3D cephalometric landmark detection by multiple stage deep reinforcement learning

The lengthy time needed for manual landmarking has delayed the widespread adoption of three-dimensional (3D) cephalometry. We here propose an automatic 3D cephalometric annotation system based on multi-stage deep reinforcement learning (DRL) and volume-rendered imaging. This system considers geometrical characteristics of landmarks and simulates the sequential decision process underlying human professional landmarking patterns. It consists mainly of constructing an appropriate two-dimensional cutaway or 3D model view, then implementing single-stage DRL with gradient-based boundary estimation or multi-stage DRL to dictate the 3D coordinates of target landmarks. This system clearly shows sufficient detection accuracy and stability for direct clinical applications, with a low level of detection error and low inter-individual variation (1.96 ± 0.78 mm). Our system, moreover, requires no additional steps of segmentation and 3D mesh-object construction for landmark detection. We believe these system features will enable fast-track cephalometric analysis and planning and expect it to achieve greater accuracy as larger CT datasets become available for training and testing.

Total Face Approach (TFA) 3D Cephalometry and Superimposition in Orthognathic Surgery: Evaluation of the Vertical Dimensions in a Consecutive Series

Background: Cephalometry is fundamental in diagnosis, analysis, and planning of orthodontic-surgical treatment as it reveals skeletal relationship between the upper and lower jaw as well as facial aesthetic parameters. Nevertheless, 3D cephalometry has still not become the exam of choice in orthognathic treatment even though today CBCT (Cone Beam Computed Tomography) is routinely used in other branches of dentistry. Methods: In a sample of 13 patients undergoing bimaxillary orthognathic surgery a chin-vertex CBCT exam was prescribed prior to orthodontic treatment (OT) and 12 months after surgery (T1). The DICOM files uploaded to MaterialiseSimplant Ortho software pro 2.1 (Materialise Co., Leuven, Belgium) were analyzed following the multiplane 3D Total Face cephalometry protocol (TFA). Results: Results comparing pre-op and post-op TFA 3D cephalometry, were then evaluated considering reference values reported in literature. The CBCT, carried out pre- and post-surgery, were subsequently analyzed employing the superimposition method using cranial base as reference. The aim of this study is to evaluate the advantages and disadvantages of the two methods in orthognathic surgery. Conclusions: Multiplane 3D TFA allows the clinician to locate where major or minor skeletal discrepancies are found with respect to ideal parameters and is also useful in classifying skeletal intermaxillary relation. The superimposition method is highly intuitive but does not provide information on the quantity and location of osteotomic movement.

Proposition of local automatic algorithm for landmark detection in 3D cephalometry

This study proposes a new contribution to solve the problem of automatic landmarks detection in three-dimensional cephalometry. 3D images obtained from CBCT (cone beam computed tomography) equipment were used for automatic identification of twelve landmarks. The proposed method is based on a local geometry and intensity criteria of skull structures. After the step of preprocessing and binarization, the algorithm segments the skull into three structures using the geometry information of nasal cavity and intensity information of the teeth. Each targeted landmark was detected using local geometrical information of the volume of interest containing this landmark. The ICC and confidence interval (95% CI) for each direction were 0, 91 (0.75 to 0.96) for x- direction; 0.92 (0.83 to 0.97) for y-direction; 0.92 (0.79 to 0.97) for z-direction. The mean error of detection was calculated using the Euclidian distance between the 3D coordinates of manually and automatically detected landmarks. The overall mean error of the algorithm was 2.76 mm with a standard deviation of 1.43 mm. Our proposed approach for automatic landmark identification in 3D cephalometric was capable of detecting 12 landmarks on 3D CBCT images which can be facilitate the use of 3D cephalometry to orthodontists.

Effects of Changes in the Frankfort Horizontal Plane Definition on the Three-Dimensional Cephalometric Evaluation of Symmetry

The plane formed by the intersection of bilateral porions (PoR and PoL) and left orbitale (OrL) is conventionally defined as the Frankfort horizontal (FH) plane. We aim to test the influence of the FH plane definition on a 3D cephalometric assessment. We selected 38 adult patients (20 males, 18 females; average age: 22.87 ± 5.17 years) without any gross asymmetry from retrospective records and traced and analyzed their cone-beam computed tomographic images. The findings were categorized into the following four groups: FH1: conventional; FH2: PoR, PoL, right orbitale (OrR); FH3: OrR, OrL, PoL; FH4: OrR, OrL, PoR. The average menton (Me) deviation from the MSP was statistically significant for the FH1 group (0.56 ± 0.27 mm; p < 0.001), compared to the FH3 (1.37 ± 1.23 mm) and FH4 (1.33 ± 1.16 mm) groups. The spatial orientation level (SOL) of the FH plane showed a marked difference (p < 0.05) between the FH2 (0.602° ± 0.503°) and FH4 (0.944° ± 0.778°) groups. The SOL of the MSP was comparatively small (p < 0.001) for FH2 (0.015° ± 0.023°) in comparison to both FH 3 (0.644° ± 0.546°) and FH 4 (0.627° ± 0.516°). Therefore, the FH plane definition can significantly influence the interpretation of cephalometric findings. Future studies should focus on standardization to improve the reliability and reproducibility of 3D cephalometry.

Three-Dimensional Cephalometric Analysis Using Computed Tomographic Imaging

Successful outcomes from the use of orthodontic devices are underpinned on their effective anchorage and the loading that they apply to the underlying facial structures. Anchorage plays an important role in determining the point of application of the corrective forces and subsequently the orientation of the resultant of these forces, which in-turn governs the outcome of treatment. Therefore, patient-specific design of anchors and their placement may benefit significantly from personalization using patient-specific and three-dimensional (3D) cephalometry. 3D cephalometry is therefore a first step to personalization of orthodontic treatment. In this feasibility study, we demonstrate the viability a novel image processing and surface analysis pipeline to quantify facial symmetry about the mid-sagittal facial plane, which may offer insight into optimal placement and orientation for implantation of orthodontic anchors, starting with patient-specific cone beam computed tomography (CBCT) images. Typical assessments of geometrical features/attributes of face include size, position, orientation, shape, and symmetry. Using 3D CBCT images in the DICOM image format, skull images were first segmented using a basic iso-contouring approach. To quantify symmetry, we split the skull along the mid-sagittal plane and used an iterative closest point (ICP) approach in order to rigidly co-register the left and right sides of the skull, optimizing for rotation, translation and scaling, after reflection of one half across the mid-sagittal plane. This was accomplished using an in-house plugin is developed for the open-source visualization toolkit (VTK) based 3D visualization tool, Paraview (Kitware Inc.). Finally, using a signed regional distance mapping plugin we were able to assess the regional asymmetry of regions of the skull (e.g. upper and lower jaw – specific targets for therapy) using colormaps of regional asymmetry (in terms of left-v/s-right side surface distance) and visualized the same as vector glyphs. The direction of these vectors is synonymous with anticipated regional forces required in order to achieve left-right symmetry, which in-turn may have value in surgical planning for orthodontic implantation. In sum, we demonstrate a workflow for computer-aided cephalometry to assess the symmetry of the skull, which shows promise for personalized orthodontic anchor design.