Follow-up report of fundus findings of tuberous sclerosis-associated retinal astrocytoma of two siblings.

A late adolescent with tuberous sclerosis (TS) presented with reduced vision in one eye to our tertiary care university hospital 4 years ago. Fundus examination revealed multiple retinal astrocytic hamartomas (RAHs) in both eyes. His younger sibling, who also had TS, was found to have RAH on retinal screening. The swept-source optical coherence tomography (SS-OCT) findings were typical of RAH. We further noted that some of the RAH lesions showed segmental whitening of the outer walls of the arterioles, which traversed through them. The segmental whitening may suggest the enveloping of normal retinal vessels by the tumour. En-face and B-scan SS-OCT angiography of patients with TS showed vascularity within the tumour. The vessels within the tumour appeared to be in continuity with the retinal vasculature. Both siblings were reviewed annually. At the end of 4 years, there was no change in visual acuity, tumour size, number, vascularity and behaviour.

Diagnostic Role of Oral Fluorescein Angiography in Pediatric Ambulatory Clinics.

Oral ingestion of fluorescein can be done in ambulatory pediatric clinics. We show that oral ultra-widefield fluorescein angiography is a non-invasive approach to rapidly diagnose and manage a diverse set of pediatric retinal vascular diseases.

Automated machine learning model for fundus image classification by health-care professionals with no coding experience.

To assess the feasibility of code-free deep learning (CFDL) platforms in the prediction of binary outcomes from fundus images in ophthalmology, evaluating two distinct online-based platforms (Google Vertex and Amazon Rekognition), and two distinct datasets. Two publicly available datasets, Messidor-2 and BRSET, were utilized for model development. The Messidor-2 consists of fundus photographs from diabetic patients and the BRSET is a multi-label dataset. The CFDL platforms were used to create deep learning models, with no preprocessing of the images, by a single ophthalmologist without coding expertise. The performance metrics employed to evaluate the models were F1 score, area under curve (AUC), precision and recall. The performance metrics for referable diabetic retinopathy and macular edema were above 0.9 for both tasks and CDFL. The Google Vertex models demonstrated superior performance compared to the Amazon models, with the BRSET dataset achieving the highest accuracy (AUC of 0.994). Multi-classification tasks using only BRSET achieved similar overall performance between platforms, achieving AUC of 0.994 for laterality, 0.942 for age grouping, 0.779 for genetic sex identification, 0.857 for optic, and 0.837 for normality with Google Vertex. The study demonstrates the feasibility of using automated machine learning platforms for predicting binary outcomes from fundus images in ophthalmology. It highlights the high accuracy achieved by the models in some tasks and the potential of CFDL as an entry-friendly platform for ophthalmologists to familiarize themselves with machine learning concepts.

Ixabepilone related angiographically silent macular edema.

INTRODUCTION: We present a single-eyed case with a previous diagnosis of breast cancer who had intraretinal cystoid changes associated with the systemic administration of ixabepilone in her only seeing eye. To our best knowledge, this is the first reported case describing this phenomenon related to the ixabepilone administration. CASE DESCRIPTION: A 54-year-old woman with a history of breast cancer was examined due to visual deterioration in her only good left eye. The patient had undergone cataract surgery and lens implantation in her right eye following a childhood accident, but subsequently had developed a refractory glaucoma and lost her right vision. Six cycles of 40 mg/m2 systemic ixabepilone (3-hly intravenous infusion once every 3 weeks) had been administered within the past six months. Her visual decline started two weeks following the last treatment session. She was offered intravitreal anti-vascular endothelial growth factor injection elsewhere. Fluorescein angiogram showed no dye leakage whereas spectral-domain optical coherence tomography demonstrated parafoveal intraretinal cystoid changes. En-face optical coherence tomography revealed petaloid type roundish hyporeflective areas at the level of superficial and deep vascular plexus. Ixabepilone-associated cystoid maculopathy was suspected as she received only ixabepilone for the chemotherapy in the last six months. We thus recommended her not to continue ixabepilone therapy. Ten weeks after the ixabepilone cessation, intraretinal cystoid changes had resolved completely. CONCLUSION: Angiographically silent intraretinal cystoid changes may develop in association with the use of ixabepilone. Referral to an ophthalmologist should be considered for the patients experiencing visual complaints as ixabepilone cessation may lead to visual improvement and avoid unnecessary treatment.

