Share Email Print
cover

Proceedings Paper

Graph-based surface extraction of the liver using locally adaptive priors for multimodal interventional image registration
Author(s): Samuel Kadoury; Bradford J. Wood; Aradhana M. Venkatesan; Roberto Ardon; James Jago; Jochen Kruecker
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

The 3D fusion of tracked ultrasound with a diagnostic CT image has multiple benefits in a variety of interventional applications for oncology. Still, manual registration is a considerable drawback to the clinical workflow and hinders the widespread clinical adoption of this technique. In this paper, we propose a method to allow for an image-based automated registration, aligning multimodal images of the liver. We adopt a model-based approach that rigidly matches segmented liver shapes from ultrasound (U/S) and diagnostic CT imaging. Towards this end, a novel method which combines a dynamic region-growing method with a graph-based segmentation framework is introduced to address the challenging problem of liver segmentation from U/S. The method is able to extract liver boundary from U/S images after a partial surface is generated near the principal vector from an electromagnetically tracked U/S liver sweep. The liver boundary is subsequently expanded by modeling the problem as a graph-cut minimization scheme, where cost functions used to detect optimal surface topology are determined from adaptive priors of neighboring surface points. This allows including boundaries affected by shadow areas by compensating for varying levels of contrast. The segmentation of the liver surface is performed in 3D space for increased accuracy and robustness. The method was evaluated in a study involving 8 patients undergoing biopsy or radiofrequency ablation of the liver, yielding promising surface segmentation results based on ground-truth comparison. The proposed extended segmentation technique improved the fiducial landmark registration error compared to a point-based registration (7.2mm vs. 10.2mm on average, respectively), while achieving tumor target registration errors that are statistically equivalent (p > 0.05) to state-of-the-art methods.

Paper Details

Date Published: 17 February 2012
PDF: 8 pages
Proc. SPIE 8316, Medical Imaging 2012: Image-Guided Procedures, Robotic Interventions, and Modeling, 83161O (17 February 2012); doi: 10.1117/12.911363
Show Author Affiliations
Samuel Kadoury, Philips Research North America (United States)
Bradford J. Wood, National Institutes of Health (United States)
Aradhana M. Venkatesan, National Institutes of Health (United States)
Roberto Ardon, Philips Medisys Research Lab. (France)
James Jago, Philips Healthcare (United States)
Jochen Kruecker, Philips Research North America (United States)


Published in SPIE Proceedings Vol. 8316:
Medical Imaging 2012: Image-Guided Procedures, Robotic Interventions, and Modeling
David R. Holmes; Kenneth H. Wong, Editor(s)

© SPIE. Terms of Use
Back to Top