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

Efficient pyramid channel attention network for pathological myopia recognition with pretraining-and-finetuning

Pathological myopia (PM) is the leading ocular disease for impaired vision worldwide. Clinically, the characteristics of pathology distribution in PM are global-local on the fundus image, which plays a significant role in assisting clinicians in diagnosing PM. However, most existing deep neural networks focused on designing complex architectures but rarely explored the pathology distribution prior of PM. To tackle this issue, we propose an efficient pyramid channel attention (EPCA) module, which fully leverages the potential of the clinical pathology prior of PM with pyramid pooling and multi-scale context fusion. Then, we construct EPCA-Net for automatic PM recognition based on fundus images by stacking a sequence of EPCA modules. Moreover, motivated by the recent pretraining-and-finetuning paradigm, we attempt to adapt pre-trained natural image models for PM recognition by freezing them and treating the EPCA and other attention modules as adapters. In addition, we construct a PM recognition benchmark termed PM-fundus by collecting fundus images of PM from publicly available datasets. The comprehensive experiments demonstrate the superiority of EPCA-Net over state-of-the-art methods in the PM recognition task. For example, EPCA-Net achieves 97.56% accuracy and outperforms ViT by 2.85% accuracy on the PM-fundus dataset. The results also show that our method based on the pretraining-and-finetuning paradigm achieves competitive performance through comparisons to part of previous methods based on traditional fine-tuning paradigm with fewer tunable parameters, which has the potential to leverage more natural image foundation models to address the PM recognition task in limited medical data regime.

Developing and validating a clinlabomics-based machine-learning model for early detection of retinal detachment in patients with high myopia

BACKGROUND

Retinal detachment (RD) is a vision-threatening disorder of significant severity. Individuals with high myopia (HM) face a 2 to 6 times higher risk of developing RD compared to non-myopes. The timely identification of high myopia-related retinal detachment (HMRD) is crucial for effective treatment and prevention of additional vision impairment. Consequently, our objective was to streamline and validate a machine-learning model based on clinical laboratory omics (clinlabomics) for the early detection of RD in HM patients.

METHODS

We extracted clinlabomics data from the electronic health records for 24,440 HM and 5607 HMRD between 2015 and 2022. Lasso regression analysis assessed fifty-nine variables, excluding collinear variables (variance inflation factor > 10). Four models based on random forest, gradient boosting machine (GBM), generalized linear model, and Deep Learning Model were trained for HMRD diagnosis and employed for internal validation. An external test of the models was done. Three random data sets were further processed to validate the performance of the diagnostic model. The primary outcomes were the area under the receiver operating characteristic curve (AUC) and the area under the precision-recall curve (AUCPR) to diagnose HMRD.

RESULTS

Nine variables were selected by all models. Given the AUC and AUCPR values across the different sets, the GBM model was chosen as the final diagnostic model. The GBM model had an AUC of 0.8550 (95%CI = 0.8322-0.8967) and an AUCPR of 0.5584 (95%CI = 0.5250-0.5879) in the training set. The AUC and AUCPR in the internal validation were 0.8405 (95%CI = 0.8060-0.8966) and 0.5355 (95%CI = 0.4988-0.5732). During the external test evaluation, it reached an AUC of 0.7579 (95%CI = 0.7340-0.7840) and an AUCPR of 0.5587 (95%CI = 0.5345-0.5880). A similar discriminative capacity was observed in the three random data sets. The GBM model was well-calibrated across all the sets. The GBM-RD model was implemented into a web application that provides risk prediction for HM individuals.

CONCLUSION

GBM algorithms based on nine features successfully predicted the diagnosis of RD in patients with HM, which will help ophthalmologists to establish a preliminary diagnosis and to improve diagnostic accuracy in the clinic.

