Repeatability and reproducibility of landmarks-a three-dimensional computed tomography study

The aim of this study was to investigate the repeatability and reproducibility of the placement of anthropological cephalometric landmarks on three-dimensional computed tomography (3D CT) cranial reconstructions derived from volume data sets. In addition, the influence of the observer's experience on the repeatability of landmark setting was also explored. Twenty patients without any craniofacial deformity (11 females and 9 males; age range 6.1-16 years) were selected retrospectively from CT volume data sets already available from 695 patients of Dental Clinic 3, Orthodontics of Erlangen University Hospital. The CT examination was performed with the SOMATOM Sensation (R) 64 (Siemens AG Medical Solutions, Erlangen, Germany). The program VoXim (R) 6.1 (IVS Solutions AG, Chemnitz, Germany) was used for 3D reconstruction of the volume data sets. A total of 28 landmarks were examined in the skeleton module of the program VoXim (R) 6.1. The randomly sorted data sets were analysed by two orthodontists and two postgraduate students. Each data set was analysed twice by each observer at an interval of 3 weeks. The analysis of variance regarding random effects was used to calculate the intraobserver and interobserver components of standard deviation (SD) of depiction of individual landmarks as measures of repeatability and reproducibility, respectively. Median intraserial SD and interserial SD of 0.46 mm (range: 0.14-2.00 mm) and 0.20 mm (range: 0.02-2.47 mm), respectively, were obtained depending on the landmark and plane. This study included systematic analysis of extreme values (outliers) in the assessment of the quality of measurements obtained. Descriptive statistics revealed qualitative differences in the depiction of different landmarks. The landmarks nasion and infradentale revealed a minor SD in all three spatial coordinates with the smallest SD for infradentale (SD = 0.18 mm) in the transverse plane. However, no systematic trend was identified with regard to the influence of the observer's level of experience affecting the repeatability of landmark positioning. Thus, the repeatability and reproducibility of placements of landmarks with 3D CT were found to be acceptable for a majority of anatomical positions.