BMC Oral Health最新文献

Background: The integration of artificial intelligence (AI) in healthcare and medical education has advanced rapidly, with conversational AI systems gaining attention for their potential in academic assessment and clinical reasoning. This study aimed to evaluate AI chatbots' performance on restorative dentistry questions from the Turkish Dental Specialty Examination (DUS), a high-stakes national exam assessing theoretical and clinical knowledge.

Methods: An in silico, cross-sectional, comparative design was employed. A total of 190 multiple-choice questions (MCQs) from 19 DUS sessions between 2012 and 2025 were obtained from the Assessment, Selection, and Placement Center (ÖSYM) website. After excluding annulled items, 188 questions were analyzed. Eight AI chatbots (ChatGPT-3.5, ChatGPT-4o Free, ChatGPT-4o Plus, Claude Sonnet 4, Microsoft Copilot, DeepSeek, Gemini 1.5, and Gemini Advanced) were tested using a standardized single-attempt protocol in Turkish. Performance measures included accuracy, response length, and response time. Questions were categorized by year, content domain, and length for subgroup analyses. Statistical analyses were conducted in Python using standard libraries. Descriptive statistics and Pearson's correlation were calculated, while comparisons involved the Shapiro-Wilk test, Levene's test, Kruskal-Wallis test, and Dunn's post hoc test, with significance set at p < 0.05.

Results: No significant difference was found in overall accuracy (p = 0.18). However, response time and word count differed significantly (p < 0.001). Gemini Advanced showed the highest accuracy (96.28%), followed by ChatGPT-4o Plus (93.62%). Gemini 1.5 produced the longest yet fastest responses, while DeepSeek had the lowest accuracy and slowest responses. Accuracy remained stable across years but varied by topic, with lower performance in complex areas such as cavity preparation. In case-based questions, Gemini Advanced, Gemini 1.5, and ChatGPT-4o Plus achieved 100% accuracy. Performance in image-based questions was inconsistent, underscoring limitations in visual reasoning.

Conclusions: AI chatbots demonstrated high accuracy in answering restorative dentistry exam questions, with Gemini Advanced, ChatGPT-4o Plus, and Gemini 1.5 showing superior performance. Despite differences in response time and content length, their potential as supplementary tools in dental education is evident, warranting further validation across specialties and contexts.

Trial registration: Not applicable.

Background: Restorative procedures following endodontic treatment are crucial for preserving the functional integrity of the tooth, preventing coronal leakage, and protecting the root canal system from contamination. The present study aims to compare the effects of different marginal elevation strategies and restorative material combinations on stress distribution and fatigue life, using three-dimensional (3D) finite element analysis (FEA) of mandibular first molar models prepared with DO, MO, and MOD cavity designs.

Methods: The anatomical structure of tooth number 36 was scanned using cone beam computed tomography. Three different cavity types were then created on the tooth. In the same model, deep margin elevation (DME) was applied in single and double layers of filler materials, each with 1 mm step increments. For the FEA, bulk-fill composite, conventional composite, hybrid composite, and flowable composite (Smart Dentin Replacement; SDR) were used as filling materials, whereas polymer-infiltrated ceramic networks (PICN) and lithium disilicate (LDS) were used as endocrown materials. After modeling, the stress values obtained from the analysis were evaluated according to the von Mises stress criterion and the results were reported in megapascals (MPa).

Results: The DME technique provided uniform stress distribution, increased fracture resistance, and increased the incidence of repairable fractures. Among the restorative materials studied, SDR demonstrated superior stress management and marginal adaptation. PICN restorations provided more uniform load distribution, supporting overall restoration integrity. This in silico study suggests that the application of the DME technique, combined with the appropriate selection of restorative and crown materials, may have a positive influence on stress distribution and fracture behavior in endodontically treated molars. However, as this study was conducted solely based on FEA and relies on an in vitro modeling approach, the findings need to be supported by long-term clinical data. Further laboratory and clinical studies are required to validate these mechanical findings under real-world clinical conditions.

Conclusions: The use of DME combined with appropriate material selection can significantly improve the mechanical behavior of restorative treatments. These findings can guide clinicians in selecting materials and techniques for optimal long-term outcomes.