Comparison of 21 artificial intelligence algorithms in automated diabetic retinopathy screening using handheld fundus camera.

BACKGROUND: Diabetic retinopathy (DR) is a common complication of diabetes and may lead to irreversible visual loss. Efficient screening and improved treatment of both diabetes and DR have amended visual prognosis for DR. The number of patients with diabetes is increasing and telemedicine, mobile handheld devices and automated solutions may alleviate the burden for healthcare. We compared the performance of 21 artificial intelligence (AI) algorithms for referable DR screening in datasets taken by handheld Optomed Aurora fundus camera in a real-world setting. PATIENTS AND METHODS: Prospective study of 156 patients (312 eyes) attending DR screening and follow-up. Both papilla- and macula-centred 50° fundus images were taken from each eye. DR was graded by experienced ophthalmologists and 21 AI algorithms. RESULTS: Most eyes, 183 out of 312 (58.7%), had no DR and mild NPDR was noted in 21 (6.7%) of the eyes. Moderate NPDR was detected in 66 (21.2%) of the eyes, severe NPDR in 1 (0.3%), and PDR in 41 (13.1%) composing a group of 34.6% of eyes with referable DR. The AI algorithms achieved a mean agreement of 79.4% for referable DR, but the results varied from 49.4% to 92.3%. The mean sensitivity for referable DR was 77.5% (95% CI 69.1-85.8) and specificity 80.6% (95% CI 72.1-89.2). The rate for images ungradable by AI varied from 0% to 28.2% (mean 1.9%). Nineteen out of 21 (90.5%) AI algorithms resulted in grading for DR at least in 98% of the images. CONCLUSIONS: Fundus images captured with Optomed Aurora were suitable for DR screening. The performance of the AI algorithms varied considerably emphasizing the need for external validation of screening algorithms in real-world settings before their clinical application.

What is already known on this topic? Diabetic retinopathy (DR) is a common complication of diabetes. Efficient screening and timely treatment are important to avoid the development of sight-threatening DR. The increasing number of patients with diabetes and DR poses a challenge for healthcare.What this study adds? Telemedicine, mobile handheld devices and artificial intelligence (AI)-based automated algorithms are likely to alleviate the burden by improving efficacy of DR screening programs. Reliable algorithms of high quality exist despite the variability between the solutions.How this study might affect research, practice or policy? AI algorithms improve the efficacy of screening and might be implemented to clinical use after thorough validation in a real-life setting.

Association between the arm-to-choroidal circulation time and clinical profile in patients with polypoidal choroidal vasculopathy.

PURPOSE: To investigate the association between the arm-to-choroidal circulation time (ACT) on indocyanine green angiography (IA) and clinical profile in patients with polypoidal choroidal vasculopathy (PCV). STUDY DESIGN: Single-center retrospective study. METHODS: We included 38 eyes of 38 patients with PCV diagnosed using multimodal imaging and did not undergo previous treatment. All patients were treated with monthly aflibercept injections for 3 months and treat-and-extend regimens for the subsequent 12 months. Posterior vortex vein ACT was assessed on the first visit using Heidelberg IA. The patients were divided into two groups: ACT ≥20 s (L group; eight eyes) and ACT <20 s (S group; 30 eyes). The clinical profiles before and after treatment were analyzed to assess associations with ACT. RESULTS: The mean ACT was 16.39±3.3 s (L group: 21.25±1.49 s, women:men=2:6, mean age: 77.3±6.5 years; S group: 15.10±2.17 s, women:men=7:23, mean age: 75.5±6.9 years). No significant difference was observed in the mean subfoveal choroidal thickness between the L and the S groups (176±75 µm vs. 230±79 µm, P=0.10). However, there were significant differences between the L and S groups in retinal fluid accumulation and hemorrhage recurrence (eight/eight eyes, 100% vs. 13/30 eyes, 43%, P<0.001), mean aflibercept injections (8.8±1.6 vs. 7.0±1.6, P<0.01) during the 12-month period, and the number of polypoidal lesions (1.8±0.7 vs. 1.3±0.5, P<0.05). CONCLUSION: Patients with PCV and ACT >20 s are more likely to experience exudative change recurrence in the retina during treatment because they have more polypoidal lesions.

