Dento maxillo facial radiology最新文献

Objectives: This study aimed to develop a robust and accurate deep learning network for detecting the posterior superior alveolar artery (PSAA) in dental cone-beam CT (CBCT) images, focusing on the precise localization of the centre pixel as a critical centreline pixel.

Methods: PSAA locations were manually labelled on dental CBCT data from 150 subjects. The left maxillary sinus images were horizontally flipped. In total, 300 datasets were created. Six different deep learning networks were trained, including 3D U-Net, deeply supervised 3D U-Net (3D U-Net DS), multi-scale deeply supervised 3D U-Net (3D U-Net MSDS), 3D Attention U-Net, 3D V-Net, and 3D Dense U-Net. The performance evaluation involved predicting the centre pixel of the PSAA. This was assessed using mean absolute error (MAE), mean radial error (MRE), and successful detection rate (SDR).

Results: The 3D U-Net MSDS achieved the best prediction performance among the tested networks, with an MAE measurement of 0.696 ± 1.552 mm and MRE of 1.101 ± 2.270 mm. In comparison, the 3D U-Net showed the lowest performance. The 3D U-Net MSDS demonstrated a SDR of 95% within a 2 mm MAE. This was a significantly higher result than other networks that achieved a detection rate of over 80%.

Conclusions: This study presents a robust deep learning network for accurate PSAA detection in dental CBCT images, emphasizing precise centre pixel localization. The method achieves high accuracy in locating small vessels, such as the PSAA, and has the potential to enhance detection accuracy and efficiency, thus impacting oral and maxillofacial surgery planning and decision-making.

MRI is a noninvasive, ionizing radiation-free imaging modality that has become an indispensable medical diagnostic method. The literature suggests MRI as a potential diagnostic modality in dentomaxillofacial radiology. However, current MRI equipment is designed for medical imaging (eg, brain and body imaging), with general-purpose use in radiology. Hence, it appears expensive for dentists to purchase and maintain, besides being complex to operate. In recent years, MRI has entered some areas of dentistry and has reached a point in which it can be provided following a tailored approach. This technical report introduces a dental-dedicated MRI (ddMRI) system, describing how MRI can be adapted to fit dentomaxillofacial radiology through the appropriate choice of field strength, dental radiofrequency surface coil, and pulse sequences. Also, this technical report illustrates the possible application and feasibility of the suggested ddMRI system in some relevant diagnostic tasks in dentistry. Based on the presented cases, it is fair to consider the suggested ddMRI system as a feasible approach to introducing MRI to dentists and dentomaxillofacial radiology specialists. Further studies are needed to clarify the diagnostic accuracy of ddMRI considering the various diagnostic tasks relevant to the practice of dentistry.

Objectives: Different factors can affect the discrepancy between the gray value (GV) measurements obtained from CBCT and the Hounsfield unit (HU) derived from multidetector CT (MDCT), which is considered the gold-standard density scale. This study aimed to explore the impact of region of interest (ROI) location and field of view (FOV) size on the difference between these two scales as a potential source of error.

Methods: Three phantoms, each consisting of a water-filled plastic bin containing a dry dentate human skull, were prepared. CBCT scans were conducted using the NewTom VGi evo system, while MDCT scans were performed using Philips system. Three different FOV sizes (8 × 8 cm, 8 × 12 cm, and 12 × 15 cm) were used, and the GVs obtained from eight distinct ROIs were compared with the HUs from the MDCT scans. The ROIs included dental and bony regions within the anterior and posterior areas of both jaws. Statistical analyses were performed using SPSS v. 26.

Results: The GVs derived from CBCT images were significantly influenced by both ROI location and FOV size (p < 0.05 for both factors). Following the comparison between GVs and HUs, the anterior mandibular bone ROI represented the minimum error, while the posterior mandibular teeth exhibited the maximum error. Moreover, the 8 × 8 cm and 12 × 15 cm FOVs resulted in the lowest and highest degrees of GV error, respectively.

Conclusions: The ROI location and the FOV size can significantly affect the GVs obtained from CBCT images. It is not recommended to use the GV scale within the posterior mandibular teeth region due to the potential for error. Additionally, selecting smaller FOV sizes, such as 8 × 8 cm, can provide GVs closer to the gold-standard numbers.