Background: The purpose of this study is to develop a deep learning model that can identify the maxillary sinus, posterior superior alveolar artery(PSAA), and alveolar ridge, and evaluate its diagnostic performance. Based on this, relevant parameters for preoperative design of maxillary sinus elevation can be measured to achieve intelligent preoperative design for maxillary posterior tooth implantation surgery.

Methods: A total of 2400 CBCT slices from patients with maxillary posterior tooth loss was selected as the initial dataset. Anatomical structure annotation and enhanced YOLOv11 architecture were used for model training to achieve segmentation of maxillary sinus, PSAA, and alveolar ridge. Intersection over union (IoU), average precision (AP), average recall (AR) and the Euclidean distance were used to evaluate the accuracy of structure segmentation. On the basis of the segmentation of the three important anatomical structures mentioned above, five anatomical parameters (A1-A5) related to maxillary posterior tooth implantation were set, and their errors were statistically analyzed.

Results: The median IoU for maxillary sinus segmentation was 0.945 (IQR: 0.934-0.951, 95%CI: 0.935-0.941), while the median IoU for PSAA segmentation was 0.991 (IQR: 0.982-1.000, 95%CI: 0.948-0.974). The model achieved an average precision of 0.902 ± 0.023 and a recall of 0.937 ± 0.024 for PSAA segmentation. For alveolar crest localization, the mean Euclidean distance errors between predicted and ground-truth landmarks were 0.50 ± 0.31 mm and 0.38 ± 0.24 mm for the two key points, respectively. 95% of AI prediction errors for A1-A4 were within 1 mm, while 95% of AI prediction errors for A5 were within 10 mm².

Conclusions: The enhanced YOLOv11 framework reliably and autonomously identifies critical anatomical structures for maxillary sinus elevation including the maxillary sinus, PSAA, and maxillary alveolar crest in CBCT images. This model enables the acquisition of reliable clinical parameters, demonstrating its potential for future intelligent assisted preoperative evaluation and design of maxillary posterior dental implant surgery.

Objectives: To compare the clinical, radiographic, and patient-reported outcomes of dental implant placement using static computer-assisted implant surgery (sCAIS) versus a robotic-assisted surgery (RAS) system over a 12-month follow-up period.

Materials and methods: This retrospective comparative study was designed as an exploratory analysis of clinical, radiographic, and patient-reported outcomes. It included 36 patients receiving 103 implants, allocated to either the sCAIS group (14 patients, 54 implants) or the RAS group (22 patients, 49 implants). Assessments included implant placement accuracy, marginal bone loss (MBL) at 6 and 12 months, pocket depth and bleeding on probe for peri-implant soft tissue health, implant survival/success rates, and patient-reported outcomes via the OHIP-14 questionnaire and a Visual Analog Scale (VAS) for intraoperative satisfaction.

Results: The RAS group demonstrated significantly higher placement accuracy, with lower deviations in vertical and horizontal dimensions at both the implant shoulder and apex (all p < 0.05). While MBL was comparable at 6 months, at the 12-month follow-up the RAS group showed significantly less bone loss on the mesial (p = 0.0117), distal (p = 0.0114), and lingual/palatal aspects (p = 0.0060). Peri-implant soft tissue health and cumulative success rates (98.1% for sCAIS vs. 100% for RAS) were not significantly different. Patients reported significantly higher intraoperative satisfaction with sCAIS (p < 0.001), but the RAS group reported a significantly better postoperative quality of life (lower OHIP-14 scores, p = 0.0010).

Conclusion: While both robotic and static guided systems yield high clinical success rates, robotic-assisted surgery provided greater placement accuracy and suggested a potential for better long-term preservation of peri-implant bone. These preliminary findings, which require confirmation through larger prospective studies, came at the cost of lower intraoperative satisfaction, even as patients reported a better postoperative quality of life.

Background: Orthognathic surgery aims to improve jaw function and facial aesthetics through bilateral sagittal split osteotomy and Le Fort I osteotomy. Recent treatment goals emphasize careful evaluation of aesthetic outcomes, particularly in the nasolabial area, as repositioning the upper jaw can lead to significant soft tissue changes. This study investigates whether nasotracheal intubation affects nostril symmetry in patients undergoing Le Fort I osteotomy with/without cinch sutures.

Methods: A retrospective analysis of adult patients (ages 18-30, ASA I-II) who underwent Le Fort I surgery with nasotracheal intubation at Erciyes University from 2012 to 2020 was conducted. Preoperative and at six months postoperative, 3D images were analyzed to measure nostril width (NW). Patients were categorized into Group I (with cinch sutures) and Group II (without cinch sutures). Soft tissue changes were assessed using the 3dMD imaging system.

