Fusing Results of Several Deep Learning Architectures for Automatic Classification of Normal and Diabetic Macular Edema in Optical Coherence Tomography

<a>Objective:</a> Diabetic macular edema (DME) is the primary cause of vision loss among individuals with diabetes mellitus (DM). We developed, validated, and tested a deep-learning (DL) system for classifying DME using images from three common commercially available optical coherence tomography (OCT) devices. Research Design and Methods: We trained and validated two versions of a multi-task convolution neural network (CNN) to classify DME (center-involved DME [CI-DME], non-CI-DME, or absence of DME) using three-dimensional (3D) volume-scans and two-dimensional (2D) B-scans respectively. For both 3D and 2D CNNs, we employed the residual network (ResNet) as the backbone. For the 3D CNN, we used a 3D version of ResNet-34 with the last fully connected layer removed as the feature extraction module. A total of 73,746 OCT images were used for training and primary validation. External testing was performed using 26,981 images across seven independent datasets from Singapore, Hong Kong, the US, China, and Australia. Results: In classifying the presence or absence of DME, the DL system achieved area under the receiver operating characteristic curves (AUROCs) of 0.937 (95% CI 0.920–0.954), 0.958 (0.930–0.977), and 0.965 (0.948–0.977) for primary dataset obtained from Cirrus, Spectralis, and Triton OCTs respectively, in addition to AUROCs greater than 0.906 for the external datasets. For the further classification of the CI-DME and non-CI-DME subgroups, the AUROCs were 0.968 (0.940–0.995), 0.951 (0.898–0.982), and 0.975 (0.947–0.991) for the primary dataset and greater than 0.894 for the external datasets. Conclusion: We demonstrated excellent performance with a DL system for the automated classification of DME, highlighting its potential as a promising second-line screening tool for patients with DM, which may potentially create a more effective triaging mechanism to eye clinics.

Download Full-text

A Multitask Deep-Learning System to Classify Diabetic Macular Edema for Different Optical Coherence Tomography Devices: A Multicenter Analysis

10.2337/figshare.14710284 ◽

2021 ◽

Author(s):

Fangyao Tang ◽

Xi Wang ◽

An-ran Ran ◽

Carmen KM Chan ◽

Mary Ho ◽

...

Keyword(s):

Optical Coherence Tomography ◽

Deep Learning ◽

Macular Edema ◽

Diabetic Macular Edema ◽

Vision Loss ◽

Three Dimensional ◽

Learning System ◽

Optical Coherence ◽

Automated Classification

<a>Objective:</a> Diabetic macular edema (DME) is the primary cause of vision loss among individuals with diabetes mellitus (DM). We developed, validated, and tested a deep-learning (DL) system for classifying DME using images from three common commercially available optical coherence tomography (OCT) devices. Research Design and Methods: We trained and validated two versions of a multi-task convolution neural network (CNN) to classify DME (center-involved DME [CI-DME], non-CI-DME, or absence of DME) using three-dimensional (3D) volume-scans and two-dimensional (2D) B-scans respectively. For both 3D and 2D CNNs, we employed the residual network (ResNet) as the backbone. For the 3D CNN, we used a 3D version of ResNet-34 with the last fully connected layer removed as the feature extraction module. A total of 73,746 OCT images were used for training and primary validation. External testing was performed using 26,981 images across seven independent datasets from Singapore, Hong Kong, the US, China, and Australia. Results: In classifying the presence or absence of DME, the DL system achieved area under the receiver operating characteristic curves (AUROCs) of 0.937 (95% CI 0.920–0.954), 0.958 (0.930–0.977), and 0.965 (0.948–0.977) for primary dataset obtained from Cirrus, Spectralis, and Triton OCTs respectively, in addition to AUROCs greater than 0.906 for the external datasets. For the further classification of the CI-DME and non-CI-DME subgroups, the AUROCs were 0.968 (0.940–0.995), 0.951 (0.898–0.982), and 0.975 (0.947–0.991) for the primary dataset and greater than 0.894 for the external datasets. Conclusion: We demonstrated excellent performance with a DL system for the automated classification of DME, highlighting its potential as a promising second-line screening tool for patients with DM, which may potentially create a more effective triaging mechanism to eye clinics.

Download Full-text

Identifying Diabetic Macular Edema and Other Retinal Diseases by Optical Coherence Tomography Image and Multiscale Deep Learning

Diabetes Metabolic Syndrome and Obesity Targets and Therapy ◽

10.2147/dmso.s288419 ◽

2020 ◽

Vol Volume 13 ◽

pp. 4787-4800

Author(s):

Quan Zhang ◽

Zhiang Liu ◽

Jiaxu Li ◽

Guohua Liu

Keyword(s):

Optical Coherence Tomography ◽

Deep Learning ◽

Macular Edema ◽

Diabetic Macular Edema ◽

Optical Coherence Tomography Image ◽

Retinal Diseases ◽

Optical Coherence ◽

Tomography Image

Download Full-text

Optical coherence tomography findings predictive of response to treatment in diabetic macular edema

Journal of International Medical Research ◽

10.1177/0300060518798503 ◽

2018 ◽

Vol 46 (11) ◽

pp. 4455-4464 ◽

Cited By ~ 5

Author(s):

Young Joo Cho ◽

Dong Hyun Lee ◽

Min Kim

Keyword(s):

Optical Coherence Tomography ◽

Macular Edema ◽

Diabetic Macular Edema ◽

Serous Retinal Detachment ◽

Intravitreal Bevacizumab ◽

Internal Limiting Membrane ◽

Response To Treatment ◽

Optical Coherence ◽

Sd Oct

Objective To evaluate the short-term efficacy of intravitreal bevacizumab (IVB) and posterior sub-tenon triamcinolone injections (PSTI) on the basis of spectral-domain optical coherence tomography (SD-OCT) patterns in diabetic macular edema (DME). Methods We retrospectively reviewed 73 eyes of 73 patients with DME. Based on the presence of serous retinal detachment (SRD), eyes were categorized into two groups, and either IVB or PSTI treatment was performed. Central macular thickness (CMT) and the degree of SRD were assessed preoperatively and 1 month postoperatively. The severity of intraretinal edema was approximated based on the distance from the external limiting membrane to the internal limiting membrane. Results In eyes with SRD, reduction of SRD was greater with IVB than with PSTI. Moreover, reduction of intraretinal edema was greater with PSTI than with IVB. In eyes without SRD, PSTI achieved greater CMT reduction, compared with IVB. Conclusions In DME patients with SRD, IVB achieved greater reduction of SRD, compared with PSTI; however, intraretinal edema responded more favorably to PSTI, regardless of the presence of SRD. Our results suggest that the classification of DME based on OCT findings may be useful to predict responses to IVB or PSTI treatments.

Download Full-text