Dento maxillo facial radiology最新文献

{"title":"Impact of labelling inaccuracy and image noise on tooth segmentation in panoramic radiographs using federated, centralized, and local learning.","authors":"Johan Andreas Balle Rubak, Khuram Naveed, Sanyam Jain, Lukas Esterle, Alexandros Iosifidis, Ruben Pauwels","doi":"10.1093/dmfr/twag001","DOIUrl":"10.1093/dmfr/twag001","url":null,"abstract":"<p><strong>Objectives: </strong>Federated learning (FL) may mitigate privacy constraints, heterogeneous data quality, and inconsistent labelling in dental artificial intelligence (AI). FL was compared with centralized (CL) and local learning (LL) for tooth segmentation in panoramic radiographs across multiple data corruption scenarios.</p><p><strong>Methods: </strong>An Attention U-Net was trained on 2066 radiographs from 6 institutions across 4 settings: baseline (unaltered data), label manipulation (dilated/missing annotations), image-quality manipulation (additive Gaussian noise), and exclusion of one faulty client with corrupted data. FL was implemented via the Flower AI framework. Per-client training and validation loss trajectories were monitored for anomaly detection and a set of metrics (Dice, IoU, HD, HD95, and ASSD) were evaluated on a hold-out test set. From these metrics significance results were reported through Wilcoxon signed-rank test. CL and LL served as comparators.</p><p><strong>Results: </strong>Baseline: FL achieved a median Dice of 0.949 (ASSD: 1.332), slightly better than CL at 0.947 (ASSD: 1.371) and LL at 0.936-0.940 (ASSD: 1.519-1.698). Label manipulation: FL maintained the best median Dice score at 0.949 (ASSD: 1.465) versus CL's 0.942 (ASSD: 1.757) and LL's 0.930-0.940 (ASSD: 1.519-2.115). Similar performance was observed when 2 faulty clients were introduced. Image noise: FL led with a Dice at 0.949 (ASSD: 1.311); CL had a Dice of 0.948 (ASSD: 1.361); LL ranged from 0.932 to 0.940 (ASSD: 1.519-1.774). Similar performance was observed when 2 faulty clients were introduced, with CL performing slightly better than FL. Faulty client exclusion: FL showed a Dice of 0.948 (ASSD: 1.331) better than CL's 0.946 (ASSD: 1.393). Loss curve monitoring reliably flagged the corrupted site.</p><p><strong>Conclusions: </strong>FL matches or exceeds CL and outperforms LL across corruption scenarios while preserving privacy. Per-client loss trajectories provide an effective anomaly-detection mechanism and support FL as a practical, privacy-preserving approach for scalable clinical AI development.</p>","PeriodicalId":11261,"journal":{"name":"Dento maxillo facial radiology","volume":" ","pages":"336-353"},"PeriodicalIF":2.9,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145988718","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"When to perform cone beam computed tomography (CBCT) in primary root canal treatment? A CBCT-based cross-sectional study on the prevalence of MB2 canal in maxillary first molars.","authors":"Bledar Lilaj, Elias Salzmann, Gernot Paul Hönigl, Rinet Dauti, Anton Dobsak, Sophie Pock, Barbara Cvikl","doi":"10.1093/dmfr/twag008","DOIUrl":"https://doi.org/10.1093/dmfr/twag008","url":null,"abstract":"<p><strong>Objectives: </strong>CBCT provides insights into the presence of an additional root canal in maxillary first molars. Since the maxillary first molar is one of the teeth most likely requiring root canal treatment, and young people are more sensitive to radiation exposure, this study aimed to evaluate the prevalence of MB2 canals in healthy, untreated maxillary first molars using CBCT.</p><p><strong>Methods: </strong>CBCT scans of maxillary first molars of 167 patients were categorized according to the classification systems of Wein, Vertucci, and Ahmed. The number of root canals and the presence of an MB2 canal were analyzed descriptively. The associations between the presence of an MB2 canal and various factors, including gender, position in the dental arch, and age group, were examined using a chi-squared test at a significance level of P < .05. All analyses were performed using SPSS.