Performance of artificial intelligence for the detection of pathological myopia from colour fundus images: a systematic review and meta-analysis

BACKGROUND

Pathological myopia (PM) is a major cause of worldwide blindness and represents a serious threat to eye health globally. Artificial intelligence (AI)-based methods are gaining traction in ophthalmology as highly sensitive and specific tools for screening and diagnosis of many eye diseases. However, there is currently a lack of high-quality evidence for their use in the diagnosis of PM.

METHODS

A systematic review and meta-analysis of studies evaluating the diagnostic performance of AI-based tools in PM was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidance. Five electronic databases were searched, results were assessed against the inclusion criteria and a quality assessment was conducted for included studies. Model sensitivity and specificity were pooled using the DerSimonian and Laird (random-effects) model. Subgroup analysis and meta-regression were performed.

RESULTS

Of 1021 citations identified, 17 studies were included in the systematic review and 11 studies, evaluating 165,787 eyes, were included in the meta-analysis. The area under the summary receiver operator curve (SROC) was 0.9905. The pooled sensitivity was 95.9% [95.5%-96.2%], and the overall pooled specificity was 96.5% [96.3%-96.6%]. The pooled diagnostic odds ratio (DOR) for detection of PM was 841.26 [418.37-1691.61].

CONCLUSIONS

This systematic review and meta-analysis provides robust early evidence that AI-based, particularly deep-learning based, diagnostic tools are a highly specific and sensitive modality for the detection of PM. There is potential for such tools to be incorporated into ophthalmic public health screening programmes, particularly in resource-poor areas with a substantial prevalence of high myopia.

Insights into artificial intelligence in myopia management: from a data perspective

Given the high incidence and prevalence of myopia, the current healthcare system is struggling to handle the task of myopia management, which is worsened by home quarantine during the ongoing COVID-19 pandemic. The utilization of artificial intelligence (AI) in ophthalmology is thriving, yet not enough in myopia. AI can serve as a solution for the myopia pandemic, with application potential in early identification, risk stratification, progression prediction, and timely intervention. The datasets used for developing AI models are the foundation and determine the upper limit of performance. Data generated from clinical practice in managing myopia can be categorized into clinical data and imaging data, and different AI methods can be used for analysis. In this review, we comprehensively review the current application status of AI in myopia with an emphasis on data modalities used for developing AI models. We propose that establishing large public datasets with high quality, enhancing the model's capability of handling multimodal input, and exploring novel data modalities could be of great significance for the further application of AI for myopia.

Initiation of China Alliance of Research in High Myopia (CHARM): protocol for an AI-based multimodal high myopia research biobank

INTRODUCTION

High myopia is a pressing public health concern due to its increasing prevalence, younger trend and the high risk of blindness, particularly in East Asian countries, including China. The China Alliance of Research in High Myopia (CHARM) is a newly established consortium that includes more than 100 hospitals and institutions participating across the nation, aiming to promote collaboration and data sharing in the field of high myopia screening, classification, diagnosis and therapeutic development.

METHODS AND ANALYSIS

The CHARM project is an ongoing study, and its initiation is distinguished by its unprecedented scale, encompassing plans to involve over 100 000 Chinese patients. This initiative stands out not only for its extensive scope but also for its innovative application of artificial intelligence (AI) to assist in diagnosis and treatment decisions. The CHARM project has been carried out using a 'three-step' strategy. The first step involves the collection of basic information, refraction, axial length and fundus photographs from participants with high myopia. In the second step, we will collect multimodal imaging data to expand the scope of clinical information, for example, optical coherence tomography and ultra-widefield fundus images. In the final step, genetic testing will be conducted by incorporating patient family histories and blood samples. The majority of data collected by CHARM is in the form of images that will be used to detect and predict the progression of high myopia through the identification and quantification of biomarkers such as fundus tessellation, optic nerve head and vascular parameters.

ETHICS AND DISSEMINATION

The study has received approval from the Ethics Committee of Beijing Tongren Hospital (TREC2022-KY045). The establishment of CHARM represents an opportunity to create a collaborative platform for myopia experts and facilitate the dissemination of research findings to the global community through peer-reviewed publications and conference presentations. These insights can inform clinical decision-making and contribute to the development of new treatment modalities that may benefit patients worldwide.

