Radiographic evaluation of the distance between the restoration margin and the alveolar bone crest in dental implant patients: A retrospective study.

OBJECTIVES: The recently introduced Implant Disease Risk Assessment (IDRA) identifies a restoration margin-alveolar bone crest (RM-AC) distance of less than 1.5 mm as a key risk factor for peri­implant disease among eight major risk factors. This study evaluated the impact of the RM-AC distance on marginal bone loss (MBL) through radiographic analysis. METHODS: This retrospective cross-sectional study included 77 partially edentulous patients (39 females and 38 males, aged 22 to 76 years) with 202 platform-switched conical connection implants, cement-retained, implant-supported fixed restorations, and bone-level implants placed between 2016 and 2021. Dental implants were followed for least 6 to 36 months at follow up functional loading. Study participants were categorized into Group A (RM-AC distance ≤ 1.5 mm, n = 69) and Group B (RM-AC distance > 1.5 mm, n = 133). Twelve patients in Group B and five patients in Group A had no history of periodontal disease. The MBL was measured radiographically from the most coronal point of the implant shoulder to the alveolar bone, and the RM-AC distance was measured from the restoration margin to the alveolar crest. Multinomial logistic regression analysis was used for statistical evaluation. RESULTS: The incidence of MBL in Group A was statistically significant and 3.42 times higher than that in Group B. The rate of MBL in periodontitis Stage 4 was found to be 26.31 times higher than that in periodontitis Stage 2. The incidence of MBL was 6.097 and 5.02 times higher with increasing implant diameter and length, respectively. CONCLUSION: This study conclusively demonstrates that RM-AC distance ≤ 1.5 significantly increases the risk of MBL, particularly in patients with a history of periodontal disease. CLINICAL SIGNIFICANCE: This study highlights the critical role of maintaining an RM-AC distance greater than 1.5 mm in the prevention of MBL, particularly in patients with a history of periodontal disease. Since implant diameter and length have a significant impact on the risk of MBL, it emphasizes that implant demographics should also be carefully evaluated.

Comparison of deep learning methods for the radiographic detection of patients with different periodontitis stages.

OBJECTIVES: The objective of this study is to assess the accuracy of computer-assisted periodontal classification bone loss staging using deep learning (DL) methods on panoramic radiographs and to compare the performance of various models and layers. METHODS: Panoramic radiographs were diagnosed and classified into 3 groups, namely "healthy," "Stage1/2," and "Stage3/4," and stored in separate folders. The feature extraction stage involved transferring and retraining the feature extraction layers and weights from 3 models, namely ResNet50, DenseNet121, and InceptionV3, which were proposed for classifying the ImageNet dataset, to 3 DL models designed for classifying periodontal bone loss. The features obtained from global average pooling (GAP), global max pooling (GMP), or flatten layers (FL) of convolutional neural network (CNN) models were used as input to the 8 different machine learning (ML) models. In addition, the features obtained from the GAP, GMP, or FL of the DL models were reduced using the minimum redundancy maximum relevance (mRMR) method and then classified again with 8 ML models. RESULTS: A total of 2533 panoramic radiographs, including 721 in the healthy group, 842 in the Stage1/2 group, and 970 in the Stage3/4 group, were included in the dataset. The average performance values of DenseNet121 + GAP-based and DenseNet121 + GAP + mRMR-based ML techniques on 10 subdatasets and ML models developed using 2 feature selection techniques outperformed CNN models. CONCLUSIONS: The new DenseNet121 + GAP + mRMR-based support vector machine model developed in this study achieved higher performance in periodontal bone loss classification compared to other models in the literature by detecting effective features from raw images without the need for manual selection.