Accuracy and reliability of tooth and root lengths measured on cone-beam computed tomographs

In this study, we evaluated the accuracy and reliability of tooth-length and root-length measurements derived from cone-beam computed tomography (CBCT) volumetric data.

CBCT scans were made of 7 fresh porcine heads. The scans were made with an i-CAT machine (Imaging Sciences International, Hatfield, Pa) at 0.2, 0.3, and 0.4 mm voxel sizes. Two film-acquired periapical radiographs were also taken of selected incisors and premolars, 52 of which (28 premolars, 24 incisors) were included in this study. By using Dolphin imaging software (version 10.5, Dolphin Imaging Systems, Chatsworth, Calif), the CBCT scans were oriented twice for each tooth (ie, 2 trials) using the mesial, distal, labial, and lingual cementoenamel junctions as reference points. Root and tooth lengths were derived from these points and compared with actual measurements of the teeth made with digital calipers after all surrounding bone had been carefully removed.

CBCT tooth-length and root-length measurements were not significantly different from the actual lengths; the mean differences were less than 0.3 mm. The periapical measurements significantly (P = 0.001) underestimated root lengths (mean difference, 2.58 mm) and overestimated tooth lengths (mean difference, 2.58 mm; P = 0.056). Mean differences between the 3 CBCT voxel sizes were all less than 0.25 mm. Within-trial method errors were almost 2 times greater for the periapical radiographs than for the CBCT scans. Between-trial method errors were greatest for the 0.4-mm CBCT scans, which were within 0.1 mm of the periapical radiograph method errors. The intraclass correlations for the periapical and CBCT measurements were all above 0.995.

CBCT scans are at least as accurate and reliable as periapical radiographs for tooth-length and root-length determinations.