-
Ophthalmology and Therapy Jun 2024The aim of this work is to estimate the sensitivity, specificity, and misclassification rate of an automated retinal image analysis system (ARIAS) in diagnosing active...
INTRODUCTION
The aim of this work is to estimate the sensitivity, specificity, and misclassification rate of an automated retinal image analysis system (ARIAS) in diagnosing active diabetic macular edema (DME) and to identify factors associated with true and false positives.
METHODS
We conducted a cross-sectional study of prospectively enrolled patients with diabetes mellitus (DM) referred to a tertiary medical retina center for screening or management of DME. All patients underwent two-field fundus photography (macula- and disc-centered) with a true-color confocal camera; images were processed by EyeArt V.2.1.0 (Woodland Hills, CA, USA). Active DME was defined as the presence of intraretinal or subretinal fluid on spectral-domain optical coherence tomography (SD-OCT). Sensitivity and specificity and their 95% confidence intervals (CIs) were calculated. Variables associated with true (i.e., DME labeled as present by ARIAS + fluid on SD-OCT) and false positives (i.e., DME labeled as present by ARIAS + no fluid on SD-OCT) of active DME were explored.
RESULTS
A total of 298 eyes were included; 92 eyes (31%) had active DME. ARIAS sensitivity and specificity were 82.61% (95% CI 72.37-89.60) and 84.47% (95% CI 78.34-89.10). The misclassification rate was 16%. Factors associated with true positives included younger age (p = 0.01), shorter DM duration (p = 0.006), presence of hard exudates (p = 0.005), and microaneurysms (p = 0.002). Factors associated with false positives included longer DM duration (p = 0.01), worse diabetic retinopathy severity (p = 0.008), history of inactivated DME (p < 0.001), and presence of hard exudates (p < 0.001), microaneurysms (p < 0.001), or epiretinal membrane (p = 0.06).
CONCLUSIONS
The sensitivity of ARIAS was diminished in older patients and those without DME-related fundus lesions, while the specificity was reduced in cases with a history of inactivated DME. ARIAS performed well in screening for naïve DME but is not effective in surveillance inactivated DME.
PubMed: 38587776
DOI: 10.1007/s40123-024-00929-8 -
Computer Methods and Programs in... Jun 2024Early detection and grading of Diabetic Retinopathy (DR) is essential to determine an adequate treatment and prevent severe vision loss. However, the manual analysis of...
BACKGROUND AND OBJECTIVE
Early detection and grading of Diabetic Retinopathy (DR) is essential to determine an adequate treatment and prevent severe vision loss. However, the manual analysis of fundus images is time consuming and DR screening programs are challenged by the availability of human graders. Current automatic approaches for DR grading attempt the joint detection of all signs at the same time. However, the classification can be optimized if red lesions and bright lesions are independently processed since the task gets divided and simplified. Furthermore, clinicians would greatly benefit from explainable artificial intelligence (XAI) to support the automatic model predictions, especially when the type of lesion is specified. As a novelty, we propose an end-to-end deep learning framework for automatic DR grading (5 severity degrees) based on separating the attention of the dark structures from the bright structures of the retina. As the main contribution, this approach allowed us to generate independent interpretable attention maps for red lesions, such as microaneurysms and hemorrhages, and bright lesions, such as hard exudates, while using image-level labels only.
METHODS
Our approach is based on a novel attention mechanism which focuses separately on the dark and the bright structures of the retina by performing a previous image decomposition. This mechanism can be seen as a XAI approach which generates independent attention maps for red lesions and bright lesions. The framework includes an image quality assessment stage and deep learning-related techniques, such as data augmentation, transfer learning and fine-tuning. We used the architecture Xception as a feature extractor and the focal loss function to deal with data imbalance.
RESULTS
The Kaggle DR detection dataset was used for method development and validation. The proposed approach achieved 83.7 % accuracy and a Quadratic Weighted Kappa of 0.78 to classify DR among 5 severity degrees, which outperforms several state-of-the-art approaches. Nevertheless, the main result of this work is the generated attention maps, which reveal the pathological regions on the image distinguishing the red lesions and the bright lesions. These maps provide explainability to the model predictions.
