Dento maxillo facial radiology最新文献

Objectives: Periodontitis is a serious periodontal infection that damages the soft tissues and bone around teeth and is linked to systemic conditions. Accurate diagnosis and staging, complemented by radiographic evaluation, are vital. This systematic review (PROSPERO ID: CRD42023480552) explores Artificial Intelligence (AI) applications in assessing alveolar bone loss and periodontitis on dental panoramic and periapical radiographs.

Methods: Five databases (Medline, Embase, Scopus, Web of Science, and Cochran's Library) were searched from January 1990 to January 2024. Keywords related to 'artificial intelligence', 'Periodontal bone loss/Periodontitis', and 'Dental radiographs' were used. Risk of bias and quality assessment of included papers were performed according to the APPRAISE-AI Tool for Quantitative Evaluation of AI Studies for Clinical Decision Support. Meta analysis was carried out via the "metaprop" command in R V3.6.1.

Results: Thirty articles were included in the review, where ten papers were eligible for meta-analysis. Based on quality scores from the APPRAISE-AI critical appraisal tool of the 30 papers, 1 (3.3%) were of very low quality (score < 40), 3 (10.0%) were of low quality (40 ≤ score < 50), 19 (63.3%) were of intermediate quality (50 ≤ score < 60), and 7 (23.3%) were of high quality (60 ≤ score < 80). No papers were of very high quality (score ≥ 80). Meta-analysis indicated that model performance was generally good, e.g.: sensitivity 87% (95% CI: 80% to 93%), specificity 76% (95% CI: 69% to 81%), and accuracy 84% (95% CI: 75% to 91%).

Conclusion: Deep Learning shows much promise in evaluating periodontal bone levels, although there was some variation in performance. AI studies can lack transparency and reporting standards could be improved.

Title: The Influence of a Deep Learning Tool on the Performance of Oral and Maxillofacial Radiologists in the Detection of Apical Radiolucencies.

Objectives: This study aimed to assess the impact of a deep learning model on oral radiologists' ability to detect periapical radiolucencies on periapical radiographs. The secondary objective was to conduct a regression analysis to evaluate the effects of years of experience, time to diagnose, and specialty.

Methods: This study used an annotated dataset and a beta-version of a deep learning model (Denti.AI). The testing subset comprised 68 intraoral periapical radiographs confirmed with cone-beam computed tomography for presence/absence of apical radiolucencies. Four oral radiologists participated in a crossover reading scenario, analyzing the radiographs under two conditions: initially without AI assistance and later with AI predictions. The study evaluated reader performance using AFROC-AUC, sensitivity, specificity, and ROC-AUC per case. It also assessed sensitivity per lesion. Regression analysis investigated how experience, time spent on images, and specialty influenced reader performance.

Results: No statistically significant differences were found in AFROC-AUC, sensitivity, specificity, and ROC-AUC. Regression analysis identified factors influencing diagnostic outcomes: unaided reading significantly prolonged diagnostic time (Beta = 12, 95% CI [11, 13], p < 0.001), while radiologists' professional status was positively associated with diagnostic accuracy (Beta = 0.02, 95% CI [0.00, 0.04], p = 0.015). These findings underscore the impact of AI on diagnostic efficiency and the critical role of radiologists' experience in diagnostic accuracy.

Conclusion: AI did not significantly enhance radiologists' overall diagnostic accuracy. However, it showed potential to enhance efficiency, particularly advantageous for non-expert clinicians. The expertise of radiologists remains vital for accuracy, underscoring the complementary role of AI in dental diagnostics.

Advances in knowledge: AI algorithms may have more notable effects on radiologists' workflow than on the accuracy of detecting apical radiolucencies.

Objectives: Cysts in jaws may have similar radiographic features. However, it is important to clarify the diagnosis prior to surgery. The aim of this study was to compare the radiomic features of radicular cysts (RCs), dentigerous cysts (DCs) and odontogenic keratocysts (OKCs) as a non-invasive diagnostic alternative to biopsy.

