Comparative performance of deep learning architectures for diabetic peripheral neuropathy detection using corneal confocal microscopy: a retrospective single-centre study
Objectives This study aims to develop a deep learning algorithm (DLA) using the InceptionV3 architecture for effective diabetic peripheral neuropathy (DPN) screening via corneal confocal microscopy (CCM) images.Design Retrospective study.Setting Ophthalmology Centre of General Hospital.Participants...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
BMJ Publishing Group
2025-08-01
|
| Series: | BMJ Open |
| Online Access: | https://bmjopen.bmj.com/content/15/8/e095342.full |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849229031879213056 |
|---|---|
| author | Yuyang Deng Wenqu Chen Weihuang Xu Jianzhang Hu |
| author_facet | Yuyang Deng Wenqu Chen Weihuang Xu Jianzhang Hu |
| author_sort | Yuyang Deng |
| collection | DOAJ |
| description | Objectives This study aims to develop a deep learning algorithm (DLA) using the InceptionV3 architecture for effective diabetic peripheral neuropathy (DPN) screening via corneal confocal microscopy (CCM) images.Design Retrospective study.Setting Ophthalmology Centre of General Hospital.Participants 127 participants enrolled: 33 healthy participants, 57 diabetic patients with DPN (DPN+) and 37 diabetic patients without DPN (DPN−).Interventions Not applicable.Main outcome measures The CCM image dataset, which was collected from participants (with five images per eye), was randomly divided into training, validation and test subsets in a 7:1:2 ratio. The images were preprocessed, augmented and used to train the InceptionV3 model. We compared its performance against the ResNet, DenseNet and Swin Transformer models. Performance was evaluated using accuracy, recall, F1 score and area under the curve (AUC) metrics.Results For single-participant predictions, the InceptionV3 model achieved the highest accuracy (0.9231), recall (0.8846), F1 score (0.9020) and AUC (0.9534) compared with the other models. For single-image predictions in the three-class classification task of CCM images, the InceptionV3 model achieved a precision of 0.8385, a recall of 0.9083, an F1 score of 0.8720 and an AUC of 0.8769 for predicting DPN+.Conclusions The InceptionV3-based DLA model achieved superior performance compared with traditional convolutional neural network architectures like ResNet and DenseNet, and the Swin transformer model, highlighting its potential for effective DPN screening. |
| format | Article |
| id | doaj-art-9aa58f60dfe84ba7a9b6058c3fc68c05 |
| institution | Kabale University |
| issn | 2044-6055 |
| language | English |
| publishDate | 2025-08-01 |
| publisher | BMJ Publishing Group |
| record_format | Article |
| series | BMJ Open |
| spelling | doaj-art-9aa58f60dfe84ba7a9b6058c3fc68c052025-08-22T09:30:12ZengBMJ Publishing GroupBMJ Open2044-60552025-08-0115810.1136/bmjopen-2024-095342Comparative performance of deep learning architectures for diabetic peripheral neuropathy detection using corneal confocal microscopy: a retrospective single-centre studyYuyang Deng0Wenqu Chen1Weihuang Xu2Jianzhang Hu3Department of Ophthalmology, Fujian Medical University Union Hospital, Fuzhou, ChinaDepartment of Ophthalmology, Fujian Medical University Union Hospital, Fuzhou, ChinaDepartment of Ophthalmology, Fujian Medical University Union Hospital, Fuzhou, ChinaDepartment of Ophthalmology, Fujian Medical University Union Hospital, Fuzhou, ChinaObjectives This study aims to develop a deep learning algorithm (DLA) using the InceptionV3 architecture for effective diabetic peripheral neuropathy (DPN) screening via corneal confocal microscopy (CCM) images.Design Retrospective study.Setting Ophthalmology Centre of General Hospital.Participants 127 participants enrolled: 33 healthy participants, 57 diabetic patients with DPN (DPN+) and 37 diabetic patients without DPN (DPN−).Interventions Not applicable.Main outcome measures The CCM image dataset, which was collected from participants (with five images per eye), was randomly divided into training, validation and test subsets in a 7:1:2 ratio. The images were preprocessed, augmented and used to train the InceptionV3 model. We compared its performance against the ResNet, DenseNet and Swin Transformer models. Performance was evaluated using accuracy, recall, F1 score and area under the curve (AUC) metrics.Results For single-participant predictions, the InceptionV3 model achieved the highest accuracy (0.9231), recall (0.8846), F1 score (0.9020) and AUC (0.9534) compared with the other models. For single-image predictions in the three-class classification task of CCM images, the InceptionV3 model achieved a precision of 0.8385, a recall of 0.9083, an F1 score of 0.8720 and an AUC of 0.8769 for predicting DPN+.Conclusions The InceptionV3-based DLA model achieved superior performance compared with traditional convolutional neural network architectures like ResNet and DenseNet, and the Swin transformer model, highlighting its potential for effective DPN screening.https://bmjopen.bmj.com/content/15/8/e095342.full |
| spellingShingle | Yuyang Deng Wenqu Chen Weihuang Xu Jianzhang Hu Comparative performance of deep learning architectures for diabetic peripheral neuropathy detection using corneal confocal microscopy: a retrospective single-centre study BMJ Open |
| title | Comparative performance of deep learning architectures for diabetic peripheral neuropathy detection using corneal confocal microscopy: a retrospective single-centre study |
| title_full | Comparative performance of deep learning architectures for diabetic peripheral neuropathy detection using corneal confocal microscopy: a retrospective single-centre study |
| title_fullStr | Comparative performance of deep learning architectures for diabetic peripheral neuropathy detection using corneal confocal microscopy: a retrospective single-centre study |
| title_full_unstemmed | Comparative performance of deep learning architectures for diabetic peripheral neuropathy detection using corneal confocal microscopy: a retrospective single-centre study |
| title_short | Comparative performance of deep learning architectures for diabetic peripheral neuropathy detection using corneal confocal microscopy: a retrospective single-centre study |
| title_sort | comparative performance of deep learning architectures for diabetic peripheral neuropathy detection using corneal confocal microscopy a retrospective single centre study |
| url | https://bmjopen.bmj.com/content/15/8/e095342.full |
| work_keys_str_mv | AT yuyangdeng comparativeperformanceofdeeplearningarchitecturesfordiabeticperipheralneuropathydetectionusingcornealconfocalmicroscopyaretrospectivesinglecentrestudy AT wenquchen comparativeperformanceofdeeplearningarchitecturesfordiabeticperipheralneuropathydetectionusingcornealconfocalmicroscopyaretrospectivesinglecentrestudy AT weihuangxu comparativeperformanceofdeeplearningarchitecturesfordiabeticperipheralneuropathydetectionusingcornealconfocalmicroscopyaretrospectivesinglecentrestudy AT jianzhanghu comparativeperformanceofdeeplearningarchitecturesfordiabeticperipheralneuropathydetectionusingcornealconfocalmicroscopyaretrospectivesinglecentrestudy |