Comparison of widefield swept-source optical coherence tomographic angiography and fluorescein fundus angiography for detection of retinal neovascularization with diabetic retinopathy.

BACKGROUND: To compare vitreous angiomosaic images (VAMIs), obtained by widefield swept-source optical coherence tomographic angiography (wfSS-OCTA) and the image of fluorescein fundus angiography (FFA) in the identification of retinal neovascularization (NV) in patients with diabetic retinopathy (DR). METHODS: In this prospective observational study, severe non-proliferative diabetic retinopathy (NPDR) or proliferative DR (PDR) patients were included. All patients underwent FFA and wfSS-OCTA. The number of NVs identified by wfSS-OCTA VAMIs using five fixations 12 × 12 mm montage scans and the resembling FFA images were compared. RESULTS: Fifty-three eyes of 29 patients were enrolled. NVs were detected in 25 eyes by using FFA, including 9 NVs of the disc (NVDs) and 72 NVs elsewhere (NVEs), and in 29 eyes by OCTA, including 11 NVDs and 90 NVEs. The detection rate of NV and NVD of OCTA was comparable to that of FFA (p > 0.05), and the level of agreement was excellent (κ = 0.850, κ = 0.754). Using FFA as the gold standard, the sensitivity for detection of NV by OCTA was 100.0%, specificity was 85.7%, the positive-predictive value was 86.2%, and the negative-predictive value was 100.0%. Compared with FFA, OCTA was superior in terms of the number of NVEs identified (p = 0.024). When we excluded images of patients treated with anti-vascular endothelial growth factor (VEGF) intravitreal therapy for < 3 months, OCTA was comparable to FFA in terms of the number of NVEs discovered (p = 0.203), with excellent agreement (intraclass correlation coefficient = 0.941). CONCLUSIONS: WfSS-OCTA is an independent non-invasive alternative to FFA for NV discovery, NVD detection, and individual NVE identification, particularly in patients with PDR who have a history of prior treatment with anti-VEGF.

MINet: Multi-scale input network for fundus microvascular segmentation.

Vessel segmentation in fundus images is a key procedure in the diagnosis of ophthalmic diseases, which can play a role in assisting doctors in diagnosis. Although current deep learning-based methods can achieve high accuracy in segmenting fundus vessel images, the results are not satisfactory in segmenting microscopic vessels that are close to the background region. The reason for this problem is that thin blood vessels contain very little information, with the convolution operation of each layer in the deep network, this part of the information will be randomly lost. To improve the segmentation ability of the small blood vessel region, a multi-input network (MINet) was proposed to segment vascular regions more accurately. We designed a multi-input fusion module (MIF) in the encoder, which is proposed to acquire multi-scale features in the encoder stage while preserving the microvessel feature information. In addition, to further aggregate multi-scale information from adjacent regions, a multi-scale atrous spatial pyramid (MASP) module is proposed. This module can enhance the extraction of vascular information without reducing the resolution loss. In order to better recover segmentation results with details, we designed a refinement module, which acts on the last layer of the network output to refine the results of the last layer of the network to get more accurate segmentation results. We use the HRF, CHASE_DB1 public dataset to validate the fundus vessel segmentation performance of the MINet model. Also, we merged these two public datasets with our collected Ultra-widefield fundus image (UWF) data as one dataset to test the generalization ability of the model. Experimental results show that MINet achieves an F1 score of 0.8324 on the microvessel segmentation task, achieving a high accuracy compared to the current mainstream models.

Multi-label classification of retinal disease via a novel vision transformer model.

Introduction: The precise identification of retinal disorders is of utmost importance in the prevention of both temporary and permanent visual impairment. Prior research has yielded encouraging results in the classification of retinal images pertaining to a specific retinal condition. In clinical practice, it is not uncommon for a single patient to present with multiple retinal disorders concurrently. Hence, the task of classifying retinal images into multiple labels remains a significant obstacle for existing methodologies, but its successful accomplishment would yield valuable insights into a diverse array of situations simultaneously. Methods: This study presents a novel vision transformer architecture called retinal ViT, which incorporates the self-attention mechanism into the field of medical image analysis. To note that this study supposed to prove that the transformer-based models can achieve competitive performance comparing with the CNN-based models, hence the convolutional modules have been eliminated from the proposed model. The suggested model concludes with a multi-label classifier that utilizes a feed-forward network architecture. This classifier consists of two layers and employs a sigmoid activation function. Results and discussion: The experimental findings provide evidence of the improved performance exhibited by the suggested model when compared to state-of-the-art approaches such as ResNet, VGG, DenseNet, and MobileNet, on the publicly available dataset ODIR-2019, and the proposed approach has outperformed the state-of-the-art algorithms in terms of Kappa, F1 score, AUC, and AVG.

