BMC Oral Health最新文献

Objective: This study evaluated multiple deep learning approaches for automated classification of dental conditions in panoramic radiographs, comparing custom convolutional neural networks (CNNs), hybrid CNN-machine learning models, and fine-tuned pre-trained architectures, comparing the performance of custom convolutional neural networks (CNNs), hybrid CNN-machine learning models, and fine-tuned pre-trained architectures for detecting fillings, cavities, implants, and impacted teeth.

Methods: A dataset of 1,512 panoramic X-ray images with 11,137 manually annotated bounding boxes for four dental conditions (fillings, cavities, implants, and impacted teeth) was analyzed, with regions of interest extracted using expert annotations for subsequent AI-based classification. Class imbalance was addressed through random downsampling, creating a balanced dataset of 894 samples per condition. Multiple approaches were evaluated via 5-fold cross-validation: a custom CNN, hybrid models combining CNN features with traditional classifiers (Support Vector Machine, Decision Tree, Random Forest), and fine-tuned pre-trained networks (VGG16, Xception, ResNet50). Performance was assessed using accuracy, precision, recall, and F1-score metrics.

Results: The hybrid CNN-Random Forest model achieved the highest accuracy of 85.4 ± 2.3% with macro-F1 score of 0.843 ± 0.028, representing an 11% point improvement over the custom CNN (74.29% accuracy, 0.724 macro-F1). VGG16 demonstrated superior pre-trained architecture performance (82.3 ± 2.0% accuracy, 0.817 macro-F1), followed by Xception (80.9 ± 2.3%) and ResNet50 (79.5 ± 2.7%). CNN + Random Forest exhibited exceptional fillings detection (F1: 0.860 ± 0.033) with balanced multi-class performance. Systematic misclassifications between morphologically similar conditions revealed inherent diagnostic challenges.

Conclusion: Hybrid CNN-based approaches combining feature extraction with Random Forest classification provide superior discriminative capability for dental condition detection on manually annotated regions compared to standalone architectures. While computationally efficient hybrid models show promise as supportive diagnostic tools, observed misclassification patterns indicate these AI systems should serve as adjuncts to clinical expertise, requiring prospective validation studies.

Background: The retrieval of fractured high-speed handpiece burs during mandibular third molar surgery, particularly those adjacent to the mandibular canal or complicated by infection, presents a significant clinical challenge. Conventional techniques often fail in such scenarios owing to anatomical constraints and limited resources in primary care settings. This case report highlights the innovative application of a modified mild steel probe technique for managing this rare but serious complication, demonstrating its adaptability across diverse clinical environments.

Case presentation: We present two cases with deeply embedded bur fragments. Patient 1 involved a 28-year-old female with a 2.8-mm fragment near the mandibular canal. Using a novel, real-time shapable (≤ 120°) unquenched mild steel probe and a "three-point" localization method under panoramic radiography, the fragment was successfully retrieved in 35 min, with resolved neurosensory deficits at the 1-month follow-up. Patient 2 involved a 27-year-old female with a larger fragment (3.2 mm × 1.5 mm) complicated by infection and bone destruction. Under CBCT guidance, an upgraded probe was shaped into a 135° reverse hook and used with a "layered dissection" technique, achieving retrieval in 25 min. The patient experienced significant symptom relief by day 3 and near-complete bone regeneration at 3 months.

Conclusions: This case report demonstrates that a malleable, non-quenched mild steel probe technique may offer a material-independent, angle-adaptive, and cost-effective alternative for retrieving deeply embedded bur fragments. Its core innovation lies in overcoming the inflexibility of conventional rigid instruments and the material limitations of magnetic systems, providing a practical solution particularly in resource-conscious settings or for non-ferromagnetic fragments.

Background: To investigated the combined effect of adhesive type and film thickness on dentin bond strength by comparing one and two-step universal adhesives applied in single and double layers.

Methods: Forty caries-free human third molars were collected and standardized Class I cavities were prepared. Specimens were restored using one-step (Clearfil Tri-S Bond Universal, Japan) and two-step (G2-Bond Universal, Japan) universal adhesive systems. Each group was subdivided into single and double-layer applications, light-cured, and restored with a universal composite resin. Sticks with a cross-sectional 1 × 1 mm² were prepared for testing. Adhesive film thickness was measured under a stereomicroscope at ×80 magnification. Micro-tensile bond strength (µTBS) was evaluated either immediately or after 5000 thermocycles (TC). Failure modes were assessed under ×40 magnification. Wilcoxon Signed Ranks Test was used for paired comparisons of immediate and after TC tensile bond strengths. Mann-Whitney U test compared single vs. double layers of the adhesives. For adhesive film thickness, Independent Samples t-test was applied for group comparisons (p < 0.05).

Results: Double-layer application produced significantly thicker adhesive films for Clearfil Tri-S Bond and G2-Bond Universal compared with single-layer application (p < 0.05). However, no significant differences in µTBS were observed between single and double-layer groups of either adhesive, regardless of aging. Thermocycling significantly reduced bond strength for both systems (p < 0.05). Overall, the two-step adhesive demonstrated higher µTBS values than the one-step adhesive.

Conclusion: Although double-layer application increased adhesive film thickness, it did not improve dentin bond strength. The two-step universal adhesive system showed superior bonding performance compared with the one-step system, both immediately and after thermocycling.