Alexander W. Keedy BS, Abstract Co-Author: Nothing to Disclose

Rizwan Aslam MBChB, Abstract Co-Author: Nothing to Disclose

Chris Metchnikoff BS, Abstract Co-Author: Nothing to Disclose

Judy Yee MD, Abstract Co-Author: Research grant, General Electric Company

Le Lu PhD, Presenter: Employee, Siemens AG

To evaluate the accuracy of polyp measurement in a colon phantom using two-dimensional (2D), multiplanar reformation (MPR), and three-dimensional (3D) endoluminal displays. The effects of reader experience, phantom angulation, and tube current on polyp measurement accuracy were also assessed.

Computed tomography data was obtained on a colon phantom model containing 24 simulated polyps (5.8-41.0 mm; 14 spherical, 6 hemispherical, and 4 irregularly shaped). The phantom was scanned with a slice thickness of 1.25 mm, varying tube current (50 and 100 mAs), and varying angles relative to the z-axis of the phantom and the CT scanner (0, 45, 90, and 135 degrees). Three readers with different levels of expertise in CT colonography measured the maximum height and width of the 24 polyps using axial, coronal, and sagittal MPR displays and a 3D endoluminal cube display. Accuracy was assessed with student’s t-test and repeated measures analysis of variance. Interobserver agreement was assessed with Pearson’s correlation coefficient.

For all polyps, mean measurement error (± SD) for 2D MPR and 3D endoluminal displays was 0.9 mm ± 1.3 and 0.6 mm ± 0.7 (p < 0.001). The mean 2D and 3D error for spherical, hemispherical, and irregular polyps was 0.58 mm ± 0.44 and 0.38 mm ± 0.37 (p < 0.001), 0.76 mm ± 1.05 and 0.58 mm ± 0.54 (p = 0.015); and 2.48 mm ± 2.13 and 1.36 mm ± 0.98 (p < 0.001), respectively. Reader experience (p = 0.47), colon angulation (p = 0.26), and tube current (p = 0.13) did not have a statistically significant effect on measurement accuracy. Interobserver agreement was higher for 3D (p < 0.001) than for 2D measurements.

For all polyps, measurements obtained on the 3D endoluminal display were significantly more accurate and reproducible than those made on the 2D MPR displays. This effect was amplified for irregularly shaped polyps, highlighting the importance of performing polyp measurement on a 3D endoluminal cube display.

Measurement of polyps on CT colonography is more accurate when obtained on the 3D endolumninal cube display than 2D MPR display and should be the technique used to assess polyp size.

Keedy, A, Aslam, R, Metchnikoff, C, Yee, J, Lu, L, Determinants of Polyp Measurement Accuracy on CT Colonography.  Radiological Society of North America 2007 Scientific Assembly and Annual Meeting, November 25 - November 30, 2007 ,Chicago IL. http://archive.rsna.org/2007/5005032.html