Purpose: This study compared the accuracy of detection of incomplete vertical root fractures (VRFs) in filled and unfilled teeth on cone-beam computed tomography images with and without a metal artefact reduction (MAR) algorithm. Materials and Methods: Forty single-rooted maxillary premolars were selected and, after endodontic instru- mentation, were categorized as unfilled teeth without fractures, filled teeth without fractures, unfilled teeth with fractures, or filled teeth with fractures. Each VRF was artificially created and confirmed by operative microscopy.The teeth were randomly arranged, and images were acquired with and without the MAR algorithm. The images were evaluated with OnDemand software (Cybermed Inc., Seoul, Korea). After training, 2 blinded observers each assessed the images for the presence and absence of VRFs 2 times separated by a 1-week interval. P-values<0.05 were considered to indicate significance. Results: Of the 4 protocols, unfilled teeth analysed with the MAR algorithm had the highest accuracy of incomplete VRF diagnosis (0.65), while unfilled teeth reviewed without MAR were associated with the least accurate diagnosis (0.55). With MAR, an unfilled tooth with an incomplete VRF was 4 times more likely to be identified as having an incomplete VRF than an unfilled tooth without this condition, while without MAR, an unfilled tooth with an incomplete VRF was 2.28 times more likely to be identified as having an incomplete VRF than an unfilled tooth without this condition. Conclusion The use of the MAR algorithm increased the diagnostic accuracy in the detection of incomplete VRF onimages of unfilled teeth.