Fundus Photography Research Articles

VISTA: vision improvement via split and reconstruct deep neural network for fundus image quality assessment

Widespread eye conditions such as cataracts, diabetic retinopathy, and glaucoma impact people worldwide. Ophthalmology uses fundus photography for diagnosing these retinal disorders, but fundus images are prone to image quality challenges. Accurate diagnosis hinges on high-quality fundus images. Therefore, there is a need for image quality assessment methods to evaluate fundus images before diagnosis. Consequently, this paper introduces a deep learning model tailored for fundus images that supports large images. Our division method centres on preserving the original image’s high-resolution features while maintaining low computing and high accuracy. The proposed approach encompasses two fundamental components: an autoencoder model for input image reconstruction and image classification to classify the image quality based on the latent features extracted by the autoencoder, all performed at the original image size, without alteration, before reassembly for decoding networks. Through post hoc interpretability methods, we verified that our model focuses on key elements of fundus image quality. Additionally, an intrinsic interpretability module has been designed into the network that allows decomposing class scores into underlying concepts quality such as brightness or presence of anatomical structures. Experimental results in our model with EyeQ, a fundus image dataset with three categories (Good, Usable, and Rejected) demonstrate that our approach produces competitive outcomes compared to other deep learning-based methods with an overall accuracy of 0.9066, a precision of 0.8843, a recall of 0.8905, and an impressive F1-score of 0.8868. The code is publicly available at https://github.com/saifalkhaldiurv/VISTA_-Image-Quality-Assessment.

Laser L ight I nduced R etinopathy in a Young Boy: A Case Report from Eye Care Hospital, Karachi

A case report of a 14 year-old-boy who came with a history of direct gazing to laser pointer at Khairun'nisa Eye Hospital, Karachi. A comprehensive eye examination including Slit lamp biomicroscopy, Fundoscopy, optical coherence tomography and Fundus photography was done. Main outcome measure were right eye foveal scarring with normal periphery. Fundoscopy showed right Eye old scar at fovea, exactly like solar retinopathy. Solar retinopathy is a type of retinopathy which is caused by directly looking at sunlight. In this report we are presenting the classical case of retinopathy exactly like solar retinopathy due to gazing directly at laser light for a long period of time.

A Review of the Utility and Limitations of Artificial Intelligence in Retinal Disorders and Pediatric Ophthalmology.

Artificial intelligence (AI) is reshaping ophthalmology by enhancing diagnostic precision and treatment strategies, particularly in retinal disorders and pediatric ophthalmology. This review examines AI's efficacy in diagnosing conditions such as diabetic retinopathy (DR) and age-related macular degeneration (AMD) using imaging techniques, such as optical coherence tomography (OCT) and fundus photography. AI also shows promise in pediatric care, aiding in the screening of retinopathy of prematurity (ROP) and the management of conditions, including pediatric cataracts and strabismus. However, the integration of AI in ophthalmology presents challenges, including ethical concerns regarding algorithm biases, privacy issues, and limitations in data set quality. Addressing these challenges is crucial to ensure AI's responsible and effective deployment in clinical settings. This review synthesizes current research, underscoring AI's transformative potential in ophthalmologywhile highlighting critical considerations for its ethical use and technological advancement.

Role of fundus autofluorescence imaging in the management of submacular hemorrhage

PurposeTo evaluate the baseline characteristics of fundus autofluorescence (FAF) in patients with submacular hemorrhage (SMH).MethodsThis retrospective study included patients diagnosed with treatment-naive, foveal-involving subretinal hemorrhage (size > 2-disc diameters) of any etiology, presenting between June 2017 and June 2023. Only cases with good-quality color fundus photographs, optical coherence tomography (OCT) scans, and blue-light FAF images at baseline were included. SMH imaging characteristics were documented and correlated with treatment outcomes. A successful treatment outcome was defined as the reduction, displacement or clearance of the SMH from beneath the fovea.ResultsNineteen cases of SMH (13 males, 6 females), ranging from 14 to 85 years, were analyzed. Neovascular age-related macular degeneration (nAMD) was the most common etiology (n = 11, 58%). Baseline visual acuity ranged from 6/9 to counting fingers at ½ meter, with a median presentation time of 7 days from symptom onset (range: 1–57 days). Treatment success was observed in 13 eyes (68%). Hypoautofluoroscence on FAF was significantly associated with SMH resolution (p = 0.021). However, no association was found between treatment success and clinical hemorrhage characteristics (p = 0.222), OCT findings (p = 0.222), or specific treatments (p > 0.05). Hypoautofluoroscence on FAF was the sole predictor of treatment success, as demonstrated by Spearman’s correlation (r = 0.637; p = 0.003) and linear regression analysis (p = 0.003).ConclusionFAF, in conjunction with color fundus photography and OCT, may provide valuable insights for clinicians in formulating treatment strategies for patients with SMH. Hypoautofluoroscence on FAF was a significant predictor of successful SMH resolution in this study.