Vasoproliferative Tumor Secondary to Sarcoidosis-Associated Intermediate Uveitis.

We report the visual and clinical outcomes of a middle-aged woman who presented with exudative retinal detachment (ERD) secondary to a vasoproliferative tumor (VPT) in an eye with sarcoidosis-associated intermediate uveitis. A 55-year-old woman previously diagnosed with sarcoidosis presented with decreased vision in the left eye (LE). Visual acuity in the LE was counting fingers. She had active vitritis, and a peripheral retinal vascular mass was noted in the superotemporal periphery. The mass was associated with ERD involving the posterior pole. The patient was managed with systemic and intravitreal steroids, and cyclosporine was subsequently added as a steroid-sparing agent. Because of recurrence of ERD, the patient underwent pars plana vitrectomy, and cryotherapy and laser photocoagulation were applied to the VPT. Two months postoperatively, visual acuity in the LE improved to 6/10. There was marked regression of the VPT and total resolution of the ERD. In conclusion, we report a favorable visual and clinical outcome in a patient with VPT-associated ERD who responded to a combination of medical therapy and surgical intervention. VPT may lead to different remote complications, so timely diagnosis of these tumors and proper management of their complications is warranted.

Swept-source optical coherence tomography angiography findings in a case of primary vitreoretinal lymphoma over a three-year follow-up.

BACKGROUND: Vitreoretinal lymphoma (VRL) still represents a diagnostic challenge for retinal specialists. Early diagnosis and treatment are critical for a better prognosis. Several diagnostic tools have proven helpful in the identification of VRL abnormalities. However, swept-source OCT angiography (SS-OCT-A) findings and their long-term follow-up are yet to be explored. CASE PRESENTATION: a 42-year-old man presented with blurred vision in his left eye for 2 weeks. He denied any systemic symptoms. A multimodal imaging examination was performed, raising the clinical suspicion of VRL and guiding the ensuing diagnostic procedures. The patient underwent treatment and at the last FU visit three years later, no disease signs were present on fundus examination, nor on oncologic evaluation. Some novel SS-OCT-A features were identified, and uncommonly reported findings were examined over a long-term follow-up. At baseline multiple hyperreflective alterations were detected on the enface outer retina slabs and choriocapillary analysis revealed low reflectance areas in the foveal and parafoveal areas. One month after the first presentation, multiple hyperreflective retinal lesions in a vertical shape were detected on OCT which appeared on midretinal slabs of enface SS-OCT-A as hyperreflective spots mainly located near second-order retinal vessels. These alterations remarkably reduced after treatment. CONCLUSION: SS-OCT-A may be a useful imaging technique in the detection of VRL, providing ophthalmologists additional findings that assist the diagnosis and follow-up of this disease. This may prove useful for a more timely and precise diagnosis, prompt therapy, and treatment response monitoring. The original aspects found in this case may provide grounds for future studies, ultimately fostering a better understanding of the disease.

A novel vector field analysis for quantitative structure changes after macular epiretinal membrane surgery.

