Semi-Automated Three-Dimensional Condylar Reconstruction:

Automatic segmentation and visualization of cortical and marrow bone in mandibular condyle on CBCT: a preliminary exploration of clinical application

OBJECTIVES

To develop a deep learning-based automatic segmentation method for cortex and marrow in mandibular condyle on cone-beam computed tomography (CBCT) images and explore its clinical application.

METHODS

825 condyles of 490 CBCT images from 3 centers of Stomatology hospital affliated to Zhejiang University School of Medicine were collected. A deep learning model was developed for simultaneous segmentation of cortex and marrow in mandibular condyle. It included a region of interest extraction network and a segmentation network based on 3D U-net, with modifications made to improve the segmentation boundaries. To evaluate its clinical potential, the model's segmentation efficiency and accuracy were compared with those of both junior and senior oral and maxillofacial radiologists. Additionally, the model's ability to assist junior radiologists in diagnosis through visualization and quantitative analysis of the generated 3D model was also assessed.

RESULTS

The Dice similarity coefficient of the deep learning model was 0.901 (cortex), 0.969 (marrow), and 0.982 (entire condyle). Hausdorff distance was 0.755 mm (cortex), 0.826 mm (marrow), and 0.760 mm (entire condyle). The model outperformed radiologists across all segmentation metrics, completing the task in merely 15.06 s. With the assistance of visualization and quantitative analysis generated from the model's segmentation, the diagnostic accuracy of junior radiologists significantly improved.

CONCLUSIONS

The proposed deep learning-based model achieved accurate and efficient segmentation for mandibular condylar cortex and marrow. It possessed capability to generate precise 3D models, facilitating visual quantitative measurement and aiding in the diagnosis of condylar bony changes. This model holds potential for clinical applications in orthognathic surgery, orthodontic treatment, and other TMJ-related interventions.

Evaluating cranial landmarks for yaw orientation in natural head position: a 3D study

OBJECTIVES

To assess the symmetry of various cranial anthropometric points used as references for yaw orientation in the natural head position (NHP), relative to the mid-sagittal plane.

MATERIALS AND METHODS

A prospective analysis using tomography data from 55 patients was conducted. Radiopaque markers, placed on patients in NHP, facilitated head position recording in three planes, with subsequent digital transfer for orientation analysis. Symmetry of eight points (zygomaxillare, zygion, ectoconchion, frontozygomatic suture, stephanion, porion, mastoidale, condylion laterale) was measured against the mid-sagittal plane.

RESULTS

Significant asymmetry was observed in the stephanion, frontozygomatic suture, and ectoconchion points (p < 0.05). No significant differences were found in the symmetry of other points (p > 0.05).

CONCLUSIONS

Findings suggest that stephanion, frontozygomatic suture, and ectoconchion are unreliable for yaw orientation in NHP. Other points, combined with clinical measurements, may offer better reliability.

Anatomical position of the mandibular condyle after open versus closed treatment of unilateral fractures: A three-dimensional analysis

This study aimed to compare open and closed treatment for unilateral mandibular condyle neck and base fractures by final three-dimensional (3D) condylar position at 6 months follow-up. 3D position was associated with mandibular functioning and pain. A total of 21 patients received open (n = 11) or closed (n = 10) treatment. 3D positions were assessed on cone-beam computed tomography scans. Volume differences, root mean square, translations, and rotations were obtained related to the pursued anatomical position and compared between treatment groups by the Mann-Whitney U test. The 3D position parameters were associated with the maximum interincisal opening (MIO), mixing ability test (MAT), Mandibular Function Impairment Questionnaire (MFIQ), and pain based on Spearman correlation coefficients (rs). Translation in the medial-lateral direction was smaller after open treatment (P = 0.014). 3D position was not associated with the MAT; however, worse position was associated with a smaller MIO. A larger pitch rotation was associated with a worse MFIQ (rs = 0.499, P = 0.025). Volume reduction of the affected condyle was associated with more pain (rs = -0.503, P = 0.020). In conclusion, after unilateral condylar fractures, worse 3D position is associated with a smaller mouth opening and worse patient-reported outcomes. This is independent of the chosen treatment, despite a better anatomical reduction after open treatment.

Impact of bruxism on craniomandibular morphology: A cone-beam computed tomographic study

OBJECTIVE

The main goal of this work was to address craniomorphological characteristics of patients with bruxism when compared to those without bruxism using cone beam computed tomography (CBCT) imaging.

METHODS

Seventy CBCT images of an equal number (n = 35) of orthodontic patients with and without bruxism (age range, 18-44 years) were retrospectively analyzed. Sagittal evaluation, mandibular shape, and skeletal asymmetry were systematically assessed in both groups.

RESULTS

Significant differences (p < 0.05) were observed between groups for Right Articular Fossa (AF) - Axial Plane (AP) and Left AF - AP (B > NB), Right Gonial Angle (GA), Left GA, Sella-Nasion and Occlusal Plane (B < NB). Age (13%), Condyle-Gonion (18%), AF to AP (67%), and Mental-AP (16%) had the greatest impact on bruxers.

