Chairs:
danisia haba
150: SEGMENTATION OF PULP AND PULP STONES WITH AUTOMATIC DEEP LEARNING IN PANORAMIC RADIOGRAPHS: AN ARTIFICIAL INTELLIGENCE STUDY
M. Firincioglulari1, M. Boztuna2, O. Mirzaei3, T. Karanfiller4, N. Akkaya5, K. Orhan6
1Cyprus International University, Faculty of Dentistry, Department of Dentomaxillofacial Radiology, Nicosia, Cyprus, 2Cyprus International University, Faculty of Dentistry, Department of Dentomaxillofacial Surgery, Nicosia, Cyprus, 3Near East University, Department of Biomedical Engineering, Faculty of Engineering, Nicosia, Cyprus, 4Cyprus International University, Department of Management Information Systems, School of Applied Sciences, Nicosia, Cyprus, 5Near East University, Department of Computer Engineering, Applied Artificial Intelligence Research Center, Nicosia, Cyprus, 6Ankara University, Faculty of Dentistry, Department of Dentomaxillofacial Radiology, Ankara, Turkey
Aim: Pulp stones are calcified masses of various sizes that may be found in dental pulp which can impact dental procedures. The purpose of this study was to measure the diagnostic accuracy of detecting pulp and pulp stone calcifications using artificial intelligence algorithms on panoramic radiographs.
Material and Methods: 713 panoramic radiographs that included at least one pulp stone were identified retrospectively and included in the study. 5,085 pulps and 4,675 pulp stones were marked on these radiographs using CVAT labeling software.
Results: In the test dataset, the AI model segmented 220 panoramic radiographs for pulp and 480 panoramic radiographs for pulp stone. The segmentation of pulp and pulp stone for precision and accuracy were 0.85, 0.998, 0.771, and 1, respectively. Artificial intelligence systems have the potential to overcome clinical problems. AI can facilitate the assessment of pulp stones and pulp based on panoramic radiographs.
Conclusion: Pulp and pulp stones can successfully be identified using artificial intelligence algorithms. This study provides evidence that artificial intelligence software using deep learning algorithms can be valuable adjunct tools in aiding clinicians in radiographic diagnosis. Further studies with larger data sets are required to increase the diagnostic accuracy rates of the artificial intelligence models.
190: DOSE MEASUREMENT AND ANALYSIS OF THE DOSE DISTRIBUTION TO RELEVANT RADIOSENSITIVE ORGANS DURING THE PRODUCTION OF DIGITAL PANORAMIC TOMOGRAPHS
F. Schindler1, H. Karle2, R. Schulze3
1Universitätsmedizin Mainz, Polyclinic for Dental Prosthetics and Materials Science, Mainz, Germany, 2Universitätsmedizin Mainz, Clinic and Polyclinic for Radiooncology and Radiotherapy, Mainz, Germany, 3Dental Clinics of the University of Bern (ZMK), Oral Diagnostic Sciences, Bern, Switzerland
Aim: This study aimed to provide data on the accumulated radiation doses in critical organs in the human skull. A better understanding of the exposure geometries of modern panoramic X-ray units and insights into radiation exposure profiles are crucial for optimizing radiological procedures and minimizing radiation risk to patients.
Material and Methods: The research involved direct measurement of absorbed X-ray radiation using an ionization chamber placed in a female Alderson-Rando X-ray phantom. The measurements were performed on four different panoramic X-ray units. Radiochromic X-ray films were placed in the phantom to visually display and analyze the dose distribution. The blackening patterns of the films provided information about the spatial distribution of the dose within the skull. In the following, statistical methods, such as Mann-Withney-U-Tests, were used to derive conclusions on identified doses.
Results: The study revealed clear differences in the radiation exposure levels of various panoramic X-ray devices. In particular, one of the analyzed devices showed a significantly lower radiation exposure compared to others. In addition, the analysis of the X-ray films provided insights into the spatial dose distribution during a panoramic slice exposure.
Conclusion: The research highlights the need to carefully evaluate radiation exposure when selecting diagnostic equipment to minimize patient risks. The measured dose differences between panoramic radiographs underscore the need to refine radiological procedures. The results of this study deepen the understanding of how radiation dose is distributed in the human head and provides insights into the imaging geometries of panoramic radiographs.
