Diagnosis Of Eye Diseases Research Articles

Results of comprehensive diagnostics of feline uveal melanoma

According to available statistics, melanocytic neoplasia is the most common form of ocular tumor in cats, accounting for 67% of cases in an analysis of 2614 cases of primary ocular neoplasia. Feline diffuse iris melanoma (FDIM) is by far the most common form of melanocytic neoplasia of the eye. The purpose of the study was to conduct a comprehensive diagnosis of eye diseases and identify the features of the course of iris melanoma in cats. Study cats with iris hyperpigmentation underwent a complete ophthalmological examination - slit-lamp biomicroscopy, gonioscopy and indirect ophthalmoscopy. As a result of the research, it was found that 5 animals had hyperpigmented lesions, which look like flat brown spots on the surface of the iris. These precursor lesions are considered benign and are known as iris melanosis. In 10 animals, thickening of the iris, dyscoria, decreased pupil mobility, and dispersion of pigment in the anterior chamber were additionally detected; in 2 of them, involvement of the iridocorneal angle in the pathological process was observed, in 4 - secondary glaucoma, due to tumor infiltration of the iridocorneal angle. According to the results of computed tomography, 3 cats had metastases in the liver. During our studies, one cat underwent an iris biopsy to differentiate melanoma from pleomorphic lymphoma. In the remaining cases, a biopsy was not performed, since 9 patients with obvious pathological changes required enucleation, and in 5 patients monitoring of the pathological process over time was prescribed. Cats presenting for evaluation of hyperpigmented iris lesions should undergo a complete clinical and ophthalmologic examination, including gonioscopy.

Deep-Ocular: Improved Transfer Learning Architecture Using Self-Attention and Dense Layers for Recognition of Ocular Diseases.

It is difficult for clinicians or less-experienced ophthalmologists to detect early eye-related diseases. By hand, eye disease diagnosis is labor-intensive, prone to mistakes, and challenging because of the variety of ocular diseases such as glaucoma (GA), diabetic retinopathy (DR), cataract (CT), and normal eye-related diseases (NL). An automated ocular disease detection system with computer-aided diagnosis (CAD) tools is required to recognize eye-related diseases. Nowadays, deep learning (DL) algorithms enhance the classification results of retinograph images. To address these issues, we developed an intelligent detection system based on retinal fundus images. To create this system, we used ODIR and RFMiD datasets, which included various retinographics of distinct classes of the fundus, using cutting-edge image classification algorithms like ensemble-based transfer learning. In this paper, we suggest a three-step hybrid ensemble model that combines a classifier, a feature extractor, and a feature selector. The original image features are first extracted using a pre-trained AlexNet model with an enhanced structure. The improved AlexNet (iAlexNet) architecture with attention and dense layers offers enhanced feature extraction, task adaptability, interpretability, and potential accuracy benefits compared to other transfer learning architectures, making it particularly suited for tasks like retinograph classification. The extracted features are then selected using the ReliefF method, and then the most crucial elements are chosen to minimize the feature dimension. Finally, an XgBoost classifier offers classification outcomes based on the desired features. These classifications represent different ocular illnesses. We utilized data augmentation techniques to control class imbalance issues. The deep-ocular model, based mainly on the AlexNet-ReliefF-XgBoost model, achieves an accuracy of 95.13%. The results indicate the proposed ensemble model can assist dermatologists in making early decisions for the diagnosing and screening of eye-related diseases.

Topical insulin, a novel corneal epithelial regeneration agent in dry eye disease.