Methods: In total, 161 odontogenic cysts diagnosed histopathologically (55 RCs, 53 DCs and 53 OKCs) were included in the present study. Each cyst was semi-automatically segmented on CBCT images, and radiomic features were extracted by an observer. A second observer repeated 20% of the evaluations and the radiomic features. Those achieving an inter-observer agreement level above 0.850 were included in the study. Consequently, 5 shape-based and 22 textural features were investigated in the study. Statistical analysis was performed comparing both three cyst features and making pairwise comparisons.

Results: All features included in the study showed statistical differences between cysts, with the exception of one textural feature (NGTDM coarseness) (p < 0.05). However, only one shape-based feature (shericity) and one textural feature (GLSZM large area emphasis) were statistically different in pairwise comparisons of all three cysts (p < 0.05).

Conclusion: Radiomics features of the RCs, DCs and OKCs showed significant differences, and may have the potential to be used as a non-invasive method in the differential diagnosis of cysts.

Objective: To develop an accurate method for converting dose-area product (DAP) to patient dose for dental cone-beam computed tomography (CBCT) using deep learning.

Methods: 24,384 CBCT exposures of an adult phantom were simulated with PCXMC 2.0, using permutations of tube voltage, filtration, source-isocenter distance, beam width/height and isocenter position. Equivalent organ doses as well as DAP values were recorded. Next, using the aforementioned scan parameters as inputs, neural networks (NN) were trained using Keras for estimating the equivalent dose per DAP for each organ. Two methods were explored for positional input features: (1) 'Coordinate' mode, which uses the (continuous) XYZ-coordinates of the isocenter, and (2) 'AP/JAW' mode, which uses the (categorical) anteroposterior and craniocaudal position. Each network was trained, validated and tested using a 3/1/1 data split. Effective dose (ED) was calculated from the combination of NN outputs using ICRP 103 tissue weighting factors. The performance of the resulting NN models for estimating ED/DAP was compared with that of a multiple linear regression (MLR) model as well as direct conversion coefficients (CC).

Results: The mean absolute error (MAE) for organ dose/DAP on the test data ranged from 0.18% (bone surface) to 2.90% (oesophagus) in 'Coordinate' mode and from 2.74% (red bone-marrow) to 14.13% (brain) in 'AP/JAW' mode. The MAE for ED was 0.23% and 4.30%, respectively, for the two modes, vs. 5.70% for the MLR model and 20.19%-32.67% for the CCs.

Conclusion: NNs allow for an accurate estimation of patient dose based on DAP in dental CBCT.

Objectives: To compare the accuracy, duration, and certainty of diagnosing approximal caries in bitewing radiographs displayed in three monitors under two luminance conditions.

Methods: A total of 39 teeth without evident caries were selected from 11 patients undergoing partial jaw resection. Before the operation, 13 bitewing radiographs were captured by a digital imaging system. Eight observers evaluated the images under the dark (9 lux) and bright (200 lux) conditions, using two medical-grade monitors and a commercial monitor. Using histological results as the gold standard, the areas under the receiver operating characteristic curves under different conditions were compared using the Z-test. Multivariate analysis of variance was conducted to assess the impact of various factors on diagnostic duration, while ordinal logistic regression was used to examine factors influencing diagnostic certainty level. It was considered significant when P＜0.05.

Results: No significant difference was found in the diagnostic accuracy or duration for diagnosis approximal caries under two luminance conditions with the three distinct monitors (P > 0.05). Ambient light, clinical experience and the pathological grade of approximal caries have influence on the degree of diagnostic confidence (P＜0.05).

Conclusions: Different monitors and ambient luminance didn't influence the diagnostic accuracy or evaluation duration. Ambient luminance, clinical experience, and the depth of caries affect the degree of diagnostic confidence.

Advances in knowledge: The study employing bitewing radiographs from real patients indicates that ambient luminance, clinical experience, and the depth of caries affect the degree of diagnostic confidence.

Objectives: The purpose of this study was to compare the image quality of ultra-high-resolution computed tomography (U-HRCT) with that of conventional multi-detector row CT (convCT) and demonstrate its usefulness in the dentomaxillofacial region.