</p><p><strong>Results: </strong>A second mesiobuccal root canal was found in 63% (Weine) and 66% (Vertucci) of the examined maxillary first molars, respectively. Significantly fewer women exhibited a second mesiobuccal root canal compared to men. No statistical differences were observed regarding the position in the dental arch or the different age groups. The most frequently observed root canal morphologies were Vertucci Type I and IV (34% each), followed by Type II (23%), Type VI (6%), Type V (2%), and Type III (1%). In the Weine classification, Types I and II (35% each) were most common, followed by Type III (29%) and Type IV (1%). In the Ahmed classification, the most common configuration was 3MB2-2DB1P1.</p><p><strong>Conclusions: </strong>A MB2 canal was present in about two-thirds of maxillary first molars, with significantly lower prevalence in women but no association with tooth position or age.</p><p><strong>Advances in knowledge: </strong>This study provides updated prevalence and morphological data on MB2 canals, showing they occur in about two-thirds of maxillary first molars and differ by gender. These findings may guide future diagnostic approaches and support more targeted use of CBCT in endodontic treatment planning.</p>","PeriodicalId":11261,"journal":{"name":"Dento maxillo facial radiology","volume":"55 4","pages":"381-391"},"PeriodicalIF":2.9,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147812349","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Cleaning of photostimulable phosphor plates with chlorhexidine-based substances does not affect image quality of radiographs.","authors":"Carolina Paes Borge, Iago Filipe Correia-Dantas, Débora Costa Ruiz, Thamiles Gonzalez-Passos, Deborah Queiroz Freitas, Amanda Farias-Gomes","doi":"10.1093/dmfr/twag004","DOIUrl":"10.1093/dmfr/twag004","url":null,"abstract":"<p><strong>Objectives: </strong>To evaluate image quality of radiographs obtained with photostimulable phosphor (PSP) plates after cleaning with 2% chlorhexidine.</p><p><strong>Methods: </strong>Eight PSP plates were used: 4 from Express system and 4 from VistaScan system. Two plates from each system were wiped with 2% chlorhexidine in aqueous solution (CHX AQ), and the other 2 plates from each system with 2% chlorhexidine in alcoholic solution (CHX AL). Before cleaning and after every 5 cleaning cycles (up to 100 cleanings), 2 radiographs of an acrylic block and 2 of an aluminium step wedge were acquired with each PSP plate. The image's gray values were measured to assess image brightness, noise, uniformity, and contrast. Two-way repeated measures ANOVA followed by Sidak's post hoc test analysed the influence of the number of cleanings and substance type on these parameters (α = 0.05).</p><p><strong>Results: </strong>Considering the number of cleanings, image brightness varied for CHX AL in the Express system and CHX AQ in the VistaScan system, with no consistent pattern (P < .05). Noise varied for both substances in the VistaScan system, also without a clear pattern (P < .05). In the Express system, regarding the substance type, CHX AL resulted in lower brightness and CHX AQ in lower noise after 75 cleanings (P < .05). Contrast and uniformity were not affected by either the number of cleanings or the substance type (P > .05). Despite numerical variations in some objective parameters, overall image quality remained stable across all cleanings for both substances. No artefacts were observed in the images.</p><p><strong>Conclusion: </strong>Regarding image quality, 2% chlorhexidine-whether aqueous or alcohol-based-can be recommended for PSP plates cleaning.</p>","PeriodicalId":11261,"journal":{"name":"Dento maxillo facial radiology","volume":" ","pages":"362-369"},"PeriodicalIF":2.9,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145970688","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Impact of rubber dam thickness on image quality parameters in photostimulable phosphor plates.","authors":"Şelale Özel, Hakan Amasya, Deniz Yanık Nalbantoğlu","doi":"10.1093/dmfr/twag012","DOIUrl":"10.1093/dmfr/twag012","url":null,"abstract":"<p><strong>Objective: </strong>The aim of this study is to investigate the effects of different rubber dam thicknesses on image quality in 2 exposure protocols.