To evaluate the efficacy of insulin eye drops for dry eye disease in reducing corneal staining and improving symptoms. In this retrospective case series, patients with dry eye disease treated with off-label use of insulin eye drops were collected. The main inclusion criterion was diagnosis of dry eye disease with epithelial damage and acceptance of the off-label use of topical insulin. Age, sex, type of dry eye disease, time since diagnosis, previous ocular surgeries, concomitant treatment, best corrected visual acuity, symptoms, conjunctival hyperemia and corneal staining were recorded. Data from the 1 and 3-month visit were included. 16 patients (32 eyes) were treated with insulin (14 females and 2 males; mean age 61.3 ± 16.8 years). 12 patients (71%) were also on autologous serum and 10 patients (63%) on cyclosporine. Symptoms were 3.4 ± 1.3 (range 2-5) when scaled from 0 to 5. Mean hyperemia was 1.0 ± 0.9 (range 0-3) and corneal staining was 2.5 ± 1.3 (range 0-5). After 3 months, 5 patients (31%) referred to be much better, 6 (38%) better, 3 (19%) slightly better and 2 patients (13%) were subjectively similar, mean symptoms being 2.3 ± 1.0 (range 1-4; p = 0.001). Hyperemia was 0.3 ± 0.4 (range 0-1) and corneal staining was 1.1 ± 1.0 (range 0-3; both p < 0.001). Topical insulin was well tolerated with no adverse events. The excellent results presented in these case series illustrate topical insulin as a promising treatment in dry eye disease with refractory epithelial damage.

Application of artificial intelligence in ophthalmology for solving the problem of semantic segmentation of fundus images

The paper presents main aspects of the application of artificial intelligence in ophthalmology for the diagnosis and treatment of eye diseases, considering the problem of semantic segmentation of fundus images as an example. The classic approach to semantic segmentation on the basis of textural features is compared to the proposed approach based on neural networks. Basic problems of using the neural network approach in biomedicine are formulated. We propose a new method for selecting an optimal zone of laser exposure for laser coagulation based on two neural networks. The first network is used for detecting anatomical objects in the fundus and the second one is used for selecting the area of macular edema. The region of interest is formed from the edema area while taking into account the location of anatomical objects in it. A comparative analysis of sev-eral architectures of neural networks for solving the problem of selecting the edema area is carried out. The best results in the selection of the edema area are shown by the neural network architecture of Unet++.

A-355 Eyescore, A Point of Care Smartphone App for Early Dry Eye Diagnosis and Management

Abstract Current diagnosis of dry eye disease (DED) has significant challenges with limited accuracy and poor correlation of clinical symptoms. The goal of this study is to develop a smartphone App, namely “EyeScore”, to provide a point of care (POC) and digital solution for early diagnosis of DED. The authors tested the hypothesis of measuring eye blink rate and partial blink count as early clinical biomarkers for the calculation of so called “eye healthiness score”, which allows a convenient, rapid eye exam at home with low resource settings. EyeScore uses an iPhone as an imaging and sensing device for in-App recordings of eyelid movements. The use of facial landmark recognition and eye aspect ratio (EAR) enabled comprehensive digital analysis of video frames for determination of the eye opening/closed states. The smartphone videos from 10 DED patients and 10 healthy controls were tested to optimize EAR derived thresholds for accurate measurements of blink rates and partial blink counts. The authors formulated a clinically relevant algorithm for the calculation of “eye healthiness score”. This 10-point scale score can be easily measured anytime with a non-invasive manner and remotely shared with the patient's eye doctor. As a result, both patients and doctors can monitor the eye conditions over time. Our results showed that EyeScore confirmed the diagnosis of all 10 DED patients. Importantly, Eyescore also identified three individuals with “pre-DED” conditions from 10 healthy controls, demonstrating its potential clinical values for early diagnosis and management of DED. Materials and Methods An Apple MacBook Pro installed with iOS software (iOS 16.1.2) and Xcode 14.0 was used for coding and debugging. Visionkit SDK from Apple Developer was adopted to extract coordinate data from six eye facial landmarks. EAR values were calculated based on the previously published formula, with a dynamic EAR developed as a threshold for the measurement of blink rate and half blink count. For eye healthiness score, a 10-point formula was formulated (0–3, normal; 4–5, Pre-DED; 6–10, DED) with several contributing factors including blink rate, half blink count, eye discomfort and demographic characteristics. Results A total of 20 iPhone videos from DED patients and healthy controls were analyzed and evaluated with EyeScore App. All DED patients obtained scores above 7, confirming their DED diagnosis. On the other hand, three controls obtained scores of 4 or 5, suggesting they were in the Pre-DED state with no noticeable clinical DED symptoms. All other controls obtain scores blow 3 with normal eye conditions. All tests were performed with the consents from the test participants. Conclusions Smart phone-based platform offers great potentials for point of care diagnostic methods. A POC enabled EyeScore App has been successfully developed for early diagnosis of DED. In the future, the authors plan to test EyeScore with a large-scale sample set of DED patients and controls to further validate the accuracy and predictability of eye healthiness score for its broader clinical applications.