251: FRACTAL ANALYSIS OF MANDIBULAR BONE STRUCTURE AFTER RADIOTHERAPY: A COMPARATIVE STUDY
U. Seki1, A. Kuran1, A. Üzel1, S. Çelik1, E.A. Sinanoğlu1
1Kocaeli University, Kocaeli, Turkey
Aim: The effects of radiation on bone tissue and the mechanisms of osteoradionecrosis (ORN) are not fully understood. Fractal analysis (FA) quantifies complex structures and fractal dimension values describe bone quality. The aim of this study was to investigate the effect of radiotherapy on mandibular bone tissue in patients without ORN who had gone radiotherapy for head and neck cancer using FA.
Materials-Methods: Panoramic radiographs (OPG)s of patients who had undergone radiotherapy (RTG) were retrieved from archives of Department of Oral and Maxillofacial Radiology, Faculty of Dentistry, Kocaeli University between 2019 and 2023. OPGs with any signs of ORN were excluded. Control group (CG) was formed with age and sex matched healthy patients. FA was conducted on the OPGs with the box-counting algorithm in three regions of interests (ROI) as ROI1: the upper part of the ramus, ROI2: the angulus and ROI3: the anterior part of the mental foramen.
Results: A total of 27 OPGs of RTG and their 27 age and sex matched OPGs of CG were analysed. Although FA was lower in RTG compared to CG in ROI2 and ROI3, there was no significant difference between the FA values of RTG and CG (p>0.05).
Conclusion: Our results indicate that there was no significant difference in the evaluation with FA to differentiate radiographic trabecular changes between patients receiving head and neck radiotherapy and their controls. Further studies with larger populations are required to explore FA‘s role in detecting osteoporotic changes in these patients before ORN.
25: NEURAL DEFORMABLE CBCT
R.K. Schulze1, L. Birklein2, E. Schoemer2, R. Brylka3, U. Schwanecke3
1University of Bern, Division of Oral Diagnostic Sciences, Bern, Switzerland, 2Johannes Gutenberg University, Institute of Computer Science, Mainz, Germany, 3RheinMain University of Applied Sciences, Labor für Computer Vision und Mixed Reality, Wiesbaden, Germany
Objectives: To use neural deformable radiance fields for marker-free motion correction in CBCT
Material and Methods: in 2021 Mildenhall et al. [1] proposed a framework for synthesizing novel views of complex scenes by optimizing an underlying continuous function from few input images. Using this technology, we developed a motion correction and reconstruction method combining inverse neural rendering and neural field deformation to enhance CBCT reconstruction. The method is capable to represent different motion patterns such as head movements, separate jaw motions and swallowing. It was evaluated using synthetic scenarios as well as real-patient images.
Results: Using the method termed «neural deformable CBCT» we achieved reconstructions of remarkable quality. The method also worked well in a motion deformed 4D-CBCT dataset of the lung.
Conclusions; The proposed framework produces high-quality 3D and 4D state-of-the-art reconstructions from motion-beset CBCT data.
171: ORAL MUCOSA AND SALIVARY GLANDS‘ RADIATION DOSE IN THREE-DIMENSIONAL TOMOGRAPHIC IMAGING OFIMPACTED WISDOM TEETH
S. Efeoğlu1
1Ankara University, Faculty of Dentistry, Ankara, Turkey
Objective (Aim): The aim of this study is to examine the absorption doses of the oral mucosa and salivary glands in two different cone beam computed tomography devices and different acquisition modes, resulting from 3D imaging used to evaluate the pre-surgical positions of impacted wisdom teeth and their relationships with anatomical structures.
Material and Methods: The PCXMX 2.0 software simulated the total dose received by the patient. Newtom7G and Newtom GO devices were used with different FOV areas and imaging protocols (best, regular, and low dose). CBCT images were obtained with the lowest voxel size according to the protocols.
Results: When the data obtained from both devices areevaluated, the highest absorbed radiation dose in thepatient’s oral mucosa and salivary gland is seen in Newom7G’s best mode.The lowest absorbed radiation dose is seenin Newtom CT-Go low mode.
Conclusion: Comparing radiation doses from imagesobtained in all modes, Newtom 7G caused greater radiationexposure to tissues than Newtom GO. Oral mucosa wasidentified as the most affected tissue in terms of radiationexposure.
71: THREE-DIMENSIONAL FACIAL IMAGING: A QUALITATIVE ANALYSIS OF FACIAL SCANNERS
T. Jindanil1, R. Ponbuddhichai1, C. Massant1, L. Xu1,2, R. Cavalcante Fontenele1, M. Cadenas de Llano Perula3, R. Jacobs1,4
1KU Leuven, Department of Imaging and Pathology, Leuven, Belgium, 2Huazhong University of Science and Technology, Department of Stomatology, Wuhan, China, 3KU Leuven, Department of Oral Health Sciences – Orthodontics, Leuven, Belgium, 4Karolinska Institute, Department of Dental Medicine, Stockholm, Sweden
Aim: To qualitatively assess the perceptions of medical professionals and non-professional observers in terms of similarity and clinical applicability of facial scanning machines based on the principles of stereophotogrammetry and structured light.
