Chairs:
erwin berkhout
ralf schulze
166: CBCT AFTER OPG FOR HIGH-RISK WISDOM TOOTH SURGERY REDUCES PATIENT ANXIETY AND ALLOWS A BETTER UNDERSTANDING OF SURGICAL RISKS.
R. Alhyari1,2, H. Anderson1, K. Khalaf1, A. Lalli1
1University of Aberdeen, Institute of Dentistry, Aberdeen, United Kingdom, 2Zarqa University, Faculty of Dentistry, Zarqa, Jordan
Aim: To determine any additional consequences of requiring an additional CBCT after an OPG for assessment of high-risk wisdom teeth.
Methodology: A cross-sectional hospital-based study was approved by the NHS Grampians‘ Quality Improvement and Assurance Team. Two independent cohorts were included, one sampled when having the CBCT and the other at the time of surgery. Two short surveys were designed, validated, and administered to participants to record their anxiety levels and evaluate their experiences and perspectives on the additional CBCT scan after already having an OPG. Data analysis involved descriptive analysis for quantitative data and thematic analysis for qualitative data.
Results: Our findings are the first to show markedly reduced pre-surgical anxiety levels and improved patient satisfaction associated with having a CBCT. All participants reported that the CBCT was worthwhile and valuable, even if it resulted in a delay in undergoing surgery and a further visit to the hospital. The majority found the CBCT helpful in better understanding surgery-associated risks. One-third of participants reported a change in their treatment decision, influencing whether to opt for a total or partial wisdom tooth removal. Qualitative findings revealed that the predominant theme of patients’ reported reasons for the need for additional CBCT, identified in 100% of the responses, was the assessment of wisdom tooth-nerve position.
Conclusion: In the absence of definitive evidence of the CBCT impact on the clinical outcomes of wisdom tooth surgery, our findings suggest CBCTs may be indicated where patients struggle to understand the associated risks and/or are extremely anxious.
70: NOVEL AI-BASED TOOL FOR PROSTHETIC CROWN SEGMENTATION SERVING AUTOMATED CBCT-IOS REGISTRATION IN CHALLENGING HIGH ARTIFACT SCENARIOS
B.M. Elgarba1,2, S. Ali3, R.C. Fontenele1, J. Meeus4, R. Jacobs1,5
1KU Leuven, OMFS-IMPATH Research Group, Department of Imaging and Pathology, Leuven, Belgium, 2Tanta University, Department of Prosthodontics, Tanta, Egypt, 3King Hussein Medical Center, Amman, Jordan, 4University Hospitals Leuven, Department of Oral and Maxillofacial Surgery, Leuven, Belgium, 5Karolinska Institute, Department of Dental Medicine, Stockholm, Sweden
Aim: To validate a cloud-based platform trained for AI-driven segmentation of prosthetic crowns on cone-beam computed tomography (CBCT) and subsequent multimodal intraoral scan (IOS) registration in the presence of high artifact expression.
Material and methods: A dataset of 30 IOS & CBCT jaw scans, each containing a minimum of four prosthetic crowns, was compiled. Cotton rolls were placed between cheeks and teeth during CBCT acquisition to enhance soft tissue delineation. AI-automated (AA) segmentation of prosthetic crowns and IOS-CBCT registration were compared to their corresponding semi-automatic (SA) approach for each case. Quantitative assessment compared AA‘s median surface deviation (MSD) and root mean square (RMS) in crown segmentation and IOS registration to those of SA. Additionally, segmented crown STL files were voxel-wise analyzed for comparison between AA and SA. Furthermore, qualitative assessment of crown segmentation evaluated the need for refinement, while registration assessment scrutinized registered-IOS alignment with CBCT teeth and soft tissue contours. Ultimately, the study compared time efficiency and consistency of both methods.
Results: Quantitatively, AA excelled with a 90% DSC for crown segmentation and an MSD of 0.04±0.03mm for CBCT-IOS registration, while also achieving 91% perfect matching of teeth and gingiva on CBCT, surpassing SA›s 80%. Furthermore, AA was significantly faster than SA in both segmentation (200 times) and registration (4.5 times) and demonstrated excellent consistency (Intraclass correlation coefficient: segmentation=0.99, registration=1).
Conclusion: The novel cloud-based platform demonstrated accurate, consistent, and time-efficient prosthetic crown segmentation, as well as IOS-CBCT registration in clinical scenarios with high artifact expression.
