Automatic segmentation and classification of frontal sinuses for sex determination from CBCT scans using a two-stage anatomy-guided attention network.

Sex determination is essential for identifying unidentified individuals, particularly in forensic contexts. Traditional methods for sex determination involve manual measurements of skeletal features on CBCT scans. However, these manual measurements are labor-intensive, time-consuming, and error-prone. The purpose of this study was to automatically and accurately determine sex on a CBCT scan using a two-stage anatomy-guided attention network (SDetNet). SDetNet consisted of a 2D frontal sinus segmentation network (FSNet) and a 3D anatomy-guided attention network (SDNet). FSNet segmented frontal sinus regions in the CBCT images and extracted regions of interest (ROIs) near them. Then, the ROIs were fed into SDNet to predict sex accurately. To improve sex determination performance, we proposed multi-channel inputs (MSIs) and an anatomy-guided attention module (AGAM), which encouraged SDetNet to learn differences in the anatomical context of the frontal sinus between males and females. SDetNet showed superior sex determination performance in the area under the receiver operating characteristic curve, accuracy, Brier score, and specificity compared with the other 3D CNNs. Moreover, the results of ablation studies showed a notable improvement in sex determination with the embedding of both MSI and AGAM. Consequently, SDetNet demonstrated automatic and accurate sex determination by learning the anatomical context information of the frontal sinus on CBCT scans.

Deep learning for sex determination: Analyzing over 200,000 panoramic radiographs.

The objective of this study is to assess the performance of an innovative AI-powered tool for sex determination using panoramic radiographs (PR) and to explore factors affecting the performance of the convolutional neural network (CNN). The study involved 207,946 panoramic dental X-rays and their corresponding reports from 15 clinical centers in São Paulo, Brazil. The PRs were acquired with four different devices, and 58% of the patients were female. Data preprocessing included anonymizing the exams, extracting pertinent information from the reports, such as sex, age, type of dentition, and number of missing teeth, and organizing the data into a PostgreSQL database. Two neural network architectures, a standard CNN and a ResNet, were utilized for sex classification, with both undergoing hyperparameter tuning and cross-validation to ensure optimal performance. The CNN model achieved 95.02% accuracy in sex estimation, with image resolution being a significant influencing factor. The ResNet model attained over 86% accuracy in subjects older than 6 years and over 96% in those over 16 years. The algorithm performed better on female images, and the area under the curve (AUC) exceeded 96% for most age groups, except the youngest. Accuracy values were also assessed for different dentition types (deciduous, mixed, and permanent) and missing teeth. This study demonstrates the effectiveness of an AI-driven tool for sex determination using PR and emphasizes the role of image resolution, age, and sex in determining the algorithm's performance.

Computer-aided design expertise affects digital wax patterns of CAD/CAM laminate veneers more than single crowns.

AIM: There is controversy in the literature regarding clinical outcomes of CAD/CAM laminate veneers. The aim of the present study was to assess the impact of different levels of CAD expertise and different software programs on the reliability and reproducibility of digital wax patterns of laminate veneers and single crowns. MATERIALS AND METHODS: The present preliminary in vitro study was performed on 10 prepared maxillary central incisors available in dental study models. Of the total, five central incisors were prepared with shoulder finish lines for single crowns, whereas the other five underwent incisal shoulder preparation for laminate veneers. The models were scanned using an intraoral scanner. Four dentists (group DENT) and four CAD dental technicians (group CAD) with expertise in different software programs performed digital diagnostic waxing on all prepared teeth. The resulting digital wax patterns (n = 80) were exported as standard tessellation language (STL) files and superimposed on gold standard digital wax patterns (obtained from the original shape of the teeth before preparations). 3D mesh deviations at the cervical margins as well as distal, mesial, and incisal/palatal surfaces between each STL and the gold standard digital wax patterns were calculated in millimeters using a CAD software program. The mean time required by each operator to perform digital waxing was also recorded. Paired comparisons between groups DENT and CAD as well as between nondental and dental software programs were performed for the crowns and laminate veneers using the Wilcoxon signed-rank and paired t tests (α = 0.05). RESULTS: For group DENT, median deviations for single crowns were 0.15 mm (range: 0.08 to 1.05 mm) and for laminate veneers they were 0.15 mm (range: 0.08 to 0.76 mm). For group CAD, median deviations for single crowns were 0.16 mm (range: 0.09 to 0.73 mm) and for laminate veneers they were 0.10 mm (range: 0.06 to 0.53 mm). The Wilcoxon signed-rank test revealed a statistically significant difference between groups DENT and CAD (P = 0.041) and between the software programs (P = 0.029) for laminate veneers, but not for single crowns (P > 0.05). Furthermore, mean times required for group CAD and for dental software programs were significantly shorter than those for group DENT (P = 0.001) and for nondental software programs (P = 0.001), respectively. CONCLUSION: Within the limitations of the present study, the findings suggest that CAD expertise and the software program significantly affect digital wax patterns for laminate veneers, but not for single crowns. (Int J Comput Dent 2022;25(4):361-0; doi: 10.3290/j.ijcd.b3555819).

