Artificial intelligence can assist with diagnosing retinal vein occlusion

AIM: To assist with retinal vein occlusion (RVO) screening, artificial intelligence (AI) methods based on deep learning (DL) have been developed to alleviate the pressure experienced by ophthalmologists and discover and treat RVO as early as possible. METHODS: A total of 8600 color fundus photographs (CFPs) were included for training, validation, and testing of disease recognition models and lesion segmentation models. Four disease recognition and four lesion segmentation models were established and compared. Finally, one disease recognition model and one lesion segmentation model were selected as superior. Additionally, 224 CFPs from 130 patients were included as an external test set to determine the abilities of the two selected models. RESULTS: Using the Inception-v3 model for disease identification, the mean sensitivity, specificity, and F1 for the three disease types and normal CFPs were 0.93, 0.99, and 0.95, respectively, and the mean area under the curve (AUC) was 0.99. Using the DeepLab-v3 model for lesion segmentation, the mean sensitivity, specificity, and F1 for four lesion types (abnormally dilated and tortuous blood vessels, cotton-wool spots, flame-shaped hemorrhages, and hard exudates) were 0.74, 0.97, and 0.83, respectively. CONCLUSION: DL models show good performance when recognizing RVO and identifying lesions using CFPs. Because of the increasing number of RVO patients and increasing demand for trained ophthalmologists, DL models will be helpful for diagnosing RVO early in life and reducing vision impairment.

Correlation of Hemoglobin A1c Test (HbAlc) with Different Grades of Diabetic Retinopathy

Diabetic retinopathy is a disorder which generally affects the retina and disturbs the microvasculature of it and is the most dreaded complication of diabetes. This study included 50 patients with diabetic retinopathy, out of which 4% of patients infected with Non-proliferative diabetic retinopathy (NPDR), 48% with mild and 20% with very high NPDR. 8% of cases had very severe NPDR while the rest 20% had PDR. Our results which showed a higher prevalence of CSME in patients with HBA 1c of 8. 7 % and above. From the finding the elevated lipid levels in serum are associated with high risk of CSME and retinal hard exudates.

OCT-angiography for the diagnosis of radiation maculopathy in patients treated with proton beam therapy: A 2-year prospective study

Purpose Radiation maculopathy (RM) is the leading cause of visual acuity (VA) loss after proton beam therapy (PBT) of choroidal melanoma. The aim of this study was to assess the value of optical coherence tomography-angiography (OCT-A) for the diagnosis of RM in patients with choroidal melanoma treated with PBT. Materials & methods This 2-year prospective, descriptive, single-center study included patients treated with PBT for choroidal melanoma. VA measurement, retinography, OCT and OCT-A were performed. Vascular density (VD) in the superficial capillary plexus (SCP), peri-foveal anastomotic ring changes and foveal avascular zone (FAZ) enlargement were studied. Results Nineteen patients were included in the study. The median baseline melanoma thickness was 5.7 [3.6–8.1] mm. The median melanoma-to-macula distance was 3.5 [2.6–4.6] mm. The earliest signs of RM identified on retinography were hard exudates developing at 12 [12–24] months, followed by retinal hemorrhages at 18 [12–30] months, found in 88.9% and 77.8% of patients respectively. On OCT, the earliest sign was the onset/progression of cystoid macular edema (CME) at 12 [6–12] months, found in 10 patients (52.6%). On OCT-A, 100% of patients presented with a discontinuity of the perifoveal anastomotic ring and a FAZ enlargement after 12 [6–24] months. After 12 months, a VD loss in the SCP by 11.7% and 10.8% compared to baseline, was found in the macular and foveal areas respectively. A significant negative correlation was found between the VA and the VD in the macular SCP (R = −0.43; p = 0.029). Conclusion OCT-A is a reliable and effective diagnostic tool for RM in patients with choroidal melanoma treated with PBT.

Diagnosis of type 3 macular neovascularization

Type 3 macular neovascularization is characterized by a complex of pathological vessels located in the sensory retina. Fundus oculi examination reveals intraretinal hemorrhages, macular edema, hard exudates and pigment epithelial detachments. Indocyanine and fluorescein angiography, OCT and angio-OCT are used for diagnosis and treatment monitoring. The treatment efficacy depends on the disease severity and the therapy applied.

