VII Center for Visual Informatics and Intelligence wsu
Home  NSF: CRI arrow Topics

Topics

Intra-Patient Supine-Prone Colon Registration in CT Colonography Using Shape Spectrum



Problem Definition


  • CT colonography (CTC) is a minimally invasive screening technique for colorectal polyps and colon cancer. Most CTC protocols acquire both prone and supine images to improve the visualization of the lumen wall and to reduce false positives. Comparisons between the prone and supine images can be facilitated by computerized registration between the scans.

Algorithms

  • Developed a fully automatic method for registering colon surfaces extracted from prone and supine images.
  • Uses shape spectrum to extract the shape characteristics which are employed as the surface signature to find the correspondent regions between the prone and supine lumen surfaces.
  • Applied the algorithm to the real CTC datasets and the experiments demonstrate the excellent accuracy of our registration results.
  • Integrate the registration component into the virtual colonoscopy (VC) software for FP reduction, which shows that it is an excellent tool aiding polyp diagnosis in CTC.

Results

  • In order to evaluate the performance of the registration we proposed,a doctor experienced in the interpretation of CTC cases evaluated the 2D and 3D visualizations of the colon by use of our software tool plugged into the VC software plateform by Yulonn Medical Imaging Company. The following figure is the illustration of false positive reduction. In (a), in supine position it has a false positive which the green arrow pointing at. In (b), in prone position it confirms that it is a pseudo-polyp but not a true polyp. (c) and (d) is the virtual colonoscopy of (a) and (b). (e) and (f) show another example which has false positives in prone position and can be confirmed as pseudo-polyps in supine position.
  • We compared our method with the deformable model method based on our datasets. The comparison result is shown in the following table. Our method outperforms the deformable model approach, and more importantly, our method does not create any artifacts resulting in pseudo-polyps but the deformable model would.

Related Publications


Top

    Contact Webmaster
    Updated by 09/28/2011