INTRODUCTION: This study compared the diagnostic efficacies of cone-beam computed tomographic (CBCT) imaging and periapical radiography (PR) in the detection of retained separated instruments located at the apical third of filled root canals. METHODS: Sixty single-rooted extracted human teeth were instrumented to size #25 and were randomly divided to a simulated 2-mm #30 K-file (stainless steel or nickel-titanium) segment separation at the apical third of the canal (n = 40) or a control group without a separated instrument (n = 20). The canals were obturated to the separated instrument or the working length for the teeth without an instrument using gutta-percha with AH26 (Dentsply DeTrey GmbH, Konstanz, Germany) or Roth sealer (Roth International Ltd, Chicago, IL). The teeth were invested in a mandible model simulating the bone density and imaged using CBCT imaging and PR. The images were evaluated separately by 2 calibrated observers twice with an interval of 4 weeks. Cohen kappa was used to evaluate the observer agreement. Receiver operating characteristic curve analysis was used to evaluate the discrimination ability. RESULTS: The intraobserver kappa was 0.744 and 0.627, and between the observers, it was 0.593 and 0.275 for PR and CBCT imaging, respectively. Using PR, the mean sensitivity was 71.25%, and the specificity was 93.75%. Using CBCT imaging, the sensitivity and specificity were 41.25% and 71.25%, respectively. Although for PR the area under the curve values ranged between 0.75 and 0.91 (P < .05), for CBCT they ranged between 0.48 and 0.60 (P > .05), regardless of the instrument or the sealer type. CONCLUSIONS: PR performed better than CBCT imaging for the detection of retained separated instruments located at the apical third of extracted human root canal-filled teeth.