Purpose: To improve the sensitivity of computed tomography ( CT) colonography in the detection of polyps by comparing the 3D reconstruction tool " colon dissection'' and endoluminal view ( virtual colonoscopy) with axial 2D reconstructions. Material and Methods: Forty- eight patients ( 22 M, 26 F, mean age 57 +/- 21) were studied after intra- anal air insufflation in the supine and prone positions using a 16- slice helical CT ( 16 x 0.625 mm, pitch 1.7; detector rotation time 0.5 s; 160 mAs und 120 kV) and conventional colonoscopy. Two radiologists blinded to the results of the conventional colonoscopy analyzed the 3D reconstruction in virtual- endoscopy mode, in colon-dissection mode, and axial 2D slices. Results: Conventional colonoscopy revealed a total of 35 polyps in 15 patients; 33 polyps were disclosed by CT methods. Sensitivity and specificity for detecting colon polyps were 94% and 94%, respectively, when using the " colon dissection'', 89% and 94% when using " virtual endoscopy'', and 62% and 100% when using axial 2D reconstruction. Sensitivity in relation to the diameter of colon polyps with " colon dissection'', `` virtual colonoscopy'', and axial 2D- slices was: polyps with a diameter > 5.0 mm, 100%, 100%, and 71%, respectively; polyps with a diameter of between 3 and 4.9 mm, 92%, 85%, and 46%; and polyps with a diameter < 3 mm, 89%, 78%, and 56%. The difference between `` virtual endoscopy'' and " colon dissection'' in diagnosing polyps up to 4.9 mm in diameter was statistically significant. Conclusion: 3D reconstruction software " colon dissection'' improves sensitivity of CT colonography compared with the endoluminal view.