A Brief Review on Adult-Onset Coats' Disease.

Adult-onset Coats disease is an uncommon and vision-threatening disease characterized by the development of abnormal blood vessels in the retina. Coats' disease commonly affects children in the first decade of life, but very rarely manifests in adults after the third decade of life, or who characteristically present with unilateral vision loss. Despite being a sight-threatening disease, the etiology remains inconclusive and various genetic and vascular abnormalities are implicated. Diagnosis relies on ophthalmologic examination, fundus photography, fluorescein angiography, and optical coherence tomography. Treatment modalities include laser photocoagulation, intravitreal injections of anti-vascular endothelial growth factor agents, and, in advanced cases, surgical interventions are needed and the treatment is aimed at avoiding complications like retinal detachment and neovascular glaucoma, which were comparatively rare in adult-onset Coats' disease. Despite therapeutic advancements, the prognosis varies, with some patients experiencing significant visual impairment. This review outlines the clinical features, diagnosis, management, and prognosis of adult-onset Coats' disease, underscoring the importance of early detection and intervention in optimizing visual outcomes.

Convolutional Neural Networks (CNNs) for detection of Eye Diseases in an Effective Manner

Introduction: This thesis concentrates on advancing a deep learning system for the precise classification of eye illnesses. Employing the Convolutional Neural Network (CNN) algorithm and Tkinter for GUI, the interface safeguards an intuitive and user-friendly experience. The system effectively identifies various eye conditions, including glaucoma, cataracts, diabetic retinopathy, among others. Objectives: The main goal of this thesis is to offer a reliable and efficient solution for classifying eye diseases through deep learning techniques by addressing the critical need for timely and accurate diagnosis. The significant contributions of this thesis include: Publication of new datasets beneficial to researchers in the field. Enhancement of a robust segmentation approach, applicable to various retinal conditions, including diabetic retinopathy. Methods: A supervised learning approach adopted to train the CNN algorithm with an extensive dataset of eye images. The system achieved an impressive accuracy in classifying eye diseases, underscoring its effectiveness in identification and diagnosis. Development of a hybrid characteristic extraction method from segmented objects, utilized by the classifier for detecting glaucoma and other eye diseases. Results: This system will be able to detect and diagnose the diseases which are related with Eys which will help patients to take care of eyes from further complications. Furthermore, with advancements in portable retinal cameras, integrating the proposed method can facilitate the diagnosis of glaucoma and other eye diseases, thereby enhancing the detection rate, especially in growing nations. Conclusions: The developed system significantly enhances the accuracy and efficiency of eye disease diagnosis, enabling early detection and treatment. The integration of Tkinter for GUI and the color-blind test feature improves the user experience, making the system more accessible to a broader audience. This thesis underscores the potential of deep learning techniques in medical diagnosis, presenting a valuable tool for healthcare professionals and patients alike.

