Oral Radiology最新文献

Objectives: To qualitatively assess the continuity of the mandibular canal cortices at different segments (ramus, angle, and body) on panoramic radiographs and to correlate these findings with bone mineral density (BMD) status determined by dual-energy X-ray absorptiometry (DXA), using a novel visual index: the Cortical Index of the Mandibular Canal (CIMC) for osteoporosis screening.

Methods: Panoramic radiographs and DXA results were retrospectively selected. Women were grouped according to consistent DXA classifications in all evaluated sites. The mandibular canal was segmented into three anatomical regions, and the continuity of its cortices was visually assessed using the CIMC. The Mandibular Cortical Index (MCI) was also applied for comparative evaluation.

Results: Ninety postmenopausal women aged 40 years or older were included. The mandibular body showed the strongest correlation between CIMC classification and BMD status. CIMC in the mandibular body also demonstrated higher accuracy (AUC = 0.731) compared to the MCI (AUC = 0.650). Accuracy of both indices improved when combined with body mass index (BMI).

Conclusions: The CIMC applied to the mandibular body is a simple visual index for screening individuals with low BMD. Loss of cortical continuity in this region may indicate systemic bone loss.

Objectives: The use of deep learning in detecting teeth with open apices can prevent the need for additional radiographs for patients. The presented study aims to detect open-apex teeth using You Only Look Once (YOLO)-based deep learning models and compare these models.

Methods: A total of 966 panoramic radiographs were included in the study. Open-apex teeth in panoramic radiographs were labeled. During the labeling process, they were divided into 6 classes in the maxilla and mandible, namely incisors, premolars, and molars. AI models YOLOv3, YOLOv4, and YOLOv5 were used. To evaluate the performance of the three detection models, both overall and separately for each class in the test dataset, precision, recall, average precision (mAP), and F1 score were calculated.

Results: YOLOv4 achieved the highest overall performance with a mean average precision (mAP) of 87.84% at IoU (Intersection over Union) 0.5 (mAP@0.5), followed by YOLOv5 with 85.6%, and YOLOv3 with 84.46%. Regarding recall, YOLOv4 also led with 90%, while both YOLOv3 and YOLOv5 reached 89%. Moreover, the F1 score was the highest for YOLOv4 (0.87), followed by YOLOv3 (0.86) and YOLOv5 (0.85).

Conclusions: In this study, YOLOv3, YOLOv4, and YOLOv5 were evaluated for the detection of open-apex teeth, and their mAP, recall, and F1 scores exceeded 84%. Deep learning-based systems can provide faster and more accurate results in the detection of open-apex teeth. This may help reduce the need for additional radiographs from patients and aid dentists by saving time.

Giant cell tumor of the bone (GCTB) is a benign osseous neoplasm that primarily occurs in the epiphyses. Approximately 2% of GCTBs occur in the head and neck regions. Herein, we report a rare case of GCTB in the skull, which was discovered during a dental consultation. The patient was a 74-year-old man who had discomfort in the left temporomandibular joint for 3 months. During the initial visit, there were no remarkable clinical or radiological findings, and the patient was placed under observation. After 3 months, the patient was referred owing to a rapid, progressive swelling on the left side of the head. CT images revealed a slight area of contrast enhancement around the temporomandibular joint, where the patient initially experienced symptoms, but no definite mass was detected, whereas an expansive lesion with cortical thinning and erosion was identified in the left temporal region. Contrast-enhanced magnetic resonance imaging (MRI) revealed heterogeneous enhancement of the temporal lesion and a separate 10-mm lesion with similar enhancement on the inferior surface of the greater wing of the sphenoid bone. Aspiration biopsies of both lesions confirmed the diagnosis of GCTB. Initially, imaging raised the possibility of multicentric GCTB. Nevertheless, while the temporal lesion clearly originated from the bone, a characteristic feature of GCTB, the sphenoid lesion showed no evidence of an intraosseous origin on either CT or MRI. Therefore, this is a rare case of GCTB of the skull with a distinct progression pattern.