CONCLUSIONS
Our results suggest that our framework is effective to automatically grade DR. The separate attention approach has proven useful for optimizing the classification. On top of that, the obtained attention maps facilitate visual interpretation for clinicians. Therefore, the proposed method could be a diagnostic aid for the early detection and grading of DR.
Topics: Humans; Diabetic Retinopathy; Artificial Intelligence; Deep Learning; Image Interpretation, Computer-Assisted; Fundus Oculi; Diabetes Mellitus
PubMed: 38583290
DOI: 10.1016/j.cmpb.2024.108160 -
Artificial Intelligence in Medicine Apr 2024This paper introduces a novel one-stage end-to-end detector specifically designed to detect small lesions in medical images. Precise localization of small lesions...
This paper introduces a novel one-stage end-to-end detector specifically designed to detect small lesions in medical images. Precise localization of small lesions presents challenges due to their appearance and the diverse contextual backgrounds in which they are found. To address this, our approach introduces a new type of pixel-based anchor that dynamically moves towards the targeted lesion for detection. We refer to this new architecture as GravityNet, and the novel anchors as gravity points since they appear to be "attracted" by the lesions. We conducted experiments on two well-established medical problems involving small lesions to evaluate the performance of the proposed approach: microcalcifications detection in digital mammograms and microaneurysms detection in digital fundus images. Our method demonstrates promising results in effectively detecting small lesions in these medical imaging tasks.
Topics: Mammography; Fundus Oculi
PubMed: 38553147
DOI: 10.1016/j.artmed.2024.102842 -
Oman Journal of Ophthalmology 2024The purpose of this study was to evaluate prediabetic patients for microvascular changes using optical coherence tomography-angiography (OCT-A) and compare with diabetic...
PURPOSE
The purpose of this study was to evaluate prediabetic patients for microvascular changes using optical coherence tomography-angiography (OCT-A) and compare with diabetic patients and healthy controls.
METHODS
OCT-A images of 60 eyes of 30 patients with diabetes mellitus (DM), 72 eyes of 36 prediabetic patients, and 108 eyes of 54 healthy controls were retrospectively evaluated and compared in this study. A swept-source OCTA (Triton, Topcon) instrument was used for collecting OCT-A images. Duration of the diabetic or prediabetic period, glycated hemoglobin, fasting blood glucose level, postprandial glucose (PPG) level, high-density lipoprotein, low-density lipoprotein, triglyceride, and creatinine values of all participants were recorded.
RESULTS
Microaneurysm, nonperfusion areas, perifoveal capillary disruption, and capillary network disorganization were detected in both prediabetics and diabetics but statistically more common in diabetic patients. Neovascularization and intraretinal microvascular anomalies were detected only in diabetic patients.
CONCLUSIONS
OCT-A seemed to be effective in detecting microvascular changes in diabetic patients. More importantly, results showed us that in prediabetic patients, microvascular changes can be seen before the onset of DM and before or concurrently with neurodegenerative changes.
PubMed: 38524313
DOI: 10.4103/ojo.ojo_197_22 -
Microvascular Research Jul 2024Dysfunctional pericytes and disruption of adherens or tight junctions are related to many microvascular diseases, including diabetic retinopathy. In this context,... (Comparative Study)
Comparative Study
Dysfunctional pericytes and disruption of adherens or tight junctions are related to many microvascular diseases, including diabetic retinopathy. In this context, visualizing retinal vascular architecture becomes essential for understanding retinal vascular disease pathophysiology. Although flat mounts provide a demonstration of the retinal blood vasculature, they often lack a clear view of microaneurysms and capillary architecture. Trypsin and elastase digestion are the two techniques for isolating retinal vasculatures in rats, mice, and other animal models. Our observations in the present study reveal that trypsin digestion impacts the association between pericytes and endothelial cells. In contrast, elastase digestion effectively preserves these features in the blood vessels. Furthermore, trypsin digestion disrupts endothelial adherens and tight junctions that elastase digestion does not. Therefore, elastase digestion emerges as a superior technique for isolating retinal vessels, which can be utilized to collect reliable and consistent data to comprehend the pathophysiology of disorders involving microvascular structures.
Topics: Animals; Pancreatic Elastase; Trypsin; Retinal Vessels; Mice, Inbred C57BL; Pericytes; Endothelial Cells; Tight Junctions; Mice; Male
PubMed: 38521153
DOI: 10.1016/j.mvr.2024.104682 -
American Journal of Ophthalmology Case... Jun 2024To report a case of a refractory foveal microaneurysm (MA) that was successfully treated by use of a new surgical procedure.