5099 Capturing the Patient Voice for Diabetes Eye Exams

Abstract Disclosure: S. Du: None. S. Lambrechts: None. M. Freeby: None. M. Han: None. Background: Diabetic retinopathy (DR) is a leading cause of blindness among adults 20-74 years old. Ophthalmic exams are recommended every 1-2 years for type 2 diabetes patients with minimal to no DR. The 50th percentile Integrated Healthcare Association benchmark for DR screening is 49%, but at our academic medical center, the completion rate is 30%. Potential barriers to screening noted in the literature include lower levels of education, lack of knowledge about DR, and financial barriers. Objective: To understand motivating factors and barriers to DR screening among patients in our health system. Methods: We utilized the Patient Voice qualitative research method to uncover themes and identify interventions to improve care delivery. The literature was synthesized to develop a structured interview guide. From our urban academic health system, 15 patients were randomly selected for interview, six of whom had completed the “Diabetes Eye Exam” care gap and nine that had not. Interviews were transcribed and coded for thematic analysis. Results: Six opportunity areas were identified. 1) There are currently only six retinal cameras located across primary and specialty care clinics serving more than 38,000 diabetes patients. 67% of patients felt the current clinic locations were inconvenient, with distant locations and long travel times. 2) 20% of patients expressed that they would like increased flexibility of appointments offered, such as evening, weekend, and same day appointments. 3) Patients were motivated by automated diabetes eye exam specific reminder messages and appreciated the ability to self-schedule within the patient portal. 47% of patients wanted yearly reminder messages, while 40% had no preference. 4) 53% of patients expressed that they did not fully understand DR or were confused about the difference between an optometrist exam and diabetes eye exam. 5) 40% of patients completed their eye exam with their local optometrist and found this to be a more convenient option. One patient expressed that she wished her doctor had informed her that she could complete her diabetes eye exam at a local optometry clinic. 6) If patients do choose to complete their exam with their local optometrist, they would like reminders to send their records to their primary care physician or endocrinologist. A more streamlined processes to upload outside records through the patient portal would decrease this barrier. Conclusion: It is important to engage patient stakeholders and understand their needs to improve overall experience and close care gaps. Through the Patient Voice, we learned that factors such as driving distance, traffic, scheduling, and issues with uploading records are major barriers to care gap completion. Convenient clinic locations, education, and reminder messages are motivating factors. Future work will focus on developing and implementing solutions in each opportunity area. Presentation: 6/1/2024

Геоинформационная система глазного дна

Purpose. To develo fundus operational geoinformation system for accurate topographic identification of the pathological process characteristics. Material and methods. An operational digital fundus geoinformation system was developed using available program «Transporter Online». A digital fundus photograph of the patient with active retinopathy of prematurity (ROP), obtained with a «RetCam» retinal camera, was placed in a cartesian field. Next, taking the center of the optic nerve head (ONH) as the zero point, we built parallels with a step equal to one optic disc radius and meridians with a step of 10 degrees by analogy with the TABO scale (Technische Ausschuss fur Brillen – Optik). Results. As a result, a geoinformation system of the fundus was built in the form of a coordinate grid comparable to a geographic coordinate system. Using the developed system, it is possible to measure many parameters on the retina. As the developed system is used, the number and information content of the parameters will undergo changes depending on the assigned tasks. It is possible to use the fundus geoinformation system for any retinal pathology, in particular for active ROP. To eliminate the influence of the scale factor and the reliability of quantitative indicators of image magnification, digitization is carried out based on the results of measuring angles and linear dimensions in «units» of the optic disc radius, which determines the individuality and originality of the proposed system. Conclusion. The proposed fundus geoinformation system will provide accurate topographic identification of the pathological process characteristics on the retina surface. The obtained coordinates of the fundus structures, which are of research interest, will form an operational navigation system that will significantly facilitate the work of ophthalmologists in transmitting information, analyzing the dynamics and outcomes of the pathological process, and can also be as the basis for a further transition to artificial intelligence use. Key words: retina, fundus, topography, geoinformation system

Associations between Disc Hemorrhage and Primary Open-Angle Glaucoma Based on Genome-Wide Association and Mendelian Randomization Analyses.

