Morphology Of Retinal Vessels Research Articles

HRD-Net: High resolution segmentation network with adaptive learning ability of retinal vessel features

Retinal segmentation is a crucial step in the early warning of human health conditions. However, retinal blood vessels possess complex curvature, irregular distribution, and contain multi-scale fine structures, which make the limited receptive field of regular convolution challenging to process their vascular details efficiently. Additionally, the encoder-decoder based network leads to irreversible spatial information loss because of multiple downsampling, resulting in over-segmentation and missed segmentation of the vessels. For this reason, we develop a high-resolution network based on Deformable Convolution v3, called HRD-Net. By constructing a high-resolution representation, the network allows special attention to be paid to the details of tiny blood vessels. The proposed feature enhancement cascade module based on Deformable Convolution v3 can flexibly adapt and capture the ever-changing morphology and intricate connections of retinal blood vessels, ensuring the continuity of vessel segmentation. In the output phase of the network, the proposed global aggregation module integrates full-resolution feature maps while suppressing redundant features, achieving an effective fusion of high-level semantic information and spatial detail information. In addition, we have re-examined the selection criteria for activation and normalization methods, and also refine the network architectures from a spatial domain perspective to release redundant computational loads. Testing on the DRIVE, STARE, and CHASE_DB1 datasets indicates that HRD-Net, with fewer parameters, outperforms existing segmentation methods on several evaluation metrics such as F1, ACC, SE, SP, AUC, and IOU.

Analysis of the relationship between axial length, optic disc morphology, and regional variations in retinal vessel density in young adults with healthy eyes.

To investigate the relationship between optic disc morphology, axial length, and regional distribution of retinal vessels in healthy eyes of young adults. Nine hundred and two healthy eyes were enrolled in this university-based, cross-sectional study. Spectral-domain optical coherence tomography angiography was used to measure the parapapillary retinal vessel density. We automated the process of calculating optic disc tilt and rotation by using a program written in Python. Relationships between optic disc rotation, optic disc tilt, parapapillary vessel density, and other ocular parameters were analyzed using regression models. As axial length increased, optic disc morphology became more tilted and rotated inferiorly. The superficial vessel density (SVD) and radial peripapillary capillary density (RPCD) gradually decreased in all regions except for the temporal quadrant. Increased temporal SVD (OR [95% CI] = 1.081 [1.039, 1.124], p < 0.001), reduced nasal SVD (OR [95% CI] = 0.898 [0.861, 0.937], p < 0.01), and short relative lens position (OR [95% CI] = 0.126 [0.032, 0.495], p = 0.003) were significantly associated with the presence of a tilted optic disc. Inferior disc rotation was associated with decreased superior deep vessel density (DVD) and increased inferior DVD and temporal DVD after adjusting for sex and axial length. The tilted and rotated optic discs were associated with the distribution of SVD and DVD, respectively. We should fully consider the influence of optic disc morphology on parapapillary vessel density in eyes with myopia.

A lightweight network guided with differential matched filtering for retinal vessel segmentation

The geometric morphology of retinal vessels reflects the state of cardiovascular health, and fundus images are important reference materials for ophthalmologists. Great progress has been made in automated vessel segmentation, but few studies have focused on thin vessel breakage and false-positives in areas with lesions or low contrast. In this work, we propose a new network, differential matched filtering guided attention UNet (DMF-AU), to address these issues, incorporating a differential matched filtering layer, feature anisotropic attention, and a multiscale consistency constrained backbone to perform thin vessel segmentation. The differential matched filtering is used for the early identification of locally linear vessels, and the resulting rough vessel map guides the backbone to learn vascular details. Feature anisotropic attention reinforces the vessel features of spatial linearity at each stage of the model. Multiscale constraints reduce the loss of vessel information while pooling within large receptive fields. In tests on multiple classical datasets, the proposed model performed well compared with other algorithms on several specially designed criteria for vessel segmentation. DMF-AU is a high-performance, lightweight vessel segmentation model. The source code is at https://github.com/tyb311/DMF-AU.

