The Reproducibility of Landmark Identification on Digital Lateral Cephalometric Radiography

Objectives: to compare the reproducibility of landmark identification on digital hardcopy with monitor-displayed images of digital lateral cephalometric radiography. Methods: The subjects consisted of 50 randomly selected patients. Two observers identified the landmarks on two image modalities 3 times respectively. Each landmark was expressed as the coordinate (x,y). Statistical comparisons between the two image modalities and the two observers were performed using repeated measures analysis of variance. Results: There was no statistically significant error in repeated identification of landmark. Porion(x), Pogonion(y), R1(y) reach statistical significant level in inter-observer error. A statistically significant difference was observed from inter-observer than image modality. In the interaction between repeated identification of landmarks and inter-observer, error of Pterygoid point(x) was statistically significant. In the interaction among repeated identification of landmarks, inter-observer and image modality, error of Orbitale(x) was statistically significant. Conclusion: the landmark identification on monitor-displayed images of digital lateral cephalometric radiography is as reproducible as that on hardcopy images.