AIM To compare the ability of cone-beam computed tomography (CBCT) and digital periapical radiographs (PR) to detect simulated tissue-occupied recesses in root canals. METHODOLOGY A standard canal was created in 30 extracted mandibular premolar roots. Each root was longitudinally split into buccal and lingual halves. In 20 teeth, a standard groove, 4mm in length, 0.5mm deep and 0.3mm wide, was prepared on each root half and filled with radiolucent plasticine (a modelling clay), simulating tissue-occupied buccal and lingual recesses. In the other 10 teeth, no grooves were cut. Each root was reassembled and filled with laterally compacted gutta-percha and sealer. PR and CBCT were used to detect the plasticine-filled grooves. The images were pooled and blindly evaluated by three calibrated examiners (A, B and C). A chi-square test was used to analyse the data. RESULTS Examiner A, B, C detected grooves on CBCT scans in 20, 20 and 23 teeth, respectively, of which 19, 18 and 19 were true positives. The accuracy of CBCT in diagnosing plasticine-filled grooves was 82-92% for three examiners, significantly higher than the accuracy of PR (30-33%; P<0.001). CONCLUSIONS Cone-beam computed tomography accurately detected simulated tissue-occupied buccal and lingual recesses.