CPMF-Net: Multi-Feature Network Based on Collaborative Patches for Retinal Vessel Segmentation

As an important basis of clinical diagnosis, the morphology of retinal vessels is very useful for the early diagnosis of some eye diseases. In recent years, with the rapid development of deep learning technology, automatic segmentation methods based on it have made considerable progresses in the field of retinal blood vessel segmentation. However, due to the complexity of vessel structure and the poor quality of some images, retinal vessel segmentation, especially the segmentation of Capillaries, is still a challenging task. In this work, we propose a new retinal blood vessel segmentation method, called multi-feature segmentation, based on collaborative patches. First, we design a new collaborative patch training method which effectively compensates for the pixel information loss in the patch extraction through information transmission between collaborative patches. Additionally, the collaborative patch training strategy can simultaneously have the characteristics of low occupancy, easy structure and high accuracy. Then, we design a multi-feature network to gather a variety of information features. The hierarchical network structure, together with the integration of the adaptive coordinate attention module and the gated self-attention module, enables these rich information features to be used for segmentation. Finally, we evaluate the proposed method on two public datasets, namely DRIVE and STARE, and compare the results of our method with those of other nine advanced methods. The results show that our method outperforms other existing methods.

Retinal Vessel Segmentation With Skeletal Prior and Contrastive Loss.

The morphology of retinal vessels is closely associated with many kinds of ophthalmic diseases. Although huge progress in retinal vessel segmentation has been achieved with the advancement of deep learning, some challenging issues remain. For example, vessels can be disturbed or covered by other components presented in the retina (such as optic disc or lesions). Moreover, some thin vessels are also easily missed by current methods. In addition, existing fundus image datasets are generally tiny, due to the difficulty of vessel labeling. In this work, a new network called SkelCon is proposed to deal with these problems by introducing skeletal prior and contrastive loss. A skeleton fitting module is developed to preserve the morphology of the vessels and improve the completeness and continuity of thin vessels. A contrastive loss is employed to enhance the discrimination between vessels and background. In addition, a new data augmentation method is proposed to enrich the training samples and improve the robustness of the proposed model. Extensive validations were performed on several popular datasets (DRIVE, STARE, CHASE, and HRF), recently developed datasets (UoA-DR, IOSTAR, and RC-SLO), and some challenging clinical images (from RFMiD and JSIEC39 datasets). In addition, some specially designed metrics for vessel segmentation, including connectivity, overlapping area, consistency of vessel length, revised sensitivity, specificity, and accuracy were used for quantitative evaluation. The experimental results show that, the proposed model achieves state-of-the-art performance and significantly outperforms compared methods when extracting thin vessels in the regions of lesions or optic disc. Source code is available at https://www.github.com/tyb311/SkelCon.

ResDO-UNet: A deep residual network for accurate retinal vessel segmentation from fundus images

For the clinical diagnosis, it is essential to obtain accurate morphology data of retinal blood vessels from patients, and the morphology of retinal blood vessels can well help doctors to judge the patient’s condition and give targeted therapeutic measures. Conventional manual retinal blood vessel segmentation by the doctors from the fundus images is time-consuming and laborious, while it also requires the rich doctor’s expertise. With the strong context feature expression ability of deep convolutional neural networks (DCNN), it has shown a promising performance on retinal blood vessel segmentation, specially U-shape network (U-Net) and its variant. However, due to the information loss issue caused by multiple pooling operations and insufficient process issue of local context features by skip connections, most of segmentation methods still exist a certain shortcoming on accurate fine vessel detection. To address this issue, based on the encoder–decoder framework, a novel retinal vessel segmentation network, called ResDO-UNet, is proposed to provide an automatic and end-to-end detection scheme from fundus images. To enhance feature extraction capabilities, combined with depth-wise over-parameterized convolutional layer (DO-conv), a residual DO-conv (ResDO-conv) network is proposed to act as the backbone network to acquire strong context features. In addition, to reduce the effect of information loss caused by multiple pooling operations, taking advantages of max pooling and average pooling layers, a pooling fusion block (PFB) is proposed to realize nonlinear fusion pooling. Meanwhile, faced with insufficient process of local context features by skip connections, an attention fusion block (AFB) is proposed to realize effective multi-scale feature expression. Combined with the three public available data sets on retinal vessel segmentation, including DRIVE, STARE and CHASE_DB1, the proposed segmentation network could reach a state-of-the-art detection performance compared to other related advanced work.