Optical coherence tomography and optical coherence tomography with angiography in the diagnosis of eye diseases

Optical coherence tomography and optical coherence tomography with angiography in the diagnosis of eye diseases

Automated Segmentation of Optical Coherence Tomography Images of the Human Tympanic Membrane Using Deep Learning.

Optical Coherence Tomography (OCT) is a light-based imaging modality that is used widely in the diagnosis and management of eye disease, and it is starting to become used to evaluate for ear disease. However, manual image analysis to interpret the anatomical and pathological findings in the images it provides is complicated and time-consuming. To streamline data analysis and image processing, we applied a machine learning algorithm to identify and segment the key anatomical structure of interest for medical diagnostics, the tympanic membrane. Using 3D volumes of the human tympanic membrane, we used thresholding and contour finding to locate a series of objects. We then applied TensorFlow deep learning algorithms to identify the tympanic membrane within the objects using a convolutional neural network. Finally, we reconstructed the 3D volume to selectively display the tympanic membrane. The algorithm was able to correctly identify the tympanic membrane properly with an accuracy of ~98% while removing most of the artifacts within the images, caused by reflections and signal saturations. Thus, the algorithm significantly improved visualization of the tympanic membrane, which was our primary objective. Machine learning approaches, such as this one, will be critical to allowing OCT medical imaging to become a convenient and viable diagnostic tool within the field of otolaryngology.

A foundation model for generalizable disease detection from retinal images

Medical artificial intelligence (AI) offers great potential for recognizing signs of health conditions in retinal images and expediting the diagnosis of eye diseases and systemic disorders1. However, the development of AI models requires substantial annotation and models are usually task-specific with limited generalizability to different clinical applications2. Here, we present RETFound, a foundation model for retinal images that learns generalizable representations from unlabelled retinal images and provides a basis for label-efficient model adaptation in several applications. Specifically, RETFound is trained on 1.6 million unlabelled retinal images by means of self-supervised learning and then adapted to disease detection tasks with explicit labels. We show that adapted RETFound consistently outperforms several comparison models in the diagnosis and prognosis of sight-threatening eye diseases, as well as incident prediction of complex systemic disorders such as heart failure and myocardial infarction with fewer labelled data. RETFound provides a generalizable solution to improve model performance and alleviate the annotation workload of experts to enable broad clinical AI applications from retinal imaging.

Eye diseases diagnosis using deep learning and multimodal medical eye imaging

Eye diseases diagnosis using deep learning and multimodal medical eye imaging

A Comparison of the Tortuosity Phenomenon in Retinal Arteries and Veins Using Digital Image Processing and Statistical Methods

The tortuosity of retinal blood vessels is an important phenomenon, and it can act as a biomarker in the diagnosis of several eye diseases. The study of abnormalities in the tortuosity of retinal arteries and veins provides ophthalmologists with important information for disease diagnosis. Our study aims to compare the tortuosity relation between retinal arteries and veins by quantifying the vessels’ tortuosity in the retina using 14 tortuosity measures applied to the AV-classification retinal dataset. Two feature sets are created, one for arteries and the other for veins. The comparison between the tortuosity of arteries and veins is based on a two-sample T-test statistical method, a regression analysis between the quantified tortuosity features, principal component analysis at the dataset level, and the introduction of the arteriovenous length ratios concept to compare the variations in these new ratios to see the tortuosity behavior in each image. The methods’ results have shown that the tortuosity of retinal arteries and veins is similar. The result of the two-sample T-test supports the research hypothesis, as the P-value obtained was greater than 0.05. Furthermore, the regression analysis between arteries and veins features showed a high correlation (r2 = 89.39% and 89.11%) for arteries and veins, respectively. The study concludes that the retinal vessel type has no statistical significance in the tortuosity calculation results.