Advances in the treatment of intraocular malignancies: A literature review.

Intraocular malignant tumors including primary and metastatic tumors, are mainly found in Retina and uvea, and very few cases originate from the sclera and optic nerve. Intraocular tumors can endanger the patient's vision and even life, and proper treatment is vital. There have been several traditional treatments for intraocular tumors, such as radiotherapy, chemotherapy and surgery. In recent years, new methods have been developed in clinical applications including anti-VEGF and gene therapy. This paper aims to provide a timely review about recent progress in the treatment of intraocular malignant tumor.

Quantitative Assessment Parameters of Peripapillary Regions with Branch Retinal Vein Occlusion by Using Optical Coherence Tomography Angiography.

Purpose: To investigate the baseline parameters of peripapillary regions in both eyes of patients with unilateral branch retinal vein occlusion (BRVO) using optical coherence tomography angiography (OCTA) and their association with best-corrected visual acuity (BCVA). Methods: Forty-eight unilateral BRVO patients were enrolled. The 4.5 × 4.5 mm disc angiogram was acquired in the BRVO eyes and fellow eyes using the OCTA. Radial peripapillary capillary (RPC), whole vessel density, and retinal nerve fiber layer (RNFL) thickness in different regions and optic nerve head (ONH) analysis were automatically calculated. The partition includes the whole image, peripapillary, superior hemifield, inferior hemifield, eight equally divided sectors, and nine evenly divided square areas. Results: All vessel density and capillary vessel density in the whole and peripapillary regions of BRVO eyes were significantly lower than those in fellow eyes. The RNFL peripapillary thickness in BRVO eyes was significantly higher than those in fellow eyes. In eyes with supertemporal vein occlusion, all vessel density both in the superior hemifield and in the G12 region was significantly reduced compared with that in the fellow eyes. The capillary vessel density was significantly lower in the superior hemifield, superior temporal (ST), superior nasal (SN), and temporal superior (TS) areas than in the fellow eyes. The RNFL thickness in the NI, IN, TI, and TS sectors was significantly higher than in fellow eyes (all P < 0.05). Conclusions: OCTA provided quantitative information on peripapillary vascular density and RNFL thickness changes in BRVO. Branch retinal vein occlusion not only affects the blood vessel density in the macular area but also decreases the radial peripapillary capillaries. The capillary density is mainly affected in the affected hemifield but not in the unaffected hemifield.

Leveraging non-expert crowdsourcing to segment the optic cup and disc of multicolor fundus images.

Multicolor scanning laser imaging (MCI) images have broad application potential in the diagnosis of fundus diseases such as glaucoma. However, the performance level of automatic aided diagnosis systems based on MCI images is limited by the lack of high-quality annotations of numerous images. Producing annotations for vast amounts of MCI images will be a prolonged process if we only employ experts. Therefore, we consider non-expert crowdsourcing, which is an alternative approach to produce useful annotations efficiently and low cost. In this work, we aim to explore the effectiveness of non-expert crowdsourcing on the segmentation of the optic cup (OC) and optic disc (OD), which is an upstream task for glaucoma diagnosis, using MCI images. To this end, desensitized MCI images are independently annotated by four non-expert annotators, constructing a crowdsourcing dataset. To profit from crowdsourcing, we propose a model consisting of coupled regularization network and segmentation network. The regularization network generates learnable pixel-wise confusion matrices (CMs) that reflects preferences of each annotator. During training, the CMs and segmentation network are simultaneously optimized to enable dynamic trade-offs for non-expert annotations and generate reliable predictions. Crowdsourcing learning using our method have an average Mean Intersection Over Union ( M ) of 91.34%, while the average M of model trained by expert annotations is 91.72%. In addition, comparative experiments show that in our segmentation task non-expert crowdsourcing can be on a par with the expert who annotates 90% of data. Our work suggests that crowdsourcing in the segmentation of OC and OD using MCI images has the potential to be a substitute to expert annotation, which will accelerate the construction of large datasets to facilitate the application of deep learning in clinical diagnosis using MCI images.

