Search

Purpose: To evaluate the performance of a disease activity (DA) model developed to detect DA in participants with neovascular age-related macular degeneration (nAMD). Design: Post hoc analysis. Participants: Patient dataset from the phase III HAWK and HARRIER (H&H) studies. Methods: An artificial intelligence (AI)-based DA model was developed to generate a DA score based on measurements of OCT images and other parameters collected from H&H study participants. Disease activity assessments were classified into 3 categories based on the extent of agreement between the DA model’s scores and the H&H investigators’ decisions: agreement (“easy”), disagreement (“noisy”), and close to the decision boundary (“difficult”). Then, a panel of 10 international retina specialists (“panelists”) reviewed a sample of DA assessments of these 3 categories that contributed to the training of the final DA model. A panelists’ majority vote on the reviewed cases was used to evaluate the accuracy, sensitivity, and specificity of the DA model. Main Outcome Measures: The DA model’s performance in detecting DA compared with the DA assessments made by the investigators and panelists’ majority vote. Results: A total of 4472 OCT DA assessments were used to develop the model; of these, panelists reviewed 425, categorized as “easy” (17.2%), “noisy” (20.5%), and “difficult” (62.4%). False-positive and false negative rates of the DA model’s assessments decreased after changing the assessment in some cases reviewed by the panelists and retraining the DA model. Overall, the DA model achieved 80% accuracy. For “easy” cases, the DA model reached 96% accuracy and performed as well as the investigators (96% accuracy) and panelists (90% accuracy). For “noisy” cases, the DA model performed similarly to panelists and outperformed the investigators (84%, 86%, and 16% accuracies, respectively). The DA model also outperformed the investigators for “difficult” cases (74% and 53% accuracies, respectively) but underperformed the panelists (86% accuracy) owing to lower specificity. Subretinal and intraretinal fluids were the main clinical parameters driving the DA assessments made by the panelists. Conclusions: These results demonstrate the potential of using an AI-based DA model to optimize treatment decisions in the clinical setting and in detecting and monitoring DA in patients with nAMD. Financial Disclosure(s): Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

Artificial intelligence has transformed medical diagnostic capabilities, particularly through medical image analysis. AI algorithms perform well in detecting abnormalities with a strong performance, enabling computer-aided diagnosis by analyzing the extensive amounts of patient data. The data serve as a foundation upon which algorithms learn and make predictions. Thus, the importance of data cannot be underestimated, and clinically corresponding datasets are required. Many researchers face a lack of medical data due to limited access, privacy concerns, or the absence of available annotations. One of the most widely used diagnostic tools in ophthalmology is Optical Coherence Tomography (OCT). Addressing the data availability issue is crucial for enhancing AI applications in the field of OCT diagnostics. This review aims to provide a comprehensive analysis of all publicly accessible retinal OCT datasets. Our main objective is to compile a list of OCT datasets and their properties, which can serve as an accessible reference, facilitating data curation for medical image analysis tasks. For this review, we searched through the Zenodo repository, Mendeley Data repository, MEDLINE database, and Google Dataset search engine. We systematically evaluated all the identified datasets and found 23 open-access datasets containing OCT images, which significantly vary in terms of size, scope, and ground-truth labels. Our findings indicate the need for improvement in data-sharing practices and standardized documentation. Enhancing the availability and quality of OCT datasets will support the development of AI algorithms and ultimately improve diagnostic capabilities in ophthalmology. By providing a comprehensive list of accessible OCT datasets, this review aims to facilitate better utilization and development of AI in medical image analysis.