Material and Methods: Facial images of 20 subjects (12 females and 8 males) were obtained using a professional camera (Nikon D800) and facial scanners (Vectra H1, RAYFace RFS200, and iReal 2E). Subjects rated the Overall Similarity Score (OSS), as well as the comfort, while practitioners measured the time and rated the user-friendliness of each scanner. The Instant Similarity Score (ISR) and Similarity Score (SS) were evaluated by seven professionals and six non-professionals, with a reliability assessment conducted. The machines were coded as: A (Vectra H1), B (RAYFace RFS200), C (iReal 2E).
Results: Scanner A had the best ISR, followed by scanners B and C (p<0.05). All scanners showed good OSS without statistically significant difference. Strong intra-observer (72.9) and inter-observer (78.8) agreements were recorded. Scanner B had the fastest total scanning time, while Scanners A and C took the same amount of time. The retake rates for Scanners A, B, and C were 10%, 15%, and 35%, respectively. The comfort and user-friendliness of all machines ranged from moderate to good.
Conclusion: All scanners showed good performance and suitability for clinical use with varying pros and cons. Professionals and non-professionals had similar perceptions on similarity, without scanner preference noted for subjects. Clinicians should consider machine capabilities alongside time, retake rates, user-friendliness, and patient comfort to optimally balance clinical needs.
112: MEDICAL – DENTOMAXILLOFACIAL RADIOLOGIST COLLABORATION IN ANGIOGRAPHY FOR LIPS: A CASE REPORT
F.R. Ramadhan1, K. Krisnandi2, M. Priaminiarti1
1Universitas Indonesia, Jakarta, Indonesia, 2Private Hospital, Bogor, Indonesia
Aim: The objective of this article is to report patient with abnormality in lip using Computed Tomography – Angiography (CT-A) examination and to describe collaborative work between medical radiologist and dentomaxillofacial radiologist.
Material methods: A 25-year-old woman was referred from the oral surgeon’s private clinic to the radiology department’s private hospital for an angiography with a complaint of swelling on the lower lip. The examination showed a lump on the lower lip’s mucosa with no distinctive colour from the surrounding area. The procedure began with several preparation steps in patients, followed by the injection of contrast IV and the CT scanning. The CT-A result showed an alteration on the diameter branch of the inferior labial artery pars inferior m. orbicularis oris. There was no other abnormality finding.
Results: A dentomaxillofacial radiologist currently has the competency in expertizing abnormalities found in dental and maxillofacial area from imaging modalities. Teamwork between medical and dentomaxillofacial radiologist including discussion, supervision, interpreting as well as establishing a radiodiagnosis were performed to have a better exam result. On the other hand, CTA in lips was rarely reported in article. This modality could be an option as soft tissue imaging within the mouth might be challenging.
Conclusion: Based on the radiographic features, this lip anomaly in this case leads to a suspect of vascular aneurism. Multidisciplinary approach involving medical and dentomaxillofacial radiologist could be done when dealing with this particular case.
146: COMPUTER SIMULATION FOR THE TREATMENT OF POST-TRAUMATIC OCCLUSAL SEQUELAE
M.A. Habi1, A. Bourihane2
1Oran Faculty of medecine, Oral and maxillo facial surgery, Oran, Algeria, 2Algiers Faculty of medecine, Algiers, Algeria
Aim: Facial traumatology occupies an important aspect and constitutes in itself a primary interest for the maxillofacial surgeon, Therefore, initial treatment must meet functional and morphological criteria, otherwise the after-effects complicate the therapeutic plan and compromise the prognosis. Computer planning also makes it possible to make a diagnosis based on a cephalometric study as well as a precise therapeutic simulation.
Material and methods: a clinical case concerning a 42-year-old man victim of isolated facial trauma whose initial intervention was followed by an occlusal disorder as well as difficulty in nasal breathing. The therapeutic project consisted of a cephalometric study, an osteotomy simulation and a custom-made retention splint.
Results: the application of data saves time during surgery as well as stability of the result.
Conclusion: the computer tool has now become an important step in the therapeutic arsenal, in that it facilitates the diagnostic and therapeutic approach in orthognathic surgery without supplanting previous methods.