Objectives: The radiographic diagnosis of dental caries is often complicated by the presence of radiographic optical effects, such as the Mach band effect and triangular-shaped radiolucencies (TSR). These phenomena can give rise to false-positive diagnoses, especially in bite-wing radiographs, thus influencing clinical decision-making and possibly leading to overdiagnosis and overtreatment. The aim of this systematic review and meta-analysis was to clarify the real prevalence of these radiographic optical effects in dental radiographs and to quantify their influence on diagnostic errors in caries detection.

Methods: This systematic literature review follow the PRISMA guidelines and registered in PROSPERO (CRD420251083823), prior to its execution. Electronic searches were conducted on PubMed, Scopus and the Cochrane Library, with the addition of Grey literature and references of the main reviews on the topic. Inclusion criteria included clinical, in vitro and ex vivo studies evaluating the impact of TSR and Mach band effects on intraoral radiographs (bitewing or periapical), reporting diagnostic performance, prevalence or misinterpretation rates. Data extraction and risk of bias assessment (AXIS, QUADAS-2) were performed independently by two reviewers. Meta-analyses were performed using random effects models.

Results: Of 640 identified reports, only five studies were included. The overall prevalence of non-carious TSR on maxillary molars was 26.44%, 270/1021. The meta-analysis showed that effects (TSR or Mach band) led to false positive diagnoses of caries or fractures, in approximately 13% of observations (60/464). Heterogeneity was high (I² > 90%) and the certainty of the evidence was classified as low to very low due to the type of studies included and the consistency indices.

Conclusions: Radiographic optical effects, especially TSR and Mach band effect, are highly prevalent in bitewing images and significantly increase the risk of false positive caries diagnoses, especially in children. Given the low certainty of the available evidence, clinicians should interpret radiographic findings with caution and always correlate them with a thorough clinical examination. Further high-quality research is needed to develop standardized diagnostic criteria and strategies to mitigate radiographic optical effects in dental practice.

Objective: To evaluate the diagnostic performance of artificial intelligence (AI) models for detecting facial bone fractures on computed tomography (CT), cone-beam CT (CBCT), and plain radiographs.

Methods: Original studies applying machine learning or deep learning algorithms for facial fracture detection in humans were included if they reported diagnostic accuracy metrics such as sensitivity, specificity, or area under the curve (AUC). PubMed-MEDLINE, Scopus, and Web of Science databases were searched up to June 3, 2025. Risk of bias was assessed using the QUADAS-2 tool. The review followed PRISMA 2020 guidelines and was registered in PROSPERO (CRD420251085644).

Results: A total of 23 studies were included. Object detection models such as YOLOv5 and Faster R-CNN-demonstrated high diagnostic accuracy in localizing facial fractures. Classification models such as ResNet and Swin Transformer achieved AUCs frequently exceeding 0.90. Segmentation and hybrid frameworks further improved anatomical specificity. However, the generalizability of findings was constrained by predominantly retrospective, single-centre study designs, limited sample sizes, inconsistent annotation practices, and the absence of external or prospective validation.

Conclusion: AI models show high diagnostic performance for detecting facial fractures across multiple anatomical regions and imaging modalities. Further multicentre prospective studies and the integration of explainable AI are essential for clinical adoption.

Clinical relevance: AI-assisted diagnostic models have the potential to enhance facial fracture detection accuracy, especially in emergency and resource-limited settings. Their integration into radiology workflows could reduce interpretation time, support less experienced clinicians, and improve patient outcomes.

Background: Artificial intelligence (AI) especially deep learning (DL) has significantly revolutionized medical image analysis, which include dental diagnostics. In endodontics, the combination of AI and CBCT provides automatic detection, classification, and segmentation of periapical lesions.

Method: This scoping review used the Arksey and O'Malley model, as updated by Levac et al., and Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. An extensive search strategy was developed to identify pertinent studies in various electronic databases, including PubMed, Scopus, Web of Science, and Google Scholar. The search was conducted for 2010-2025 articles. Studies were included if they applied AI models to CBCT imaging for detecting, classifying, or segmenting periapical lesions. Data were charted on model types, imaging approaches, diagnostic metrics, and clinical translation.