Retinal vessel segmentation with constrained-based nonnegative matrix factorization and 3D modified attention U-Net

Due to the complex morphology and characteristic of retinal vessels, it remains challenging for most of the existing algorithms to accurately detect them. This paper proposes a supervised retinal vessels extraction scheme using constrained-based nonnegative matrix factorization (NMF) and three dimensional (3D) modified attention U-Net architecture. The proposed method detects the retinal vessels by three major steps. First, we perform Gaussian filter and gamma correction on the green channel of retinal images to suppress background noise and adjust the contrast of images. Then, the study develops a new within-class and between-class constrained NMF algorithm to extract neighborhood feature information of every pixel and reduce feature data dimension. By using these constraints, the method can effectively gather similar features within-class and discriminate features between-class to improve feature description ability for each pixel. Next, this study formulates segmentation task as a classification problem and solves it with a more contributing 3D modified attention U-Net as a two-label classifier for reducing computational cost. This proposed network contains an upsampling to raise image resolution before encoding and revert image to its original size with a downsampling after three max-pooling layers. Besides, the attention gate (AG) set in these layers contributes to more accurate segmentation by maintaining details while suppressing noises. Finally, the experimental results on three publicly available datasets DRIVE, STARE, and HRF demonstrate better performance than most existing methods.

Establishing a mouse model of choroidal neovascularization to study the therapeutic effect of levotinib and its mechanism.

ObjectiveTo study the therapeutic effect and mechanism of levotinib on choroidal neovascularization (CNV) in mice.Methods45 healthy C57BL/6 mice were selected and randomly divided into three groups: control group (group A), model group (group B) and levotinib group (group C). The model of CNV in mice was established. The fluorescence leakage of choroidal lesions in mice was observed by fundus fluorescein angiography. The morphological changes of retinal vessels in mice were observed by retinal slice preparation, the pathological changes of eyeball tissues in mice were observed by hematoxylin-eosin (HE) staining, the expression of vascular endothelial growth factor (VEGF) in mice retina was detected by real-time quantitative fluorescence PCR, and the protein expression of VEGF in mice retina was detected by Western blotting.ResultOn the 7th, 14th and 21st day after modeling, compared with group B, the fluorescence leakage area of group C mice was significantly reduced, and the difference was statistically significant (P < 0.05). The morphology of retinal vessels in group A was normal. In group B, the retinal vessels showed large areas of ischemia without perfusion and abundant neovascularization clusters and capillaries. Compared with group B, the morphology of retinal vessels in group C was significantly improved. Group A mice had normal eyeball structure, group B mice had visible spindle-like damage to the inner and outer retina, while group C mice had significantly less spindle-like damage than group B. Compared with group A, group B mice had significantly higher expression of retinal VEGF and the difference was statistically significant (P < 0.05), but compared with group B mice, the expression of VEGF in the retina of mice in group C was significantly decreased, and the difference was statistically significant (P < 0.05). Compared with group A, the expression of VEGF in retina of group B mice was significantly increased, and the difference was statistically significant (P < 0.05). Compared with group B, the expression of VEGF in retina of group C mice was significantly decreased, and the difference was statistically significant (P < 0.05).ConclusionLevatinib has obvious therapeutic effect on CNV, which may be achieved by inhibiting the high expression of VEGF in CNV.

SUD-GAN: Deep Convolution Generative Adversarial Network Combined with Short Connection and Dense Block for Retinal Vessel Segmentation.

Since morphology of retinal blood vessels plays a key role in ophthalmological disease diagnosis, retinal vessel segmentation is an indispensable step for the screening and diagnosis of retinal diseases with fundus images. In this paper, deep convolution adversarial network combined with short connection and dense block is proposed to separate blood vessels from fundus image, named SUD-GAN. The generator adopts U-shape encode-decode structure and adds short connection block between convolution layers to prevent gradient dispersion caused by deep convolution network. The discriminator is all composed of convolution block, and dense connection structure is added to the middle part of the convolution network to strengthen the spread of features and enhance the network discrimination ability. The proposed method is evaluated on two publicly available databases, the DRIVE and STARE. The results show that the proposed method outperforms the state-of-the-art performance in sensitivity and specificity, which were 0.8340 and 0.9820, and 0.8334 and 0.9897 respectively on DRIVE and STARE, and can detect more tiny vessels and locate the edge of blood vessels more accurately.

SWEPT-SOURCE OPTICAL COHERENCE TOMOGRAPHY ANGIOGRAPHY REVEALS INTERNAL LIMITING MEMBRANE PEELING ALTERS DEEP RETINAL VASCULATURE.

