RSNA 2005 

Abstract Archives of the RSNA, 2005


SSK17-07

Improved Electronic Colon Cleansing with Less-Stressful Bowel Preparation for Computer-aided Detection of Polyps in CT Colonography

Scientific Papers

Presented on November 30, 2005
Presented as part of SSK17: Physics (Computer-aided Detection with Colonography)

Participants

Jerome Z. Liang PhD, Presenter: Nothing to Disclose
Zigang Wang PhD, Abstract Co-Author: Nothing to Disclose
Bin Li PhD, Abstract Co-Author: Nothing to Disclose
Hongbing Lu PhD, Abstract Co-Author: Nothing to Disclose

PURPOSE

Our previous electronic colon cleansing (ECC), which was integrated into the commerical Viatronix V3D colon navigation system, has shown good performance in clinical trial as reported in the NEJM 2003. This work aims to improve the ECC.

METHOD AND MATERIALS

Twenty patient studies were selected retrospectively to demonstrate the improved ECC. These patients underwent a less-stressful bowel preparation with low-residue diet and oral contrast tagging of the colonic material. A single-slice spiral CT was used with 5mm collimation and 1mm reconstruction, covering the entire abdomen during a single breath-holding period. Both supine and prone scans resulted in 40 datasets. Among these datasets, nine contained at least one polyp larger than 5mm in diameter. The polyp sizes and positions were reported by optical colonoscopy and verified by the V3D navigation system. The datasets were first processed by our previous ECC, which labels concerned voxels into air, tagged material, and colon wall (neglecting the partial volume (PV) effect) and cleanses the lumen by grouping the air and tagged voxels. The PV boundaries between air and tagged material and between tagged material and colon wall are detected through ray tracing, which starts at a selected air voxel and moves along the gravity force. Then the datasets were processed by the improved ECC, which segments the tissue mixture in each voxel. All the voxels associated with air, tagged material and their mixture are identified for the lumen. The mixtures of colon wall with air and colon wall with tagged material are extracted for the mucosa layer. The processed results were assessed by both the V3D navigation system and computer aided detection of polyps (CADpolyp).

RESULTS

The improved ECC was confirmed by both visual judgment via the V3D navigation and visualization and our developed CADpolyp. For 100% sensitivity, the previous ECC retained 154 false positives (FPs) per study v.s. 148 FPs of the improved ECC for surface-based CADpolyp and 2.2FPs v.s. 2.0FPs for texture-based CADpolyp.

CONCLUSION

Improved ECC with less-stressful bowel preparation is demonstrated which is an essential step toward screening of polyps by CT colonography.

DISCLOSURE

J.Z.L.,Z.W.,B.L.,H.L.: This work was partly supported by NIH Grant #CA82402 of the National Cancer Institute. Dr. H. Lu is supported by Grant 30170278 of the National Nature Science Foundation of China.

Cite This Abstract

Liang, J, Wang, Z, Li, B, Lu, H, Improved Electronic Colon Cleansing with Less-Stressful Bowel Preparation for Computer-aided Detection of Polyps in CT Colonography.  Radiological Society of North America 2005 Scientific Assembly and Annual Meeting, November 27 - December 2, 2005 ,Chicago IL. http://archive.rsna.org/2005/4415157.html