Imaging Science in Dentistry最新文献

Purpose: The aim of this study was to compare the accuracy of AI-based AudaxCeph software, Dolphin software, and the manual technique for identifying orthodontic landmarks and tracing lateral cephalograms.

Materials and methods: In this cross-sectional study, 23 anatomical landmarks were identified on 60 randomly selected lateral cephalograms, and 5 dental indices, 4 skeletal indices, and 1 soft tissue index were measured. Each cephalogram was traced using 4 different methods: manually, with the Dolphin software, with the AudaxCeph software automatically, and with the AudaxCeph software in semi-automatic mode. The intra-class correlation coefficient (ICC) and Bland-Altman plots were used to evaluate the agreement between methods. Inter-observer and intra-observer agreements, calculated using the ICC, confirmed the accuracy, reliability, and reproducibility of the results.

Results: There was strong agreement among the AudexCeph (semi-automated or automated) AudaxCeph, Dolphin, and manual methods in measuring orthodontic indices, with ICC values consistently above 0.90. Bland-Altman plots confirmed satisfactory agreement between both versions of AudaxCeph (semi-automated and automated) with the manual method, with mean differences close to 0 and about 95% of data points within the limits of agreement. However, the semi-automated AudaxCeph showed greater agreement and precision than the automated version, as indicated by narrower limits of agreement. The ICC values for inter-observer and intra-observer agreements exceeded 0.98 and 0.99, respectively.

Conclusion: The semi-automated AudaxCeph software offers a robust and cost-effective solution for cephalometric analysis. Its high accuracy and affordability make it a compelling alternative to Dolphin software and the manual method.

Purpose: This study was performed to assess the performance and accuracy of artificial intelligence (AI) in the detection and diagnosis of maxillary sinus pathologies using computed tomography (CT)/cone-beam computed tomography (CBCT) imaging.

Materials and methods: A comprehensive literature search was conducted across 4 databases: Google Scholar, BioMed Central (BMC), ProQuest, and PubMed. Combinations of keywords such as "DCNN," "deep learning," "convolutional neural network," "machine learning," "predictive modeling," and "data mining" were used to identify relevant articles. The study included articles that were published within the last 5 years, written in English, available in full text, and focused on diagnostic accuracy.

Results: Of an initial 530 records, 12 studies with a total of 3,349 patients (7,358 images) were included. All articles employed deep learning methods. The most commonly tested pathologies were maxillary rhinosinusitis and maxillary sinusitis, while the most frequently used AI models were convolutional neural network architectures, including ResNet and DenseNet, YOLO, and U-Net. DenseNet and ResNet architectures have demonstrated superior precision in detecting maxillary sinus pathologies due to their capacity to handle deeper networks without overfitting. The performance in detecting maxillary sinus pathology varied, with an accuracy ranging from 85% to 97%, a sensitivity of 87% to 100%, a specificity of 87.2% to 99.7%, and an area under the curve of 0.80 to 0.91.

Conclusion: AI with various architectures has been used to detect maxillary sinus abnormalities on CT/CBCT images, achieving near-perfect results. However, further improvements are needed to increase accuracy and consistency.

Purpose: The bisecting angle technique (BAT) encounters difficulties in positioning the X-ray cone and aligning the central beam with the tooth. To address this, a rectangular laser featuring a pointed light was integrated into the intraoral X-ray cone. This study evaluated its effectiveness in improving radiograph quality and minimizing errors.

Materials and methods: Fifty fifth-year Bachelor of Dental Surgery students were divided into 3 groups. Group 1 (n=16) used the paralleling technique, group 2 (n=17) employed the conventional BAT, and group 3 (n=17) utilized the laser-assisted BAT on mannequins. Two independent oral radiologists assessed the quality of the radiographs, categorizing the images as either diagnostically acceptable or not acceptable. Inter-group comparisons of quality and error rates were conducted using the chi-square test (significance level: P<0.05).

Results: The paralleling technique group produced 77.5% diagnostically acceptable radiographs and 22.5% that were not diagnostically acceptable. These percentages were 65.3% and 34.7%, respectively, in the conventional BAT group and 75.3% and 24.7%, respectively, in the laser-assisted BAT group, showing results similar to the paralleling technique group. The quality of radiographs differed significantly among the groups (P<0.05). The percentage of error-free radiographs was 38.1% in the paralleling technique group, 20.6% in the conventional BAT group, and 40.0% in the laser-assisted BAT group, with these differences being statistically significant (P<0.05).