CONCLUSION

In this study, the CBCT 3D image showed significant differences in craniofacial morphology, particularly in the mandibular structure of bruxers compared with non-bruxers.

Analysis of the Volumetric Asymmetry of the Mandibular Condyles Using CBCT

Objectives

The aim of this study was to analyse volumetric asymmetries between the right and left condyles in relation to age, gender, and dental status.

Materials and methods

A retrospective analysis of 150 cone beam computed tomography (CBCT) scans was conducted. A single investigator performed the volumetric analysis of the CBCT scans using Vesalius 3D software. The volumetric data were analysed in relation to the gender, age, and dental status.

Results

The mean right condylar volume was significantly higher (P < .01) than the left condylar volume. Right and left condylar volumes were significantly higher (P < .01) in male study participants when compared to female study participants. There was no significant difference (P = .47) in the volumetric asymmetry between the male and female study participants. The volumetric asymmetry was significantly higher (P < .01) in the older age groups when compared to the younger age groups. The volumetric asymmetry was significantly higher (P < .01) in the partially and completely edentulous patients when compared to the dentate study participants. The condylar volume on the side having a partially edentulous area was significantly lower than the condylar volume of the contralateral dentate side (P < .001).

Conclusions

The volumetric asymmetry between the right and left condyle significantly increases with age and edentulousness. The result of the study encourages the clinicians to perform volumetric evaluation of the condyles in cases of radiographically evident condylar asymmetries to obtain a more accurate diagnosis.

Automatic Segmentation of Mandibular Ramus and Condyles

In order to diagnose TMJ pathologies, we developed and tested a novel algorithm, MandSeg, that combines image processing and machine learning approaches for automatically segmenting the mandibular condyles and ramus. A deep neural network based on the U-Net architecture was trained for this task, using 109 cone-beam computed tomography (CBCT) scans. The ground truth label maps were manually segmented by clinicians. The U-Net takes 2D slices extracted from the 3D volumetric images. All the 3D scans were cropped depending on their size in order to keep only the mandibular region of interest. The same anatomic cropping region was used for every scan in the dataset. The scans were acquired at different centers with different resolutions. Therefore, we resized all scans to 512×512 in the pre-processing step where we also performed contrast adjustment as the original scans had low contrast. After the pre-processing, around 350 slices were extracted from each scan, and used to train the U-Net model. For the cross-validation, the dataset was divided into 10 folds. The training was performed with 60 epochs, a batch size of 8 and a learning rate of 2×10^-5. The average performance of the models on the test set presented 0.95 ± 0.05 AUC, 0.93 ± 0.06 sensitivity, 0.9998 ± 0.0001 specificity, 0.9996 ± 0.0003 accuracy, and 0.91 ± 0.03 F1 score. This study findings suggest that fast and efficient CBCT image segmentation of the mandibular condyles and ramus from different clinical data sets and centers can be analyzed effectively. Future studies can now extract radiomic and imaging features as potentially relevant objective diagnostic criteria for TMJ pathologies, such as osteoarthritis (OA). The proposed segmentation will allow large datasets to be analyzed more efficiently for disease classification.

3D Auto-Segmentation of Mandibular Condyles

Temporomandibular joints (TMJ) like a hinge connect the jawbone to the skull. TMJ disorders could cause pain in the jaw joint and the muscles controlling jaw movement. However, the disease cannot be diagnosed until it becomes symptomatic. It has been shown that bone resorption at the condyle articular surface is already evident at initial diagnosis of TMJ Osteoarthritis (OA). Therefore, analyzing the bone structure will facilitate the disease diagnosis. The important step towards this analysis is the condyle segmentation. This article deals with a method to automatically segment the temporomandibular joint condyle out of cone beam CT (CBCT) scans. In the proposed method we denoise images and apply 3D active contour and morphological operations to segment the condyle. The experimental results show that the proposed method yields the Dice score of 0.9461 with the standards deviation of 0.0888 when it is applied on CBCT images of 95 patients. This segmentation will allow large datasets to be analyzed more efficiently towards data sciences and machine learning approaches for disease classification.

Minimally invasive intraoral proportional condylectomy with a three-dimensionally printed cutting guide

The aim of this study was to describe the steps of a minimally invasive surgical technique used to perform a proportional intraoral condylectomy with a three-dimensionally (3D) printed cutting guide. The technique consists of two steps: virtual surgical planning and intraoral condylectomy. During virtual surgical planning, the mandibular ramus was measured bilaterally, the height of the proportional condylectomy was planned virtually, and a cutting guide was 3D printed. In the intraoral condylectomy, the mandibular condyle was approached intraorally, the 3D printed cutting guide was positioned in the sigmoid notch, and the proportional condylectomy was performed. The protocol reported in this technical note is the sum of knowledge acquired from a series of studies published previously by the authors, who have jointly developed a surgical technique that is both minimally invasive and accurate for the treatment of condylar hyperplasia.