TRIAL REGISTRATION NUMBER

ChiCTR2300071219.

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

BACKGROUND/OBJECTIVE

Pathologic myopia (PM) is a major cause of severe visual impairment and blindness, and current applications of artificial intelligence (AI) have covered the diagnosis and classification of PM. This meta-analysis and systematic review aimed to evaluate the overall performance of AI-based models in detecting PM and related complications.

METHODS

We searched PubMed, Scopus, Embase, Web of Science and IEEE Xplore for eligible studies before Dec 20, 2022. The methodological quality of included studies was evaluated using the Quality Assessment for Diagnostic Accuracy Studies (QUADAS-2). We calculated the pooled sensitivity (SEN), specificity (SPE) and the summary area under the curve (AUC) using a random effects model, to evaluate the performance of AI in the detection of PM based on fundus or optical coherence tomography (OCT) images.

RESULTS

22 studies were included in the systematic review, and 14 of them were included in the quantitative analysis. Of all included studies, SEN and SPE ranged from 80.0% to 98.7% and from 79.5% to 100.0% for PM detection, respectively. For the detection of PM, the summary AUC was 0.99 (95% confidence interval (CI) 0.97 to 0.99), and the pooled SEN and SPE were 0.95 (95% CI 0.92 to 0.96) and 0.97 (95% CI: 0.94 to 0.98), respectively. For the detection of PM-related choroid neovascularization (CNV), the summary AUC was 0.99 (95% CI: 0.97 to 0.99).

CONCLUSION

Our review demonstrated the excellent performance of current AI algorithms in detecting PM and related complications based on fundus and OCT images.

Evaluation of a novel deep learning based screening system for pathologic myopia

AIM

To evaluate the clinical application value of the artificial intelligence assisted pathologic myopia (PM-AI) diagnosis model based on deep learning.

METHODS

A total of 1156 readable color fundus photographs were collected and annotated based on the diagnostic criteria of Meta-pathologic myopia (PM) (2015). The PM-AI system and four eye doctors (retinal specialists 1 and 2, and ophthalmologists 1 and 2) independently evaluated the color fundus photographs to determine whether they were indicative of PM or not and the presence of myopic choroidal neovascularization (mCNV). The performance of identification for PM and mCNV by the PM-AI system and the eye doctors was compared and evaluated via the relevant statistical analysis.

RESULTS

For PM identification, the sensitivity of the PM-AI system was 98.17%, which was comparable to specialist 1 (P=0.307), but was higher than specialist 2 and ophthalmologists 1 and 2 (P<0.001). The specificity of the PM-AI system was 93.06%, which was lower than specialists 1 and 2, but was higher than ophthalmologists 1 and 2. The PM-AI system showed the Kappa value of 0.904, while the Kappa values of specialists 1, 2 and ophthalmologists 1, 2 were 0.968, 0.916, 0.772 and 0.730, respectively. For mCNV identification, the AI system showed the sensitivity of 84.06%, which was comparable to specialists 1, 2 and ophthalmologist 2 (P>0.05), and was higher than ophthalmologist 1. The specificity of the PM-AI system was 95.31%, which was lower than specialists 1 and 2, but higher than ophthalmologists 1 and 2. The PM-AI system gave the Kappa value of 0.624, while the Kappa values of specialists 1, 2 and ophthalmologists 1 and 2 were 0.864, 0.732, 0.304 and 0.238, respectively.

CONCLUSION

In comparison to the senior ophthalmologists, the PM-AI system based on deep learning exhibits excellent performance in PM and mCNV identification. The effectiveness of PM-AI system is an auxiliary diagnosis tool for clinical screening of PM and mCNV.