Conclusion: The device produced higher acceptability and fewer radiographic errors than the conventional BAT technique, suggesting accurate adjustment of the X-ray cone and central beam to the desired teeth.

Purpose: The aim of this study was to identify, compile, and report the technical specifications of current and historical intraoral digital radiographic systems and recommend standardised reporting practices for production companies.

Materials and methods: A comprehensive report was prepared on 150 intraoral digital radiographic systems, comprising 105 sensor-based (70%) and 45 phosphor storage plate (PSP)-based systems (30%). Technical specifications were obtained from official company sources and scientific articles to ensure a complete collection of available data.

Results: These systems were produced by 55 companies across 11 countries, with the United States leading (35.3%), followed by France (12%). Among the sensor systems, 76.2% used complementary metal-oxide-semiconductor (CMOS) technology, with notable variations in sizes and resolutions. PSP systems were available in 7 plate sizes and displayed diverse resolutions and scanning times. Twenty-one companies produced both sensor- and PSP-based systems, 33 produced only sensor-based systems, and 1 produced exclusively PSP-based systems.

Conclusion: This report identified 150 digital radiographic systems, revealing wide variability in technical specifications and a lack of standardised reporting protocols. The comprehensive summary and recommendations for consistent documentation provided here can help clinicians make informed decisions and encourage manufacturers and production companies to adopt uniform reporting standards aligned with local regulatory frameworks.

Purpose: This study was conducted to compare the objective image quality of radiographs acquired with a handheld X-ray device to those obtained with a wall-mounted X-ray device.

Materials and methods: Brightness, noise, uniformity, and contrast were evaluated. To assess the first 3 parameters, radiographs of an acrylic block were acquired with an unused photostimulable phosphor (PSP) plate from the VistaScan system (Mini Easy, Dürr Dental, Bietigheim-Bissingen, Germany). Initially, 6 radiographs were taken with a Focus X-ray wall-mounted device (Instrumentarium, Tuusula, Finland) operating at 60 kVp, 7 mA, and 0.125 s. Another 6 radiographs were captured using an Eagle handheld X-ray device (Alliage, São Paulo, Brazil) at 60 kVp, 2.5 mA, and 0.35 s. The means and standard deviations of the gray values for all radiographs were calculated using ImageJ (National Institutes of Health, Bethesda, MD, USA). For contrast assessment, radiographs of an aluminum step wedge were obtained using the same PSP plate, X-ray devices, and acquisition parameters. The percentage of contrast variation was determined. The impacts of the devices on image quality were compared using the Student t-test, with a significance level of 5% (P<0.05).

Results: Compared with the wall-mounted device, the handheld device produced radiographs with higher brightness and noise, as indicated by mean values of 6.57 (0.49) and 3.49 (0.02), respectively. Furthermore, it demonstrated lower uniformity and contrast, with respective means of 3.75 (0.02) and 35.48 (0.09) (P<0.05).

Conclusion: Radiographs obtained using a handheld X-ray device exhibit lower theoretical image quality than those acquired with a wall-mounted device.

Purpose: Periarticular and generalized osteoporosis are well-known comorbidities of rheumatoid arthritis (RA), associated with either the disease itself or glucocorticoid therapy. This study was performed to quantitatively evaluate changes in the jawbones of patients with RA using fractal analysis (FA).

Materials and methods: The study comprised 186 participants, including 144 women and 42 men. Of these, 93 were patients with RA and 93 were healthy controls. For the RA group, disease duration, laboratory findings, and medication use were recorded. Measurements of fractal dimension (FD), lacunarity (Lac), bone area fraction (BA/TA), and trabecular thickness (Tb.Th) were taken at the mandibular condyle and angle. These measurements were obtained from panoramic radiographs using ImageJ version 1.52p (National Institutes of Health, Bethesda, MD, USA).

Results: Relative to control participants, patients with RA exhibited higher FD values for both mandibular condyle and angle regions (P<0.05). Conversely, the RA group displayed lower Lac values for both regions (P<0.05). Similarly, Tb.Th values were significantly lower in the RA patient group for both the condyle and the angle (P<0.05). Furthermore, a significant negative correlation was identified between disease duration and FD at the mandibular condyle (P<0.05).

