Introduction: 3D imaging has become the standard for orthognathic surgery planning. Manual segmentation and landmark localization are the foundation for assessing deformities and planning of the corrective surgery. However, in 3D imaging, these manual procedures are complex and time consuming. Deep learning has emerged as a promising solution to automate this workflow.

Objective: To develop an automated segmentation and landmark localization workflow to reduce time and efforted required for orthognathic surgery planning using CBCT.

Methods: The dataset consisted of 57 presurgical Cone Beam CT (CBCT) scans. Manual segmentations were created for the mandible, maxilla, skin, mandibular canals (MC) left and right. Additionally, 43 landmarks (13 mandibular, 10 maxillary, 10 dental, and 10 surgical) were manually annotated. Automated landmark localization was approached as a segmentation task, with spherical segmentation of adjacent tissue with radii of 4mm, 5mm, or 6mm around landmarks. The dataset was split into 64% train, 16% validation 20% testset. Seven nnU-Net models were trained: one for segmentation and six (two models for each radius configuration) for landmark localization. Landmark positions were determined by calculating the center of mass of predicted segmentations. Performance was evaluated by comparing results to manual ground truth segmentations and landmark locations.

Results: The nnU-Net was successfully trained to identify five segmentations and 43 landmarks. The overall median [Q1 – Q3] volumetric dice coefficient (vDSC) was 0.91 [0.79 - 0.96] for the segmentation model. The overall median radial error (MRE) for the landmarking models using 4mm, 5mm and 6mm spherical segmentations (each based on two models) were 0.98 [0.58 - 1.60] mm, 1.03 [0.63 - 1.67] mm, and 1.08 [0.66 - 1.82] mm, respectively. The successful detection rate below the clinical acceptability threshold of < 2mm ranged from 78.6% to 81.6%. The inference workflow required 17.7 minutes per patient on average.

Conclusion: The nnU-Net achieved accurate craniomaxillofacial (CMF) anatomy segmentations and precise localization of landmarks, maintaining a clinical acceptability level with an error margin of less than 2mm. This fully automated workflow has the potential to enhance the efficiency of CMF planning.

Currently, morphological measurements on conventional radiographs are used to assess calcaneal fractures. The two most commonly used angles for determining the degree of depression and displacement of such a fracture are Böhler’s angle (BA) and the Critical angle of Gissane (CAG). However, conventional radiographs only provide limited amount of information and are increasingly being omitted, while computerized tomography (CT) scans are increasingly being used as they contain much more information regarding fracture configuration and displacement. It is therefore important that morphological measurements can be performed using CT-scans. The purpose of this thesis were to design a novel method to perform semi-automated morphological measurements based on 3D models of the calcaneus. The results of the novel method were compared with conventional 2D measurements of the same calcaneus.

119 lateral foot radiographs and CT-scan of the same foot were obtained. The data consisted out of three populations: 40 non-fractured, 39 fractured and 40 reconstructed calcanei. Based on the CT-scans, 3D models were reconstructed through interactive segmentation processes. Measurements on the radiographs and 3D models were performed in a specially designed Graphical User Interface. Three observers measured BA and the CAG, after which they repeated the measurements after a minimum of one week. The inter- and intra-observers reliability were calculated separately for all three populations, using the intra-class correlation coefficient (ICC). Additionally, the frequency of consensus given an allowed discrepancy was calculated.

The ICC for inter- and intra-observer reliability of the 3D measurements showed an improvement for BA and the CAG for all populations compared with the ICC of the 2D measurements. The ICC for inter- and intra-observer reliability of the 3D measured BA ranged from 0.97 to 1.00, indicating an almost perfect agreement between the observers. For the CAG, the inter- and intra-observer ICC ranged from 0.89 to 0.92 and from 0.87 to 0.96, respectively. To achieve 80% consensus for the BA among the observers, an allowable discrepancy of 20 degrees was needed for the 2D measurements, while only 5 degrees was needed for the 3D measurements. To achieve 80% of consensus for the 2D CAG measurements, an allowable discrepancy of at least 20 degrees was needed. For the fractured and reconstructed population, even an allowed discrepancy of 30 degrees was needed. For the 3D CAG measurements, 80% of consensus was achieved with an allowable discrepancy of 5 degrees, 20 degrees and 10 degrees for the non-fractured, fractured and reconstructed calcanei populations, respectively.

3D morphological measurements based on 3D models of the calcaneus showed better reliability and repeatability compared to the 2D measurements. However, the novel method of performing the measurements is not yet safe for use in clinical practice, as the measurements first should be validated. Nevertheless, the method is well suited to be employed as a reference for the development of further (semi-)automated methods to perform morphological measurements.