Volume-rendered AVM (left) and extracted vessels color-coded by subtree (right).

Supported by R01-CA67812 NIH-NCI and R01-EB000219 NIH-NIBIB. 3D-DSA image provided by Siemens.

Vessels are ubiquitous throughout the human body. A method of defining vessels from 3D image data and of defining the connections and flow patterns of these vessels would therefore be of use to multiple applications.

Dr. Stephen Aylward has developed methods of extracting tubular objects from any type of 3D image data. The approach can define even small vessels that approach the level of image resolution, and has been applied successfully to images obtained by MR, CT, ultrasound, 3D-DSA, and confocal microscopy. Once extracted, vessels can be linked together to form vessel trees.

Individual research projects (links at left) have used these vessel segmentations to form statistical atlases of human vascular anatomy, to assess disease (particularly cancer) in both human subjects and genetically engineered mice, and to help guide open and endovascular operations.

The figure above shows a complex tangle of abnormal blood vessels (AVM) extracted from a 3D-DSA. The ability to manipulate connected groups of vessels by color-coding or by turning groups "off" makes it possible to understand the vascular feeding patterns in a way not otherwise possible.

Aylward S, Bullitt E (2002) Initialization, noise, singularities and scale in height ridge traversal for tubular object centerline extraction. IEEE-TMI 21:61-75. <pdf>

Bullitt E, Aylward S, Smith K, Mukherji S, Jiroutek M, Muller K (2001) Symbolic Description of Intracerebral Vessels Segmented from MRA and Evaluation by Comparison with X-Ray Angiograms. Medical Image Analysis 5:157-169. <pdf>

Bullitt E, Aylward SR. Visualizing blood vessel trees in three dimensions: Clinical Applications (2005) Keynote speech SPIE 2005; Eckstein MP, Jiang Y (eds) Image Perception, Observer Performance, and Technology Assessment, Proceedings SPIE 2005 vol 5749 <pdf>.