Result: AI models used were CNN-based models including U-Net, DenseNet, custom architectures such as PALNet, and commercial packages such as Diagnocat. High reported diagnostic performance was sensitivity of 67.3% to 97.1% and AUC of up to 0.98. While several studies demonstrated high diagnostic accuracy and potential for clinical decision support, the majority were retrospective, used small or homogenous datasets, and lacked external validation or standard ground truth comparisons (e.g., histological correlation).

Conclusion: AI demonstrates promising potential in enhancement of the diagnostic accuracy and efficiency of CBCT-based periapical lesion assessment. DL models such as U-Net, PAL-Net, and commercial software as of Diagnocat, have shown good performance on the tasks of segmentation and classification. However, additional prospective validation, model calibration, and real-world clinical use, studies are necessary to establish reliability and generalizability.

Clinical significance: The evidence from the current review leans towards an increasing body of evidence in support of AI use in CBCT-based periapical lesion detection. Further, reported the clinical potential of AI to decrease diagnostic time, increase consistency, and assist less experienced operators.

Objectives: The aim of this study is to investigate the morphological differences in temporomandibular joint (TMJ) structures according to the disc displacement status using cone-beam computed tomography (CBCT)-based three-dimensional (3D) reconstructed models.

Methods: Overall, 100 TMJs from 50 adult patients were retrospectively analyzed using magnetic resonance imaging (MRI) and categorized into three groups: normal, anterior disc displacement with reduction (ADcR), and without reduction (ADsR). CBCT data were reconstructed into 3D models using the Mimics software (Materialise, Leuven, Belgium), and anatomical parameters, including articular eminence angle, axial condylar inclination, anteroposterior position of the condyle, and vertical position of the condyle, were measured and statistically compared.

Results: The articular eminence angle was highest in normal (33.11 ± 9.26°) and lowest in ADsR (26.86 ± 8.90°), with a significant difference (p = 0.011). The axial condylar inclination and vertical position of the condyle significantly differed between the groups (p < 0.05). These parameters decreased in the following order: Normal, ADcR, and ADsR. No significant differences were observed in the anteroposterior condyle positions.

Conclusions: TMJ disc displacement is associated with measurable structural changes in joint morphology. CBCT-based 3D analysis of TMJ structures may serve as a useful adjunct method to MRI for the diagnosis and understanding of internal TMJ derangement.

Objectives: Novice dentists or those at the onset of their professional careers require assistance in diagnosing cases that necessitate fillings or root canal treatments. In this study, we propose an innovative recommendation system based on deep learning to assist dentists in identifying the type of tooth caries (requiring filling and necessitating root canal treatment) and determining the appropriate treatment for detected caries. Correctly identifying the type of caries in teeth with no caries, only one type of caries, or more than one type of caries is important for determining the type of treatment to be applied.

Methods: We utilized 1253 bitewing images augmented with various variations, employing three different segmentation methods to automatically detect caries types in the first molar teeth. Furthermore, this study introduces a novel recommendation system for determining the treatment type required for the detected caries type, which represents a significant contribution to this field. The YOLOv8, U-Net, and Detectron-2 networks were evaluated for their efficacy in detecting various types of caries and recommending appropriate treatment methods.

Results: The pixel-label-based comparative results generated by these methods on data labeled by experienced dentists were as follows: 95.03% for Detectron2, 90.88% for U-Net, and 89.23% for YOLOv8. The determination of the type of caries and the recommendation of the type of treatment differ from each other. In terms of treatment recommendations, the success rates of the three methods were as follows: Detectron-2, 88.09%; YOLOv8 70.23%; and U-Net, 61.90%. Consequently, Detectron-2 produced the most successful outcome among the three methods. These results are acceptable for the auxiliary treatment recommendation system.

Conclusion: The system can serve as a supportive tool for less-experienced dentists and as a diagnostic aid for experienced practitioners. The learning-based segmentation method shows great promise for clinical use in the recommendation of treatment.