Potential Prognostic Indicators for Patients With Retinal Vein Occlusion.

The second most prevalent cause of retinal vascular disease is retinal vein occlusion (RVO). RVO raises intravascular pressure in the capillary and veins, triggering vessel barrier collapse and subsequent leaking of blood or plasma components into the tissue (edema). Macular edema (ME) is a major complication of RVO that results in significant visual impairment. Laser therapy, intravitreal steroid injections, and vascular endothelial growth factor (VEGF) inhibitors are the major therapeutic techniques. Different therapies reduce ME of RVO and improve visual activity. However, some people have no impact on the resolution of ME, while others have a poor visual prognosis despite full ME cure. There are many investigators who studied the relationship between indicators of various instruments with visual activity. However, a summary of those findings is currently lacking. Therefore, we will focus on the predictive factors of different studies associated with positive visual activity outcomes, which would be very useful and important to help address both treatment expectations and methods for patients with RVO.

Multicolor image classification using the multimodal information bottleneck network (MMIB-Net) for detecting diabetic retinopathy.

Multicolor (MC) imaging is an imaging modality that records confocal scanning laser ophthalmoscope (cSLO) fundus images, which can be used for the diabetic retinopathy (DR) detection. By utilizing this imaging technique, multiple modal images can be obtained in a single case. Additional symptomatic features can be obtained if these images are considered during the diagnosis of DR. However, few studies have been carried out to classify MC Images using deep learning methods, let alone using multi modal features for analysis. In this work, we propose a novel model which uses the multimodal information bottleneck network (MMIB-Net) to classify the MC Images for the detection of DR. Our model can extract the features of multiple modalities simultaneously while finding concise feature representations of each modality using the information bottleneck theory. MC Images classification can be achieved by picking up the combined representations and features of all modalities. In our experiments, it is shown that the proposed method can achieve an accurate classification of MC Images. Comparative experiments also demonstrate that the use of multimodality and information bottleneck improves the performance of MC Images classification. To the best of our knowledge, this is the first report of DR identification utilizing the multimodal information bottleneck convolutional neural network in MC Images.

Deep multispectral image registration network.

Multispectral imaging (MSI) of the ocular fundus provides a sequence of narrow-band images to show the different depths in the retina and choroid. One challenge in analyzing MSI images comes from the image-to-image spatial misalignment, which occurs because the acquisition time of eye MSI images is commonly longer than the natural time scale of the eye's saccadic movement. It is necessary to align images because ophthalmologists usually overlay two of the images to analyze specific features when analyzing MSI images. In this paper, we propose a weakly supervised MSI image registration network, called MSI-R-NET, for multispectral fundus image registration. Compared to other deep-learning-based registration methods, MSI-R-NET utilizes the blood vessel segmentation label to provide spatial correspondence. In addition, we employ a feature equilibrium module to connect the aggregating layers better, and propose a multiresolution auto-context structure to adapt the registration task. In the testing stage, given a new pair of MSI images, the trained model can predict the pixelwise spatial correspondence without labeled blood vessel information. The experimental results demonstrate that the proposed segmentation-driven registration method is highly accurate.

Evaluation of Microvascular Structure Changes after Conbercept Treatment on Macular Edema Secondary to Retinal Vein Occlusion.