As society ages, the importance of age-related eye diseases such as age-related macular degeneration (AMD) increases. The intravitreal treatment of neovascular AMD (nAMD), which accounts for about 10% of all AMD cases, revolutionized the world of ophthalmology about 15 years ago. This form of therapy, which can be used to treat - in addition to nAMD - other vascular retinal diseases, has led to a significant reduction in the incidence of vision loss and blindness caused by AMD and diabetes worldwide. Current developments in ophthalmology may shape the future of AMD treatment and include advancements in home-monitoring technologies like simpler, more cost-effective optical coherence tomography (OCT) devices, application of artificial intelligence for more detailed OCT, development and approval of novel therapeutics with longer-lasting effects on the eye or new mechanisms of action as well as new application systems using nanotechnological “Sustained-Delivery Systems” or suprachoroidal treatment with viral vectors for gene therapy. Overall, these exciting ongoing developments will change the treatment landscape in AMD and pave the way to increased quality of life for our elderly patients. Moreover, home monitoring technologies will ease the burden on our healthcare systems. At the same time, novel therapeutic options that require intravitreal surgical interventions may increase it.

[This corrects the article DOI: 10.1371/journal.pone.0271156.].

AIM: To evaluate a high-resolution functional imaging device that yields quantitative data regarding macular blood flow and capillary network features in eyes with diabetic retinopathy (DR). METHODS: Prospective, cross-sectional comparative case-series in which blood flow velocities (BFVs) and non-invasive capillary perfusion maps (nCPMs) in macular vessels were measured in patients with DR and in healthy controls using the Retinal Functional Imager (RFI) device. RESULTS: A total of 27 eyes of 21 subjects were studied [9 eyes nonproliferative diabetic retinopathy (NPDR), 9 eyes proliferative diabetic retinopathy (PDR) and 9 controls]. All diabetic patients were type 2. All patients with NPDR and 5 eyes with PDR also had diabetic macular edema (DME). The NPDR group included eyes with severe (n=3) and moderate NPDR (n=6), and were symptomatic. A significant decrease in venular BFVs was observed in the macular region of PDR eyes when compared to controls (2.61±0.6 mm/s and 2.92±0.72 mm/s in PDR and controls, respectively, P=0.019) as well as PDR eyes with DME compared to NPDR eyes (2.36±0.51 mm/s and 2.94±1.09 mm/s in PDR with DME and NPDR, respectively, P=0.01). CONCLUSION: The RFI, a non-invasive imaging tool, provides high-resolution functional imaging of the retinal microvasculature and quantitative measurement of BFVs in visually impaired DR patients. The isolated diminish venular BFVs in PDR eyes compared to healthy eyes and PDR eyes with DME in comparison to NPDR eyes may indicate the possibility of more retinal vein compromise than suspected in advanced DR.

Purpose There is abundant evidence on the benefits of physical activity on cardiovascular health. However, there are only few data on the acute effects of physical exercise on the retina and choroid. Our aim was the in vivo examination of chorioretinal alterations following short intense physical activity by spectral domain optical coherence tomography (SD-OCT). Methods Twenty-one eyes of 21 healthy, young subjects (mean age 22.5 ± 4.1 years, 15 males and 6 females) were recruited. Macular scanning with a SD-OCT was performed before and following a vita maxima-type physical strain exercise on a rowing ergometer until complete fatigue. Follow-up OCT scans were performed 1, 5, 15, 30 and 60 minutes following the exercise. The OCT images were exported and analyzed using our custom-built OCTRIMA 3D software and the thickness of 7 retinal layers was calculated, along with semi-automated measurement of the choroidal thickness. One-way ANOVA analysis was performed followed by Dunnett post hoc test for the thickness change compared to baseline and the correlation between performance and thickness change has also been calculated. The level of significance was set at 0.001. Results We observed a significant thinning of the total retina 1 minute post-exercise (-7.3 ± 0.6 μm, p < 0.001) which was followed by a significant thickening by 5 and 15 minutes (+3.6 ± 0.6 μm and +4.0 ± 0.6 μm, respectively, both p Conclusions Our study implies that in young adults, intense physical exercise has an acute effect on the granular layers of the retina, resulting in thinning followed by rebound thickening before normalization. We could not identify any clear correlation with either choroidal changes or performance that might explain our observations, and hence the exact mechanism warrants further clarification. We believe that a combination of vascular and mechanic changes is behind the observed trends.