Influence of receiver bandwidth on MRI artifacts caused by orthodontic brackets composed of different alloys.

PURPOSE: The aim of this in vitro study was to assess the role of bandwidth on the area of magnetic resonance imaging (MRI) artifacts caused by orthodontic appliances composed of different alloys, using different pulse sequences in 1.5 T and 3.0 T magnetic fields. MATERIALS AND METHODS: Different phantoms containing orthodontic brackets (ceramic, ceramic bracket with a stainless-steel slot, and stainless steel) were immersed in agar gel and imaged in 1.5 T and 3.0 T MRI scanners. Pairs of gradient-echo (GE), spin-echo (SE), and ultrashort echo time (UTE) pulse sequences were used differing in bandwidth only. The area of artifacts from orthodontic devices was automatically estimated from pixel value thresholds within a region of interest (ROI). Mean values for similar pulse sequences differing in bandwidth were compared at 1.5 T and 3.0 T using analysis of variance. RESULTS: The comparison of groups revealed a significant inverse association between bandwidth values and artifact areas of the stainless-steel bracket and the self-ligating ceramic bracket with a stainless-steel slot (P<0.05). The areas of artifacts from the ceramic bracket were the smallest, but were not reduced significantly in pulse sequences with higher bandwidth values (P<0.05). Significant differences were also observed between 1.5 T and 3.0 T MRI using SE and UTE, but not using GE 2-dimensional or 3-dimensional pulse sequences. CONCLUSION: Higher receiver bandwidth might be indicated to prevent artifacts from orthodontic appliances in 1.5 T and 3.0 T MRI using SE and UTE pulse sequences.

Evaluation of cervical peri-implant optical density in longitudinal control of immediate implants in the anterior maxilla region.

OBJECTIVE: This research aimed to longitudinally evaluate the optical density of peri-implant alveolar bone. The data acquired from study participants previously treated with 37 osseointegrated implants were analyzed utilizing the radiographic subtraction technique. METHODS AND MATERIALS: The radiographic follow-ups were performed five times: at the implantation of the prostheses and after 15, 90, 180 and 360 days. Intraoral radiographs were obtained by the paralleling technique using individualized Hanshin-type positioners to guarantee the standardization of the images. The obtained digital images were aligned and equalized before they were submitted to the radiographic subtraction procedure. RESULTS: A significant difference was found between the distal region of Group I (patients treated with osseointegrated implants who required extraction of the dental element) and the 360 day follow-up and the distal region of Group II (patients with healed alveolar sockets) in all follow-up analyses (p < 0.05). We did not observe a significant difference between the groups analyzed and other follow-ups concerning the subcrestal and middle third regions for both the mesial and distal variables (p > 0.05). There was a statistically significant difference in the distal sites [χ2 = 5,745,, p = 0.03], showing a significant association between time and the presence of bone resorption. This association was not shown on the mesial surface (p = 0.16). CONCLUSION: We concluded that there was no statistically significant difference between groups I and II. Using this technique, we were able to quantitatively and qualitatively evaluate the changes in the proximal sites on the digital radiographic images for the analyzed data. Digital subtraction technology to measure peri-implant bone density is an accurate and reproducible technique for quantifying peri-implant bone reactions to different therapeutic modalities.