Results: Eighty-five patients were included. Significant changes in nostril width were observed between preoperative and six-month postoperative assessments in both groups. Right intubation led to increased right nostril diameter in both groups, while the left nostril showed significant change only in the cinch group. For left intubation, no significant changes were observed in nostril dimensions in the cinched group. The findings indicate that intubation side significantly influences nostril symmetry, particularly in cases of right nasotracheal intubation. The use of cinch sutures does not fully mitigate the widening effect, suggesting that the timing of cinch suture placement may be crucial.

Conclusion: This study demonstrates that the nasotracheal intubation side may influence postoperative nostril width following Le Fort I osteotomy, particularly in patients receiving alar cinch sutures. The findings suggest that the physical presence of a nasotracheal tube during wound closure could interfere with the accurate assessment of alar base width.

Background: The proper and reliable functioning of the bonding mechanism in the presence of saliva is a challenging process. The aim of this study was to evaluate the effect of decontamination methods on surface roughness, contact angle, and shear bond strength of various framework materials bonded to composite resin cement.

Methods: A total of 180 specimens were prepared, with 60 specimens each from zirconia and two different production methods of CoCr alloys. The specimens were randomly separated to 6 subgroups for each material; no saliva contamination served as a control group. Group CO: saliva contamination, group AW: saliva contamination + air dried and water rinsed, group S: saliva contamination + air abrasion, group K and group Z were treated with different cleaning solutions after saliva contamination. Prior to resin cement application, zirconia primer was applied to the zirconia specimens, while alloy primer was used for the CoCr alloys specimens. Shear bond strength test was measured with universal test machine (3340, Instron Corp., Wycombe, UK). Fracture analysis was evaluated with stereomicroscope. Water contact angle was measured using a goniometer, and surface roughness was assessed with a profilometer. Statistical analysis was conducted using ANCOVA and multiple comparison tests at a 0.05 significance level.

Results: All decontamination methods enhanced SBS values compared to the saliva-contaminated groups. However, only sandblasting (Group S) revealed significantly higher SBS values than the other decontamination methods for all frameworks (p < 0.05). Decontamination methods altered the surface roughness and contact angle values of all framework materials.

Conclusions: Decontamination process after saliva contamination enhanced the bond strength and wettability. Cleaning solution may be the alternative decontamination technique to sandblasting for framework materials.

Objectives: This study investigated the accuracy and learning curve of dental postgraduate trainees in robot-assisted implant surgery (r-CAIS) and examined the characteristics of learning curve.

Materials and methods: After receiving comprehensive training, eleven dental trainees used r-CAIS to place three implants, 3 trials per day over four experimental days, with each experimental day occurring every 72 h. Trueness was assessed by comparing postoperative model scans with preoperative planning. The total time was recorded during the experiment and subsequently categorized into preparation and surgery times.

Results: The average deviations for platform, apex, and angular were 0.91 ± 0.39 mm, 0.93 ± 0.38 mm, and 1.38 ± 0.61°, respectively. The mean total time, preparation time, and surgery time were 954.36 ± 222.69 s, 321.36 ± 119.30 s, and 632.53 ± 140.30 s, respectively. Analysis indicated that there were no statistically significant differences in trueness between different experimental days and trials for platform, apex, and angular (p-values: Day: 0.0924, 0.2735, 0.3223; Trial: 0.6647, 0.7873, 0.5270). However, all time measures (total, preparation, and surgery time) showed significant reductions as days and trials increased (p-values < 0.0001 for all). Additionally, there was a significant interaction between days and trials for surgery time (p = 0.0123). Normality tests revealed that the trueness data and surgery time were not normally distributed on day 1, but conformed to a normal distribution on the rest days.

Conclusion: The results demonstrated stability in trueness during the learning phase of r-CAIS, although there could still be a learning curve in terms of time metrics. Additionally, the accumulation of experience may exert varying influences during the preparation and surgical phases.

Clinical significance: The r-CAIS can enable dental trainees to achieve high trueness during their initial exposure to surgical procedures. Understanding the differing learning processes associated with various r-CAIS procedures can help develop more effective training programs and promote the broader clinical applications of r-CAIS.