Background/Objectives: We aimed to investigate the genetic loci related to disc hemorrhage (DH) and the relationship of causation between DH and primary open-angle glaucoma (POAG) using a genome-wide association study (GWAS) in East Asian individuals. Methods: The GWAS included 8488 Koreans who underwent ocular examination including fundus photography to determine the presence of DH and POAG. We performed a GWAS to identify significant single-nucleotide polymorphisms (SNPs) associated with DH and analyzed the heritability of DH and genetic correlation between DH and POAG. The identified SNPs were utilized as instrumental variables (IVs) for two-sample Mendelian randomization (MR) analysis. The POAG outcome dataset was adopted from Biobank Japan data (n = 179,351). Results: We found that the rs62463744 (TMEM270;ELN), rs11658281 (CCDC42), and rs77127203 (PDE10A;LINC00473) SNPs were associated with DH. The SNP heritability of DH was estimated to be 6.7%, with an absence of a genetic correlation with POAG. MR analysis did not reveal a causal association between DH and POAG for East Asian individuals. Conclusions: The novel loci underlying DH in the Korean cohort revealed SNPs in the ELN, CCDC41, and LINC00473 genes. The absence of a causal association between DH and POAG implies that DH is a shared risk factor, rather than an independent culprit factor, and warrants further investigation.

Cardiovascular Risk Factors as Independent Predictors of Diabetic Retinopathy in Type II Diabetes Mellitus: The Development of a Predictive Model.

Background: Diabetic retinopathy (DR) is a leading cause of blindness in patients with type 2 diabetes mellitus (T2DM). Cardiovascular risk factors, such as hypertension, obesity, and dyslipidemia, may play a crucial role in the development and progression of DR, though the evidence remains mixed. This study aimed to assess cardiovascular risk factors as independent predictors of DR and to develop a predictive model for DR progression in T2DM patients. Methods: A retrospective cross-sectional study was conducted on 377 patients with T2DM who underwent a comprehensive eye exam. Clinical data, including blood pressure, lipid profile, BMI, and smoking status, were collected. DR staging was determined through fundus photography and classified as No DR, Non-Proliferative DR (NPDR), and Mild, Moderate, Severe, or Proliferative DR (PDR). A Multivariate Logistic Regression was used to evaluate the association between cardiovascular risk factors and DR presence. Several machine learning models, including Random Forest, XGBoost, and Support Vector Machines, were applied to assess the predictive value of cardiovascular risk factors and identify key predictors. Model performance was evaluated using accuracy, precision, recall, and ROC-AUC. Results: The prevalence of DR in the cohort was 41.6%, with 34.5% having NPDR and 7.1% having PDR. A multivariate analysis identified systolic blood pressure (SBP), LDL cholesterol, and body mass index (BMI) as independent predictors of DR progression (p < 0.05). The Random Forest model showed a moderate predictive ability, with an AUC of 0.62 for distinguishing between the presence and absence of DR XGBoost showing a better performance, featuring a ROC-AUC of 0.68, while SBP, HDL cholesterol, and BMI were consistently identified as the most important predictors across models. After tuning, the XGBoost model showed a notable improvement, with an ROC-AUC of 0.72. Conclusions: Cardiovascular risk factors, particularly BP and BMI, play a significant role in the progression of DR in patients with T2DM. The predictive models, especially XGBoost, showed moderate accuracy in identifying DR stages, suggesting that integrating these risk factors into clinical practice may improve early detection and intervention strategies for DR.

Macular hole associated Retinal Detachment in presumed Autosomal Recessive Bestrophinopathy (ARB)- A retrospective observational series of two cases.

Surgical outcome of presumed Autosomal Recessive Bestrophinopathy (ARB)with full thickness macular hole and retinal detachment managed by pars plana vitrectomy. A retrospective observational case series of two eyes of two patients who underwent 25 G pars plana vitrectomy with peeling of internal limiting membrane and SF6 tamponade and were assessed with respect to best corrected visual acuity, fundus photographs and spectral domain optical coherence tomography. Type 1 closure of macular hole was achieved along with resolution of retinal detachment in both patients. 25 G pars plana vitrectomy with peeling of internal limiting membrane and SF6 tamponade can result in favorable outcome in macular hole associated retinal detachment in patients with presumed Autosomal recessive Bestrophinopathy.

Branch retinal vein occlusion and peripheral nonperfusion associated with astrocytic hamartoma of the optic nerve head in a patient with tuberous sclerosis.