Artificial intelligence-aided diagnosis and treatment in the field of optometry

With the rapid development of computer technology, the application of artificial intelligence (AI) to ophthalmology has gained prominence in modern medicine. As modern optometry is closely related to ophthalmology, AI research on optometry has also increased. This review summarizes current AI research and technologies used for diagnosis in optometry, related to myopia, strabismus, amblyopia, optical glasses, contact lenses, and other aspects. The aim is to identify mature AI models that are suitable for research on optometry and potential algorithms that may be used in future clinical practice.

Intelligent Diagnosis of Multiple Peripheral Retinal Lesions in Ultra-widefield Fundus Images Based on Deep Learning

INTRODUCTION

Compared with traditional fundus examination techniques, ultra-widefield fundus (UWF) images provide 200° panoramic images of the retina, which allows better detection of peripheral retinal lesions. The advent of UWF provides effective solutions only for detection but still lacks efficient diagnostic capabilities. This study proposed a retinal lesion detection model to automatically locate and identify six relatively typical and high-incidence peripheral retinal lesions from UWF images which will enable early screening and rapid diagnosis.

METHODS

A total of 24,602 augmented ultra-widefield fundus images with labels corresponding to 6 peripheral retinal lesions and normal manifestation labelled by 5 ophthalmologists were included in this study. An object detection model named You Only Look Once X (YOLOX) was modified and trained to locate and classify the six peripheral retinal lesions including rhegmatogenous retinal detachment (RRD), retinal breaks (RB), white without pressure (WWOP), cystic retinal tuft (CRT), lattice degeneration (LD), and paving-stone degeneration (PSD). We applied coordinate attention block and generalized intersection over union (GIOU) loss to YOLOX and evaluated it for accuracy, sensitivity, specificity, precision, F1 score, and average precision (AP). This model was able to show the exact location and saliency map of the retinal lesions detected by the model thus contributing to efficient screening and diagnosis.

RESULTS

The model reached an average accuracy of 96.64%, sensitivity of 87.97%, specificity of 98.04%, precision of 87.01%, F1 score of 87.39%, and mAP of 86.03% on test dataset 1 including 248 UWF images and reached an average accuracy of 95.04%, sensitivity of 83.90%, specificity of 96.70%, precision of 78.73%, F1 score of 81.96%, and mAP of 80.59% on external test dataset 2 including 586 UWF images, showing this system performs well in distinguishing the six peripheral retinal lesions.

CONCLUSION

Focusing on peripheral retinal lesions, this work proposed a deep learning model, which automatically recognized multiple peripheral retinal lesions from UWF images and localized exact positions of lesions. Therefore, it has certain potential for early screening and intelligent diagnosis of peripheral retinal lesions.

Development and validation of a routine blood parameters-based model for screening the occurrence of retinal detachment in high myopia in the context of PPPM

Background/aims

Timely detection and treatment of retinal detachment (RD) could effectively save vision and reduce the risk of progressing visual field defects. High myopia (HM) is known to be associated with an increased risk of RD. Evidently, it should be clearly discriminated the individuals with high or low risk of RD in patients with HM. By using multi-parametric analysis, risk assessment, and other techniques, it is crucial to create cutting-edge screening programs that may be utilized to improve population eye health and develop person-specific, cost-effective preventative, and targeted therapeutic measures. Therefore, we propose a novel, routine blood parameters-based prediction model as a screening program to help distinguish who should offer detailed ophthalmic examinations for RD diagnosis, prevent visual field defect progression, and provide personalized, serial monitoring in the context of predictive, preventive, and personalized medicine (PPPM/3 PM).

Methods

This population-based study included 20,870 subjects (HM = 19,284, HMRD = 1586) who underwent detailed routine blood tests and ophthalmic evaluations. HMRD cases and HM controls were matched using a nested case-control design. Then, the HMRD cases and HM controls were randomly assigned to the discovery cohort, validation cohort 1, and validation cohort 2 maintaining a 6:2:2 ratio, and other subjects were assigned to the HM validation cohort. Receiver operating characteristic curve analysis was performed to select feature indexes. Feature indexes were integrated into seven algorithm models, and an optimal model was selected based on the highest area under the curve (AUC) and accuracy.

