This study investigated the use of cone-beam computed tomography (CBCT) as a diagnostic tool for identifying root canal systems (RCSs) when compared with images obtained by using charged coupled device (CCD) and photostimulable phosphor plate (PSP) digital radiography in vitro. Seventy-two extracted teeth were exposed with CCD, PSP, and CBCT radiography. Specimens included 24 maxillary molars, mandibular premolars, and mandibular incisors each. Tooth-type selections were based on the greater possibility of multiple RCSs. Three board-certified endodontists evaluated CCD and PSP images to determine the number of RCSs. CBCT images were used to establish "ground truth" for the comparisons and were evaluated by a board-certified oral and maxillofacial radiologist who determined the number of RCSs. Descriptive statistical analysis revealed that the endodontists consistently agreed among themselves. When compared with CBCT evaluation, the endodontists, on average, failed to identify 1 or more RCSs in 41% of the teeth with CCD and 40% of the time with PSP. CBCT evaluations identified an average of 3.58 RCSs per maxillary molar, 1.21 per mandibular premolar, and 1.5 per mandibular incisor. Evaluation of CCD images demonstrated an average number of 1.0 RCSs per mandibular incisor, 1.0 per mandibular first premolar, and 3.1 per maxillary molar. Evaluation of PSP images demonstrated an average number of 1.3 RCSs per mandibular incisor, 1.1 per mandibular first premolar, and 3.0 per maxillary molar. In summary, endodontist evaluators with either CCD or PSP methods failed to identify at least 1 RCS in approximately 4 of 10 teeth, which can result in a less optimal healing outcome if a missed RCS is left uninstrumented and unobturated.