186: REFERENCE VALUES FOR THE ELASTICITY OF SUBMANDIBULAR LYMPHADENOPATHY CAUSED BY INFLAMMATORY DISEASES OF THE MAXILLOFACIAL REGION: A PRELIMINARY STUDY
B. Yılmaz1, M. Özaydın1, F. Aşantoğrol1
1Gaziantep University, Oral and Maxillofacial Radiology, Gaziantep, Turkey
Aim: This study aimed to determine an average elasticity value for reactive lymph nodes by examining submandibular lymphadenopathies caused by inflammatory diseases of the maxillofacial region using shear-wave elastography (SWE).
Material and methods: The study encompassed 43 patients older than 18 years with a total of 100 submandibular lymphadenopathies caused by inflammation in the maxillofacial region. Elastography was performed on all patients using a GE Logiq S8 ultrasound machine (GE Healthcare, Little Chalfont, UK). For each lymph node, 3 circular regions of interest, each 3 mm in diameter, SWE elasticity index values were calculated from a total of 9 relevant areas obtained from 3 consecutive sections. The 9 values were averaged and reported as mean SWE.
Results: Of the patients included in our study, 22 were male and 21 were female. The mean age of all individuals was 25.90±8.62 years. The mean SWE value for all participants was 21.37±11.10 kPa, while it was 20.87±9.95 kPa for males and 21.87±12.13 kPa for females. No statistically significant difference was found between the mean SWE values of males and females (p= 0.838).
Conclusion: Determination of reference values for the elasticity of submandibular lymphadenopathy using SWE is expected to be a guide in the diagnosis and treatment planning of lymph node pathologies in the oral and maxillofacial region, helping to distinguish between malignant lymph nodes, reactive lymph nodes caused by inflammation and other benign lymph nodes.
188: A NOVEL CONVOLUTIONAL NEURAL NETWORK-BASED TOOL FOR AUTOMATED SEGMENTATION OF PULP CAVITY STRUCTURES IN SINGLE-ROOTED TEETH USING CBCT
M.L. Slim1,2, R. Cavalcante Fontenele1,3, A. Oliveira Santos-Junior4,1, F. Sampaio Neves1,5, R. Jacobs1,3,6
1Katholieke Universiteit Leuven, OMFS IMPATH Research Group, Department of Imaging and Pathology, Faculty of Medicine, Leuven, Belgium, 2Saint Joseph University of Beirut, Department of Endodontics, Beirut, Lebanon, 3University Hospitals Leuven, Department of Oral and Maxillofacial Surgery, Leuven, Belgium, 4São Paulo State University (UNESP), Department of Restorative Dentistry, São Paulo, Brazil, 5Federal University of Bahia (UFBA), Department of Propedeutics and Integrated Clinic, Division of Oral Radiology, Bahia, Brazil, 6Karolinska Institutet, Department of Dental Medicine, Stockholm, Sweden
Aim: To develop and validate an artificial-intelligence (AI)-driven tool using convolutional neural networks for automated segmentation of pulp cavity structures in single-rooted teeth on cone-beam computed tomography (CBCT).
Material and Methods: After ethical approval, a dataset comprising 69 clinical CBCT scans was retrospectively gathered from the hospital’s database and divided into training (n=31, 88 teeth), validation (n=8, 15 teeth), and testing (n=30, 120 teeth) sets. Two endodontists and one dentomaxillofacial radiologist performed manual segmentation of CBCT-based pulp cavity structures to establish the ground truth, employing the Virtual Patient Creator platform (Relu, Belgium). The AI model underwent testing on 30 CBCT scans (n= 120 teeth), with the resulting three-dimensional (3D) models being qualitatively assessed and revised (R-AI) by a dentomaxillofacial radiologist. Comparative analysis involved AI, R-AI and manually segmented 3D models. Segmentation time for each method was recorded in seconds (s).
Results: The AI tool demonstrated highly accurate segmentation (Dice similarity coefficient [DSC] ranging from 89% to 93%; 95% Hausdorff distance [HD] ranging from 0.10 to 0.13 mm). AI-based segmentation outperformed human expert performance (p<.05), showing higher DSC and lower 95% HD values. Manual segmentation required significantly more time (2262.4 ± 679.1 s) compared to R-AI (94 ± 64.7 s) and AI (41.8 ± 12.2 s) methods (p<.05).
Conclusion: The novel AI-powered tool showed highly accurate and efficient automatic segmentation of pulp cavity structures in single-rooted teeth on CBCT, outperforming human expert performance. The current AI opens perspectives towards automated pulp canal segmentation in existing CBCTs for further use by dentists.