Results

Six feature indexes were selected: lymphocyte, basophil, mean platelet volume, platelet distribution width, neutrophil-to-lymphocyte ratio, and lymphocyte-to-monocyte ratio. Among the algorithm models, the algorithm of conditional probability (ACP) showed the best performance achieving an AUC of 0.79, a diagnostic accuracy of 0.72, a sensitivity of 0.71, and a specificity of 0.74 in the discovery cohort. A good performance of the ACP model was also observed in the validation cohort 1 (AUC = 0.81, accuracy = 0.72, sensitivity = 0.71, specificity = 0.73) and validation cohort 2 (AUC = 0.77, accuracy = 0.71, sensitivity = 0.70, specificity = 0.72). In addition, ACP model calibration was found to be good across three cohorts. In the HM validation cohort, the ACP model achieved a diagnostic accuracy of 0.81 for negative classification.

Conclusion

We have developed a routine blood parameters-based model with an ACP algorithm that could potentially be applied in the clinic with a PPPM approach for serial monitoring and predicting the occurrence of RD in HM and can facilitate the prevention of HM progression to RD. According to the current study, routine blood measures are essential in patient risk classification, predictive diagnosis, and targeted therapy. Therefore, for high-risk RD persons, novel screening programs and prompt treatment plans are essential to enhance individual outcomes and healthcare offered to the community with HM.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13167-023-00319-3.

A deep network using coarse clinical prior for myopic maculopathy grading

Pathological Myopia (PM) is a globally prevalent eye disease which is one of the main causes of blindness. In the long-term clinical observation, myopic maculopathy is a main criterion to diagnose PM severity. The grading of myopic maculopathy can provide a severity and progression prediction of PM to perform treatment and prevent myopia blindness in time. In this paper, we propose a feature fusion framework to utilize tessellated fundus and the brightest region in fundus images as prior knowledge. The proposed framework consists of prior knowledge extraction module and feature fusion module. Prior knowledge extraction module uses traditional image processing methods to extract the prior knowledge to indicate coarse lesion positions in fundus images. Furthermore, the prior, tessellated fundus and the brightest region in fundus images, are integrated into deep learning network as global and local constrains respectively by feature fusion module. In addition, rank loss is designed to increase the continuity of classification score. We collect a private color fundus dataset from Beijing TongRen Hospital containing 714 clinical images. The dataset contains all 5 grades of myopic maculopathy which are labeled by experienced ophthalmologists. Our framework achieves 0.8921 five-grade accuracy on our private dataset. Pathological Myopia (PALM) dataset is used for comparison with other related algorithms. Our framework is trained with 400 images and achieves an AUC of 0.9981 for two-class grading. The results show that our framework can achieve a good performance for myopic maculopathy grading.

Artificial intelligence technology for myopia challenges: A review

Myopia is a significant global health concern and affects human visual function, resulting in blurred vision at a distance. There are still many unsolved challenges in this field that require the help of new technologies. Currently, artificial intelligence (AI) technology is dominating medical image and data analysis and has been introduced to address challenges in the clinical practice of many ocular diseases. AI research in myopia is still in its early stages. Understanding the strengths and limitations of each AI method in specific tasks of myopia could be of great value and might help us to choose appropriate approaches for different tasks. This article reviews and elaborates on the technical details of AI methods applied for myopia risk prediction, screening and diagnosis, pathogenesis, and treatment.

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

Purpose

To propose a five-category model for the automatic detection of myopic macular lesions to help grassroots medical institutions conduct preliminary screening of myopic macular lesions from limited number of color fundus images.

