exploring pathways to accelerate converting a winning contest entry (or entries) into a publication.

Amit Chourasia, University of California, San Diego
Gordon Kindlmann, University of Chicago

The deadline for submissions is to be announced.

The details of submission procedures and the required materials are being determined.

All participants will be informed of the contest results prior to the close of early registration for IEEE VisWeek 2009.

For more information, please visit the contest website.