The aim of this study was to introduce novel vector field analysis for the quantitative measurement of retinal displacement after epiretinal membrane (ERM) removal. We developed a novel framework to measure retinal displacement from retinal fundus images as follows: (1) rigid registration of preoperative retinal fundus images in reference to postoperative retinal fundus images, (2) extraction of retinal vessel segmentation masks from these retinal fundus images, (3) non-rigid registration of preoperative vessel masks in reference to postoperative vessel masks, and (4) calculation of the transformation matrix required for non-rigid registration for each pixel. These pixel-wise vector field results were summarized according to predefined 24 sectors after standardization. We applied this framework to 20 patients who underwent ERM removal to obtain their retinal displacement vector fields between retinal fundus images taken preoperatively and at postoperative 1, 4, 10, and 22 months. The mean direction of displacement vectors was in the nasal direction. The mean standardized magnitudes of retinal displacement between preoperative and postoperative 1 month, postoperative 1 and 4, 4 and 10, and 10 and 22 months were 38.6, 14.9, 7.6, and 5.4, respectively. In conclusion, the proposed method provides a computerized, reproducible, and scalable way to analyze structural changes in the retina with a powerful visualization tool. Retinal structural changes were mostly concentrated in the early postoperative period and tended to move nasally.

GravityNet for end-to-end small lesion detection.

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.

[Preclinical diagnostics of von Hippel-Lindau syndrome in a child].

The description of the child aged 5 months with the von Hippel-Lindau syndrome without any manifestations of this syndrome is presented. The reason for the molecular genetic examination was the presence of cases of this syndrome in the family (mother and sister). The heterozygous variant c.355T>C p.F119L was found in the VHL gene. An objective examination revealed no pathology. A comprehensive laboratory and instrumental examination aimed at searching for components of the von Hippel-Lindau syndrome, including a blood test for metanephrines and normetanephrines, ultrasound of the abdominal organs, examination of the fundus, also did not reveal any abnormalities. Given the results of molecular genetic diagnosis, the child remains under observation and will undergo regular examinations to identify components of the von Hippel-Lindau syndrome, including blood/urine tests for normetanephrines.

[Clinical observation of the intraocular distribution characteristics of indocyanine green after epiretinal membrane peeling using a fluorescence detection system developed in Python].

Objective: To utilize a Python-based fluorescence area detection system to observe and quantitatively analyze the intraocular distribution characteristics and metabolic patterns of Indocyanine Green (ICG) following epiretinal membrane peeling. Methods: A prospective case series study was conducted on patients with idiopathic epiretinal membrane undergoing vitrectomy at West China Hospital of Sichuan University from March 2019 to March 2021. ICG staining was applied during surgery for peeling the epiretinal membrane and internal limiting membrane. Patients were followed up at 1 week, 1 month, 3 months, 6 months, and 12 months postoperatively, with assessments including best-corrected visual acuity, intraocular pressure, fundus photography, near-infrared fundus fluorescence imaging (NIR-FF), and optical coherence tomography (OCT). A Python-based ICG intraocular metabolism detection system was developed to measure the residual area of ICG fluorescence on NIR-FF, predict the ICG metabolic pattern equation, and correlate it with postoperative visual acuity and peripapillary retinal nerve fiber layer thickness. Results: A total of 64 patients (64 eyes) were included, with an average age of 64.6±8.4 years, including 25 males (39.1%) and 39 females (60.9%). Preoperative NIR-FF images showed no ICG strong fluorescence. At 1 week postoperatively, diffuse ICG strong fluorescence appeared in the posterior pole, and the internal limiting membrane removal area exhibited a ring-like weak fluorescence. Over time, ICG strong fluorescence was observed along the vascular arch and nerve fiber trajectory, gradually diminishing toward the optic disc, with residual ICG fluorescence still visible at the optic disc at 1 year. The Python-based ICG fluorescence area detection system effectively measured intraocular residual ICG area. A predictive equation for the 12-month residual ICG area was constructed through linear regression analysis (Residual ICG area=0.22 × Residual ICG area at 6 months, R2=16%, P=0.002). Except for a negative correlation between the ICG residual area at 1 month and postoperative visual acuity (P=0.017, r=-0.195), no correlation was found between intraocular ICG fluorescence residual area and postoperative visual acuity or peripapillary retinal nerve fiber layer thickness at other follow-up times (all P>0.05). Conclusions: In patients with idiopathic epiretinal membrane undergoing ICG staining for internal limiting membrane peeling, ICG exhibits characteristic metabolic processes in the eye, with strong fluorescence along the vascular arch and nerve fiber trajectory, gradually converging toward the optic disc over time. The Python-based ICG fluorescence area detection system provides a clear display of the intraocular distribution characteristics of ICG after epiretinal membrane peeling and serves as a tool for predicting the metabolic patterns of ICG in the eye.

