Automated detection of myopic maculopathy from color fundus photographs using deep convolutional neural networks

Abstract

Background Myopic maculopathy (MM) has become a major cause of visual impairment and blindness worldwide, especially in East Asian countries. Deep learning approaches such as deep convolutional neural networks (DCNN) have been successfully applied to identify some common retinal diseases and show great potential for the intelligent analysis of MM. This study aimed to build a reliable approach for automated detection of MM from retinal fundus images using DCNN models. Methods A dual-stream DCNN (DCNN-DS) model that perceives features from both original images and corresponding processed images by color histogram distribution optimization method was designed for classification of no MM, tessellated fundus (TF), and pathologic myopia (PM). A total of 36,515 gradable images from four hospitals were used for DCNN model development, and 14,986 gradable images from the other two hospitals for external testing. We also compared the performance of the DCNN-DS model and four ophthalmologists on 3000 randomly sampled fundus images. Results The DCNN-DS model achieved sensitivities of 93.3% and 91.0%, specificities of 99.6% and 98.7%, areas under the receiver operating characteristic curves (AUC) of 0.998 and 0.994 for detecting PM, whereas sensitivities of 98.8% and 92.8%, specificities of 95.6% and 94.1%, AUCs of 0.986 and 0.970 for detecting TF in two external testing datasets. In the sampled testing dataset, the sensitivities of four ophthalmologists ranged from 88.3% to 95.8% and 81.1% to 89.1%, and the specificities ranged from 95.9% to 99.2% and 77.8% to 97.3% for detecting PM and TF, respectively. Meanwhile, the DCNN-DS model achieved sensitivities of 90.8% and 97.9% and specificities of 99.1% and 94.0% for detecting PM and TF, respectively. Conclusions The proposed DCNN-DS approach demonstrated reliable performance with high sensitivity, specificity, and AUC to classify different MM levels on fundus photographs sourced from clinics. It can help identify MM automatically among the large myopic groups and show great potential for real-life applications.

Artificial intelligence technology for myopia challenges: A review

“…The results showed that the detection accuracy of this auxiliary classification system for all lesions except choroidal neovascularization was more than 84%, and the overall detection accuracy for myopic macular degeneration was up to 87.53%, whereas the classification accuracy of ophthalmology specialists on the same task was merely 89%. A study by Li et al (2022c) showed similar results, further confirming the effectiveness of EfficientNet. However, according to Du et al (2021) , the detection of choroidal neovascularization was only 37.07%, which might be related to the poor visualization of blood vessels in color fundus images ( Jiang et al, 2020 ; Laíns et al, 2021 ).…”

Performances of artificial intelligence in detecting pathologic myopia: a systematic review and meta-analysis

“…Figure 1 shows the flowchart of the literature eligibility process. Finally, 22 studies were included for systematic review [ 12 – 33 ], and 14 of them were included for quantitative meta-analysis [ 12 , 13 , 16 , 19 – 23 , 26 , 27 , 29 – 31 , 33 ].

Fig.

…”

Automated detection of myopic maculopathy using five-category models based on vision outlooker for visual recognition

“…Some researchers had also done research on multiple macular disorders. Li et al (2022) used DCNN-DS model to detect no myopic maculopathy, tessellated fundus, and pathologic myopia, and the validation accuracies on the two external testing datasets were 96.3 and 93.0%, respectively. Tang et al (2022) used ResNet-50 model to develop the META-PM study categorizing system, and the mean accuracy was 0.9119 ± 0.0093 on the five categories.…”