Evaluation of a deep learning system for automatic detection of proximal surface dental caries on bitewing radiographs.

OBJECTIVE: This study aimed to evaluate a deep learning (DL) system using convolutional neural networks (CNNs) for automatic detection of caries on bitewing radiographs. STUDY DESIGN: In total, 2468 bitewings were labeled by 3 dentists to create the reference standard. Of these images, 1257 had caries and 1211 were sound. The Faster region-based CNN was applied to detect the regions of interest (ROIs) with potential lesions. A total of 13,246 ROIs were generated from all 'sound' images, and 50% of 'caries' images (selected randomly) were used to train the ROI detection module. The remaining 50% of 'caries' images were used to validate the ROI detection module. Caries detection was then performed using Inception-ResNet-v2. A set of 3297 'caries' and 5321 'sound' ROIs cropped from the 2468 images was used to train and validate the caries detection module. Data sets were randomly divided into training (90%) and validation (10%) data sets. Recall, precision, specificity, accuracy, and F1 score were used as metrics to assess performance. RESULTS: The caries detection module achieved recall, precision, specificity, accuracy, and F1 scores of 0.89, 0.86, 0.86, 0.87, and 0.87, respectively. CONCLUSIONS: The proposed DL system demonstrated promising performance for detecting proximal surface caries on bitewings.

Deep learning for automated detection and numbering of permanent teeth on panoramic images.

OBJECTIVE: This study aimed to evaluate an automated detection system to detect and classify permanent teeth on orthopantomogram (OPG) images using convolutional neural networks (CNNs). METHODS: In total, 591 digital OPGs were collected from patients older than 18 years. Three qualified dentists performed individual teeth labelling on images to generate the ground truth annotations. A three-step procedure, relying upon CNNs, was proposed for automated detection and classification of teeth. Firstly, U-Net, a type of CNN, performed preliminary segmentation of tooth regions or detecting regions of interest (ROIs) on panoramic images. Secondly, the Faster R-CNN, an advanced object detection architecture, identified each tooth within the ROI determined by the U-Net. Thirdly, VGG-16 architecture classified each tooth into 32 categories, and a tooth number was assigned. A total of 17,135 teeth cropped from 591 radiographs were used to train and validate the tooth detection and tooth numbering modules. 90% of OPG images were used for training, and the remaining 10% were used for validation. 10-folds cross-validation was performed for measuring the performance. The intersection over union (IoU), F1 score, precision, and recall (i.e. sensitivity) were used as metrics to evaluate the performance of resultant CNNs. RESULTS: The ROI detection module had an IoU of 0.70. The tooth detection module achieved a recall of 0.99 and a precision of 0.99. The tooth numbering module had a recall, precision and F1 score of 0.98. CONCLUSION: The resultant automated method achieved high performance for automated tooth detection and numbering from OPG images. Deep learning can be helpful in the automatic filing of dental charts in general dentistry and forensic medicine.

Patient-Controlled Analgesia and Length of Hospital Stay in Orthognathic Surgery: A Randomized Controlled Trial.

PURPOSE: The objective of this prospective, randomized controlled pilot study of patients undergoing orthognathic surgery was to compare the hospital length of stay (LOS) in patients using intravenous patient-controlled analgesia (PCA) versus patients receiving scheduled and as-needed oral analgesia. PATIENTS AND METHODS: A total of 40 patients (19 male and 21 female patients) aged 16 to 56 years (mean, 20.73 years; standard deviation, 6.87 years) were recruited prospectively and randomized to PCA and non-PCA groups of equal size. Recording of the patient-reported pain score using a visual analog scale was commenced at 8:00 am on day 1 after surgery until discharge. The pain score from routine nursing observations during the postoperative period was recorded until the patient was discharged from the hospital. RESULTS: Randomization resulted in approximately equal proportions of male patients (45% vs 50%) and median ages (18.5 years vs 20 years) for the PCA group versus the non-PCA group; however, the PCA group was noted to have a higher proportion of double-jaw surgery (65% vs 40%). The median LOS was 2 days for both the PCA and non-PCA groups (P = .06). No statistically significant difference in pain scores was found between the 2 groups either at rest (P = .27) or on movement (P = .13). CONCLUSIONS: No evidence was found to indicate the superiority of either the PCA or non-PCA regimen with respect to LOS and pain scores; however, this is not evidence of equivalence of the 2 pain management approaches.

Open reduction and internal fixation of mandibular angle fractures: does the transbuccal technique produce fewer complications after treatment than the transoral technique?