Objectives: The purpose of this study was to evaluate the difference in performance of deep-learning (DL) models with respect to the image classes and amount of training data to create an effective DL model for detecting both unilateral cleft alveoli (UCAs) and bilateral cleft alveoli (BCAs) on panoramic radiographs.

Methods: Model U was created using UCA and normal images, and Model B was created using BCA and normal images. Models C1 and C2 were created using the combined data of UCA, BCA, and normal images. The same number of CAs was used for training Models U, B, and C1, whereas Model C2 was created with a larger amount of data. The performance of all four models was evaluated with the same test data and compared with those of two human observers.

Results: The recall values were 0.60, 0.73, 0.80, and 0.88 for Models A, B, C1, and C2, respectively. The results of Model C2 were highest in precision and F-measure (0.98 and 0.92) and almost the same as those of human observers. Significant differences were found in the ratios of detected to undetected CAs of Models U and C1 (p = 0.01), Models U and C2 (p < 0.001), and Models B and C2 (p = 0.036).

Conclusions: The DL models trained using both UCA and BCA data (Models C1 and C2) achieved high detection performance. Moreover, the performance of a DL model may depend on the amount of training data.

Objectives: This study aims to determine the reproducibility and location-stability of cone-beam computed tomography (CBCT) radiomic features.

Methods: Centrifugal tubes with six concentrations of K₂HPO₄ solutions (50, 100, 200, 400, 600, and 800 mg ml^-1) were imaged within a customized phantom. For each concentration, images were captured twice as test and retest sets. Totally, 69 radiomic features were extracted by LIFEx. The reproducibility was assessed between the test and retest sets. We used the concordance correlation coefficient (CCC) to screen qualified features and then compared the differences in the numbers of them under 24 series (four locations groups * six concentrations). The location-stability was assessed using the Kruskal-Wallis test under different concentration sets; likewise, the numbers of qualified features under six test sets were analyzed.

Results: There were 20 and 23 qualified features in the reproducibility and location-stability experiments, respectively. In the reproducibility experiment, the performance of the peripheral groups and high-concentration sets was significantly better than the center groups and low-concentration sets. The effect of concentration on the location-stability of features was not monotonic, and the number of qualified features in the low-concentration sets was greater than that in the high-concentration sets. No features were qualified in both experiments.

Conclusions: The density and location of the target object can affect the number of reproducible radiomic features, and its density can also affect the number of location-stable radiomic features. The problem of feature reliability should be treated cautiously in radiomic research on CBCT.

Objectives: To develop and validate a novel artificial intelligence (AI) tool for automated segmentation of mandibular incisive canal on cone beam computed tomography (CBCT) scans.

Methods: After ethical approval, a data set of 200 CBCT scans were selected and categorized into training (160), validation (20), and test (20) sets. CBCT scans were imported into Virtual Patient Creator and ground truth for training and validation were manually segmented by three oral radiologists in multiplanar reconstructions. Intra- and interobserver analysis for human segmentation variability was performed on 20% of the data set. Segmentations were imported into Mimics for standardization. Resulting files were imported to 3-Matic for analysis using surface- and voxel-based methods. Evaluation metrics involved time efficiency, analysis metrics including Dice Similarity Coefficient (DSC), Intersection over Union (IoU), Root mean square error (RMSE), precision, recall, accuracy, and consistency. These values were calculated considering AI-based segmentation and refined-AI segmentation compared to manual segmentation.

Results: Average time for AI-based segmentation, refined-AI segmentation and manual segmentation was 00:10, 08:09, and 47:18 (284-fold time reduction). AI-based segmentation showed mean values of DSC 0.873, IoU 0.775, RMSE 0.256 mm, precision 0.837 and recall 0.890 while refined-AI segmentation provided DSC 0.876, IoU 0.781, RMSE 0.267 mm, precision 0. 852 and recall 0.902 with the accuracy of 0.998 for both methods. The consistency was one for AI-based segmentation and 0.910 for manual segmentation.

Conclusions: An innovative AI-tool for automated segmentation of mandibular incisive canal on CBCT scans was proofed to be accurate, time efficient, and highly consistent, serving pre-surgical planning.

Objectives: To identify clinical and local radiographic predictors for medication-related osteonecrosis of the jaws (MRONJ) by the assessment of pre-operative CBCT images of oncologic patients treated with anti-resorptive drugs (ARDs) undergoing tooth extractions.