To report a case of branch retinal vein occlusion (BRVO) causing retinal neovascularization (NV) and vitreous hemorrhage associated with astrocytic hamartoma (AH) involving the optic nerve head in a patient with tuberous sclerosis complex (TSC). We review the natural history of the clinical presentation and posit its etiology. Electronic health records were reviewed for patient history and demographics and multimodal ophthalmic imaging including fundus photography, fluorescein angiography, and optical coherence tomography (OCT). Neovascularization emanating superiorly and inferiorly from a large AH of the optic nerve head was identified as the cause of a longstanding vitreous hemorrhage. Retinal nonperfusion was also found in superior peripheral quadrants, most notably distal to the superior NV. OCT was useful in visualizing small retinal AH in the fellow eye that was otherwise undetected on prior dilated fundus examinations. The patient was treated with intravitreal anti-VEGF and superior sectoral PRP in the involved eye. BRVO and associated NV may represent rare vision-threatening sequela of AH in TSC and manifested in this case with diffuse peripheral retinal nonperfusion secondary to BRVO-induced ischemia. OCT is a valuable resource in identifying and monitoring AH of the retina.

Affordable ultra-widefield smartphone PedCam for comprehensive pediatric retinal examination.

Widefield fundus photography is critical for the detection, documentation, and management of pediatric eye diseases. Existing clinical pediatric fundus cameras offer a limited field of view (FOV) and suboptimal image contrast, hindering comprehensive peripheral retina examination. Additionally, the high cost and lack of portability of commercial devices restrict their use in resource-limited settings. We introduce a cost-effective smartphone-based pediatric camera (PedCam) that provides a 180° eye angle (126° visual angle) snapshot FOV. Utilizing trans-pars planar illumination, the device enables nonmydriatic imaging by allocating the pupil exclusively for imaging, eliminating the need for pharmacological pupillary dilation. By adjusting the optical axis of the PedCam relative to the ocular axis, the effective FOV can be expanded up to 240° eye angle (180° visual angle), enabling complete retinal evaluation. This innovative smartphone PedCam represents a significant advancement in affordable telemedicine for the screening, monitoring, and management of retinopathy of prematurity and other pediatric eye conditions.

Evaluation of the different methods of optic disc examination in glaucomatous patients

Background Glaucoma is a leading cause of blindness worldwide. Many patients discover glaucoma accidentally during regular checkups. Purpose To evaluate the agreement using slit lamp biomicroscopy, stereoscopic fundus photography, and optical coherence tomography (OCT) in the evaluation of the optic disc area, the cup area, and the vertical cup disc ratio (VCDR) in glaucomatous patients. Patients and methods This cross-sectional observational study includes 28 eyes of glaucomatous patients. All patients underwent slit lamp biomicroscopy using a +90 D lens, stereoscopic fundus photography using a computerized program to measure the optic disc parameters, and OCT for optic disc evaluation. Three examiners did the analysis. Results Comparing results obtained on stereoscopic fundus photography and OCT, examiner one reported moderate reliability regarding disc area, cup area, and the VCDR with an intraclass coefficient (ICC) of 0.67, 0.67, and 0.50 and P values of 0.68, 0.61, and 0.01, respectively. Examiner two reported moderate reliability regarding disc area, cup area, and VCDR with an ICC of 0.66, 0.63, and 0.54 and P values of 0.16, 0.69, and 0.03, respectively. Examiner three reported moderate reliability regarding disc area and VCDR with an ICC of 0.65, 0.71, and P values of 0.69 and 0.07, respectively. A good reliability for the cup area was reported with an ICC of 0.82 and a P value of 0.02. The slit lamp and stereoscopic fundus photos showed moderate reliability compared with fundus photography values. Conclusion There is an agreement between values obtained by slit lamp biomicroscopy, stereoscopic fundus photography, and OCT.

Retro mode imaging signatures in choroidal nevus: A case series

To report the retro mode (RM) imaging signatures of choroidal nevus. This was a consecutive case series of individuals with choroidal nevus, a benign choroidal tumor. Fundus photography and RM imaging were performed using the Mirante system ((Nidek Co. Ltd, Gamagori, Japan). Four eyes of four patients were included. Choroidal nevus was observed as hyporetroreflective in all four patients. RM image was able to distinctly delineate the margins of the nevus. Overlying drusen was variably imaged with RM, appearing as isoretroreflective in one eye and not visualized in another. RM imaging was able to depict choroidal nevus as a dark lesion with well-delineated margins. It may have a role in the documentation and follow-up of eyes with choroidal nevus.