Methods

First, 1,750 fundus images of non-myopic retinal lesions and four categories of pathological myopic maculopathy were collected, graded, and labeled. Subsequently, three five-classification models based on Vision Outlooker for Visual Recognition (VOLO), EfficientNetV2, and ResNet50 for detecting myopic maculopathy were trained with data-augmented images, and the diagnostic results of the different trained models were compared and analyzed. The main evaluation metrics were sensitivity, specificity, negative predictive value (NPV), positive predictive value (PPV), area under the curve (AUC), kappa and accuracy, and receiver operating characteristic curve (ROC).

Results

The diagnostic accuracy of the VOLO-D2 model was 96.60% with a kappa value of 95.60%. All indicators used for the diagnosis of myopia-free macular degeneration were 100%. The sensitivity, NPV, specificity, and PPV for diagnosis of leopard fundus were 96.43, 98.33, 100, and 100%, respectively. The sensitivity, specificity, PPV, and NPV for the diagnosis of diffuse chorioretinal atrophy were 96.88, 98.59, 93.94, and 99.29%, respectively. The sensitivity, specificity, PPV, and NPV for the diagnosis of patchy chorioretinal atrophy were 92.31, 99.26, 97.30, and 97.81%, respectively. The sensitivity, specificity, PPV, and NPV for the diagnosis of macular atrophy were 100, 98.10, 84.21, and 100%, respectively.

Conclusion

The VOLO-D2 model accurately identified myopia-free macular lesions and four pathological myopia-related macular lesions with high sensitivity and specificity. It can be used in screening pathological myopic macular lesions and can help ophthalmologists and primary medical institution providers complete the initial screening diagnosis of patients.

Global trends and frontiers of research on pathologic myopia since the millennium: A bibliometric analysis

Background and purpose

Pathologic myopia (PM) is an international public health issue. This study aimed to analyze PM research trends by reporting on publication trends since 2000 and identifying influential journals, countries, authors, and keywords involved in PM.

Methods

A bibliometric analysis was performed to evaluate global production and development trends in PM since 2000 and the keywords associated with PM.

Results

A total of 1,435 publications were retrieved. PM has become a fascinating topic (with relative research interest ranging from 0.0018% in 2000 to 0.0044% in 2021) and a global public health issue. The top three countries with the highest number of publications were China, the USA, and Japan. The journals, authors, and institutions that published the most relevant literature came from these three countries. China exhibited the most rapid increase in the number of publications (from 0 in 2000 to 69 in 2021). Retina published the most papers on PM. Kyoko Ohno-Matsui and Tokyo Medical and Dental University contributed the most publications among authors and institutions, respectively. Based on keyword analysis, previous research emphasized myopic choroidal neovascularization and treatment, while recent hotspots include PM changes based on multimodal imaging, treatment, and pathogenesis. Keyword analysis also revealed that deep learning was the latest hotspot and has been used for the detection of PM.

Conclusion

Our results can help researchers understand the current status and future trends of PM. China, the USA, and Japan have the greatest influence, based on the number of publications, top journals, authors, and institutions. Current research on PM highlights the pathogenesis and application of novel technologies, including multimodal imaging and artificial intelligence.

Novel Uses and Challenges of Artificial Intelligence in Diagnosing and Managing Eyes with High Myopia and Pathologic Myopia

Myopia is a global health issue, and the prevalence of high myopia has increased significantly in the past five to six decades. The high incidence of myopia and its vision-threatening course emphasize the need for automated methods to screen for high myopia and its serious form, named pathologic myopia (PM). Artificial intelligence (AI)-based applications have been extensively applied in medicine, and these applications have focused on analyzing ophthalmic images to diagnose the disease and to determine prognosis from these images. However, unlike diseases that mainly show pathologic changes in the fundus, high myopia and PM generate even more data because both the ophthalmic information and morphological changes in the retina and choroid need to be analyzed. In this review, we present how AI techniques have been used to diagnose and manage high myopia, PM, and other ocular diseases and discuss the current capacity of AI in assisting in preventing high myopia.