AIMS: To confirm the therapeutic efficacy of conbercept for the treatment of macular edema (ME) secondary to retinal vein occlusion (RVO) by using optical coherence tomography angiography (OCTA) and to find out the differences in therapeutic efficacy between ischemic and nonischemic retinal vein occlusion (iRVO or non-iRVO) after conbercept treatment. METHODS: In this prospective, randomized, and comparative study, 60 unilateral eyes suffered from RVO combined with macular edema were included and fellow eye as controls. After an initial intravitreal injection of conbercept (IVIC), a pro re nata (PRN) strategy was adopted, and the follow-up time was 6 months. The foveal avascular zone (FAZ), vascular density of superficial capillary plexus (SCP), and vascular density of deep retinal capillary plexus (DCP), nonperfused areas (NPAs) were evaluated with OCTA on baseline and after treatment. RESULTS: The mean intravitreal injection number was 2.9 ± 0.89 times during six months in iRVO patients and 2.1 ± 0.86 times in non-iRVO patients, with statistically significant difference (p < 0.05). On baseline, central macular thickness (CMT) and FAZ were significantly thickened and enlarged compared to those of healthy fellow eyes; the vascular density of SCP and DCP were significantly decreased, and the differences were statistically significant (p < 0.05). Compared to baseline, after treatment, the best-corrected visual acuity (BCVA) was improved in either iRVO or non-iRVO (-0.601 ± 0.387, -0.241 ± 0.341 logMAR, p < 0.05). In iRVO, the improvement was more substantial than that of the non-iRVO group. FAZ in the non-iRVO group had significantly decreased compared to that in iRVO group (-0.044 ± 0.040 versus 0.014 ± 0.043 mm2, p < 0.05). CMT, the vascular density of SCP, and DCP had no significant difference. CONCLUSIONS: The changes of microvascular structure can be quantitatively evaluated by using OCTA for the patients with RVO. Conbercept had a significant effect on treatment of RVO with macular edema. A more profound effect was achieved in the iRVO group on visual improvement and FAZ reduction in the non-iRVO group after conbercept treatment.

Segmenting Diabetic Retinopathy Lesions in Multispectral Images Using Low-Dimensional Spatial-Spectral Matrix Representation.

Multispectral imaging (MSI) provides a sequence of en-face fundus spectral slices and allows for the examination of structures and signatures throughout the thickness of retina to characterize diabetic retinopathy (DR) lesions comprehensively. Manual interpretation of MSI images is commonly conducted by qualitatively analyzing both the spatial and spectral properties of multiple spectral slices. Meanwhile, there exist few computer-based algorithms that can effectively exploit the spatial and spectral information of MSI images for the diagnosis of DR. We propose a new approach that can quantify the spatial-spectral features of MSI retinal images for automatic DR lesion segmentation. It combines a generalized low-rank approximation of matrices with a supervised regularization term to generate low-dimensional spatial-spectral representations using the feature vectors in all spectral slices. Experimental results showed that the proposed approach is very effective for the segmentation of DR lesions in MSI images, which suggests it as an interesting tool for assisting ophthalmologists in diagnosing, analyzing, and managing DR lesions in MSI.

Distinct downstream signaling and the roles of VEGF and PlGF in high glucose-mediated injuries of human retinal endothelial cells in culture.

Vascular endothelial growth factor (VEGF) and placental growth factor (PlGF) plays a crucial role in breakdown of the blood-retinal barrier due to hyperpermeability in diabetic retinopathy (DR). However, the distinct signaling driven by VEGF and PlGF in the pathogenesis of DR remains unclear. In this study, we investigated VEGF- and PlGF- related signaling pathways and their roles in cultured human microvascular retinal endothelial cells (hRECs) under high glucose conditions (HG; 25 mM). Apoptotic cell death was evaluated, and FITC conjugated bovine serum albumin across monolayer hRECs served as an index of permeability. Western blots were used to assess the protein levels of VEGF and PlGF, as well as the phosphorylation of p38MAPK, STAT1 and Erk1/2. Knockdown of VEGF and PlGF was performed by using siRNA. Following HG treatment, increases of VEGF and PlGF as well as PKC activity were detected in hRECs. Increased phosphorylations of p38MAPKThr180/Thr182, STAT1Ser727, and Erk1/2Tyr202/Tyr185 as well as VEGFR1Tyr1213 and VEGFR2Tyr1175 were also detected in HG-treated hRECs. Inhibition of PKC activity by Go 6976 prevented HG-induced increases of phosphor-Erk1/2 and nitric oxide synthase (NOS1) expressions as well as hyperpermeability, whereas inhibition of p38MAPK pathway by SB203580 selectively suppressed activation of STAT1 and decreased apoptotic cell death under HG conditions. Moreover, VEGF knockdown predominantly inhibited activation of VEGFR2, and phosphorylation of p38MAPK and STAT1, as well as apoptotic cell death in HG-treated hRECs. Nevertheless, PlGF knockdown mainly suppressed phosphorylation of VEGFR1, PKC, and Erk1/2, as well as NOS1 expressions and hyperpermeability. Taken together, we provide evidence demonstrating that HG-induced elevation of PlGF is responsible for hyperpermeability mainly through increasing activation of PKC-Erk1/2-NOS axis via VEGFR1, while HG-induced elevation of VEGF is associated with induction of apoptotic cell death mainly through increasing activation of p38MAPK/STAT1 signaling via VEGFR2.