MODELING AND ENHANCING LOW-QUALITY RETINAL FUNDUS IMAGES

Retinal Fundus images are widely used for the clinical screening and diagnosis of eye diseases. However, fundus images captured by operators with various levels of experience have a large variation in quality. Low-quality fundus images increase uncertainty in clinical observation and lead to the risk of misdiagnosis. However, due to the special optical beam of fundus imaging and structure of the retina, natural image enhancement methods cannot be utilized directly to address this. The proposed system analyzed the main contribution is a new set of shape features, called Dynamic Shape Features, which do not require precise segmentation of the regions to be classified. Then, based on the degradation model, a clinically oriented fundus enhancement network (cofe-Net) is proposed to suppress global degradation factors, while simultaneously preserving anatomical retinal structures and pathological characteristics for clinical observation and analysis. Moreover, the system show that the fundus correction method can benefit medical image analysis applications, e.g., retinal vessel segmentation and optic disc/cup detection. Key Words: retinal fundus, retinal vessel segmentation, dynamic shape features, optic disc/cup detection

A fundus image classification framework for learning with noisy labels.

Fundus images are widely used in the screening and diagnosis of eye diseases. Current classification algorithms for computer-aided diagnosis in fundus images rely on large amounts of data with reliable labels. However, the appearance of noisy labels degrades the performance of data-dependent algorithms, such as supervised deep learning. A noisy label learning framework suitable for the multiclass classification of fundus diseases is presented in this paper, which combines data cleansing (DC), adaptive negative learning (ANL), and sharpness-aware minimization (SAM) modules. Firstly, the DC module filters the noisy labels in the training dataset based on the prediction confidence. Then, the ANL module modifies the loss function by choosing complementary labels, which are neither the given labels nor the labels with the highest confidence. Moreover, for better generalization, the SAM module is applied by simultaneously optimizing the loss and its sharpness. Extensive experiments on both private and public datasets show that our method greatly promotes the performance for classification of multiple fundus diseases with noisy labels.

Artifact-Tolerant Clustering-Guided Contrastive Embedding Learning for Ophthalmic Images in Glaucoma.

Ophthalmic images, along with their derivatives like retinal nerve fiber layer (RNFL) thickness maps, play a crucial role in detecting and monitoring eye diseases such as glaucoma. For computer-aided diagnosis of eye diseases, the key technique is to automatically extract meaningful features from ophthalmic images that can reveal the biomarkers (e.g., RNFL thinning patterns) associated with functional vision loss. However, representation learning from ophthalmic images that links structural retinal damage with human vision loss is non-trivial mostly due to large anatomical variations between patients. This challenge is further amplified by the presence of image artifacts, commonly resulting from image acquisition and automated segmentation issues. In this paper, we present an artifact-tolerant unsupervised learning framework called EyeLearn for learning ophthalmic image representations in glaucoma cases. EyeLearn includes an artifact correction module to learn representations that optimally predict artifact-free images. In addition, EyeLearn adopts a clustering-guided contrastive learning strategy to explicitly capture the affinities within and between images. During training, images are dynamically organized into clusters to form contrastive samples, which encourage learning similar or dissimilar representations for images in the same or different clusters, respectively. To evaluate EyeLearn, we use the learned representations for visual field prediction and glaucoma detection with a real-world dataset of glaucoma patient ophthalmic images. Extensive experiments and comparisons with state-of-the-art methods confirm the effectiveness of EyeLearn in learning optimal feature representations from ophthalmic images.

