Convolutional neural network for automated classification of osteonecrosis and related mandibular trabecular patterns
Objective: The present study aimed to develop and validate a tool for the automated classification of normal, affected, and osteonecrosis mandibular trabecular bone patterns in panoramic radiographs using convolutional neural networks (CNNs). Methods: A dataset of 402 panoramic images from 376 patie...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2022-12-01
|
| Series: | Bone Reports |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2352187222004661 |
| _version_ | 1811294154063347712 |
|---|---|
| author | Soroush Baseri Saadi Catalina Moreno-Rabié Tim van den Wyngaert Reinhilde Jacobs |
| author_facet | Soroush Baseri Saadi Catalina Moreno-Rabié Tim van den Wyngaert Reinhilde Jacobs |
| author_sort | Soroush Baseri Saadi |
| collection | DOAJ |
| description | Objective: The present study aimed to develop and validate a tool for the automated classification of normal, affected, and osteonecrosis mandibular trabecular bone patterns in panoramic radiographs using convolutional neural networks (CNNs). Methods: A dataset of 402 panoramic images from 376 patients was selected, comprising 112 control radiographs from healthy patients and 290 images from patients treated with antiresorptive drugs (ARD). The latter was subdivided in 70 radiographs showing thickening of the lamina dura, 128 with abnormal bone patterns, and 92 images of clinically diagnosed osteonecrosis of the jaw (ONJ). Four pre-trained CNNs were fined-tuned and customized to detect and classify the different bone patterns. The best performing network was selected to develop the classification tool. The output was arranged as a colour-coded risk index showing the category and their odds. Classification performance of the networks was assessed through evaluation metrics, receiver operating characteristic curves (ROC), and a confusion matrix. Furthermore, Gradient-weighted Class Activation Mapping (Grad-CAM) was employed to visualise class-discriminative regions. Results: All networks correctly detected and classified the mandibular bone patterns with optimal performance metrics. InceptionResNetV2 showed the best results with an accuracy of 96 %, precision, recall and F1-score of 93 %, and a specificity of 98 %. Overall, most misclassifications occurred between normal and abnormal trabecular bone patterns. Conclusion: CNNs offer reliable potentials for automatic classification of abnormalities in the mandibular trabecular bone pattern in panoramic radiographs of antiresorptive treated patients. Clinical significance: A novel method that supports clinical decision making by identifying sites at high risk for ONJ. |
| first_indexed | 2024-04-13T05:13:14Z |
| format | Article |
| id | doaj.art-bfb0946f641b4dc2af24c095c1b69703 |
| institution | Directory Open Access Journal |
| issn | 2352-1872 |
| language | English |
| last_indexed | 2024-04-13T05:13:14Z |
| publishDate | 2022-12-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Bone Reports |
| spelling | doaj.art-bfb0946f641b4dc2af24c095c1b697032022-12-22T03:00:59ZengElsevierBone Reports2352-18722022-12-0117101632Convolutional neural network for automated classification of osteonecrosis and related mandibular trabecular patternsSoroush Baseri Saadi0Catalina Moreno-Rabié1Tim van den Wyngaert2Reinhilde Jacobs3OMFS-IMPATH Research Group, Department of Imaging and Pathology, Faculty of Medicine, University of Leuven, Leuven, Belgium; Department of Oral and Maxillofacial Surgery, University Hospitals Leuven, Leuven, BelgiumOMFS-IMPATH Research Group, Department of Imaging and Pathology, Faculty of Medicine, University of Leuven, Leuven, Belgium; Department of Oral and Maxillofacial Surgery, University Hospitals Leuven, Leuven, BelgiumDepartment of Nuclear Medicine, Antwerp University Hospital, Edegem, Belgium; Faculty of Medicine and Health Sciences, University of Antwerp, Antwerp, BelgiumOMFS-IMPATH Research Group, Department of Imaging and Pathology, Faculty of Medicine, University of Leuven, Leuven, Belgium; Department of Oral and Maxillofacial Surgery, University Hospitals Leuven, Leuven, Belgium; Department of Dental Medicine, Karolinska Institutet, Stockholm, Sweden; Corresponding author at: Karolinska Institutet, Stockholm, Sweden.Objective: The present study aimed to develop and validate a tool for the automated classification of normal, affected, and osteonecrosis mandibular trabecular bone patterns in panoramic radiographs using convolutional neural networks (CNNs). Methods: A dataset of 402 panoramic images from 376 patients was selected, comprising 112 control radiographs from healthy patients and 290 images from patients treated with antiresorptive drugs (ARD). The latter was subdivided in 70 radiographs showing thickening of the lamina dura, 128 with abnormal bone patterns, and 92 images of clinically diagnosed osteonecrosis of the jaw (ONJ). Four pre-trained CNNs were fined-tuned and customized to detect and classify the different bone patterns. The best performing network was selected to develop the classification tool. The output was arranged as a colour-coded risk index showing the category and their odds. Classification performance of the networks was assessed through evaluation metrics, receiver operating characteristic curves (ROC), and a confusion matrix. Furthermore, Gradient-weighted Class Activation Mapping (Grad-CAM) was employed to visualise class-discriminative regions. Results: All networks correctly detected and classified the mandibular bone patterns with optimal performance metrics. InceptionResNetV2 showed the best results with an accuracy of 96 %, precision, recall and F1-score of 93 %, and a specificity of 98 %. Overall, most misclassifications occurred between normal and abnormal trabecular bone patterns. Conclusion: CNNs offer reliable potentials for automatic classification of abnormalities in the mandibular trabecular bone pattern in panoramic radiographs of antiresorptive treated patients. Clinical significance: A novel method that supports clinical decision making by identifying sites at high risk for ONJ.http://www.sciencedirect.com/science/article/pii/S2352187222004661OsteonecrosisPanoramic radiographyDiagnostic imagingArtificial intelligenceConvolutional neural network |
| spellingShingle | Soroush Baseri Saadi Catalina Moreno-Rabié Tim van den Wyngaert Reinhilde Jacobs Convolutional neural network for automated classification of osteonecrosis and related mandibular trabecular patterns Bone Reports Osteonecrosis Panoramic radiography Diagnostic imaging Artificial intelligence Convolutional neural network |
| title | Convolutional neural network for automated classification of osteonecrosis and related mandibular trabecular patterns |
| title_full | Convolutional neural network for automated classification of osteonecrosis and related mandibular trabecular patterns |
| title_fullStr | Convolutional neural network for automated classification of osteonecrosis and related mandibular trabecular patterns |
| title_full_unstemmed | Convolutional neural network for automated classification of osteonecrosis and related mandibular trabecular patterns |
| title_short | Convolutional neural network for automated classification of osteonecrosis and related mandibular trabecular patterns |
| title_sort | convolutional neural network for automated classification of osteonecrosis and related mandibular trabecular patterns |
| topic | Osteonecrosis Panoramic radiography Diagnostic imaging Artificial intelligence Convolutional neural network |
| url | http://www.sciencedirect.com/science/article/pii/S2352187222004661 |
| work_keys_str_mv | AT soroushbaserisaadi convolutionalneuralnetworkforautomatedclassificationofosteonecrosisandrelatedmandibulartrabecularpatterns AT catalinamorenorabie convolutionalneuralnetworkforautomatedclassificationofosteonecrosisandrelatedmandibulartrabecularpatterns AT timvandenwyngaert convolutionalneuralnetworkforautomatedclassificationofosteonecrosisandrelatedmandibulartrabecularpatterns AT reinhildejacobs convolutionalneuralnetworkforautomatedclassificationofosteonecrosisandrelatedmandibulartrabecularpatterns |