Methods: Phantoms were helically scanned with U-HRCT and convCT scanners using clinical protocols. In U-HRCT, phantoms were scanned in super-high-resolution (SHR) mode, and hybrid iterative reconstruction (HIR) and filtered-back projection (FBP) techniques were performed using a bone kernel (FC81). The FBP technique was performed using the same kernel as in convCT (reference). Two observers independently evaluated the 54 resulting images using a 5-point scale (5: excellent diagnostic image quality; 4: above average; 3: average; 2: subdiagnostic; and 1: unacceptable). The system performance function (SPF) was calculated for a comprehensive evaluation of the image quality using the task transfer function and noise power spectrum. Statistical analysis using the Kruskal-Wallis test was performed to compare the image quality among the three protocols.

Results: The observers assigned higher scores to images acquired with the SHRHIR and SHRFBP protocols than to those acquired with the reference (p < 0.0001 and p < 0.0001, respectively). The relative SPF value at 1.0 cycles/mm in SHRHIR and SHRFBP compared to the reference protocol were 151.5 and 45.6%, respectively.

Conclusions: Through phantom experiments, this study demonstrated that U-HRCT can provide superior-quality images compared to conventional CT in the dentomaxillofacial region. The development of a better image reconstruction method is required to improve image quality and optimise the radiation dose.

Objectives: To evaluate the reliability of 3 T MRI nerve-bone fusion in assessing the lingual nerve (LN) and its anatomical relationship to the lingual cortical plate prior to the impacted mandibular third molar (IMTM) extraction.

Methods: The MRI nerve and bone sequences used in this study were 3D-T2-weighted fast field echo (3D-T2-FFE) and fast field echo resembling a CT using restricted echo-spacing (FRACTURE), respectively. Both sequences were performed in 25 subjects, and the resulting 3D-T2-FFE/FRACTURE fusion images were assessed by two independent observers. Semi-quantitative analyses included assessments of overall image quality, image artifacts, nerve continuity, and the detectability of five intermediate points (IPs). Quantitative analyses included measurements of the lingual cortical plate thickness (LCPT), vertical distance (V1* and V2*) and the closest horizontal distance (CHD) between the LN and the lingual cortical plate. Reliability was evaluated using weighted Cohen's kappa coefficient (κ), intraclass correlation coefficient (ICC) and Bland-Altman plots. Differences in LCPT between 3D-T2-FFE/FRACTURE fusion images and one-beam computed tomography (CBCT) were compared using independent samples T-tests or Mann-Whitney U tests.

Results: The fusion images demonstrated that the LN continuity score was 3.00 (1.00) (good), with 88% (44/50) of LNs displayed continuously to the IMTM level. Intra-reader agreement for nerve continuity was moderate (κ = 0.527), as was inter-reader agreement (κ = 0.428). The intra-reader and inter-reader agreement for LCPT measurements at the neck, mid-root and apex of the IMTM were all moderate (ICC > 0.60). Intra-reader agreements for V1*, V2* and CHD were moderate to excellent (ICC = 0.904, 0.967 and 0.723, respectively), and inter-reader agreements for V1*, V2* and CHD were also moderate to excellent (ICC = 0.948, 0.941 and 0.623, respectively). The reliability of LCPT measurements between 3D-T2-FFE/FRACTURE fusion and CBCT was moderate (ICC = 0.609-0.796).

Conclusions: The 3D-T2-FFE/FRACTURE fusion technique demonstrated potential feasibility for the identification of the LN and its relationship to the lingual cortical plate, as well as for the measurement of LCPT.

Advances in knowledge: This study has generated a dataset that is capable of simultaneously defining the LN and LCPT.

Objectives: To develop and evaluate a deep learning (DL) model to reduce metal artefacts originating from the exomass in cone-beam computed tomography (CBCT) of the jaws.