The Role of Physics &amp; Optics in Revolutionizing Eye Care

Eye care has experienced numerous significant changes throughout history, progressing from simple, basic therapies to the sophisticated, exact methods we use today. The foundation of these developments has been the development of modern optical science, which is the study of light and its interactions with matter. As a result, the diagnosis and treatment of eye diseases now have never before-seen levels of accuracy, precision, and understanding.

Clinical utility of maximum blink interval measured by smartphone application DryEyeRhythm to support dry eye disease diagnosis

The coronavirus disease (COVID-19) pandemic has emphasized the paucity of non-contact and non-invasive methods for the objective evaluation of dry eye disease (DED). However, robust evidence to support the implementation of mHealth- and app-based biometrics for clinical use is lacking. This study aimed to evaluate the reliability and validity of app-based maximum blink interval (MBI) measurements using DryEyeRhythm and equivalent traditional techniques in providing an accessible and convenient diagnosis. In this single-center, prospective, cross-sectional, observational study, 83 participants, including 57 with DED, had measurements recorded including slit-lamp-based, app-based, and visually confirmed MBI. Internal consistency and reliability were assessed using Cronbach’s alpha and intraclass correlation coefficients. Discriminant and concurrent validity were assessed by comparing the MBIs from the DED and non-DED groups and Pearson’s tests for each platform pair. Bland–Altman analysis was performed to assess the agreement between platforms. App-based MBI showed good Cronbach’s alpha coefficient, intraclass correlation coefficient, and Pearson correlation coefficient values, compared with visually confirmed MBI. The DED group had significantly shorter app-based MBIs, compared with the non-DED group. Bland–Altman analysis revealed minimal biases between the app-based and visually confirmed MBIs. Our findings indicate that DryEyeRhythm is a reliable and valid tool that can be used for non-invasive and non-contact collection of MBI measurements, which can assist in accessible DED detection and management.

The effects of female sex hormones on the human cornea across a woman’s life cycle

The cornea is a hormone-responsive tissue that responds to changing levels of female sex hormones. This review focuses on the structural and functional changes in the human cornea associated with the hormonal milestones of menarche, pregnancy, and menopause, as well as consequences stemming from the use of exogenous sex hormones for fertility control and replacement. Articles were identified by searching PubMed without language or region restrictions. The primary outcomes evaluated were changes in central corneal thickness (CCT), intraocular pressure (IOP), and quality of the ocular tear film. The potential impact of hormone-associated changes on the diagnosis and surgical management of common eye diseases, as well as the potential use of sex hormones as therapeutic agents is also considered. Understanding the physiological effects of female sex hormones on the cornea is important because that knowledge can shape the management decisions physicians and women face about ocular health across their life stages.

Diagnosa Penyakit Mata Berdasarkan Citra Ocular Disease Intelligent Recognition (ODIR) Dengan Gabor Filter Klasifikasi Levenberg-Marquardt

Light captured by both eyeballs is passed on to the pupil, focused through the sensitive part of the eyeball. The retinal organ of the eye converts the light capture into nerve impulses, and delivers them to the brain through the nerve fibers contained in the impulses. There are several complications of disease in the sense of sight that can be diagnosed from the arrangement of the retina of the eye. Image segmentation stages are needed so that the fiber contained in the retinal optics can be changed. The initial stage is with the initialization of the arrangement for the contour movement stages. Using the Levenberg-Marquardt algorithm is able to improve the following results obtained as a basis for previously collected data. In this study eye disease only diagnosed eye disease in complaints of the retina of the eye. The purpose of this research is to diagnose early sight diseases that appear based on the symptoms shown by the composition of the retinal parts of the human eyeball and classify the level of danger of the results of eye disease diagnoses.

