Automatic Three-Dimensional Cephalometric Landmarking via Deep Learning

Abstract

The increasing use of three-dimensional (3D) imaging by orthodontists and maxillofacial surgeons to assess complex dentofacial deformities and plan orthognathic surgeries implies a critical need for 3D cephalometric analysis. Although promising methods were suggested to localize 3D landmarks automatically, concerns about robustness and generalizability restrain their clinical use. Consequently, highly trained operators remain needed to perform tedious manual landmarking. In this study, we aimed to train and evaluate a deep learning (DL) pipeline based on SpatialConfiguration-Net for automatic localization of 3D cephalometric landmarks on computed tomography (CT) scans. A retrospective sample of consecutive presurgical CT scans was randomly distributed between a training/validation set (n = 160) and a test set (n = 38). The reference data consisted in 33 landmarks, manually localized once by 1 operator (n = 178) or twice by 3 operators (n = 20, test set only). After inference on the test set, one CT scan showed "very low" confidence level predictions; we excluded it from the overall analysis but still assessed and discussed the corresponding results. The model performance was evaluated by comparing the predictions with the reference data; the outcome set included localization accuracy, cephalometric measurements and comparison to manual landmarking reproducibility. On the hold-out test set, the mean localization error was 1.0 ± 1.3mm, while success detection rates for 2.0, 2.5 and 3.0mm were 90.4%, 93.6% and 95.4%, respectively. Mean errors were -0.3 ± 1.3° and -0.1 ± 0.7mm for angular and linear measurements, respectively. When compared to manual reproducibility, the measurements were within the Bland-Altman 95% limits of agreement for 91.9% and 71.8% of skeletal and dentoalveolar variables, respectively. To conclude, while our DL method still requires improvement, it provided highly accurate 3D landmark localization on a challenging test set, with a reliability for skeletal evaluation on par with what clinicians obtain.

Discrepancies in Cephalometric Analysis Results between Orthodontists and Radiologists and Artificial Intelligence: A Systematic Review

“…The results of studies comparing artificial intelligence (AI) and traditional cephalometric analysis vary. Some authors highlight numerous advantages of AI, such as increased convenience, shorter analysis time, and high accuracy [48,53]. For instance, a study by Wang et al [54] suggests that AI analysis demonstrates high precision, with deviations of up to 2 mm, and a detection rate of up to 80%.…”

Accuracy of automated 3D cephalometric landmarks by deep learning algorithms: systematic review and meta-analysis

“…Description of the sample characteristics (e.g., age, gender, craniofacial characteristics) was not provided by all the articles. Only eight studies [29,[34][35][36][37][38][39][40] indicated the execution time of Fig. 1 Prisma flowchart for the papers' selection process the identification system, and four of them [34,36,38,39] also reported the duration of the training phase.…”

“…Only eight studies [29,[34][35][36][37][38][39][40] indicated the execution time of Fig. 1 Prisma flowchart for the papers' selection process the identification system, and four of them [34,36,38,39] also reported the duration of the training phase.…”

“…Only one study [40] indicated the inter-rater agreement. For eleven studies the reported outcome was the mean error and the standard deviation between manual and automatic landmarking and they were included in the meta-analysis [29,30,[33][34][35][36][37][38][39][40][41]. Four studies didn't report the standard deviation [31,32,42,43]; thus, they were excluded.…”

A novel coarse‐to‐fine computational method for three‐dimensional landmark detection to perform hard‐tissue cephalometric analysis

“…However, the testing dataset was significantly low in this experiment. Dot et al (2022) has performed DL-based experiment for automatic cephalometric landmark detection. Thirty-three landmarks were plotted automatically on 38 testing images to demonstrate the accuracy of the method.…”