Activation of the Notch­Nox4­reactive oxygen species signaling pathway induces cell death in high glucose­treated human retinal endothelial cells.

Diabetic retinopathy (DR) occurs in almost all patients with diabetes and remains as one of the major causes of vision loss worldwide. Nevertheless, the molecular mechanisms underlying the pathogenesis of DR remain elusive. The present study aimed to investigate the role and association of Notch signaling and NADPH oxidase 4 (Nox4)­mediated oxidative stress in high glucose (HG)­treated retinal cells. Human retinal endothelial cells were cultured for various durations in RPMI­1640 medium containing 30 mM glucose (HG) or 30 mM mannitol (MN) as an osmotic control; apoptotic cell death and reactive oxygen species (ROS) levels were assessed, respectively. Alterations in the expression profiles of Nox and Notch proteins were evaluated using reverse transcription­quantitative polymerase chain reaction and western blot analysis. Knockdown of Nox4 and recombination signal­binding protein J (RBPj) was generated by transfection with specific small interfering (siRNA). Persistent activation of Notch signaling was induced via the overexpression of Notch intracellular domain (NICD). In the present study, time­dependent increases in ROS production and cell death were detected in HG­treated cells. Depletion of ROS by diphenyleneiodonium decreased HG­induced cell death, and suppressed increases in caspase 3 activity and B­cell lymphoma 2­associated X protein levels. In HG­treated cells, Nox4 expression was upregulated at the mRNA and protein levels, and inhibition of Nox4 by GKT137831 or knockdown of expression by siRNA Nox4 significantly reduced ROS levels and cell death. In the presence of HG, Notch1 expression levels were elevated, and increased NICD abundance was detected in whole cell lysates and nuclear fractions. Additionally, HG­induced cell death was decreased by treatment with γ­secretase inhibitor (GSI), but increased via the overexpression of NICD. The application of GSI or knockdown of RBPj by siRNA RBPj prevented increases in Nox4 expression within HG­treated cells. The findings of the present study demonstrated that Nox4­mediated ROS serves an important role in HG­induced retinal cell damage, in which the activation of Notch signaling may be responsible for Nox4 upregulation. Therefore, inhibition of Notch signaling or Nox4 expression may be considered as potential therapeutic targets in patients with DR.

Directional ABCA1-mediated cholesterol efflux and apoB-lipoprotein secretion in the retinal pigment epithelium.

Cholesterol-containing soft drusen and subretinal drusenoid deposits (SDDs) occur at the basolateral and apical side of the retinal pigment epithelium (RPE), respectively, in the chorioretina and are independent risk factors for late age-related macular degeneration (AMD). Cholesterol in these deposits could originate from the RPE as nascent HDL or apoB-lipoprotein. We characterized cholesterol efflux and apoB-lipoprotein secretion in RPE cells. Human RPE cells, ARPE-19, formed nascent HDL that was similar in physicochemical properties to nascent HDL formed by other cell types. In highly polarized primary human fetal RPE (phfRPE) monolayers grown in low-lipid conditions, cholesterol efflux to HDL was moderately directional to the apical side and much stronger than ABCA1-mediated efflux to apoA-I at both sides; ABCA1-mediated efflux was weak and equivalent between the two sides. Feeding phfRPE monolayers with oxidized or acetylated LDL increased intracellular levels of free and esterified cholesterol and substantially raised ABCA1-mediated cholesterol efflux at the apical side. phfRPE monolayers secreted apoB-lipoprotein preferentially to the apical side in low-lipid and oxidized LDL-feeding conditions. These findings together with evidence from human genetics and AMD pathology suggest that RPE-generated HDL may contribute lipid to SDDs.

ShenFu Preparation Protects AML12 Cells Against Palmitic Acid-Induced Injury Through Inhibition of Both JNK/Nox4 and JNK/NFκB Pathways.