To describe morphology of retinal and choroidal vessels in swept-source optical coherence tomography angiography before and after vitrectomy with the temporal inverted internal limiting membrane (ILM) flap technique for full-thickness macular holes. Prospective, observational study of 36 eyes of 33 patients with full-thickness macular holes swept-source optical coherence tomography angiography was performed in patients before and 1 month after vitrectomy. Vitrectomy with the temporal inverted ILM flap technique was performed. In this method, ILM is peeled only at one side of the fovea. An ILM flap is created to cover the macular hole. Comparison of retina vasculature in the areas of ILM peeling vs. no ILM peeling at 1 and 3 months after successful vitrectomy was performed. The study demonstrated lower density of vessels in the deep retinal plexus in the area where ILM was peeled as compared to the rest of the fovea. Visual acuity and central retinal thickness 1 month after surgery correlates with fovea avascular zone diameter in deep retinal layers at the same time point (P = 0.001). This study confirmed that ILM peeling might alter blood flow in deep retinal vessels below the peeling area in the early postoperative period. The area of the fovea avascular zone corresponds to functional results at the same time point.

Choroidal neovascularization in central serous chorioretinopathy—a new clinical entity

To describe the morphology of retinal and choroidal vessels in swept source optical coherence tomography angiography (SS-OCTA) in choroidal neovascularization (CNV) during the course of central serous chorioretinopathy (CSC). Additionally, to report intravitreal anti-VEGF (vascular endothelial growth factor) injection treatment results. A case series of patients diagnosed with chronic central serous chorioretinopathy. Fluorescein angiography, spectral domain optical coherence tomography, and SS-OCTA were performed. Analysis of retina and choroidal vasculature was performed. All patients received anti-VEGF therapy. The study demonstrated pathological vessels in the retina avascular zone and at the level of choriocapillaris during the course of choroidal neovascularization in central serous chorioretinopathy. These vessels are invisible in fluorescein angiography, as they are hidden by pooling of the dye. These vessels correspond to a slightly hyper-reflective elevation of retinal pigment epithelium visible in optical coherence tomography. Visual acuity and central retinal thickness decreased after anti-VEGF treatment. The vessels did not completely disappear despite anti-VEGF treatment. This study shows that SS-OCTA enables us to identify neovessels during the course of CSC. In diagnosing this entity, SS-OCTA appears superior to fluorescein angiography and might also enable us to decide on the best timing for anti-VEGF injections.

Development of automatic retinal vessel segmentation method in fundus images via convolutional neural networks.

The analysis of fundus photograph is one of useful diagnosis tools for diverse retinal diseases such as diabetic retinopathy and hypertensive retinopathy. Specifically, the morphology of retinal vessels in patients is used as a measure of classification in retinal diseases and the automatic processing of fundus image has been investigated widely for diagnostic efficiency. The automatic segmentation of retinal vessels is essential and needs to precede computer-aided diagnosis system. In this study, we propose the method which implements patch-based pixel-wise segmentation with convolutional neural networks (CNNs) in fundus images for automatic retinal vessel segmentation. We construct the network composed of several modules which include convolutional layers and upsampling layers. Feature maps are made by modules and concatenated into a single feature map to capture coarse and fine structures of vessel simultaneously. The concatenated feature map is followed by a convolutional layer for performing a pixel-wise prediction. The performance of the proposed method is measured on DRIVE dataset. We show that our method is comparable to the results of other state-of-the-art algorithms.

Protective Effect of Tang Wang One Decoction on the Retinal Vessels of Diabetic Rats.

Objective. This study aimed to determine the influence of Tang Wang One Decoction (TWOD) on the retinal vessels of diabetic rats. Methods. The hemorheology of diabetic rats was observed. Morphological studies of retinal vessels were conducted using optical microscopy and electron microscopy. Immunological histochemistry assay was used to measure the expression levels of MMP-9, occludin, and claudin-5. Results. Obvious pathological damage was observed in the retinal vessels of diabetic rats. TWOD positively affected the hemorheology and morphology of retinal vessels. The decoction also decreased the expression of MMP-9 and increased the expression of occludin and claudin-5. Conclusions. The results suggest that the retinal protective effects of TWOD might be related to downregulation of MMP-9 and upregulation of occludin and claudin-5.