Privacy Preserving Technology using Federated Learning and Blockchain in protecting against Adversarial Attacks for Retinal Imaging

PurposeCollaboration provides valuable data for robust artificial intelligence (AI) model development. Federated learning (FL) is a privacy enhancing technology that allows collaboration while respecting privacy via the development of models without raw data transfer. However state-of-the-art FL models still face challenges in non-independent and identically distributed (non-i.i.d.) health care settings and remain susceptible to privacy breaches. We propose a novel FL framework coupled with blockchain technology to address these challenges. DesignRetrospective multicohort study Subjects and Methods27,145 images from Singapore, China and Taiwan were used to design a novel FL aggregation method for the detection of myopic macular degeneration (MMD) from fundus photographs and macular disease from optical coherence tomography (OCT) scans in feature distribution skew and label distribution imbalance scenarios. We further performed adversarial attacks (label flipping and clean label). As proof of concept, blockchain was incorporated into FL to demonstrate secure transfer of model updates across collaborating sites. Main Outcome MeasuresWe evaluated our FL model performance in MMD and OCT classification and compared our model against state-of the-art FL and centralized models. ResultsOur FL model showed robust performance with areas under the receiving operating characteristic curves (AUC) of 0.868±0.009 for MMD detection and 0.970±0.012 for OCT macular disease classification. In label flipping attack, our FL model had an AUC of 0.861±0.019, similar to the centralized model (AUC 0.856± 0.015) and higher than other FL models (AUC 0.578–0.819) In clean label attack, our FL model had an AUC of 0.878±0.006 which was comparable to the centralized model (AUC 0.878±0.001) and superior to other state-of-the-art FL models with AUC of 0.529-0.838. Simulation showed that the additional time with blockchain in one global epoch was around 5 seconds. The addition of blockchain to the FL framework was feasible with a minimal impact on model development time. ConclusionsOur proposed FL algorithm overcomes the shortcoming of the traditional FL in non i.i.d. situations and remains robust to against adversarial attacks. The integration of blockchain adds further security during the transfer of model updates. Blockchain-enabled FL can be a trusted platform for collaborative health AI research.

The Optical Nature of Myopic Changes in Retinal Vessel Caliber

PurposeDimensional measures of retinal features are subject to the optical influence of ocular magnification. We examined the impact of ocular magnification on the association between axial length (AL) and measurements of retinal vessel calibre in fundus photographs. DesignCross-sectional study. ParticipantsEighty-two normal right eyes from healthy participants aged 16-31 years. MethodsCentral retinal arteriolar and venular equivalents (CRAE and CRVE) were derived from colour fundus photographs using semi-automated software. Ordinary least squares linear regression was used to assess the influence of AL (independent variable) on CRAE and CRVE, controlling for age, sex and ethnicity, both before and after magnification correction using different formulae. These formulae estimate magnification based on different ocular parameters: AL only (Bennnett’s formula), refractive error only (Bengtsson’s formula) and refractive error combined with keratometry (Littmann’s formula). Previous research has primarily relied on Bengtsson’s formula, which is less accurate than Bennett’s formula. We also examined the impact of treating the non-telecentric fundus camera used in this study as telecentric when applying these magnification correction formulae. Main Outcome MeasuresCentral retinal arteriolar and venular equivalents (in pixels). ResultsBefore magnification correction, increasing AL was associated with decreasing CRAE (β: -0.49, 95% CI: -0.89 to -0.09, p=0.02) and CRVE (β: -0.91, 95% CI: -1.62 to -0.20, p=0.01). After magnification correction, this observation was no longer evident, regardless of the correction formula applied. When inappropriately assuming the fundus camera to be telecentric, we observed a bias towards increasing magnification-corrected CRAE and CRVE with increasing AL (β coefficients were positive or became more positive), reaching statistical significance (p<0.05) for CRAE corrected using Bennett’s or Littmann’s formula, and for CRVE corrected using Bennett’s formula. ConclusionsFailing to correct for ocular magnification results in apparent narrowing of vessels in longer eyes, while inappropriate assumptions about telecentricity during magnification correction introduce an optical artifact that causes apparent widening of vessels. These findings suggest that myopic changes in retinal vessel calibre are optical (not biological) in nature. Proper correction of this effect to accurately derive dimensional measures is a crucial—yet often overlooked—methodological consideration in “oculomics” research investigating retinal biomarkers of systemic conditions.