Methods: Five porcine mandibles, each featuring six tubes filled with a radiopaque solution, were scanned using four CBCT units before and after the incremental insertion of up to three titanium, titanium-zirconium, and zirconia dental implants in the exomass of a small field of view. A conditional denoising diffusion probabilistic model, using DL techniques, was employed to correct images from exomass-related metal artefacts across the CBCT units and implant scenarios. Three examiners independently scored the image quality of all datasets, including those without an implant (ground truth), with implants in the exomass (original), and DL-generated ones. Quantitative analysis compared contrast-to-noise ratio (CNR) to validate artefact reduction using repeated measures analysis of variance in a factorial design followed by Tukey test (α = 0.05).

Results: The visualisation of the hard tissues and overall image quality was reduced in the original and increased in the DL-generated images. The score variation observed in the original images was not observed in the DL-generated images, which generally scored higher than the original images. DL-generated images revealed significantly greater CNR than both the ground truth and their corresponding original images, regardless of the material and quantity of dental implants and the CBCT unit (p < 0.05). Original images revealed significantly lower CNR than the ground truth (p < 0.05).

Conclusions: The developed DL model demonstrated promising performance in correcting exomass-related metal artefacts in CBCT of the jaws.

Objectives: To compare the performance of the convolutional neural network (CNN) with the vision transformer (ViT) and the gated multilayer perceptron (gMLP) in the classification of radiographic images of dental structures.

Methods: Retrospectively collected 2-dimensional images derived from cone beam computed tomographic volumes were used to train CNN, ViT and gMLP architectures as classifiers for 4 different cases. Cases selected for training the architectures were the classification of the radiographic appearance of maxillary sinuses, maxillary and mandibular incisors, presence or absence of the mental foramen and the positional relationship of the mandibular third molar to the inferior alveolar nerve canal. The performance metrics (sensitivity, specificity, precision, accuracy and f1-score) and area under curve (AUC) - receiver operating characteristic and precision-recall curves were calculated.

Results: The ViT with an accuracy of 0.74-0.98, performed on par with the CNN model (accuracy 0.71-0.99) in all tasks. The gMLP displayed marginally lower performance (accuracy 0.65-0.98) as compared to the CNN and ViT. For certain tasks, the ViT outperformed the CNN. The AUCs ranged from 0.77-1.00 (CNN), 0.80-1.00 (ViT) and 0.73-1.00 (gMLP) for all of the 4 cases.

Conclusions: The difference in performance of the ViT, gMLP and the CNN (the current state-of-the-art) was significant in certain tasks. This difference in model performance for various tasks proves that capabilities of different architectures may be leveraged.

Advances in knowledge: The vision transformer, followed by the gated multilayer perceptron are deep learning models that exhibit comparable performance with the convolutional neural network in the classification of dental radiographic images.

Objectives: To systematically evaluate the mean gray values (MGV) and noise provided by bone and soft tissue equivalent materials and air imaged with varied acquisition parameters in nine cone-beam computed tomography (CBCT) machines.

Methods: The DIN6868-161 phantom, composed of bone and soft tissue equivalent material and air gap, was scanned in nine CBCT machines. Tube current (mA) and tube voltage (kV), field of view (FOV) size, and rotation angle were varied over the possible range. The effect of the acquisition parameters on the MGV and contrast-to-noise indicator (CNI) was analyzed by Kruskal Wallis and Dunn-Bonferroni tests for each machine independently (α = 0.05).

Results: Tube current did not influence MGV in most machines. Viso G7 and Veraview X800 presented a decrease in the MGV for increasing kV. For ProMax 3D MAX and X1, the kV did not affect the MGV. For the majority of machines, MGV decreased with increasing FOV height. In general, the rotation angle did not affect the MGV. In addition, CNI was lower with lower radiation and large FOV and did not change from 80 kV in all machines.

Conclusions: The MGV and noise provided by the tested phantom vary largely among machines. The MGV is mainly influenced by the FOV size, especially for bone equivalent radiodensity. For most machines, when the acquisition parameters affect the MGV, the MGV decrease with the increase in the acquisition parameters.

Advances in knowledge: Knowing the expected GV behavior in different exposure conditions hold potential for future calibration of MGV among CBCT machines.