Alleviating effects of α-melanocyte-stimulating hormone on retinal vessel leakage in diabetic rats

Background The primary pathogenic basis of diabetic retinopathy (DR) is microangiopathy and inflammatory procedure. Studies showed that α-melanocyte stimulating hormone (α-MSH) can play anti-oxidative stress and anti-apoptosis effects on retina via intravitreal injection and therefore mend blood-retinal barrier (BRB). However, experiment has not yet proved its mechanism. Objective This study was to investigate the protective effects of α-MSH on retinal vessel in diabetic rats. Methods Ninety clean SD rats were assigned to the normal control group, diabetes mellitus (DM) model group and α-MSH group. DM models were established by streptozocin (STZ) injection via tail vein in the rats of the DM model group and the α-MSH group, and sodium citrate buffer was injected in the same way in the normal control group. α-MSH of 3 μl (3.3 μg/μl) was intravitreally injected 1 week and 3 weeks after modeling in the rats of the α-MSH group, and normal saline solution was used in the same way in the normal control group; while no operation was performed in the DM model group. In the fifth week after modeling, Evans blue at the dose of 30 mg/ml (45 mg/kg) was injected via tail vein of rats, and retinal mount was prepared 2 hours later to examine the morphology of retinal vessels under the fluorescent microscope. The blood samples were collected from inner canthus vein for the assessment of leakage amount of Evans blue. The rat retinas were isolated after perfusion of sodium citrate buffer via the left ventricle for the observation of choroidal and retinal vessel ultrastructure by transmission electron microscopy (TEM). The relative expressions of intercellular cell adhesion molecule-1 (ICAM-1) mRNA, tumor necrosis factor-α (TNF-α), occluding mRNA and vascular endothelial growth factor (VEGF) mRNA were detected by quantitative real-time PCR. Results The body weight was lighter and the water intake was higher in the DM rats than those in the normal control rats. The blood glucose levels in the DM model group were (29.69±4.77) mmol/L at the third day and (24.64±2.72) mmol/L in the fifth week following the injection of STZ, respectively. An amount of Evans blue leaked from vessels in the rat retina of DM model group, but the retinal vessels were intact in the rats of the α-MSH group and the normal control group. The leakage amount was (10.04±8.18)μl/(g·h) in the DM model group, which was significantly more than that of (2.62±3.73)μl/(g·h) in the α-MSH group and (3.10±1.13)μl/(g·h) in the normal control group at the fifth week after modeling (P=0.035, 0.041). The ultrastructure of choriodal and retinal vessels was almost normal in the α-MSH group and the normal control group, but the swelling of choriodal vascular endothelial cells and vacuolar-like changes of the retinal pigment epithelium (RPE) were found in the DM model group. In addition, the relative expression levels of ICAM-1 mRNA and TNF-α mRNA were significantly elevated, but those of occludin mRNA were significantly declined in the DM model group compared with the α-MSH group and the normal control group (all at P<0.05). No significant difference was found in the relative expression level of VEGF mRNA among the three groups (F=0.791, P=0.466). Conclusions α-MSH ameliorates retinal vascular leakage in STZ-induced diabetic rats after intravitreal injection by improving ultrastructure of choroid and retinal vessels, down-regulating the expression of proinflammatory factors and up-regulating tight junction component in retina. Key words: α-Melanocyte-stimulating hormone; Diabetic retinopathy/therapy; Intercellular adhesion molecule-1; Tumor necrosis factor-α; Occludin; Vascular endothelial growth factor; Blood-retina barrier; Rats, Sprague Dawley

Morphology of retinal vessels in patients with optic nerve head drusen and optic disc edema.

We quantitatively investigated the peripapillary vascular morphology in patients with optic nerve head drusen (ONHD) and optic disc edema (ODE). Computer-based fundus analysis was used to investigate peripapillary vascular morphology, including length, branching, and diameter of arteries and veins calibrated by optical coherence tomography. Patients with ONHD showed significantly larger diameters of arteries without branching (P = 0.05), and arteries after primary/before secondary branching (P = 0.04) and secondary venous branching started closer to the optic disc (P = 0.03) compared to healthy controls. The ODE patients had significantly reduced number of small peripapillary veins and larger number of veins without branching compared to ONHD and controls (P = 0.02). Anomalous branching with arterial and venous trifurcation presented in the ODE and ONHD groups, with significant higher prevalence in ODE patients for venous trifurcations compared to ONHD and controls (P = 0.02). The diameter of vessels in ONHD patients were significantly larger in arteries without branching (P = 0.05), after primary branching (P = 0.04), and venous branching closer to the disc area (P = 0.03) compared to controls. The ODE patients demonstrated widening of the small peripapillary veins measured by a significantly larger number of veins without branching (P = 0.001 and P = 0.02, compared to controls and ONHD, respectively) and less small veins (P = 0.001 and P = 0.04, compared to controls and ONHD, respectively).

