Automated Cephalometric Landmark Detection Using Deep Reinforcement Learning.

Accurate cephalometric landmark detection leads to accurate analysis, diagnosis, and surgical planning. Many studies on automated landmark detection have been conducted, however reinforcement learning-based networks have not yet been applied. This is the first study to apply deep Q-network (DQN) and double deep Q-network (DDQN) to automated cephalometric landmark detection to the best of our knowledge. The performance of the DQN-based network for cephalometric landmark detection was evaluated using the IEEE International Symposium of Biomedical Imaging (ISBI) 2015 Challenge data set and compared with the previously proposed methods. Furthermore, the clinical applicability of DQN-based automated cephalometric landmark detection was confirmed by testing the DQN-based and DDQN-based network using 500-patient data collected in a clinic. The DQN-based network demonstrated that the average mean radius error of 19 landmarks was smaller than 2 mm, that is, the clinically accepted level, without data augmentation and additional preprocessing. Our DQN-based and DDQN-based approaches tested with the 500-patient data set showed the average success detection rate of 67.33% and 66.04% accuracy within 2 mm, respectively, indicating the feasibility and potential of clinical application.

Early Surgical Intervention of Congenital Syngnathia.

Congenital syngnathia is a rarely reported malformation when there is a fusion between the maxilla and the mandible. It is necessary to modify it in childhood because congenital syngnathia causes incongruity in pronunciation, diet, and esthetics during the growth process. In this case report, 1 case of syngnathia, a rare craniofacial anomaly, is presented with a review of reports. Prompt diagnosis and surgery were performed right after birth for the present case. A partial limitation point was resolved for further growth. Herein, the authors present the case of a female infant (7 d after birth) diagnosed with congenital syngnathia and treated by early surgical intervention.

A deep learning model based on concatenation approach to predict the time to extract a mandibular third molar tooth.

BACKGROUND: Assessing the time required for tooth extraction is the most important factor to consider before surgeries. The purpose of this study was to create a practical predictive model for assessing the time to extract the mandibular third molar tooth using deep learning. The accuracy of the model was evaluated by comparing the extraction time predicted by deep learning with the actual time required for extraction. METHODS: A total of 724 panoramic X-ray images and clinical data were used for artificial intelligence (AI) prediction of extraction time. Clinical data such as age, sex, maximum mouth opening, body weight, height, the time from the start of incision to the start of suture, and surgeon's experience were recorded. Data augmentation and weight balancing were used to improve learning abilities of AI models. Extraction time predicted by the concatenated AI model was compared with the actual extraction time. RESULTS: The final combined model (CNN + MLP) model achieved an R value of 0.8315, an R-squared value of 0.6839, a p-value of less than 0.0001, and a mean absolute error (MAE) of 2.95 min with the test dataset. CONCLUSIONS: Our proposed model for predicting time to extract the mandibular third molar tooth performs well with a high accuracy in clinical practice.

Reconstruction of mandibular defects in osteoradionecrosis and medication-related osteonecrosis of the jaw using fibula free flap and management of postoperative wound infections.

BACKGROUND: Complications from osteoradionecrosis (ORN) and medication-related osteonecrosis of the jaw (MRONJ) include oro-cutaneous fistulas, necrotic bone exposure, soft-tissue defects, and pathologic fractures. The fibula free flap (FFF) is a common free flap method used to reconstruct the mandible in severe cases. Recently, we have used the FFF successfully for the reconstruction of ORN and MRONJ mandibular defects. We report this method as a recommended technique for the treatment of ORN and MRONJ and the management method of postoperative infections. METHODS: Four patients who were diagnosed with ORN of the mandible and 3 patients who were diagnosed with MRONJ of the mandible were included in the study. Among the 7 patients, 3 patients also had pathologic fractures. Partial mandibulectomy and FFF reconstruction were performed at the Department of Oral and Maxillofacial Surgery, Samsung Medical Center from April 2019 to March 2021. RESULTS: All 7 patients recovered following the reconstruction of the defect by FFF. Four patients experienced infections after surgery and pus cultures were performed. All were well healed without flap damage after changing the antibiotics by consultation with infectious medicine experts. CONCLUSION: FFF is a widely used method and can provide an extensive flap to reconstruct the mandible, especially those affected by ORN or MRONJ. If an infection occurs after surgery, appropriate antibiotic changes should be made through cooperation with the infectious medicine department. Therefore, FFF is a well-established and recommended method even in cases of challenging reconstruction.

