Anno Graser MD, Presenter: Nothing to Disclose

Luca Bogoni PhD, Abstract Co-Author: Nothing to Disclose

Christoph Richard Becker MD, Abstract Co-Author: Nothing to Disclose

Maximilian Ferdinand Reiser MD, Abstract Co-Author: Nothing to Disclose

To evaluate the accuracy of a CAD system in CT Colonography of a screening population as compared to expert readers.

105 asymptomatic individuals (62 men, 53 women, mean age 59 years) underwent multidetector CT colonography (virtual colonoscopy) for colon cancer screening. All scans were performed using 16-detector row scanners (Somatom Sensation 16, Siemens Medical Solutions, Forchheim, Germany) at 0.75 mm collimation. Individuals were scanned in supine and prone positions without fecal tagging using room air for bowel distension. 1-mm axial slices were reconstructed for 3D endoluminal visualization. Two expert readers read the scans using a primary endoluminal approach. A prototype CAD system (polyp-enhanced view, PEV, Siemens Medical Solutions, Malvern, PA) was used as “second reader”. The system’s sensitivity for the detection of colorectal polyps was evaluated and the number of false-positive findings per dataset was recorded. The average running time of the CAD software was also documented. Polyps were classified as small (≤ 6mm), medium (7-9 mm) and large (≥ 10 mm).

In the screening population included in this study, a total of 98 polyps (small, 51; medium, 32; and large, 15 polyps) was found in 42 patients. CAD detected 69 of these polyps (overall sensitivity, 70 %) with an average of 3.8 false positives per dataset. Average CAD running time was 4.8 minutes. Sensitivity was 51 % (26/51) for small, 94 % (30/32) for medium and 87 % (13/15) for large polyps.

Computer-aided detection shows high sensitivity in detecting clinically significant polyps >6 mm in size with acceptable false-positive rates. These numbers are within the range of expert readers. The short calculation time of the CAD system makes it easily useable in routine MDCT colonography.

Graser, A, Bogoni, L, Becker, C, Reiser, M, Computer-aided Detection (CAD) in MDCT Colonography: Evaluation of the Performance of a Prototype System in More Than 100 Cases.  Radiological Society of North America 2005 Scientific Assembly and Annual Meeting, November 27 - December 2, 2005 ,Chicago IL. http://archive.rsna.org/2005/4409795.html