PurposeFor the training of machine learning (ML) algorithms, correctly labeled ground truth data are inevitable. In this pilot study, we assessed the performance of graders with different backgrounds in the labeling of retinal fundus image quality.MethodsColor fundus photographs were labeled using a Python-based tool using four image categories: excellent (E), good (G), adequate (A) and insufficient for grading (I). We enrolled 8 subjects (4 with and 4 without medical background, groups M and NM, respectively) to whom a tutorial was presented on image quality requirements. We randomly selected 200 images from a pool of 18,145 expert-labeled images (50/E, 50/G, 50/A, 50/I). The performance of the grading was timed and the agreement was assessed. An additional grading round was performed with 14 labels for a more objective analysis.ResultsThe median time (interquartile range) for the labeling task with 4 categories was 987.8 sec (418.6) for all graders and 872.9 sec (621.0) vs. 1019.8 sec (479.5) in the M vs. NM groups, respectively. Cohen's weighted kappa showed moderate agreement (0.564) when using four categories that increased to substantial (0.637) when using only three by merging the E and G groups. By the use of 14 labels, the weighted kappa values were 0.594 and 0.667 when assigning four or three categories, respectively.ConclusionImage grading with a Python-based tool seems to be a simple yet possibly efficient solution for the labeling of fundus images according to image quality that does not necessarily require medical background. Such grading can be subject to variability but could still effectively serve the robust identification of images with insufficient quality. This emphasizes the opportunity for the democratization of ML-applications among persons with both medical and non-medical background. However, simplicity of the grading system is key to successful categorization.

Optical coherence tomography (OCT) is a high speed, high resolution and non-invasive imaging modality that enables the capturing of the 3D structure of the retina. The fast and automatic analysis of 3D volume OCT data is crucial taking into account the increased amount of patient-specific 3D imaging data. In this work, we have developed an automatic algorithm, OCTRIMA 3D (OCT Retinal IMage Analysis 3D), that could segment OCT volume data in the macular region fast and accurately. The proposed method is implemented using the shortest-path based graph search, which detects the retinal boundaries by searching the shortest-path between two end nodes using Dijkstra's algorithm. Additional techniques, such as inter-frame flattening, inter-frame search region refinement, masking and biasing were introduced to exploit the spatial dependency between adjacent frames for the reduction of the processing time. Our segmentation algorithm was evaluated by comparing with the manual labelings and three state of the art graph-based segmentation methods. The processing time for the whole OCT volume of 496×644×51 voxels (captured by Spectralis SD-OCT) was 26.15 seconds which is at least a 2-8-fold increase in speed compared to other, similar reference algorithms used in the comparisons. The average unsigned error was about 1 pixel (∼ 4 microns), which was also lower compared to the reference algorithms. We believe that OCTRIMA 3D is a leap forward towards achieving reliable, real-time analysis of 3D OCT retinal data.

PurposeTo assess the differences in texture descriptors and optical properties of retinal tissue layers in patients with multiple sclerosis (MS) and to evaluate their usefulness in the detection of neurodegenerative changes using optical coherence tomography (OCT) image segmentation.Patients and methods38 patients with MS were examined using Stratus OCT. The raw macular OCT data were exported and processed using OCTRIMA software. The enrolled eyes were divided into two groups, based on the presence of optic neuritis (ON) in the history (MSON+ group, n = 36 and MSON- group, n = 31). Data of 29 eyes of 24 healthy subjects (H) were used as controls. A total of seven intraretinal layers were segmented and thickness as well as optical parameters such as contrast, fractal dimension, layer index and total reflectance were measured. Mixed-model ANOVA analysis was used for statistical comparisons.ResultsSignificant thinning of the retinal nerve fiber layer (RNFL), ganglion cell/inner plexiform layer complex (GCL+IPL) and ganglion cell complex (GCC, RNFL+GCL+IPL) was observed between study groups in all comparisons. Significant difference was found in contrast in the RNFL, GCL+IPL, GCC, inner nuclear layer (INL) and outer plexiform layer when comparing MSON+ to the other groups. Higher fractal dimension values were observed in GCL+IPL and INL layers when comparing H vs. MSON+ groups. A significant difference was found in layer index in the RNFL, GCL+IPL and GCC layers in all comparisons. A significant difference was observed in total reflectance in the RNFL, GCL+IPL and GCC layers between the three examination groups.ConclusionTexture and optical properties of the retinal tissue undergo pronounced changes in MS even without optic neuritis. Our results may help to further improve the diagnostic efficacy of OCT in MS and neurodegeneration.