Automated Detection and Counting of Hard Exudates for Diabetic Retinopathy by using Watershed and Double Top-Bottom Hat Filtering Algorithm

Diabetic Retinopathy (DR) is one of diabetes complications that affects our eyes. Hard Exudate (HE) are known to be the early signs of DR that potentially lead to blindness. Detection of DR automatically is a complicated job since the size of HE is very small. Besides, our community nowadays lack awareness on diabetic where they do not know that diabetes can affect eyes and lead to blindness if regular check-up is not performed. Hence, automated detection of HE known as Eye Retinal Imaging System (EyRis) was created to focus on detecting the HE based on fundus image. The purpose of this system development is for early detection of the symptoms based on retina images captured using fundus camera. Through the captured retina image, we can clearly detect the symptoms that lead to DR. In this study, proposed Watershed segmentation method for detecting HE in fundus images. Top-Hat and Bottom-Hat were use as enhancement technique to improve the quality of the image. This method was tested on 15 retinal images from the Universiti Sains Malaysia Hospital (HUSM) at three different stages: Normal, NPDR, and PDR. Ten of these images have abnormalities, while the rest are normal retinal images. The evaluation of the segmentation images would be compared by Sensitivity, F-score and accuracy based on medical expert's hand drawn ground truth. The results achieve accuracy 0.96 percent with 0.99 percent sensitivity for retinal images.

EAD-Net: A Novel Lesion Segmentation Method in Diabetic Retinopathy Using Neural Networks

Diabetic retinopathy (DR) is a common chronic fundus disease, which has four different kinds of microvessel structure and microvascular lesions: microaneurysms (MAs), hemorrhages (HEs), hard exudates, and soft exudates. Accurate detection and counting of them are a basic but important work. The manual annotation of these lesions is a labor-intensive task in clinical analysis. To solve the problem, we proposed a novel segmentation method for different lesions in DR. Our method is based on a convolutional neural network and can be divided into encoder module, attention module, and decoder module, so we refer it as EAD-Net. After normalization and augmentation, the fundus images were sent to the EAD-Net for automated feature extraction and pixel-wise label prediction. Given the evaluation metrics based on the matching degree between detected candidates and ground truth lesions, our method achieved sensitivity of 92.77%, specificity of 99.98%, and accuracy of 99.97% on the e_ophtha_EX dataset and comparable AUPR (Area under Precision-Recall curve) scores on IDRiD dataset. Moreover, the results on the local dataset also show that our EAD-Net has better performance than original U-net in most metrics, especially in the sensitivity and F1-score, with nearly ten percent improvement. The proposed EAD-Net is a novel method based on clinical DR diagnosis. It has satisfactory results on the segmentation of four different kinds of lesions. These effective segmentations have important clinical significance in the monitoring and diagnosis of DR.

Algorithm for Detection and Quantification of Hyperreflective Dots on Optical Coherence Tomography in Diabetic Macular Edema

Purpose: To develop an algorithm to detect and quantify hyperreflective dots (HRDs) on optical coherence tomography (OCT) in patients with diabetic macular edema (DME).Materials and Methods: Twenty OCTs (each OCT contains 128 b scans) from 20 patients diagnosed with DME were included in this study. Two types of HRDs, hard exudates and small HRDs (hypothesized to be activated microglia), were identified and labeled independently by two raters. An algorithm using deep learning technology was developed based on input (in total 2,560 OCT b scans) of manual labeling and differentiation of HRDs from rater 1. 4-fold cross-validation was used to train and validate the algorithm. Dice coefficient, intraclass coefficient (ICC), correlation coefficient, and Bland–Altman plot were used to evaluate agreement of the output parameters between two methods (either between two raters or between one rater and proposed algorithm).Results: The Dice coefficients of total HRDs, hard exudates, and small HRDs area of the algorithm were 0.70 ± 0.10, 0.72 ± 0.11, and 0.46 ± 0.06, respectively. The correlations between rater 1 and proposed algorithm (range: 0.95–0.99, all p < 0.001) were stronger than the correlations between the two raters (range: 0.84–0.96, all p < 0.001) for all parameters. The ICCs were higher for all the parameters between rater 1 and proposed algorithm (range: 0.972–0.997) than those between the two raters (range: 0.860–0.953).Conclusions: Our proposed algorithm is a good tool to detect and quantify HRDs and can provide objective and repeatable information of OCT for DME patients in clinical practice and studies.