Is Automated Machine Learning useful for ocular toxoplasmosis identification and classification of the inflammatory activity?

PurposeTo evaluate the performance of Automated Machine Learning (AutoML) models in diagnosing ocular toxoplasmosis (OT) and classifying its inflammatory activity from fundus photographs. DesignCross-sectional study. MethodsFundus photographs from OT patients in two Colombian referral centers and an open-source OT database were classified into active OT, inactive OT, and no OT. Image quality assessment excluded images with artifacts but included blurry images due to vitritis. Photos were uploaded to Amazon Web Services S3 and Google Cloud Bucket. Two models were developed on each platform: a binary model (active/inactive OT vs. no OT) and a multiclass model (active OT, inactive OT, and no OT). Datasets were split into 70% for training, 20% for testing, and 10% for validation. Sensitivity, specificity, precision, accuracy, F1-score, the area under the precision-recall curve (AUPRC), and Cohen's Kappa were calculated for each platform and model. An external validation using an open-source image bank was performed. ResultsThe binary model on AWS showed a sensitivity of 0.97, specificity of 0.98, and AUPRC of 1.00, while the Google Cloud binary model had a sensitivity of 0.82, specificity of 0.91, and AUPRC of 0.91. The multiclass model on AWS achieved an F1 score of 0.88, with Cohen's Kappa of 0.81, while the Google Cloud model reached an F1 score of 0.88, with Cohen's Kappa of 0.82. External validation for Google Cloud achieved an accuracy of 87.5% and 80.3% in the binary and multiclass models, respectively. ConclusionsAutoML is a powerful tool for diagnosing OT and classifying inflammatory activity, potentially guiding diagnosis and treatment decisions.

Illuminated depressor enhanced fundus photography

Illuminated depressor enhanced fundus photography

Evaluation of the Different Methods of Optic Disc Examination in Glaucomatous Patients

Abstract Background Glaucoma is the second leading cause of irreversible blindness worldwide. It affects the nerve fiber layer which forms the optic nerve. Many patients discover glaucoma accidently when they’re doing regular checkups. Glaucoma may have no specific symptoms until the advanced stages and it can be diagnosed by measuring intra ocular pressure, optic disc evaluation and visual field examination. The present work aims at evaluation of different methods in optic disc examination in glaucoma patients to detect their efficacy in diagnosis and follow up. Purpose To evaluate the agreement among slit lamp biomicroscopy, stereoscopic fundus photography and OCT in glaucomatous patients regarding Disc area, Cup area and vertical cup disc ratio (VCDR). Material and Methods 14 known glaucomatous patients were included in this cross-sectional observational descriptive. All patients under went slit lamp biomicroscopy examination by using +90 lens, stereoscopic fundus photography to obtain 3D image for optic disc and using computerized program to measure the optic disc parameters and Optical Coherence Tomography for the optic disc. Results Comparing each examiner values measured on stereoscopic fundus image in comparison with OCT. Examiner one have a moderate reliability regarding disc area, cup area and VCDR with ICC 0.67, 0.67 and 0.50 and P values 0.68, 0.61 and 0.01 respectively and examiner two have a moderate reliability regarding disc area, cup area and VCDR with ICC 0.66, 0.63 and 0.54 and P values 0.16, 0.69 and 0.03 respectively and examiner 3 has a moderate reliability regarding disc area and VCDR with ICC 0.65, 0.71 and P value 0.69, 0.07 respectively while a good reliability for cup area with ICC 0.82 and P value 0.02. Comparing values obtained by slit lamp using +90 lens and stereoscopic fundus photography regarding VCDR there is a moderate reliability with ICC 0.50 and P value 0.06 for examiner one only. Conclusion The results of the present study showed that there is an moderate agreement between stereoscopic fundus photography and OCT and slit lamp values regarding disc area, cup area and CD ratio.