Objective: In this study the apical debris score regarding the rotary use of the CanalBrush (Coltène-Whaledent) with three different tip variations was compared to that of sole needle irrigation. Method and Materials: Sixty straight roots of human teeth with one canal were instrumented to size 45/.04. Roots were split into halves, reassembled, and embedded in model resin to form separable tooth models. Artificially manufactured dentin debris was packed into the apical 2 mm of the roots to form a relatively consistent debris challenge to be assessed. Then, the debris of each root half was evaluated apically using a modified Hülsmann-score under a light microscope. After remounting the models, root canals were irrigated with NaOCl 3% (2 mL): group 1 with sole needle irrigation, in groups 2 to 4 additionally agitated with different versions of the CanalBrush for 30 s at 600 rpm; group 2, one straight bristle ("old"); group 3, two additional diagonal bristles ("new"); group 4, central straight bristle of the new version removed ("experimental"). Following this, all root halves were reevaluated. Results: Irrigation reduced debris scores in all groups (Wilcoxon tests, P .001). Mean debris scores were reduced from a pre-irrigation score of about 4 for all groups to a score of around 2 to 3 in groups 1 and 2, and to scores around 1 in groups 3 and 4. Samples of groups 3 and 4 were significantly cleaner than of groups 1 and 2 (Mann-Whitney tests, P .05). Conclusion: CanalBrushes with diagonal bristles achieved an improved apical cleaning ability. Keywords: apical cleanliness, artificial dentin debris, CanalBrush, Hülsmann-score, rotary agitation, separable tooth model