The aim of this study was to evaluate the effect of axial length (AL) on the thickness of intraretinal layers in the macula using optical coherence tomography (OCT) image analysis.Fifty three randomly selected eyes of 53 healthy subjects were recruited for this study. The median age of the participants was 29 years (range: 6 to 67 years). AL was measured for each eye using a Lenstar LS 900 device. OCT imaging of the macula was also performed by Stratus OCT. OCTRIMA software was used to process the raw OCT scans and to determine the weighted mean thickness of 6 intraretinal layers and the total retina. Partial correlation test was performed to assess the correlation between the AL and the thickness values.Total retinal thickness showed moderate negative correlation with AL (r = -0.378, p = 0.0007), while no correlation was observed between the thickness of the retinal nerve fiber layer (RNFL), ganglion cell layer (GCC), retinal pigment epithelium (RPE) and AL. Moderate negative correlation was observed also between the thickness of the ganglion cell layer and inner plexiform layer complex (GCL+IPL), inner nuclear layer (INL), outer plexiform layer (OPL), outer nuclear layer (ONL) and AL which were more pronounced in the peripheral ring (r = -0.402, p = 0.004; r = -0.429, p = 0.002; r = -0.360, p = 0.01; r = -0.448, p = 0.001).Our results have shown that the thickness of the nuclear layers and the total retina is correlated with AL. The reason underlying this could be the lateral stretching capability of these layers; however, further research is warranted to prove this theory. Our results suggest that the effect of AL on retinal layers should be taken into account in future studies.

The purpose of this study was to evaluate the possible structural changes of the macula in patients with unilateral amblyopia using optical coherence tomography (OCT) image segmentation.38 consecutive patients (16 male; mean age 32.4±17.6 years; range 6-67 years) with unilateral amblyopia were involved in this study. OCT examinations were performed with a time-domain OCT device, and a custom-built OCT image analysis software (OCTRIMA) was used for OCT image segmentation. The axial length (AL) was measured by a LenStar LS 900 device. Macular layer thickness, AL and manifest spherical equivalent refraction (MRSE) of the amblyopic eye were compared to that of the fellow eye. We studied if the type of amblyopia (strabismus without anisometropia, anisometropia without strabismus, strabismus with anisometropia) had any influence on macular layer thickness values.There was significant difference between the amblyopic and fellow eyes in MRSE and AL in all subgroups. Comparing the amblyopic and fellow eyes, we found a statistically significant difference only in the thickness of the outer nuclear layer in the central region using linear mixed model analysis keeping AL and age under control (p = 0.032). There was no significant difference in interocular difference in the thickness of any macular layers between the subgroups with one-way between-groups ANCOVA while statistically controlling for interocular difference in AL and age.According to our results there are subtle changes in amblyopic eyes affecting the outer nuclear layer of the fovea suggesting the possible involvement of the photoreceptors. However, further studies are warranted to support this hypothesis.