</p><p><strong>Methods: </strong>For the study, 3 rubber dam thicknesses were used: thin (0.14 mm), medium (0.18 mm), and heavy (0.22 mm). To mimic clinical conditions, a rubber dam was used in 2 layers. Exposure was performed using 2 different protocols: Protocol 1: 0.080 s exposure time, 65 kV, 7 mA, and Protocol 2: 0.160 s exposure time, 65 kV, 7 mA. For each thickness and protocol, 47 measurements were taken (n = 47). Radiographic images were exported in TIFF and analysed using ImageJ software. The region of interest (ROI) was determined, and signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), Michelson, and Weber contrast values were obtained. One-way ANOVA, post hoc Tukey, and intra-class correlation were used for statistical analysis.</p><p><strong>Results: </strong>No statistical difference was detected for Protocol 1 between rubber dam thicknesses in SNR, CNR, Michelson, and Weber contrasts (P > .05). For protocol 2 (longer exposure time), heavy and medium thickness had lower SNR and CNR values than the thin one (P < .05). Michelson and Weber contrasts were statistically changed in different thicknesses of rubber dam (P < .05). ICC values were good and excellent.</p><p><strong>Conclusions: </strong>The thick rubber dam reduced SNR and CNR values; likewise, Michelson and Weber contrasts were changed, which pointed a reduced image quality and negatively affected object visibility. Short exposure times are recommended to maintain image quality in clinical situations requiring the use of a thick rubber dam.</p>","PeriodicalId":11261,"journal":{"name":"Dento maxillo facial radiology","volume":" ","pages":"392-398"},"PeriodicalIF":2.9,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146212392","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Comparative evaluation of direct and digital measurements of peri-implant defects using CBCT, ultrasound, and intraoral scanning.","authors":"Mahmure Ayşe Tayman, Kıvanç Kamburoğlu, Esra Ece Çakmak","doi":"10.1093/dmfr/twag006","DOIUrl":"10.1093/dmfr/twag006","url":null,"abstract":"<p><strong>Objectives: </strong>To compare the measurement accuracy of cone-beam computed tomography (CBCT), high-resolution ultrasound (US), and intraoral scanning (IOS) with the gold-standard direct method in the measurement of peri-implant bone defects.</p><p><strong>Methods: </strong>Forty standard-threaded and 38 aggressive-threaded (Aggressor) implants-identical in diameter and length (4.3/10 mm) but differing in macro-thread design-were placed into bovine rib bones in vitro. Dehiscence, 2/3-wall, and 4-wall (circumferential) defects were prepared around the implants. Each defect was measured for maximum width, depth, and height using CBCT, US, IOS, and direct manual measurement. Analyses were performed using the general linear model (ANOVA). A P-value < .05 was considered statistically significant.</p><p><strong>Results: </strong>Intra-operator and inter-operator agreement showed high reliability (Gage R&R below 10%). For maximum width, defect type (F = 894.81, P < .001), method (F = 6.76, P < .001), and implant type (F = 5.39, P = .021) were significant. For maximum depth, defect type (F = 861.12, P < .001) and method (F = 3.39, P = .018) were significant. For maximum height, method (F = 12.62, P < .001) and defect type (F = 38.91, P < .001) were significant. The model demonstrated high explanatory power for width (R2 = 75.9%) and depth (R2 = 76.6%) measurements but lower for height (R2 = 20.7%). CBCT provided the most consistent results relative to direct measurements, followed by US, whereas IOS showed greater deviations.</p><p><strong>Conclusions: </strong>CBCT showed the highest agreement with direct measurements, followed by US, while IOS exhibited greater variability. Defect type and measurement modality were the primary determinants of accuracy. These findings indicate that CBCT and US can be considered reliable tools for assessing peri-implant bone defects.