Personalized Lens Correction Improves Quantitative Fundus Autofluorescence Analysis.

Purpose: Quantitative fundus autofluorescence (QAF) currently deploys an age-based score to correct for lens opacification. However, in elderly people, lens opacification varies strongly between individuals of similar age, and innate lens autofluorescence is not included in the current correction formula. Our goal was to develop and compare an individualized formula. Methods: One hundred thirty participants were examined cross-sectionally, and a subset of 30 participants received additional multimodal imaging 2-week post-cataract-surgery. Imaging included the Scheimpflug principle, anterior chamber optical coherence tomography (AC-OCT), lens quantitative autofluorescence (LQAF), and retinal QAF imaging. Among the subset, least absolute shrinkage and selection operator regression and backward selection was implemented to determine which lens score best predicts the QAF value after lens extraction. Subsequently, a spline mixed model was applied to the whole cohort to quantify the influence of LQAF and Scheimpflug on QAF. Results: Age and LQAF measurements were found to be the most relevant variables, whereas AC-OCT measurements and Scheimpflug were eliminated by backward selection. Both an increase in Scheimpflug and LQAF values were associated with a decrease in QAF. The prediction error of the spline model (mean absolute error [MAE] ± standard deviation) of 32.2 ± 23.4 (QAF a.u.) was markedly lower compared to the current age-based formula MAE of 96.1 ± 93.5. Both smooth terms, LQAF (P < 0.01) and Scheimpflug (P < 0.001), were significant for the spline mixed model. Conclusions: LQAF imaging proved to be the most predictive for the impact of the natural lens on QAF imaging. The application of lens scores in the clinic could improve the accuracy of QAF imaging interpretation and might allow including aged patients in future QAF studies.

Indocyanine Green Angiography.

The choroid plays an important role in the pathophysiology of the eye. Multimodal imaging offers different techniques to examine the choroid. Fundus fluorescein angiography offers limited visualization of the deep layers of the fundus due to the barrier property of the retinal pigment epithelium. Therefore, indocyanine green angiography (ICGA) is widely used in the angiographic examination of the choroidal structure. ICGA is an important component of multimodal imaging in the diagnosis and treatment of many degenerative, tumoral, and inflammatory diseases of the choroid and retina. This review presents the general characteristics of ICGA and a practical approach to its clinical use.

Effects of hemoglobin concentration on retinochoroidal vascular plexuses: an optical coherence tomography angiography study.

PURPOSE: To evaluate retinal and choroidal microvascular structures using optical coherence tomography angiography (OCTA) in patients with anemia and polycythemia vera (PV). METHODS: In this prospective study, 142 patients (142 eyes) were enrolled and divided into 3 groups: the anemia (n = 56), PV (n = 46), and healthy groups (n = 40, controls). For each patient, 6- × 6-mm macular angiography images were taken using an OCTA system (optovue, Inc., Fremont). For each eye analyzed, the software automatically measured vessel density (VD) in the superficial capillary plexus (SCP), deep capillary plexus (DCP; superior, nasal, temporal, and inferior quadrants), choriocapillaris (CC), and foveal avascular zone (FAZ). RESULTS: The OCTA analysis revealed that the VD of the DCP was significantly decreased in the superior and inferior hemispheres of the whole area, multiple quadrants of the perifovea, and CC with a 1-mm2 flow area in the anemia group compared with the PV group (p < 0.017), but the VD of the SCP did not show any significant difference in any quadrant between the two groups (p > 0.017). When compared with the healthy group, the anemia and PV groups showed a significant difference in multiple quadrants of the parafovea and temporal quadrant of the perifovea in the VD of the SCP and CC with a 2-mm2 flow area (p < 0.017). The FAZ and non-flow area did not manifest any significant difference between the groups (p > 0.017). The hemoglobin concentrations in the anemia, PV, and healthy groups were 8.11 ± 1.33, 17.5 ± 1.2, and 15.6 ± 0.73 g/dL, respectively, indicating statistically significant differences between the groups (p < 0.001). CONCLUSION: In this study, quantitative OCTA analysis revealed a higher tendency for retinal and choroidal microvascular morphological changes in patients with anemia and PV. The outcomes of the current investigation can provide new insights into the retinal and choroidal pathophysiologies found in patients with hemoglobin abnormalities.