Methods: This retrospective, longitudinal, case-control study included clinical and imaging data of 97 patients, divided into study and control group. Patients in the study group (n = 47; 87 tooth extractions) had received at least one dose of ARD, undergone tooth extraction(s), and had a pre-operative CBCT. An age-, gender-, and tooth extraction-matched control group (n = 50; 106 tooth extractions) was selected. Three calibrated, blinded, and independent examiners evaluated each tooth extraction site. Statistical analysis used χ²/Fisher's exact/Mann-Whitney U test to contrast control and study group, ARD type used, and sites with or without MRONJ development. p-value ≤ 0.05 was considered significant.

Results: From the study group, 15 patients (32%) and 33 sites (38%) developed MRONJ after tooth extraction. When controls were compared to study sites, the latter showed significantly more thickening of the lamina dura, widened periodontal ligament space, osteosclerosis, osteolysis, and sequestrum formation. In the study group, MRONJ risk significantly increased in patients who had multiple tooth extractions, were smokers, and had shorter drug holidays. Periosteal reaction and sequestrum formation may indicate latent MRONJ lesions. Additionally, patients given bisphosphonates showed considerably more osteosclerosis than those given denosumab.

Conclusions: Periosteal reaction and sequestrum formation are suspected to be pre-clinical MRONJ lesions. Furthermore, ARD induced bony changes and radiographic variations between ARD types were seen.

Objectives: Three-dimensional cone beam computed tomography (CBCT) imaging can be considered, especially in patients with complicated peri-implantitis (PI). Artifacts induced by dense materials are the drawback of CBCT imaging and the peri-implant bone condition may not be assessed reliably because the artifacts are present in the same area. This pilot study investigates the performance of the artifact reduction algorithm (ARA) of the Planmeca Viso G7 CBCT device (Planmeca, Helsinki, Finland) with three different implant materials and imaging parameters.

Methods: Three pairs of dental implants consisting of titanium, zirconia, and fiber reinforced composite (FRC) were set into a pig mandible. A vertical defect simulating peri-implantitis bone loss was made on the buccal side of one of each implant. The defect was identified and measured by two observers and compared to the actual dimensions. In addition, the bone structure and the marginal cortex visibility between the implants were estimated visually.

Results: The bone defect and its dimensions with the zirconia implant could not be identified in any image with or without the metal artifact reduction algorithm. The bone defect of titanium and FRC implants were identified with all three imaging parameters or even without ARA. The interobserver agreement between the two observers was almost perfect for all categories analyzed.

Conclusion: Peri-implantitis defect of the zirconia implant and the peri-implant bone structure of the zirconia implants cannot be recognized reliably with any ARA levels, or any imaging parameters used with the Planmeca Viso G7. The need for ARA when imaging the peri-implant bone condition of the titanium and FRC implants may be unnecessary.

Objectives: Artefacts from dental implants in three-dimensional (3D) imaging may lead to incorrect representation of anatomical dimensions and impede virtual planning in navigated implantology. The aim of this study was quantitative assessment of artefacts in 3D STL models from cone beam CT (CBCT) and multislice CT (MSCT) using different scanning protocols and titanium-zirconium (Ti-Zr) and zirconium (ZrO₂) implant materials.

Methods: Three ZrO₂ and three Ti-Zr implants were respectively placed in the mandibles of two fresh human specimens. Before (baseline) and after implant placement, 3D digital imaging scans were performed (10 repetitions per timepoint: voxel size 0.2 mm³ and 0.3 mm³ for CBCT; 80 and 140 kV in MSCT). DICOM data were converted into 3D STL models and evaluated in computer-aided design software. After precise merging of the baseline and post-op models, the surface deviation was calculated, representing the extent of artefacts in the 3D models.

Results: Compared with baseline, ZrO₂ emitted 36.5-37.3% (±0.6-0.8) artefacts in the CBCT and 39.2-50.2% (±0.5-1.2) in the MSCT models. Ti-Zr implants produced 4.1-7.1% (±0.3-3.0) artefacts in CBCT and 5.4-15.7% (±0.5-1.3) in MSCT. Significantly more artefacts were found in the MSCT vs CBCT models for both implant materials (p < 0.05). Significantly fewer artefacts were visible in the 3D models from scans with higher kilovolts in MSCT and smaller voxel size in CBCT.

Conclusions: Among the four applied protocols, the lowest artefact proportion of ZrO₂ and Ti-Zr implants in STL models was observed with CBCT and the 0.3 mm³ voxel size.