ASSOCIATION OF SERUM LIPID LEVELS WITH DIABETIC RETINOPATHY

AIM:To study the association of serum lipids with retinal hard exudates formation, occurrence of clinically signicant macular oedema (CSME) and severity of diabetic retinopathy (DR) in type 2 diabetes METHODS:Type 2 diabetic patients seeking ocular evaluation for diabetic retinopathy were included in this cross-sectional study. They were assessed for presence and severity of diabetic retinopathy (DR), presence of hard exudates, clinically signicant macular oedema (CSME) .Retinal ndings were correlated to serum lipids levels . RESULTS:Totally 100 patients were included, of which 42/100 had diabetic retinopathy of any grade. Retinal hard exudates formation was found to have statistically signicant correlation with the presence of dyslipidemia (p=0.02), increased cholesterol (p=0.002) and LDL levels (p=0.001). The occurrence of CSME showed a statistically signicant correlation with dyslipidemia (p=0.04) and increased LDL levels (p=0.04). Neither occurrence of dyslipidemia nor the increased levels of the various components of serum lipids showed a statistically signicant correlation with high Triglyceride levels or Low HDL-C or increasing severity of diabetic retinopathy CONCLUSION:Elevated serum lipids showed a signicant association with retinal hard exudate formation, CSME in type 2 diabetics. Lipid lowering agents may help in reducing the occurrence of these retinal ndings in diabetic patients.

Multimodal evaluation of hemorrhages within intraretinal cystoid spaces in a blood-fluid level presentation

Introduction: We describe characteristic findings on multimodal evaluation and the features of hemorrhage within a foveal cystoid space in a patient presenting cystoid macular edema secondary to Branch Retinal Vein Occlusion (BRVO). Case description: We report a case of a 64-year-old diabetic male patient presenting gradual blurry vision in the left eye. Fundoscopic findings were suggestive of BRVO, such as hard exudates and mild venous engorgement superotemporally and diffuse macular intraretinal hemorrhages. In the foveal area, there was cystoid edema with blood-fluid level (BFL) inside one of the cystoid spaces. Retina multimodal evaluation, including color, blue filter, and red-free fundus photography, fluorescein angiography, fundus autofluorescence, and spectral-domain optical coherence tomography (SD-OCT) B and C scan imaging, confirmed blood within foveal cystoid space. The patient underwent antiangiogenic therapy with significant improvement of macular edema and reduction of the cystoid space after 3 months. In addition, there was a resolution of visual symptoms. The cystoid space previously partially filled with blood, persisted, despite presenting smaller volume and medium reflectivity in the SD-OCT. Conclusions: Multimodal evaluation of blood-fluid level within foveal cystoid space in patients with BRVO has not been described previously. Identification of this sign may support the diagnosis of retinal vein occlusion in doubtful cases and further studies must be carried out to establish if the presence of BFL correlates with visual outcomes.

Support vector machine and deep-learning object detection for localisation of hard exudates

Hard exudates are one of the main clinical findings in the retinal images of patients with diabetic retinopathy. Detecting them early significantly impacts the treatment of underlying diseases; therefore, there is a need for automated systems with high reliability. We propose a novel method for identifying and localising hard exudates in retinal images. To achieve fast image pre-scanning, a support vector machine (SVM) classifier was combined with a faster region-based convolutional neural network (faster R-CNN) object detector for the localisation of exudates. Rapid pre-scanning filtered out exudate-free samples using a feature vector extracted from the pre-trained ResNet-50 network. Subsequently, the remaining samples were processed using a faster R-CNN detector for detailed analysis. When evaluating all the exudates as individual objects, the SVM classifier reduced the false positive rate by 29.7% and marginally increased the false negative rate by 16.2%. When evaluating all the images, we recorded a 50% reduction in the false positive rate, without any decrease in the number of false negatives. The interim results suggested that pre-scanning the samples using the SVM prior to implementing the deep-network object detector could simultaneously improve and speed up the current hard exudates detection method, especially when there is paucity of training data.