Reproducibility of three-dimensional posterior cranial base angles using low-dose computed tomography

One of the key aspects of three-dimensional (3D) craniofacial cephalometry is the measurement of posterior cranial base angle as this area is deeply involved in craniofacial development. The purpose of our retrospective study was to define the best reproducible 3D posterior cranial base angles among five 3D angles transposed from 2D cephalometry (Cousin, BL1 of Ross and Ravosa, Bjork, Delaire, CBA4 of Liberman) and seven 3D angles based on physical anthropology studies and on new concepts (R1 to R7). The null hypothesis was that all 3D posterior cranial base angles were equally reproducible. We used a preoperative low-dose computed tomography (CT) data from 20 adult patients undergoing orthognathic surgery after approval by local ethical committee. Two independent observers performed two series of 23 3D landmark identifications on 3D CT surface rendering of each patient using Maxilim software. Then, the same observers performed twice 3D cephalometric analyses (23 landmarks, 4 midpoints, 19 planes) that provided the automatic measurement of 12 posterior cranial base angles. Inter-observer correlation coefficient varied from 0.545 (Cousin) to 0.695 (CBA4 of Liberman) and from -0.177 (R2) to 0.827 (R4). The null hypothesis was rejected. The most reproducible angle was 3D angle R4 based on "basion," "superior optic" (right, left), and "crista galli inferior" landmarks. R4 angle might be used as reference 3D posterior cranial base angle in further clinical studies involving 3D cephalometry as a diagnostic tool for orthodontics and for orthognathic surgery.