Efficacy and safety of rhBMP/ß-TCP in alveolar ridge preservation: a multicenter, randomized, open-label, comparative, investigator-blinded clinical trial.

BACKGROUND: The aim of this multicenter, randomized, open-label, comparative, investigator-blinded study was to investigate the efficacy and safety of recombinant human bone morphogenetic protein 2 (rhBMP-2) combined with ß-TCP (rhBMP-2/ß-TCP) in alveolar ridge preservation. MATERIALS AND METHODS: Eighty-four subjects from three centers were enrolled in this clinical trial. After tooth extraction, rhBMP-2/ß-TCP (n = 41, test group) or ß-TCP (n = 43, control group) were grafted to the extraction socket with an absorbable barrier membrane for alveolar ridge preservation. Using computed tomography images obtained immediately after and 12 weeks after surgery, changes in the alveolar bone height and width were analyzed for each group and compared between the two groups. RESULTS: Both the test and control groups showed a significant decrease in alveolar bone height in the 12 weeks after surgery (both groups, p < 0.0001). However, the test group exhibited a significantly lower decrease in alveolar bone height than the control group (p = 0.0004). Alveolar bone width also showed significantly less resorption in the test group than in the control group for all extraction socket levels (ESL) (p = 0.0152 for 75% ESL; p < 0.0001 for 50% ESL; p < 0.0001 for 25% ESL). There were no statistically significant differences in the incidence of adverse events between the two groups. No severe adverse events occurred in either group. CONCLUSIONS: The results of this study suggest that rhBMP-2/ß-TCP is a safe graft material that provides a high alveolar bone preservation effect in patients receiving dental extraction. TRIAL REGISTRATION: Clinicaltrials.gov , NCT02714829 , Registered 22 March 2016.

Identification of de novo EP300 and PLAU variants in a patient with Rubinstein-Taybi syndrome-related arterial vasculopathy and skeletal anomaly.

Rubinstein-Taybi syndrome (RSTS) is a human genetic disorder characterized by distinctive craniofacial features, broad thumbs and halluces, and intellectual disability. Mutations in the CREB binding protein (CREBBP) and E1A binding protein p300 (EP300) are the known causes of RSTS disease. EP300 regulates transcription via chromatin remodeling and plays an important role in cell proliferation and differentiation. Plasminogen activator, urokinase (PLAU) encodes a serine protease that converts plasminogen to plasmin and is involved in several biological processes such as the proteolysis of extracellular matrix-remodeling proteins and the promotion of vascular permeability and angiogenesis. Recently, we discovered a patient who presented with RSTS-related skeletal anomaly and peripheral arterial vasculopathy. To investigate the genetic cause of the disease, we performed trio whole genome sequencing of the genomic DNA from the proband and the proband's parents. We identified two de novo variants coined c.1760T>G (p.Leu587Arg) and c.664G>A (p.Ala222Thr) in EP300 and PLAU, respectively. Furthermore, functional loss of EP300a and PLAUb in zebrafish synergistically affected the intersegmental vessel formation and resulted in the vascular occlusion phenotype. Therefore, we hypothesize that the de novo EP300 variant may have caused RSTS, while both the identified EP300 and PLAU variants may have contributed to the patient's vascular phenotype.

Neutral Position Facilitates Nasotracheal Intubation with a GlideScope Video Laryngoscope: A Randomized Controlled Trial.