Magnetic resonance imaging parameters on lacrimal gland in thyroid eye disease: a systematic review and meta-analysis

BackgroundThyroid eye disease is an extrathyroidal manifestation of Graves’ disease and is associated with dry eye disease. This is the first systematic review and meta-analysis to evaluate the role of magnetic resonance imaging lacrimal gland parameters in thyroid eye disease diagnosis, activity grading, and therapeutic responses prediction.MethodsUp to 23 August, 2022, 504 studies from PubMed and Cochrane Library were analyzed. After removing duplicates and imposing selection criteria, nine eligible studies were included. Risk of bias assessment was done. Meta-analyses were performed using random-effect model if heterogeneity was significant. Otherwise, fixed-effect model was used. Main outcome measures include seven structural magnetic resonance imaging parameters (lacrimal gland herniation, maximum axial area, maximum coronal area, maximum axial length, maximum coronal length, maximum axial width, maximum coronal width), and three functional magnetic resonance imaging parameters (diffusion tensor imaging-fractional anisotropy, diffusion tensor imaging-apparent diffusion coefficient or mean diffusivity, diffusion-weighted imaging-apparent diffusion coefficient).ResultsThyroid eye disease showed larger maximum axial area, maximum coronal area, maximum axial length, maximum axial width, maximum coronal width, diffusion tensor imaging-apparent diffusion coefficient/ mean diffusivity, and lower diffusion tensor imaging-fractional anisotropy than controls. Active thyroid eye disease showed larger lacrimal gland herniation, maximum coronal area, diffusion-weighted imaging-apparent diffusion coefficient than inactive. Lacrimal gland dimensional (maximum axial area, maximum coronal area, maximum axial length, maximum axial width, maximum coronal width) and functional parameters (diffusion tensor imaging-apparent diffusion coefficient, diffusion tensor imaging-apparent diffusion coefficient) could be used for diagnosing thyroid eye disease; lacrimal gland herniation, maximum coronal area, and diffusion-weighted imaging-apparent diffusion coefficient for differentiating active from inactive thyroid eye disease; diffusion tensor imaging parameters (diffusion tensor imaging-fractional anisotropy, diffusion tensor imaging-mean diffusivity) and lacrimal gland herniation for helping grading and therapeutic responses prediction respectively.ConclusionsMagnetic resonance imaging lacrimal gland parameters can detect active thyroid eye disease and differentiate thyroid eye disease from controls. Maximum coronal area is the most effective indicator for thyroid eye disease diagnosis and activity grading. There are inconclusive results showing whether structural or functional lacrimal gland parameters have diagnostic superiority. Future studies are warranted to determine the use of magnetic resonance imaging lacrimal gland parameters in thyroid eye disease.

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

BackgroundPathological 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.MethodsA 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.ResultsOf 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].ConclusionsThis 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.

Diagnosing dry-eye: Which tests are most accurate?

PurposeTo demonstrate how the likelihood of making a correct diagnosis of dry eye disease varies according to the clinical test methods used. MethodsThe probability of a person having dry eye, given that they return a positive test, was calculated for a range of standard tests, using the Bayes-Price rule. Global specificity and sensitivity values for each test were estimated by employing the Beta distribution to combine all relevant data obtained from a literature review. ResultsAt an assumed prevalence of 11.6%, the single test with the highest probability of a correct diagnosis was corneal staining (probability = 0.28) and the lowest was the ocular surface disease index - OSDI (0.14). The best combination of symptoms with a single test of tear film homeostasis was the 5-item dry eye questionnaire (DEQ-5) + corneal staining (0.42) while OSDI + tear film break up time (TBUT) was the worst (0.23). The simultaneous observation of conjunctival and corneal staining was associated with a probability of 0.49. The probability of a correct diagnosis increased with the number of positive tests, up to a maximum of 0.90 when all of DEQ-5, conjunctival and corneal staining, osmolarity and TBUT were positive. ConclusionA significant risk of misdiagnosis is associated with using any single test for dry eye disease, or the minimum TFOS DEWS II criterion of symptoms plus any single test of tear film homeostasis. To minimize this risk, the maximum number of tests available should be performed and the results used to inform diagnosis. The simultaneous occurrence of conjunctival and corneal staining should be considered a key outcome and be specified in future guidelines.