Occlusive Retinal Vasculitis in a Pediatric Patient With Kikuchi-Fujimoto Disease and Sickle Cell Trait.

A 13-year-old Black male patient with a history of Kikuchi-Fujimoto disease (KFD) and sickle cell trait presented with acute painless vision loss and no light perception vision (NLP) in his left eye. The examination was indicative of occlusive retinal vasculitis with near total central retinal artery occlusion (CRAO). He was started on oral steroids with dramatic reperfusion and improvement of the retinal hemorrhages. However, his vision remained at NLP. Oral steroids were tapered, and rituximab infusion was initiated. While ocular involvement is uncommon in KFD, vision-limiting complications, such as occlusive retinal vasculitis, ophthalmic artery occlusion, and CRAO can occur. Early systemic immunosuppression is key in achieving rapid remission. [Ophthalmic Surg Lasers Imaging Retina 2024;55:235-239.].

Fundus torsion following inferior oblique muscle weakening.

PURPOSE: To determine the postoperative fundus torsion following surgical inferior oblique (IO) weakening in patients with inferior oblique overaction (IOOA). METHODS: In this interventional case series, 37 patients with IOOA ≥ +1 and reliable preoperative fundus photographs were included. IO weakening was accomplished through myectomy or anterior transposition. Fundus photography was repeated at 3 months' follow-up. The primary outcome was change of disk foveal angle (DFA), measured as the angle between the line connecting the center of the fovea and optic disk and horizontal. Success or normal fundus torsion was defined as a postoperative angle of 0° to 8° of extorsion. Residual extorsion was defined as postoperative extorsion of > 8°. Postoperative intorsion was defined as any amount of induced intorsion. RESULTS: Mean age of patients (56% females) was 6.46 ± 6.79 years. At follow-up at least 3 months after surgery, the mean change of DFA was 7.79 ± 6.24° of reduction of extorsion after IO myectomy (P < 0.001) and 6.05 ± 6.07° after IO anterior transposition (P = 0.005). Normal fundus torsion was achieved in 50% of patients after myectomy and 36.4% of patients after anterior transposition. Residual extorsion was observed in 49% of patients; induced intorsion in 5%. CONCLUSIONS: Using objective measurement of fundus photographs, fundus extorsion was decreased in 94.7% of patients after myectomy or anterior transposition.

Longitudinal Characteristics of Choroidal Neovascular Membrane in Pediatric Patients.

PURPOSE: To report the clinical and imaging characteristics, including optical coherence tomography angiography (OCTA), and treatment outcomes of choroidal neovascular membranes (CNVMs) in children. DESIGN: Retrospective clinical cohort study. METHODS: Thirty eyes from 25 children (56% girls) with CNVM from 2 centers were examined from 2005 to 2022. Clinical features, imaging findings, treatment regimens, and outcomes are described. RESULTS: The most common causes of CNVM were idiopathic (48%) and inflammatory (20%). At diagnosis, most CNVMs were unilateral (80%), active (83.3%), and juxtafoveal (46.7%). Twenty-five eyes (83.3%) of 21 patients (84%) were treated. The most common first-line treatment was intravitreal injection of anti-vascular endothelial growth factor (anti-VEGF) (92%), with a retreatment rate of 52.2% at an average of 237 days. The average number of total injections per eye was 2.3. Injections were safely administered in the clinic (52.2%). A gain of 3 lines or 15 ETDRS (Early Treatment Diabetic Retinopathy Study) letters was observed at final visit. The average duration of follow-up was 56.46 ± 42.51 months. No ocular or systemic complication related to treatment was reported. Sixteen eyes (64%) had OCTA images at both presentation and final visit, which showed a decrease in CNVM vessel density and vessel-length density, and in the height of retinal pigment epithelium detachment (RPED). CONCLUSIONS: There are a variety of underlying etiologies for pediatric CNVMs, which are most often unilateral. Treatment with intravitreal anti-VEGF can be beneficial and does not often require frequent or chronic dosing. OCTA demonstrated a decrease in the CNVM vessel density and vessel-length density as well as in the height of RPED.