The optimal head position for GlideScope facilitated nasotracheal intubation has not yet been determined. We compared the neutral and sniffing positions to establish the degree of intubation difficulty. A total of 88 ASA I-II patients requiring nasotracheal intubation for elective dental surgery with normal airways were divided into two groups according to head position, neutral position (group N), and sniffing position (group S). The primary outcome was the degree of intubation difficulty according to the Intubation Difficulty Scale (IDS): Easy (IDS = 0), mildly difficult (IDS = 1 to 4), and moderately to severely difficult (IDS ≥ 5). Eighty-seven patients completed the study and their data were analyzed. The degree of intubation difficulty was significantly different between the two groups (p = 0.004). The frequency of difficult intubation (IDS > 0) was 12 (27.9%) in group N and 28 (63.6%) in group S (difference in proportion, 35.7%; 95% confidence interval [CI], 14.8 to 52.6%; p = 0.001). The neutral position facilitates nasotracheal intubation with GlideScope by aligning the nasotracheal tube and the glottis inlet more accurately than the sniffing position.

Tracking Accuracy of a Stereo Camera-Based Augmented Reality Navigation System for Orthognathic Surgery.

PURPOSE: Tracking accuracy is critical to successful augmented reality (AR) in the diagnosis and surgical correction of maxillofacial deformities. The present study investigated the tracking accuracy of an AR navigation system combined with a stereo camera during repositioning of the maxilla after a Le Fort I osteotomy using a 3-dimensional skull model and compared the tracking accuracy with that of an existing infrared (IR)-based optical tracking system (OTS). MATERIALS AND METHODS: Five maxillary surgery plans were designed using a 6 degrees-of-freedom articulator that allowed maxillary movement to be set up quantitatively (target distance, 5 mm). To evaluate the accuracy of the stereo camera AR navigation system, it was compared with a commercially available and commonly used IR-based OTS. RESULTS: The mean error was 0.0584 mm in the IR-based OTS and 0.0596 mm in the AR navigation system. The mean accuracy was 98.83% in the IR-based OTS and 98.81% in the AR navigation system. CONCLUSIONS: In this study, the stereo camera-based AR navigation system fabricated and analyzed by the authors was designed for accuracy. The experiments showed its reliability and accuracy. The hardware developed for this AR navigation system displayed accuracy similar to that of existing high-cost imported devices at a substantially lower cost. In addition to surgery, potential applications of the AR navigation system include patient communication and training for novice clinicians.

Three-dimensional analysis of changes in airway space after bimaxillary orthognathic surgery with maxillomandibular setback and their association with obstructive sleep apnea.

BACKGROUND: Bimaxillary orthognathic surgery with maxillomandibular setback is often accompanied by changes in airway space. We analyzed the changes in airway space before and after surgery and assessed their association with obstructive sleep apnea. METHODS: This study is based on the cohort of 13 adult patients (9 males, 4 females, average age 23.85 years) who underwent bimaxillary orthognathic surgery with maxillomandibular setback.We performed computed tomography and portable polysomnography before and after the surgery to assess changes in airway space and Apnea-Hypopnea Index (AHI) values (total, supine, non-supine). RESULTS: The oropharyngeal airway volume decreased by 29% after the surgery, which was statistically significant (p < .05). The upper airway volume and hypopharyngeal airway volume were decreased, but not significantly (4 and 19%, respectively). The changes in airway surface area were statistically significant at all levels examined (p < .05). Changes in the maximum anteroposterior width of the airway were also significant at all levels (p < .05). However, the changes in maximum lateral width were only statistically significant at C2 level (p < .05). AHI values were increased after the surgery but not significantly at any position. CONCLUSIONS: Although bimaxillary surgery with maxillomandibular setback significantly reduces the airway space, it does not affect AHI values or induce obstructive sleep apnea.

Are Three-Dimensional Monitors More Appropriate Than Two-Dimensional Monitors in the Three-Dimensional Analysis?