</p>","PeriodicalId":11261,"journal":{"name":"Dento maxillo facial radiology","volume":" ","pages":"370-380"},"PeriodicalIF":2.9,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13132167/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146017761","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Craniofacial measurements using zero echo time magnetic resonance imaging.","authors":"Yuka Uchimoto, Maziahtul Zawani Binti Munshi, Tadashi Sasai, Sven Kreiborg, Sanjay M Mallya, Fan-Pei Gloria Yang, Shumei Murakami","doi":"10.1093/dmfr/twag009","DOIUrl":"10.1093/dmfr/twag009","url":null,"abstract":"<p><strong>Objectives: </strong>Radiation exposure remains a major concern in dentomaxillofacial imaging, especially for children and young adults who require repeated examinations. A precise, non-ionizing alternative for craniofacial measurement would have significant implications for clinical practice. Computed tomography (CT) and cephalometry are widely used but expose patients to ionizing radiation. Zero Echo Time Magnetic Resonance Imaging (ZTE-MRI) enables visualization of short-T2 tissues, including cortical bone, without radiation by minimizing TE towards zero; however, clinical head-and-neck use has been limited. The goal of the present study was to investigate the possibility of performing precise 3D measurements of the craniofacial regions using ZTE-MRI.</p><p><strong>Methods: </strong>Twenty-nine participants (mean age 32.14 ± 17.09 years; 23 female, 6 male) underwent MDCT and ZTE-MRI. Standard cephalometric landmarks were identified, yielding 27 linear and 21 angular variables. For each variable, linear regression analysis was performed for the results on MDCT (Multi-Detector CT) images and ZTE-MRI.</p><p><strong>Results: </strong>Nineteen linear and eight angular variables had regression coefficients between 0.9 and 1.1, and 28 of the linear and 16 of the angular variables had R2 values of 0.81 or higher. Twenty-five linear and 16 angular variables showed Cohen's effect sizes within the ±0.8 range.</p><p><strong>Conclusions: </strong>Variables derived from skeletal reference points showed clinically acceptable results, indicating that ZTE-MRI is useful for these measurements. In contrast, variables involving dental reference points demonstrated lower agreement, making accurate measurement in these regions difficult.</p><p><strong>Advances in knowledge: </strong>To our knowledge, this study is the first to show that ZTE-MRI allows feasible linear and angular craniofacial measurements used in dental practice and demonstrates good agreement with MDCT for major skeletal landmarks. These findings support its potential future clinical applicability for radiation-free bone assessment; however, further research is warranted to establish its performance across all measurements.</p>","PeriodicalId":11261,"journal":{"name":"Dento maxillo facial radiology","volume":" ","pages":"399-408"},"PeriodicalIF":2.9,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146017701","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A mouse-tracking study on visualization of panoramic radiographs among Oral & Maxillofacial Radiology faculty.","authors":"Li Zhen Lim, Chun Ning Fong, Joey Jia Yi Lee, Pei Hua Cher, Yu Fan Sim, Peggy P Lee","doi":"10.1093/dmfr/twaf087","DOIUrl":"10.1093/dmfr/twaf087","url":null,"abstract":"<p><strong>Objectives: </strong>Panoramic radiographs (PANs) are a commonly used dental imaging modality and are complex to evaluate. This study aimed to use mouse-tracking technology to identify the visualization characteristics of experienced Oral & Maxillofacial Radiology (OMR) faculty when reviewing PANs and to assess relationships with detection accuracy.</p><p><strong>Methods: </strong>Seventeen OMR faculty in US dental schools with 5 to 38 years of experience were recruited. Participants were shown 17 PANs (5 \"Normal\" and 12 \"Pathology\") over Zoom and instructed to move their mouse cursor in sync with their eye movements. Data collection was based on video recordings with X, Y coordinates auto-detected using mouse-tracking algorithms. Parameters collected included detection accuracy, time taken, completeness of search coverage, revisits and search patterns used.</p><p><strong>Results: </strong>Most common search patterns were Dental to Periphery (41.5%) and Periphery to Dental (37.7%). The mean accuracy score for all cases with pathology was 84.8%. Except for 2 cases with subtle findings (condylar fracture and fibrous dysplasia), mean accuracy scores ranged from 76.5% to 100%. There were no associations between search patterns and detection accuracy. Participants took longer durations and performed more complete searches with more revisits on \"Normal\" cases. They were more likely to use shorter search times with smaller coverage and perform single searches when correctly detecting lesions.