Translational Vision Science & Technology最新文献

Purpose: We aimed to analyze retinal neurovascular unit (RNVU) alterations and function via optical coherence tomography angiography (OCTA) and full-field electroretinography (ERG) in patients with ischemic stroke (IS).

Methods: OCTA was used to measure RNVU changes in 229 participants (101 with IS and 128 healthy controls). The RETeval device was used to record full-field electroretinograms (FERGs) in 40 participants (14 with IS and 26 healthy controls). Logistic regression models for IS were constructed. Receiver operating characteristic (ROS) curves were constructed to assess the predictive value of various models for IS.

Results: Patients with ipsilateral internal carotid artery stenosis (ICAS) had a greater occurrence of IS. A decrease in the vascular density (VD) of the parafovea, FD-300, and nasal optic disc; a decrease in the thickness of the retinal nerve fiber layer (RNFL) around the nasal optic disc; and an increase in the acircularity index (AI) were observed in patients with IS (P < 0.05). An increase in the AI was identified as a risk factor for IS, whereas the other factors were found to be protective factors. The IS group presented a delayed a-wave implicit time and decreased b-wave amplitudes at the scotopic point. By incorporating traditional risk factors, the degree of ipsilateral ICAS, and OCTA parameters, a high predictive value for IS was achieved (area under the curve [AUC] = 0.933).

Conclusions: Patients with IS without visible fundus lesions presented changes in the RNVU, characterized by reductions in retinal VD and RNFL thickness, alongside dysfunction of photoreceptor cells and bipolar cells. The combination of RNVU changes with traditional risk factors can enhance the prediction of IS, which provides valuable guidance for monitoring this disease.

Translational relevance: This study demonstrated that the combination of OCTA parameters, the degree of ipsilateral ICAS, and traditional risk factors could can enhance the prediction of IS. These findings provide valuable guidance for monitoring IS by assessing RNVUs.

Purpose: This study aimed to investigate the effects of extraocular muscle recession performed as part of strabismus surgery on posterior retinal and choroidal blood flow.

Methods: A single-center prospective study was conducted on patients who underwent strabismus surgery. Optical coherence tomography, optical coherence tomography angiography, and laser speckle flowgraphy of the macula were performed before surgery and at one week, one month, and four months after surgery. Preoperative and postoperative ratios were calculated, and longitudinal changes in retinal blood flow, choroidal thickness, and choroidal blood flow were analyzed. Furthermore, the changes based on the types of resected muscle were examined.

Results: In total, 254 eyes from 127 patients were included. The subfoveal choroidal thickness increased significantly at one week and one month after surgery, with no significant change at four months after surgery. The choroidal blood flow increased significantly at one week after surgery, with no significant changes at one and four months after surgery. The retinal vessel density significantly decreased at one week after surgery, with no significant changes at one and four months after surgery. Analysis of groups that had various muscles excised showed no significant changes in any measurements. Choroidal thickness and blood flow were significantly correlated at one week after surgery.

Conclusions: Strabismus surgery decreased retinal blood flow but increased choroidal thickness and blood flow in the early postoperative period. Moreover, no significant changes were observed in the long term compared to the preoperative period.

Translational relevance: Strabismus surgery affects the retina and choroid in the early postoperative period but not in the long term.

Purpose: To assess corneal resistance to enzymatic digestion after rose bengal (RB)/green light and RB/green light followed by riboflavin (RF)/ultraviolet A (UV-A) cross-linking (CXL), with or without oxygen.

Methods: Ex vivo porcine corneal buttons (n = 144) underwent CXL with RB/green or RB/green-RF/UV-A under atmospheric 21% oxygen conditions or in a nitrogen chamber with 0.1% oxygen (hypoxic conditions) to test 10- and 15-J/cm2 fluences. After CXL, corneas were digested with 0.3% collagenase A, and mean digestion times (MDTs) were recorded.

Results: For the non-irradiated control group, the MDT was 19.75 ± 1.34 hours. Under atmospheric oxygen conditions, RB/green CXL yielded MDTs of 33.69 ± 1.4 and 34.38 ± 1.31 hours with fluences of 10 and 15 J/cm2, respectively. RB/green + RF/UV-A showed MDTs of 39.56 ± 1.93 and 51.94 ± 4.2 hours for combined fluences of 10 + 10 J/cm2 and 15 + 15 J/cm2, respectively. Hypoxic RB/green MDTs were 33.88 ± 1.02 and 34.06 ± 1.57 hours, and RB/green + RF/UV-A MDTs were 39.62 ± 2.5 and 50.35 ± 1.59 hours for the same respective fluences. No significant differences were observed between the control groups and corresponding intervention groups (all P > 0.05).

Conclusions: CXL via RB/green and RB/green-RF/UV-A significantly enhanced corneal collagenase digestion resistance, irrespective of oxygen presence. These findings could help optimize infectious keratitis therapy CXL protocols.

Translational relevance: Our findings aid the understanding of the molecular mechanisms underlying the therapeutic effect of CXL and may contribute to refining accelerated PACK-CXL protocols and other CXL treatment strategies.