PURPOSE: The study's purpose was to answer the following clinical question: in patients with mandibular angle fractures requiring open reduction and internal fixation, do those who have fixation screws inserted using a transbuccal approach compared with those with fixation screws inserted using a transoral approach have fewer complications after treatment? The investigators hypothesized that the transoral approach was associated with a higher risk of complications. MATERIALS AND METHODS: A multicenter retrospective cohort study was performed in patients who had open reduction and internal fixation of mandibular angle fractures from 2008 to 2010 within Western Australia. Patients were divided into transbuccal and transoral groups and then further subdivided into groups with and without fixation failures (primary outcome variable) and statistically compared. Binary logistic regression was used to control for possible confounders, which included patient gender, age, a wisdom tooth within the fracture not extracted, dental caries, partial dentition, bilateral/unilateral fractures, and smoking. RESULTS: In total 597 patients were in the study. Sixteen percent of patients in the transoral group had complications after treatment versus 10% in the transbuccal group. For the transoral technique, the odds of having fixation failure was 1.71 times greater than with the transbuccal technique (95% confidence interval, 1.02 to 2.93; P = .04). Incidences of all complication variables (hardware loosening/fracturing, wound dehiscence, secondary infection, surgery redo, nonunion/malunion of fracture, and removal of plate) were lower in the transbuccal group apart from plate fracture. CONCLUSION: The transbuccal technique was associated with fewer complications after treatment compared with the transoral technique.

The effect of mouthguard design on respiratory function in athletes.

OBJECTIVE: To test the hypothesis that 2 types of custom-made mouthguards will have no effect on ventilation (.V(E), L·min⁻¹), oxygen uptake (.VO2, mL·kg⁻¹·min·â»¹), and heart rate (beats per minutes) at varying exercise intensities (10 km·h⁻¹ and 12 km·h⁻¹) and at subjective maximal effort (.VO2peak) in male field hockey and water polo players. DESIGN: A randomized, prospective, crossover study. SETTING: The Physiology Testing Laboratory, School of Sports Science, Exercise and Health at the University of Western Australia, a tertiary educational institution. PARTICIPANTS: Twenty-seven male team-sport athletes. INTERVENTIONS: Each athlete participated in 3 experimental exercise sessions separated by 1-week intervals. Testing involved a graded exercise test (GXT) performed on a treadmill wearing either a custom laminated mouthguard with normal palatal surface, a custom laminated mouthguard with palatal coverage up to the gingival margin, or no mouthguard. The experimental trials were performed in a random counterbalanced order. MAIN OUTCOME MEASURES: .V(E) (L·min⁻¹) and .VO2 (mL·kg⁻¹·min·â»¹) were measured during the GXT at intensities that equated to 10 km·h⁻¹, 12 km·h⁻¹ and subjective maximal effort (.VO2peak). RESULTS: There were no significant differences between trials for .V(E) (L·min⁻¹) and .VO2 (mL·kg⁻¹·min·â»¹) at any of the intensities assessed (P < 0.05). CONCLUSIONS: The wearing of 2 different custom-made mouthguards during a GXT did not impair .V(E) or .VO2 during varying levels of exercise intensity in team sport athletes.

Pulsed low-intensity ultrasound: a new salvage procedure for delayed unions and nonunions after leg lengthening in children.

From 1998 to 2001, 112 lengthening procedures with or without deformity correction were performed in 108 children by external fixation with the Ilizarov method. Of these cases, 16.9% did not lead to a solid bone consolidation. Two children were operated the second time, mainly because of the parent's decision. Seventeen delayed unions or nonunions in 13 children were treated with low-intensity pulsed ultrasound. All 17 cases healed within 3 to 12 months without any risk of surgical intervention.

Low-intensity pulsed ultrasound: effects on nonunions.

To study the efficacy of low-intensity pulsed ultrasound (US), or LIPUS, of 85 treated nonunion cases with a minimum fracture age of 8 months, 67 cases met the study criteria. These were: no surgical intervention during 4 months before US treatment and radiographically ceased healing for 3 months before US. In a self-paired control study, the mean fracture age of the 67 patients was 39 +/- 6.2 months. After a daily 20-min US treatment at home for an average of 168 days, 85% (57 of 67) of the nonunion cases were clinically and radiographically healed. The study did not include any cases that were malaligned, grossly instable, actively infected or that had extensive bone loss. The results demonstrate that the specific US can effect heal rates similar to those achieved by surgical means, without the associated risks and complications, and to those achieved by electrical bone growth stimulation or by extracorporeal shock-wave therapy.

Hydration-driven transport of deformable lipid vesicles through fine pores and the skin barrier.

We studied aggregate transport through semipermeable, nano-porous barriers experimentally and theoretically. By measuring and modeling the effect of hydration gradient across such barriers, spontaneous transbarrier transport of suitable lipid aggregates in vesicular form was proven to be driven by partial aggregate dehydration at the application site. By generalizing the Onsager transport model we derived a set of equations that rationalize all pertinent observations. Dehydration-induced vesicle motion starts with a lag time. This corresponds to the time needed to reach the limiting vesicle hydration; both are proportional to the starting excess water volume and decrease with increasing relative humidity at application site. The rate of transbarrier transport is insensitive to these parameters but increases with vesicle deformability and volume exchange capability. Both these properties depend on membrane composition. Reversible demixing of bilayer components is the cause of nonlinear bilayer characteristics and also potentially affects the effective membrane hydrophilicity. High hydrophilicity of vesicle surface and extreme aggregate shape adaptability together are necessary for successful material transport across the skin. This demonstrates the significance of basic biophysical investigations for better understanding of biological systems and for the practical use of artificial, nature-inspired carriers in drug delivery.