In orthognathic surgery, three-dimensional (3D) program-based analysis of 3D reconstructions of computed tomography (CT) images is commonly used, and images viewed on a monitor. The authors compared the coordinates of facial landmarks on images in a 3D program displayed on a two-dimensional (2D) (standard) or 3D monitor. Facial bone CT data from 30 patients were reconstructed in 3D. Four researchers identified 33 facial landmarks, 3 times each on 2D and 3D monitors, for each patient, by their x-, y-, and z-coordinates. The time taken to complete these identifications was measured.For each set of coordinates, the average intraclass coefficient was >0.8 for 2D and 3D analyses, as well as among 4 readers. It took on average of 2 minutes 46 seconds to identify the landmarks on the 2D monitor, compared with 2 minutes 25 seconds on the 3D monitor. The variance of individual coordinates differed when measured on 2D or 3D monitor. The landmarks affected were located near the median region of the facial area, and are important for setting the reference sagittal plane during diagnosis for orthognathic surgery. Therefore, identifying facial landmarks using 3D monitors may be helpful for conducting accurate facial diagnoses.

Acoustic analysis of vowel sounds before and after orthognathic surgery.

The purpose of this study was to compare the articular structures and vowel sounds of patients with mandibular prognathism before and after bilateral sagittal split ramus osteotomy (BSSRO). Eight patients who underwent BSSRO to correct mandibular prognathism were selected for inclusion in this study. All patients were asked to read short words (vowels), and these sounds were recorded. Every utterance was repeated twice in four different sessions before the operation and at 6 weeks, 3 months, and 6 months after the operation. The data were analysed using Praat (ver. 5.1.31), and the formant frequencies (F1, F2) of the eight vowels were extracted. PlotFormant (ver. 1.0) was used to draw formant diagrams. The F1 and F2 of front-low vowels were reduced after BSSRO, and the articulating positions of the patients shifted in a posterior-superior direction after the procedure. Additionally, the area of vowel articulation was dramatically reduced after BSSRO but increased slowly over time.

Frankfort horizontal plane is an appropriate three-dimensinal reference in the evaluation of clinical and skeletal cant.

OBJECTIVES: In three-dimensional computed tomography (3D-CT), the cant is evaluated by measuring the distance between the reference plane (or line) and the tooth. The purpose of this study was to determine the horizontal skeletal reference plane that showed the greatest correlation with clinical evaluation. MATERIALS AND METHODS: The subjects were 15 patients who closed their eyes during the CT image taking process. The menton points of all patients deviated by more than 3 mm. In the first evaluation, clinical cant was measured. The distance from the inner canthus to the ipsilateral canine tip and the distance from the eyelid to the ipsilateral first molar were obtained. The distance between the left and right sides was also measured. In the second evaluation, skeletal cant was measured. Six reference planes and one line were used for the evaluation of occlusal cant: 1) FH plane R: Or.R - Or.L - Po.R; 2) FH plane L: Or.R - Or.L - Po.L; 3) F. Ovale plane R: Rt.F.Ovale - Lt.F.Ovale - Or.R; 4) F. Ovale plane L: Rt.F.Ovale - Lt.F.Ovale - Or.L; 5) FZS plane R: Rt.FZS - Lt.FZS - Po.R; 6) FZS plane R: Rt.FZS - Lt.FZS - Po.L, and; 7) FZS line: Rt.FZS - Lt.FZS. RESULTS: The clinical and skeletal cants were compared using linear regression analysis. The FH plane R, FH plane L, and FZS line showed the highest correlation (P<0.05). CONCLUSION: The FH plane R and FH plane L are the most appropriate horizontal reference plane in evaluation of occlusal cant on 3D-CT.

Occurrence of multiple myeloma in the head and neck: a report of two cases.

Multiple myeloma (MM) is a disease reported to account for 1% of all cancers and 10% of hematological malignant diseases. Unlike other malignant diseases that are transferred to the osseous tissues, MM does not show new bone formation, is associated with characteristic osteolytic lesions, and shows monoclonal protein (M-protein) on the immunohematological test, which is an important index in its diagnosis. Solitary lesions of MM are rare in the head and neck area, and, in most cases, MM of the head and neck area is related to systemic sympomts.