Improved predictive diagnosis of diabetic macular edema based on hybrid models: An observational study.

Diabetic Macular Edema (DME) is the most common sight-threatening complication of type 2 diabetes. Optical Coherence Tomography (OCT) is the most useful imaging technique to diagnose, follow up, and evaluate treatments for DME. However, OCT exam and devices are expensive and unavailable in all clinics in low- and middle-income countries. Our primary goal was therefore to develop an alternative method to OCT for DME diagnosis by introducing spectral information derived from spontaneous electroretinogram (ERG) signals as a single input or combined with fundus that is much more widespread. Baseline ERGs were recorded in 233 patients and transformed into scalograms and spectrograms via Wavelet and Fourier transforms, respectively. Using transfer learning, distinct Convolutional Neural Networks (CNN) were trained as classifiers for DME using OCT, scalogram, spectrogram, and eye fundus images. Input data were randomly split into training and test sets with a proportion of 80 %-20 %, respectively. The top performers for each input type were selected, OpticNet-71 for OCT, DenseNet-201 for eye fundus, and non-evoked ERG-derived scalograms, to generate a combined model by assigning different weights for each of the selected models. Model validation was performed using a dataset alien to the training phase of the models. None of the models powered by mock ERG-derived input performed well. In contrast, hybrid models showed better results, in particular, the model powered by eye fundus combined with mock ERG-derived information with a 91 % AUC and 86 % F1-score, and the model powered by OCT and mock ERG-derived scalogram images with a 93 % AUC and 89 % F1-score. These data show that the spontaneous ERG-derived input adds predictive value to the fundus- and OCT-based models to diagnose DME, except for the sensitivity of the OCT model which remains the same. The inclusion of mock ERG signals, which have recently been shown to take only 5 min to record in daylight conditions, therefore represents a potential improvement over existing OCT-based models, as well as a reliable and cost-effective alternative when combined with the fundus, especially in underserved areas, to predict DME.

Correction Method for Optical Scaling of Fundoscopy Images: Development, Validation, and First Implementation.

Purpose: Although fundus photography is extensively used in ophthalmology, refraction prevents accurate distance measurement on fundus images, as the resulting scaling differs between subjects due to varying ocular anatomy. We propose a PARaxial Optical fundus Scaling (PAROS) method to correct for this variation using commonly available clinical data. Methods: The complete optics of the eye and fundus camera were modeled using ray transfer matrix formalism to obtain fundus image magnification. The subject's ocular geometry was personalized using biometry, spherical equivalent of refraction (RSE), keratometry, and/or corneal topography data. The PAROS method was validated using 41 different eye phantoms and subsequently evaluated in 44 healthy phakic subjects (of whom 11 had phakic intraocular lenses [pIOLs]), 29 pseudophakic subjects, and 21 patients with uveal melanoma. Results: Validation of the PAROS method showed small differences between model and actual image magnification (maximum 3.3%). Relative to the average eye, large differences in fundus magnification were observed, ranging from 0.79 to 1.48. Magnification was strongly inversely related to RSE (R2 = 0.67). In phakic subjects, magnification was directly proportional to axial length (R2 = 0.34). The inverse relation was seen in pIOL (R2 = 0.79) and pseudophakic (R2 = 0.12) subjects. RSE was a strong contributor to magnification differences (1%-83%). As this effect is not considered in the commonly used Bennett-Littmann method, statistically significant differences up to 40% (mean absolute 9%) were observed compared to the PAROS method (P < 0.001). Conclusions: The significant differences in fundus image scaling observed among subjects can be accurately accounted for with the PAROS method, enabling more accurate quantitative assessment of fundus photography.