</p><p><strong>Conclusions: </strong>Experienced OMRs were able to detect lesions without scanning the entire PAN and within a single search, regardless of the search pattern used.</p><p><strong>Advances in knowledge: </strong>This is the first study using mouse-tracking algorithms to evaluate visualization characteristics on PANs. Experienced OMRs can conduct efficient and accurate reviews of PANs regardless of search strategy.</p>","PeriodicalId":11261,"journal":{"name":"Dento maxillo facial radiology","volume":" ","pages":"327-335"},"PeriodicalIF":2.9,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13132168/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145676799","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Comparison of dental-dedicated MRI to 2D radiographic images and cone beam CT in the assessment of lower third molars: a prospective study.","authors":"Jennifer Christensen, Rubens Spin-Neto, Louise Hauge Matzen","doi":"10.1093/dmfr/twag011","DOIUrl":"10.1093/dmfr/twag011","url":null,"abstract":"<p><strong>Objectives: </strong>The aim of this study is to show that the use of 0.55 T MRI combined with a \"dental-dedicated\" coil produces images of sufficient diagnostic value to assess lower third molars (LTMs), that are not inferior to currently utilized dental-oriented radiographic images.</p><p><strong>Methods: </strong>Dental-dedicated MRI (ddMRI) scans were acquired using a Magnetom Free.Max 0.55 T scanner combined with a dedicated dental coil. Three observers assessed all images according to predefined criteria. Twenty percent of images were assessed twice by all observers. Kappa statistics were performed to assess intra- and inter-observer agreement as well as inter-modality agreement.</p><p><strong>Results: </strong>Dental-dedicated MRI was acquired on 67 patients (89 LTMs) in addition to initial radiographic exams (intraoral, panoramic, and/or CBCT). Inter-observer agreement for each modality ranged from low to perfect (intraoral/panoramic 0.480-0.942 [average 0.74], CBCT 0.218-1.000 [average 0.69], ddMRI -0.038 to 0.889 [average 0.53]). Intra-observer agreement ranged from low to perfect (intraoral/panoramic -0.047 to 1.000 [average 0.76], CBCT 0.389-1.000 [average 0.83], ddMRI -0.025 to 1.000 [average 0.61]). Inter-modality agreement ranged from low to high (intraoral/panoramic vs CBCT -0.078 to 0.743 [average 0.32], intraoral/panoramic vs ddMRI -0.078 to 0.752 [average 0.30], CBCT vs ddMRI 0.074-0.886 [average 0.49]).</p><p><strong>Conclusion: </strong>Dental-dedicated MRI could be a feasible diagnostic modality for LTM imaging. The modality shows promise for radiation-free imaging in the future.</p><p><strong>Advances in knowledge: </strong>This article is the first to demonstrate the use of ddMRI in LTM imaging and to compare the modality to existing modalities. The added value of this radiation-free modality can be beneficial to dentists and patients in the future.</p>","PeriodicalId":11261,"journal":{"name":"Dento maxillo facial radiology","volume":" ","pages":"354-361"},"PeriodicalIF":2.9,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146212421","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Analysis of the Generalizability of An Artificial Intelligence-Based Software for Tomographic Segmentation of Posterior Teeth-An External Validation Study.","authors":"Erielma Lomba Dias Julião, Gabriel Cunha Adiverci, Fernanda Bulhões Fagundes, André Ferreira Leite, Frederico Sampaio Neves, Deborah Queiroz de Freitas, Rocharles Cavalcante Fontenele, Reinhilde Jacobs, Francisco Haiter-Neto, Sergio Lins de- Azevedo-Vaz","doi":"10.1093/dmfr/twag018","DOIUrl":"https://doi.org/10.1093/dmfr/twag018","url":null,"abstract":"<p><strong>Objective: </strong>To evaluate the generalizability of an artificial intelligence (AI)-based software for automated segmentation of posterior teeth in cone beam computed tomography (CBCT) scans and to identify the variables that influence the need for refinement of automatic segmentations (AS).