Optical coherence tomography angiography (OCTA) is a signal processing and scan acquisition approach that enables OCT devices to clearly identify vascular tissue down to the capillary scale. As originally proposed, OCTA included several important limitations, including small fields of view relative to allied imaging modalities and the presence of confounding artifacts. New approaches, including both hardware and software, are solving these problems and can now produce high-quality angiograms from tissue throughout the retina and choroid. Image analysis tools have also improved, enabling OCTA data to be quantified at high precision and used to diagnose disease using deep learning models. This review highlights these advances and trends in OCTA technology, focusing on work produced since 2020.

Purpose: To develop an intelligent grading model for myopic maculopathy based on a long-tail learning framework, using the improved loss function LTBSoftmax. The model addresses the long-tail distribution problem in myopic maculopathy data to provide preliminary grading, aiming to improve grading capability and efficiency.

Methods: This study includes a data set of 7529 color fundus photographs. Experienced ophthalmologists meticulously annotated the ground truth. A new intelligent grading model for myopic maculopathy was constructed using the improved loss function LTBSoftmax, which predicts lesions by locally enhancing feature extraction with ND Block. Standard grading metrics were selected to evaluate the LTBSoftmax model.

Results: The improved model demonstrated excellent performance in diagnosing four types of myopic maculopathy, achieving a κ coefficient of 88.89%. Furthermore, the model's size is 18.7 MB, which is relatively smaller compared to traditional models, indicating that the model not only achieves a high level of agreement with expert diagnoses but is also more efficient in terms of both storage and computational resources. These metrics further validate the model's well-conceived design and superiority in practical applications.

Conclusions: The intelligent grading system, using long-tailed learning strategies, effectively improves the classification of myopic maculopathy, offering a practical grading tool for clinicians, particularly in areas with limited resources.

Translational relevance: This model translates long-tail learning research into a practical grading tool for myopic maculopathy. It addresses data imbalance with the improved LTBSoftmax loss function, achieving high accuracy and efficiency. By enhancing feature extraction with ND Block, it provides reliable grading support for clinicians, especially in resource-limited settings.

Purpose: The Ocular Pain Assessment Survey (OPAS) has been used to quantify chronic ocular pain and quality of life (QOL). We aim to investigate the psychometric properties of individual OPAS items with the Rasch analysis in an Asian population of dry eye disease and neuropathic corneal pain (NCP).

Methods: Question responses were obtained from 196 patients; 138 with dry eye disease (DED) and 58 with NCP, at the Singapore National Eye Centre. Item hierarchy, item fit statistics, item separation, reliability indices, and Yen's Q3 values were calculated.

Results: Individual dimensions that quantify eye pain levels in the past 24 hours and QOL showed good discriminative ability according to their person separation index values. However, individual dimensions that measured eye pain in the past 2 weeks, non-eye pain, as well as aggravating and associated factors showed suboptimal person separation index values. Significant correlations were found between the individual item pairs of the aggravating factors dimension as well as between some of the items in the QOL and associated factors dimensions.

Conclusions: Two dimensions of the OPAS questionnaire were validated with the Rasch analysis. Based on these findings, we shorten the number of questions in some dimensions to improve the performance of the tool in similar Asian populations.

Translational relevance: Our study provides insights to improve the existing OPAS for real-world clinical applications and clinical trials.

Purpose: To describe a novel algorithm (MEDTEG) for dynamically adding new test locations to a perimetric grid-to provide a more personalized/comprehensive visual field (VF) assessment.

Methods: MEDTEG operates by finding the most informative new test location. First, Voronoi tessellation is used to construct a list of candidate locations (i.e., points that lie in between the current test locations, even when the current grid is sparse or irregular). Next, each candidate's probability mass function is computed using natural neighbor interpolation. Finally, the most informative candidate is determined by computing the expected reduction in entropy (after trial t + 1) and then multiplying this value by the area of its Voronoi cell, to estimate the overall volume of expected information gain. Optional weighting coefficients can be applied to encourage/restrict testing to particular spatial locations (e.g., to avoid the midline, target the macula, or prioritize regions exhibiting structural damage).

Results: Using a combination of mathematics, graphics, and MATLAB code, we describe the algorithm and simulate possible use cases. These include ways of providing more detailed evaluations of small scotomas ("enhanced perimetry"), more efficiently assessing patients with extensive loss ("personalized perimetry"), or maximizing VF information in patients with limited attention spans ("indeterminate duration perimetry").

Conclusions: Simulations of perimetric data indicate that MEDTEG provides a logical and flexible way of automatically adding test locations to an existing perimetric test grid, or of constructing entirely novel grids based on a handful of seed locations.

Translational relevance: MEDTEG may facilitate more informative VF assessments or allow testing in challenging populations.

Purpose: This study aimed to explore the potential causal relationship between genetically determined elevated levels of pro-inflammatory cytokines and iridocyclitis, including its subtypes: acute and subacute iridocyclitis (ASIR) and chronic iridocyclitis (CIR).