PURPOSE
To report a case of a refractory foveal microaneurysm (MA) that was successfully treated by use of a new surgical procedure.
OBSERVATIONS
This study involved a 79-year-old female with an active foveal MA associated with branch retinal vein occlusion in her left eye. Despite anti-vascular endothelial growth factor treatments, the MA remained active without closure, and best-corrected visual acuity (VA) gradually decreased from 20/20 to 20/200. After our new surgical procedure was explained in detail to the patient, written informed consent was obtained from the patient and the surgery was performed. Briefly, following pars plana vitrectomy, the internal limiting membrane in her left eye was peeled and the retina of the external wall of the MA was then gently incised. The exposed MA was then directly grabbed and pulled up onto the retina using 27-gauge microforceps, and photocoagulation was performed. At 3-months postoperative, closure of the MA and improvement in the retinal findings were observed, and best-corrected VA improved to 20/67.
CONCLUSIONS AND IMPORTANCE
We report a case of a refractory foveal MA that was successfully treated with a novel surgical technique that closed the MA, avoided thermal damage to the surrounding tissue, and resulted in improved postoperative VA.
PubMed: 38495594
DOI: 10.1016/j.ajoc.2024.102034 -
IEEE Journal of Biomedical and Health... Mar 2024Early-stage diabetic retinopathy (DR) presents challenges in clinical diagnosis due to inconspicuous and minute microaneurysms (MAs), resulting in limited research in...
Early-stage diabetic retinopathy (DR) presents challenges in clinical diagnosis due to inconspicuous and minute microaneurysms (MAs), resulting in limited research in this area. Additionally, the potential of emerging foundation models, such as the segment anything model (SAM), in medical scenarios remains rarely explored. In this work, we propose a human-in-the-loop, label-free early DR diagnosis framework called GlanceSeg, based on SAM. GlanceSeg enables real-time segmentation of MA lesions as ophthalmologists review fundus images. Our human-in-the-loop framework integrates the ophthalmologist's gaze maps, allowing for rough localization of minute lesions in fundus images. Subsequently, a saliency map is generated based on the located region of interest, which provides prompt points to assist the foundation model in efficiently segmenting MAs. Finally, a domain knowledge filtering (DKF) module refines the segmentation of minute lesions. We conducted experiments on two newly-built public datasets, i.e., IDRiD and Retinal-Lesions, and validated the feasibility and superiority of GlanceSeg through visualized illustrations and quantitative measures. Additionally, we demonstrated that GlanceSeg improves annotation efficiency for clinicians and further enhances segmentation performance through fine-tuning using annotations. The clinician-friendly GlanceSeg is able to segment small lesions in real-time, showing potential for clinical applications.
PubMed: 38483801
DOI: 10.1109/JBHI.2024.3377592 -
Journal of Diabetes and Its... Apr 2024To investigate the association between diabetic retinopathy (DR) and coronary artery disease (CAD) using coronary angiotomography (CCTA) and multimodal retinal imaging...
AIMS
To investigate the association between diabetic retinopathy (DR) and coronary artery disease (CAD) using coronary angiotomography (CCTA) and multimodal retinal imaging (MMRI) with ultra-widefield retinography and optical coherence tomography angiography and structural domain.
METHODS
Single-center, cross-sectional, single-blind. Patients with diabetes who had undergone CCTA underwent MMRI. Uni and multivariate analysis were used to assess the association between CAD and DR and to identify variables independently associated with DR.
RESULTS
We included 171 patients, 87 CAD and 84 non-CAD. Most CAD patients were males (74 % vs 38 %, P < 0.01), insulin users (52 % vs 38 %, p < 0.01) and revascularized (64 %). They had a higher prevalence of DR (48 % vs 22 %, p = 0.01), microaneurysms (25 % vs 13 %, p = 0.04), intraretinal cysts (22 % vs 8 %, p = 0.01) and areas of reduced capillary density (46 % vs 20 %, p < 0.01). CAD patients also had lower mean vascular density (MVD) (15.7 % vs 16.5,%, p = 0.049) and foveal avascular zone (FAZ) circularity (0.64 ± 0.1 vs 0.69 ± 0.1, p = 0.04). There were significant and negative correlations between Duke coronary score and MVD (r = -0.189; p = 0.03) and FAZ circularity (r = -0,206; p = 0.02). CAD, DM duration and insulin use independently associated with DR.