</p><p><strong>Methods: </strong>A total of 190 scans from 190 patients, acquired using five CBCT systems were imported into the Virtual Patient Creator (Relu, Leuven, Belgium) for AS. Two dental surgeons qualitatively assessed the segmentations of posterior teeth and refined those requiring correction. Manual segmentation (MS) of 20% of the sample was performed using Mimics software (Materialise, Leuven, Belgium). Performance was analyzed through voxel-by-voxel and surface-based comparisons, in addition to the evaluation of time efficiency. Associations between independent variables and the need for refinement were analyzed using mixed logistic regression (α = 5%).</p><p><strong>Results: </strong>Among the 1,005 teeth evaluated, only 12.7% required refinement. Age and the presence of brackets were significant predictors (p < 0.001). The unexplained variability was attributed mainly to the patients, with minimal influence from the CBCT systems. AS showed agreement with refined segmentations (R-AI) (IoU: 0.93-0.96; DSC: 0.96-0.98; Precision: 0.99-1.00; Recall: 0.94-0.96; Accuracy: 0.98-0.99; MAD: 0.05-0.07; RMSE: 0.06-0.14) and excellent performance compared to MS (IoU: 0.94; DSC: 0.97; Precision: 0.98; Recall: 0.95; Accuracy: 0.98; MAD: 0.05; RMSE: 0.09). AS was more time-efficient (12 [AIQ : 5]) compared to R-AI (202 [AIQ : 334]) and MS (1,726 [AIQ: 863]).</p><p><strong>Conclusion: </strong>The AI-based software demonstrated high accuracy and generalizability for automated segmentation of posterior teeth in CBCT scans.</p>","PeriodicalId":11261,"journal":{"name":"Dento maxillo facial radiology","volume":" ","pages":""},"PeriodicalIF":2.9,"publicationDate":"2026-04-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147662390","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Automatic Multi-Class Classification of 3D Relationships Between the Mandibular Third Molar and Canal on CBCT Using a Geometry-Aware Network.","authors":"Hyomin Kim, Ji Yong Han, Su Yang, Jo-Eun Kim, Kyung-Hoe Huh, Sam-Sun Lee, Min-Suk Heo, Jeong Joon Han, Joo-Young Park, Won-Jin Yi","doi":"10.1093/dmfr/twag010","DOIUrl":"https://doi.org/10.1093/dmfr/twag010","url":null,"abstract":"<p><strong>Objectives: </strong>Accurate three-dimensional (3D) evaluation of the spatial relationship between the mandibular third molar (M3) and the mandibular canal (MC) is critical for assessing the risk of nerve injury during M3 extraction. The purpose of this study was to automatically classify the diverse spatial relationships between M3 and MC into five distinct categories based on the degree of anatomical contact or involvement as well as spatial proximity on CBCT images using a geometry-aware deep learning framework.</p><p><strong>Methods: </strong>The 3D spatial relationships between M3 and MC were categorized into five distinct types. The proposed framework consisted of two stages: first, the modified mAttUNet, an improved 3D U-Net architecture augmented with an attention mechanism was used for segmentation of M3 and MC; second, the proposed DenseAttNet was developed for multi-class classification. By incorporating dense attention mechanisms and signed distance map (SDM) inputs, the proposed network effectively captured both geometric and anatomical features, leading to more accurate and reliable multi-class classification.</p><p><strong>Results: </strong>The mAttUNet outperformed other conventional models, achieving the highest segmentation performance with average precision scores of 0.96 for MC and 0.84 for M3. The proposed DenseAttNet demonstrated superior and consistent performance, achieving an overall AUC of 0.97 and maintaining high accuracy across all five relationship types, effectively and reliably distinguishing the various spatial relationships between MC and M3.</p><p><strong>Conclusions: </strong>This automated and accurate classification of M3-MC spatial relationships offers valuable clinical utility, supporting enhanced risk evaluation and optimized surgical planning, and ultimately helping to reduce complications such as inferior alveolar nerve injury.</p>","PeriodicalId":11261,"journal":{"name":"Dento maxillo facial radiology","volume":" ","pages":""},"PeriodicalIF":2.9,"publicationDate":"2026-03-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147510482","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}