BACKGROUND/AIMS: Nonalcoholic steatohepatitis includes steatosis along with liver inflammation, hepatocyte injury and fibrosis. In this study, we investigated the protective role and the potential mechanisms of a traditional Chinese medicine ShenFu (SF) preparation in an in vitro hepatic steatosis model. METHODS: In palmitic acid (PA)-induced murine hepatic AML12 cell injury, effects of SF preparation on cellular apoptosis and intracellular triglyceride (iTG) level were assessed using TUNEL and TG Colorimetric Assay. Reactive oxygen species (ROS) and mitochondrial membrane potential (MMP) levels were measured using DCF and JC-1 assay. Cytokine levels were evaluated using ELISA assay. Immunoblot was used to compare the activation level of c-Jun N terminal kinase (JNK), NADPH oxidase (Nox4), and NFκB pathways. RESULTS: Addition of SF preparation prevented PA-mediated increase of apoptosis and iTG as well as IL-8 and IL-6. In PA-treated cell, SF preparation reduced the level of Nox4 and ROS, while increasing the level of MMP and the expression of manganese superoxide dismutase (MnSOD) and catalase, indicating emendation of mitochondrial dysfunction. Nox4 inhibitor GKT137381 prevented PA-induced increase of ROS and apoptosis, while decreasing iTG slightly and not influencing the level of IL-8 and IL-6. SF preparation prevented PA-induced upregulation of phospho-JNK. JNK inhibitor SP600125 prevented PA-mediated increase of Nox4, IL-8, IL-6 and iTG. Nuclear translocation of NFκB/p65 was detected in PA-treated cells, which was prevented by SF preparation. An IκB degradation inhibitor, BAY11-7082, prevented PA-induced increase of IL-8 and IL-6 as well as iTG, whereas it only decreased ROS levels slightly and showed no influence on cellular apoptosis. CONCLUSION: SF preparation shows a beneficial role in prevention of hepatocyte injury by attenuating oxidative stress and cytokines production at least partially through inhibition of JNK/Nox4 and JNK/NFκB pathway, respectively.

Deblurring sequential ocular images from multi-spectral imaging (MSI) via mutual information.

Multi-spectral imaging (MSI) produces a sequence of spectral images to capture the inner structure of different species, which was recently introduced into ocular disease diagnosis. However, the quality of MSI images can be significantly degraded by motion blur caused by the inevitable saccades and exposure time required for maintaining a sufficiently high signal-to-noise ratio. This degradation may confuse an ophthalmologist, reduce the examination quality, or defeat various image analysis algorithms. We propose an early work specially on deblurring sequential MSI images, which is distinguished from many of the current image deblurring techniques by resolving the blur kernel simultaneously for all the images in an MSI sequence. It is accomplished by incorporating several a priori constraints including the sharpness of the latent clear image, the spatial and temporal smoothness of the blur kernel and the similarity between temporally-neighboring images in MSI sequence. Specifically, we model the similarity between MSI images with mutual information considering the different wavelengths used for capturing different images in MSI sequence. The optimization of the proposed approach is based on a multi-scale framework and stepwise optimization strategy. Experimental results from 22 MSI sequences validate that our approach outperforms several state-of-the-art techniques in natural image deblurring.

One-Year Outcome of Conbercept Therapy for Diabetic Macular Edema.

