Atlas-based Autosegmentation

Elekta recently launched ABAS (Atlas-based Autosegmentation) 2.0 software, which includes the Simultaneous Truth and Performance Level Estimation (STAPLE) algorithm to increase contouring accuracy. ABAS—considered the industry gold standard for autosegmentation accuracy—deforms atlases of anatomy previously defined on a reference image onto a new patient image, creating a new structure set fit to the patient anatomy and enhancing planning efficiency.

“The STAPLE algorithm allows multiple atlases to be calculated against a single patient,” says Randy Larson, ABAS product manager. “The benefit is increased accuracy because the user is achieving the best result for each contour based on multiple sources. ABAS 2.0 also uses a graphics processor unit [GPU], which increases contouring speed by up to 50 percent. In addition, this release includes an improved user interface featuring new tools and functionality, as well as full DICOM service, to enhance workflow and file maintenance.”

With its deformable registration algorithms, ABAS saves physician and dosimetrist time by automatically contouring new image sets based on anatomy defined in the atlas. Users may further edit and refine the new image sets.

User-definable atlases allow clinicians to select an atlas that accurately reflects a given patient in a particular clinic. Atlases also can be patient-specific to account for change in size, shape or deformation during treatment. The algorithm completes multiple phases of refinement for increased accuracy.

Read more at http://www.elekta.com/healthcare_international_press_release_20071113.php.