Methods: A two-sample Mendelian randomization (MR) analysis was conducted using genome-wide association study (GWAS) summary data for inflammatory cytokines (cases, n = 14,824), iridocyclitis (IR; cases, n = 7306 and controls, n = 357,814), and its subtypes (ASIR: cases, n = 6166 and controls, n = 357,814; and CIR: cases, n = 1401 and controls, n = 357,814). The inverse variance-weighted (IVW) method served as the primary analysis method. Supplementary analytic methods included MR Egger, Weighted median, Weighted mode, and Simple mode methods. Pleiotropy and heterogeneity were evaluated using the Cochran's Q test, MR Egger intercept test, MR Pleiotropy RESidual Sum and Outlier test (MR-PRESSO), Bayesian colocalization analysis, and Linkage disequilibrium score regression (LDSC) analysis.

Results: Genetically predicted high levels of eotaxin, fibroblast growth factor 23 (FGF23), TNF-related apoptosis-inducing ligand (TRAIL), and Neurotrophin-3 were associated with an increased risk of IR. On the contrary, a high level of interleukin (IL)-2 was associated with a decreased risk of IR. Meanwhile, the IR subgroup analysis demonstrated that high levels of eotaxin and TRAIL were also associated with an increased risk of ASIR. High levels of cystatin D, tumor necrosis factor receptor superfamily member 9 (TNFRSF9), and caspase 8 were associated with an increased risk of CIR. CCL20 and CDCP1 were associated with a decreased risk of CIR. Heterogeneity and pleiotropy tests demonstrated that our findings were stable and reliable.

Conclusions: Inflammatory cytokines are involved in the occurrence of IR and its subtypes. Further studies are warranted to elucidate the precise mechanisms of inflammatory cytokines in IR and its subtypes.

Translational relevance: The present study highlighted the role of inflammatory cytokines in the development of IR.

Purpose: The purpose of this study was to develop an artificial intelligence (AI)-based intraocular lens (IOLs) power calculation formula for improving the accuracy of IOLs power calculations in patients with congenital ectopia lentis (CEL).

Methods: A total of 651 eyes with CEL that underwent IOLs implantation surgery were included in this study. An AI-based ensemble formula-the Jiang Formula, was developed using a training dataset of 520 eyes and evaluated on a testing dataset of 131 eyes. A five-fold cross-validation and a two-layer ensemble learning model were constructed. The formula was then tested in a test set and compared against five current classic formulas.

Results: The cohort included young patients (mean age = 14.38 ± 13.35 years). The Jiang Formula showed the lowest prediction error (PE; = 0.08 ± 1.01 diopters [D]), absolute error (AE; = 0.77 ± 0.65 D), median absolute error (MedAE; = 0.66 D), and root mean square error (RMSE; = 1.02 D) among six formulas (P < 0.001). Moreover, 68.00% of the eyes in the test set had AE within 1.0 D in the Jiang Formula.

Conclusions: AI-integrated two-layer ensemble learning model demonstrates promising applications in IOLs power calculations for patients with CEL, not only providing higher predictive accuracy than current classic formulas but also accommodating extreme values and variations in surgical techniques.

Translational relevance: The Jiang Formula, an AI-integrated two-layer ensemble learning model, enhances IOLs power calculation accuracy in patients with CEL, ultimately improving surgical outcomes and supporting more effective, personalized treatment in this unique patient group.

Purpose: We present RetOCTNet, a deep learning tool to segment the retinal nerve fiber layer (RNFL) and total retinal thickness automatically from optical coherence tomography (OCT) scans in rats following retinal ganglion cell (RGC) injury.

Methods: We created unilateral RGC injury by ocular hypertension (OHT) or optic nerve crush (ONC), and contralateral eyes were not injured. We manually segmented the RNFL and total retina of 3.0-mm radial OCT scans (1000 A-scans per B-scan, 20 frames per B-scan) as ground truth (n = 192 scans). We used these segmentations for training (80%), testing (10%), and validation (10%) to optimize the F1 score. To determine the generalizability of RetOCTNet, we tested it on volumetric scans of a separate cohort at baseline and 4, 8, and 12 weeks post-ONC.

Results: RetOCTNet's segmentations achieved high F1 scores relative to the ground-truth segmentations created by human annotators: 0.88 (RNFL) and 0.98 (retinal thickness) for control eyes, 0.84 and 0.98 for OHT eyes, and 0.78 and 0.96 for ONC eyes, respectively. On volumetric scans 12 weeks post-ONC, RetOCTNet calculated thinning of 29.49% and 10.82% in the RNFL and retina at the optic nerve head (ONH) and thinning of 38.34% and 9.85% in the RNFL and retina superior to the ONH.

Conclusions: RetOCTNet can segment the RNFL and total retinal thickness of both radial and volume OCT scans. RetOCTNet can be applied to longitudinally monitor RNFL in rodent models of RGC injury.

Translational relevance: This tool automates RNFL and retinal thickness measurement for rat OCT scans following RGC injury, reducing analysis time and increasing the consistency between studies.