To evaluate macular morphology in the eyes of patients with multiple sclerosis (MS) with or without optic neuritis (ON) in previous history.Optical coherence tomography (OCT) examination was performed in thirty-nine patients with MS and in thirty-three healthy subjects. The raw macular OCT data were processed using OCTRIMA software. The circumpapillary retinal nerve fiber layer (RNFL) thickness and the weighted mean thickness of the total retina and 6 intraretinal layers were obtained for each eye. The eyes of MS patients were divided into a group of 39 ON-affected eyes, and into a group of 34 eyes with no history of ON for the statistical analyses. Receiver operating characteristic (ROC) curves were constructed to determine which parameter can discriminate best between the non-affected group and controls.The circumpapillary RNFL thickness was significantly decreased in the non-affected eyes compared to controls group only in the temporal quadrant (p = 0.001) while it was decreased in the affected eyes of the MS patients in all quadrants compared to the non-affected eyes (p

Abstract The aim of this study was to evaluate qualitative and quantitative differences in vascular density analysis of an established and a novel alternative for post-processing on optical coherence tomography angiography (OCTA) images in healthy individuals. OCTA examinations of 38 subjects were performed. After extracting the images, two semi-manual post-processing techniques, the already established Mexican hat filtering (MHF) and an alternative, the Shanbhag thresholding (ST) were applied. We assessed Vessel Density (VD), Skeleton Density (SkD) and Vessel Diameter Index (VDI). We analyzed the results in order to establish similarities or potentially relevant differences. Regarding SkD and VD, MHF generally gave higher values than ST. Simultaneously, mean values were also predominantly higher by MHF; however, standard deviations (SD) were higher by ST (range [mean ± SD]: 0.054 ± 0.038 to 0.134 ± 0.01 and 0.134 ± 0.095 to 0.362 ± 0.028 vs 0.012 ± 0.014 to 0.087 ± 0.03 and 0.039 ± 0.047 to 0.4 ± 0.095 for SkD and VD with MHF vs SkD and VD with ST, respectively). Values of VDI were considerably higher with ST than with MHF, while standard deviation was still significantly higher with ST (range [mean ± SD]: 2.459 ± 0.144 to 2.71 ± 0.084 and 2.983 ± 0.929 to 5.19 ± 1.064 for VDI with MHF and ST, respectively). The noise level reduction of the two methods were almost identical (noise levels: 65.8% with MHT and 65.24% with ST). Using MHF, the vascular network gets more fragmented by an average of 40% compared to ST. Both methods allow the segmentation of the vascular network and the examination of vascular density parameters, but they produce largely inconsistent results. To determine if these inconsistent results are clinically meaningful, and which method is more suitable for clinical use, our results provide further evidence that detailed understanding of the image analysis method is essential for reliable decision making for patients with retinal pathology. For longitudinal monitoring, use of the same image processing method is recommended.

Abstract Our aim was to analyze the intraocular pressure (IOP) changes following different intravitreous injection (IVI) procedures with or without prefilled syringes (PFS) and to elaborate their possible causes. Clinical study and laboratory assessment. 173 eyes of 141 patients. The IOP was prospectively measured pre- and postoperatively in three groups of patients receiving IVI either with ranibizumab (RP), aflibercept PFS (AP) or aflibercept vials (AV). The AP emptying volume (EV) was assessed using 40 aflibercept PFS vials: the plunger was aligned precisely (normal volume, NV) or right below the indication line (high volume, HV) and the drug was ejected with (wP) or without forced pressure (nP). Primary outcome was post-treatment IOP with type of IVI and pre-treatment IOP as fixed factors. Secondary outcome was identification of possibly confounding factors (age, sex, pathology, presence of pseudophakia, spherical error, and number of injections) and IOP > 30 mmHg post-treatment. An IOP rise above 30 mmHg was observed in 8/38 (22%), 16/51 (31%) and 35/86 (41%) cases in the RP, AV and AP groups, respectively (p = 0.129). Pre-treatment IOP was the only predictive variable for IOP rise (p

