Landmark identification on direct digital versus film-based cephalometric radiographs: A human skull study.

摘要

The purpose of this study was to investigate differences in landmark identification on vertically scanned, direct digital and conventional (18 x 24 cm) cephalometric radiographs. Eight observers, all orthodontists or postgraduate orthodontic students, recorded 6 landmarks twice on 3 digital and 3 conventional cephalograms obtained from 3 human skulls in a standardized fashion. Digital images were displayed on a 15.1-in TFT monitor in 3:1 mode (20 x 26 cm). Recordings were transferred into standardized coordinate systems and evaluated separately for each coordinate. After correcting for magnification, precision was assessed with Maloney-Rastogi tests, and intraobserver and interobserver reproducibility was calculated from squared differences. Effective magnification was larger for the digital images (x, 13%; y, 12%). Significantly different (P <.05) precision was found for nasion (N), posterior nasal spine (PNS), sella (S), supraspinale (A), and orbitale (Or), but average differences were entirely below 1 mm. Interobserver and intraobserver reproducibility did not differ significantly between the 2 image modes. Squared differences were largest for PNS and Or in both modalities. Our results indicate comparable errors in landmark recording for both evaluated machines. However, these results must be considered in the context of the specific display conditions for digital images, because no general standard exists for this purpose.