Conclusion: The results demonstrate that FA can be used to quantitatively assess changes in trabecular bone in the jawbones of patients with RA. Image analysis parameters derived from panoramic radiographs, including FA, Lac, BA/TA, and Tb.Th, hold promise in guiding patients with RA toward appropriate medical examinations.

Purpose: This review aimed to explore the scientific literature concerning the methodologies and applications of artificial intelligence (AI) in the field of endodontics. The findings may equip dentists with the necessary technical knowledge to understand the opportunities presented by AI.

Materials and methods: Articles published between 1992 and 2023 were retrieved through an electronic search of Medline via the PubMed, Scopus, and Google Scholar databases. The search, which was limited to articles published in English, aimed to identify relevant studies by employing the following keywords: "artificial intelligence," "machine learning," "deep learning," "endodontic," "root canal treatment," and "radiography." Ultimately, 71 studies that addressed the application of AI in endodontics were selected.

Results: Numerous studies have demonstrated the effectiveness of AI applications in endodontics. These uses encompass the identification of root fractures and periapical lesions, assessment of working length, investigation of root canal system anatomy, prediction of retreatment success, and evaluation of dental pulp stem cell viability.

Conclusion: AI technology is poised to advance aspects of endodontics including scheduling, patient care, management of drug-drug interactions, prognostic diagnosis, and the emerging area of robotic endodontic surgery. AI methods have demonstrated accuracy and precision in the identification, assessment, and prediction of diseases. Thus, AI can significantly improve endodontic diagnosis and treatment, increasing the overall efficacy of endodontic therapy.

Purpose: The study was performed to examine the error associated with image superimposition when computed tomography (CT) images of the dental region are substituted with dental scan data, according to the block-out material used in dental impressions.

Materials and methods: A typodont model was created by bonding orthodontic brackets to teeth on plaster dental stones using melted dental base-plate wax. In the experiment, 2 groups were compared: one using wax and the other using putty as the block-out material to prevent tearing of the impression material during the acquisition of dental impressions. In the wax group, dental cast digital data were superimposed onto the CT data of the typodont model using a tooth occlusal surface-based registration method. In the putty group, a surface-based registration method was used that incorporated not only the occlusal surfaces of the teeth, but also the area where the radiopaque putty covered the orthodontic brackets. The absolute value of the error on the dental surfaces was measured for comparisons.

Results: Analysis of the scanned areas used for replacement in dental CT images revealed fewer superimposition errors when considering only the tooth occlusal surface area, excluding the area containing putty impression material.

Conclusion: The clinical recommendation is to block out the orthodontic bracket with radiolucent wax when obtaining dental impressions. Furthermore, only dental occlusal surface-based registration should be used for the superimposition and replacement of the CT image of the dental area with scanned data from a dental cast model.

Purpose: This study investigated the impact of the field of view (FOV), voxel size, and exposure parameters on the contrast-to-noise ratio (CNR) in cone-beam computed tomography (CBCT).

Materials and methods: A SedentexCT phantom was scanned using 3D Accuitomo 170 across 3 FOVs (40×40 mm, 60×60 mm, 80×80 mm). Each FOV had 4 settings for kVp and 3 for mA. Volumes were reconstructed with voxel sizes from 80 to 250 µm. The CNR was calculated using ImageJ (ver. 1.41, National Institutes of Health, Bethesda). Statistical analyses included Pearson correlation coefficients and regression (R²).

Results: Positive correlations were observed in the 40×40 FOV between voxel size, kVp, mA, rotation degree, and CNR. The 60×60 FOV showed positive correlations between mA, kVp, and CNR, while the 80×80 FOV exhibited correlations for voxel size, kVp, and mA. In the 40×40 (R²=0.551) and 80×80 (R²=0.550) FOVs, mA, kVp, and voxel size influenced CNR. For the 60×60 FOV, mA and kVp were significant contributors (R²=0.389). Using a constant 80-µm voxel size, both mA and kVp significantly influenced CNR (R²=0.467); neither FOV nor rotation degree had substantial impacts.

Conclusion: CNR increased with higher mA, kVp, and larger voxel sizes for 40×40 and 80×80 FOVs. mA was the most influential factor across all FOVs. Regression models showed significant effects of mA and kVp on CNR with 80-µm voxels, while FOV had no effect.