Evidence is accumulating that cognitive function, and visual impairment may be related. In this pilot study, we investigated whether multifractal dimension and lacunarity analyses performed in sectoral regions of the retina may reveal changes in patients with cognitive impairment (CI) that may be masked in the study considering the whole retinal branching pattern. Prospective age-matched subjects (n = 69) with and with no CI and without the presence of any ophthalmic history were recruited (age > 55+ years). The Montreal Cognitive Assessment (MoCA) was used to measure CI, and full-field electroretinogram (ERG) was performed. Also, visual performance exams were conducted using the Rabin cone contrast test (CCT). Quantification of the retinal structure was performed in retinal fundus images [45o field of view (FOV), optic disk centered] with excellent quality for all individuals [19 healthy controls (HC) and 20 patients with CI] after evaluating the inclusion and exclusion criteria in all study participants recruited (n = 69). The skeletonized vasculature network that comprised the whole branching pattern observable in the full 45° FOV was obtained for each image and divided into nine equal regions (superotemporal, superior, superonasal, macular, optic disk, nasal, inferotemporal, inferior, and inferonasal). The multifractal behavior was analyzed by calculating the generalized dimension Dq (Do, D1, and D2), the lacunarity parameter (Λ), and singularity spectrum f(α) in the nine sectoral skeletonized images as well as in the skeletons that comprised the whole branching pattern observable in the full 45° FOV. The analyses were performed using the ImageJ program together with the FracLac plug-in. Independent sample t-tests or Mann Whitney U test and Pearson correlation coefficient were used to find associations between all parameters in both groups. The effect size (Cohen’s d) of the difference between both groups was also assessed. A p-value < 0.05 was considered statistically significant. Significant correlations between multifractal and Λ parameters with the MoCA and implicit time ERG-parameter were observed in the regional analysis. In contrast, no trend was found when considering the whole retinal branching pattern. Analysis of combined structural-functional parameters in sectoral regions of the retina, instead of individual retinal biomarkers, may provide a useful clinical marker of CI.

Previous studies have demonstrated that cognitive impairment (CI) is not limited to the brain but also affects the retina. In this pilot study, we investigated the correlation between the retinal vascular complexity and neurodegenerative changes in patients with CI using a low-cost multimodal approach. Quantification of the retinal structure and function were conducted for every subject (n = 69) using advanced retinal imaging, full-field electroretinogram (ERG) and visual performance exams. The retinal vascular parameters were calculated using the Singapore Institute Vessel Assessment software. The Montreal Cognitive Assessment was used to measure CI. Pearson product moment correlation was performed between variables. Of the 69 participants, 32 had CI (46%). We found significantly altered microvascular network in individuals with CI (larger venular-asymmetry factor: 0.7 ± 0.2) compared with controls (0.6 ± 0.2). The vascular fractal dimension was lower in individuals with CI (capacity, information and correlation dimensions: D0, D1, and D2 (mean ± SD): 1.57 ± 0.06; 1.56 ± 0.06; 1.55 ± 0.06; age 81 ± 6years) vs. controls (1.61 ± 0.03; 1.59 ± 0.03; 1.58 ± 0.03; age: 80 ± 7 years). Also, drusen-like regions in the peripheral retina along with pigment dispersion were noted in subjects with mild CI. Functional loss in color vision as well as smaller ERG amplitudes and larger peak times were observed in the subjects with CI. Pearson product moment correlation showed significant associations between the vascular parameters (artery-vein ratio, total length-diameter ratio, D0, D1, D2 and the implicit time (IT) of the flicker response but these associations were not significant in the partial correlations. This study illustrates that there are multimodal retinal markers that may be sensitive to CI decline, and adds to the evidence that there is a statistical trend pointing to the correlation between retinal neuronal dysfunction and microvasculature changes suggesting that retinal geometric vascular and functional parameters might be associated with physiological changes in the retina due to CI. We suspect our analysis of combined structural-functional parameters, instead of individual biomarkers, may provide a useful clinical marker of CI that could also provide increased sensitivity and specificity for the differential diagnosis of CI. However, because of our study sample was small, the full extent of clinical applicability of our approach is provocative and still to be determined.