Alteration of retinal vascular development by postnatal systemic lipopolysaccharide administration: a novel model of retinopathy of prematurity (832.6)

Purpose: ROP is the leading cause of blindness in the children. Inflammation is being highlighted as the major factor for the morbidity of prematurity. We intended to investigate the effect of neonatal systemic inflammation on the retinal angiogenesis in newborn rats and to show the role of inflammation in the retinopathy of prematurity.Method: Systemic inflammation was induced by the pre‐ and post‐natal intraperitoneal injections of lipopolysaccharide (LPS) in neonatal rats. Enucleation and retinal separation was performed at 1 and 2 post‐natal weeks. Analyses on morphology of retinal vessels, recruited immune cells, change of inflammatory cytokines were assessed with immunofluorescence, FACS and RT‐PCR.Result: Administration of LPS postnatlly, not prenatally, induced inflammation in the retina with recruitment of CD11c positive cells and increased expression of a set of inflammatory cytokine genes. Postnatal LPS administration also suppressed retinal vessel development and deranged the retinal vascular structures combined with abnormal vascular tuft formation that resembles features of retinopathy of prematurity.Conclusion: Neonatal systemic inflammation alters the normal development of retinal vessels by inducing retinal inflammation. Our model of systemic inflammation‐induced abnormal development of retinal vasculature can be a novel experimental model of retinopathy of prematurity.

An approach to localize the retinal blood vessels using bit planes and centerline detection

The change in morphology, diameter, branching pattern or tortuosity of retinal blood vessels is an important indicator of various clinical disorders of the eye and the body. This paper reports an automated method for segmentation of blood vessels in retinal images. A unique combination of techniques for vessel centerlines detection and morphological bit plane slicing is presented to extract the blood vessel tree from the retinal images. The centerlines are extracted by using the first order derivative of a Gaussian filter in four orientations and then evaluation of derivative signs and average derivative values is performed. Mathematical morphology has emerged as a proficient technique for quantifying the blood vessels in the retina. The shape and orientation map of blood vessels is obtained by applying a multidirectional morphological top-hat operator with a linear structuring element followed by bit plane slicing of the vessel enhanced grayscale image. The centerlines are combined with these maps to obtain the segmented vessel tree. The methodology is tested on three publicly available databases DRIVE, STARE and MESSIDOR. The results demonstrate that the performance of the proposed algorithm is comparable with state of the art techniques in terms of accuracy, sensitivity and specificity.

Morphology of retinal vessels in the optic disk in a Göttingen minipig experimental glaucoma model

To compare the morphology of normal, healthy Gottingen minipig retinal vessels of the optic disk with experimentally induced glaucomatous optic disks in order to identify the glaucomatous excavation. Present results were compared to human glaucoma findings. Sixteen eyes from eight Göttingen minipigs were studied using fundoscopic photography and fluorescein angiography. Experimental glaucoma was then induced in the left eyes over 14 months, and changes in the optic disk vessels were assessed using fundoscopic photography and fluorescein angiography. The changes were compared with those previously reported in humans. Regarding the number of vessels, the location from where they emerge and the sectors of the optic disk that they cross, arterial and retinal vessels in Göttingen minipigs present a more asymmetric layout than in humans. The central excavation is filled by the central venous ring. Changes in the glaucomatous optic disk include arteriolar incurvation, and sometimes, nasal, and peripheral displacement of the arterioles that emerge between the ganglion cell axons of the neuroretinal ring. No angiographic changes were observed in the experimental glaucoma model. The changes in the glaucomatous optic disk of the minipig imply a predominant involvement of the arterioles. However, in humans with primary open-angle glaucoma (POAG), both the arterioles and the venules are displaced, and the central excavation is easier to distinguish, because of the absence of a central venous ring.

Retinal vascular patterns. II. Human retinal vessels studied in three dimensions.

Trypsin digestion of the retina has been advantageously employed to study the morphology of retinal vessels in flat mounts.1These two-dimensional observations, however, are inadequate for the study of the relationships of the vessels, particularly of the capillaries, to each other at various depths in the retina. It is our purpose in this communication, therefore, to report experience with preparation of trypsin-digested retinas examined in three dimensions. The retinas were first treated with trypsin as previously described.1After incubation the isolated vessels were returned to 10% formalin for 10 or more hours, and then stained either by brief immersion in methylene blue or Paragon solution or by adding a few drops of the dye to an aqueous solution in which the vessels were suspended. The vessels were examined with a dissecting microscope (magnification 2080 ×) while suspended in a water medium or after they had been embedded in