CONCLUSIONS
CAD patients had higher prevalence of DR and lower MVD. CAD, DM duration and insulin use were independently associated with DR.
Topics: Male; Humans; Female; Diabetic Retinopathy; Cross-Sectional Studies; Coronary Artery Disease; Single-Blind Method; Retinal Vessels; Tomography, Optical Coherence; Insulins; Diabetes Mellitus
PubMed: 38471431
DOI: 10.1016/j.jdiacomp.2024.108721 -
Ocular Immunology and Inflammation Mar 2024Systemic sclerosis (SSc) is a chronic multisystemic disease characterized by immunological activation, diffuse vasculopathy, and generalized fibrosis exhibiting a... (Review)
Review
Systemic sclerosis (SSc) is a chronic multisystemic disease characterized by immunological activation, diffuse vasculopathy, and generalized fibrosis exhibiting a variety of symptoms. A recognized precursor of SSc is Raynaud's phenomenon, which is part of the very early disease of systemic sclerosis (VEDOSS) in combination with nailfold videocapillaroscopy (NVC) impairment. The pathophysiology of ocular involvement, alterations in internal organs, and body integumentary system involvement in SSc patients are complicated and poorly understood, with multiple mechanisms presumptively working together. The most prevalent ocular symptoms of SSc are abnormalities of the eyelids and conjunctiva as well as dry eye syndrome, due to fibroblasts' dysfunction and inflammation of the ocular surface. In particular, lagophthalmos, blepharophimosis limitation of eyelid motion, eyelid telangiectasia, and rigidity or tightening of the lids may affect up to two-third of the patients. In addition, reduction in central corneal thickness, iris defects and higher rates of glaucoma were reported. In the first reports based on retinography or fluorescein angiography, about 50% of SSc patients showed signs of vascular disease: peripheral artery occlusion, thinning of retinal pigment epithelium and choroidal capillaries, ischemic areas surrounded by intraretinal extravasation and microaneurysms, and peripheral capillary non-perfusion. Successively, thanks to the advent of optical coherence tomography angiography (OCTA), several studies highlighted significant impairment of either the choriocapillaris and retinal vascular plexuses, also correlating with NVC involvement and skin disease, even in VEDOSS disease. Given the sensitivity of this technique, ocular micro-vasculopathy may act as a tool for early SSc identification and discriminate between disease stages.
PubMed: 38466107
DOI: 10.1080/09273948.2024.2308030 -
Artificial Intelligence in Medicine Mar 2024Diabetic retinopathy (DR) is the most prevalent cause of visual impairment in adults worldwide. Typically, patients with DR do not show symptoms until later stages, by...
Diabetic retinopathy (DR) is the most prevalent cause of visual impairment in adults worldwide. Typically, patients with DR do not show symptoms until later stages, by which time it may be too late to receive effective treatment. DR Grading is challenging because of the small size and variation in lesion patterns. The key to fine-grained DR grading is to discover more discriminating elements such as cotton wool, hard exudates, hemorrhages, microaneurysms etc. Although deep learning models like convolutional neural networks (CNN) seem ideal for the automated detection of abnormalities in advanced clinical imaging, small-size lesions are very hard to distinguish by using traditional networks. This work proposes a bi-directional spatial and channel-wise parallel attention based network to learn discriminative features for diabetic retinopathy grading. The proposed attention block plugged with a backbone network helps to extract features specific to fine-grained DR-grading. This scheme boosts classification performance along with the detection of small-sized lesion parts. Extensive experiments are performed on four widely used benchmark datasets for DR grading, and performance is evaluated on different quality metrics. Also, for model interpretability, activation maps are generated using the LIME method to visualize the predicted lesion parts. In comparison with state-of-the-art methods, the proposed IDANet exhibits better performance for DR grading and lesion detection.
Topics: Adult; Humans; Diabetic Retinopathy; Neural Networks, Computer; Image Interpretation, Computer-Assisted; Diabetes Mellitus
PubMed: 38462283
DOI: 10.1016/j.artmed.2024.102782