PURPOSE: To evaluate the therapeutic efficacy of conbercept for the treatment of diabetic macular edema (DME) with different baseline visual acuity. METHODS: This is a retrospective, comparative study. A total of 107 eyes of 107 patients were included. According to the levels of baseline best corrected visual acuity (BCVA) and therapeutic regimen, the DME patients were divided into four groups: conbercept treatment subgroup with worse baseline VA (less than 69 letters, Snellen equivalent, 20/50 or worse; n = 37), untreated subgroup with worse baseline VA(n = 28); conbercept treatment subgroup with better baseline VA (78 to 69 letters, Snellen equivalent, 20/32 to 20/40; n = 25), untreated subgroup (n = 17). Patients received one initial intravitreal injection followed by re-treatments based on BCVA loss or increase of central macular thickness (CMT). RESULTS: At month 12, the mean improvement of BCVA was significantly higher in both worse baseline VA group and better baseline VA group with conbercept treatment than that of corresponding untreated controls:18(15) letters vs. -4(6) letters, P < 0.001; 7(1) letters vs. -5(5) letters; P < 0.001; respectively. At month12 the mean CMT from baseline was significantly declined in both worse baseline VA group and better baseline VA group with conbercept treatment than that of respective untreated controls (-212.8 ± 11.9 vs.-44.3 ± 35.3µm,P < 0.001; -116.1 ± 88.9vs.-33.7 ± 49.8µm, P = 0.001; respectively). At the end of twelve month follow-up, the BCVA improvement and CMT declination in worse baseline VA group were more prominent than that in better baseline group (P < 0.001). The mean numbers of injections were 6.7 ± 0.9, 6.5 ± 1.1 in worse baseline VA group and better baseline VA group, respectively (P = 0.35). The two groups have no significant difference in the number of injections. CONCLUSION: Conbercept was effective in the treatment of DME at different levels of baseline BCVA. For worse baseline VA, BCVA improvement was more prominent than that of better VA subgroup.

Measuring Spectral Inconsistency of Multispectral Images for Detection and Segmentation of Retinal Degenerative Changes.

Multispectral imaging (MSI) creates a series of en-face fundus spectral sections by leveraging an extensive range of discrete monochromatic light sources and allows for an examination of the retina's early morphologic changes that are not generally visible with traditional fundus imaging modalities. An Ophthalmologist's interpretation of MSI images is commonly conducted by qualitatively analyzing the spectral consistency between degenerated areas and normal ones, which characterizes the image variation across different spectra. Unfortunately, an ophthalmologist's interpretation is practically difficult considering the fact that human perception is limited to the RGB color space, while an MSI sequence contains typically more than ten spectra. In this paper, we propose a method for measuring the spectral inconsistency of MSI images without supervision, which yields quantitative information indicating the pathological property of the tissue. Specifically, we define mathematically the spectral consistency as an existence of a pixel-specific latent feature vector and a spectrum-specific projection matrix, which can be used to reconstruct the representative features of pixels. The spectral inconsistency is then measured using the number of latent feature vectors required to reconstruct the representative features in practice. Experimental results from 54 MSI sequences show that our spectral inconsistency measurement is potentially invaluable for MSI-based ocular disease diagnosis.

Joint alignment of multispectral images via semidefinite programming.

In this paper, we introduce a novel feature-point-matching based framework for achieving an optimized joint-alignment of sequential images from multispectral imaging (MSI). It solves a low-rank and semidefinite matrix that stores all pairwise-image feature-mappings by minimizing the total amount of point-to-point matching cost via a convex optimization of a semidefinite programming formulation. This unique strategy takes a complete consideration of the information aggregated by all point-matching costs and enables the entire set of pairwise-image feature-mappings to be solved simultaneously and near-optimally. Our framework is capable of running in an automatic or interactive fashion, offering an effective tool for eliminating spatial misalignments introduced into sequential MSI images during the imaging process. Our experimental results obtained from a database of 28 sequences of MSI images of human eye demonstrate the superior performances of our approach to the state-of-the-art techniques. Our framework is potentially invaluable in a large variety of practical applications of MSI images.

Groupwise registration of sequential images from multispectral imaging (MSI) of the retina and choroid.

Multispectral Imaging (MSI) produces a sequence of discrete spectral slices that penetrate different light-absorbing species or chromophores and is a noninvasive technology useful for the early detection of various retinal, optic nerve and choroidal diseases. However, eye movement during the image acquisition process may introduce spatial misalignment between MSI images. This potentially causes trouble in the manual/automatic interpretation of MSI, but still remains an unresolved problem to this date. To deal with this MSI misalignment problem, we present a method on the groupwise registration of sequential images from MSI of the retina and choroid. The advantage of our algorithm is at least threefold: 1) simultaneous estimation of landmark correspondences and a parametric motion model via quadratic programming, 2) enforcement of temporal smoothness on the estimated motion, and 3) inclusion of a robust matching cost function. As validated in our experiments with a database of 22 MSI sequences, our algorithm outperforms two state-of-the-art registration techniques proposed originally in other domains. Our algorithm is potentially invaluable in ophthalmologists' clinical practice regarding various eye diseases.