一种基于视网膜主血管方向的视盘定位及提取方法

为分割出眼底图像中的视盘,构建基于眼底图像的计算机辅助诊断系统,提出了一种基于视网膜主血管方向的视盘定位及提取方法。首先,利用Otsu阈值分割眼底图像R通道获取视盘候选区域;然后利用彩色眼底图像的HSV空间的H通道提取视网膜主血管并确定主血管方向;在此基础上,通过在方向图内寻找出对加权匹配滤波器响应值最高的点确定视盘中心位置;最后,利用该位置信息从视盘候选区域中"挑选"出真正的视盘。利用该方法对100幅不同颜色、不同亮度的眼底图像进行视盘分割,得到准确率98%,平均每幅图像处理时间1.3 s。结果表明:该方法稳定可靠,能快速、有效分割出眼底图像中的视盘。     

Glaucoma-Induced Optic Disc Morphometric Changes and Glaucoma Diagnostic Ability of Heidelberg Retina Tomograph II in Highly Myopic Eyes

Purpose

This study was performed to first investigate the morphological differences in the optic nerve head between highly myopic non-glaucomatous controls and highly myopic glaucomatous eyes in comparison with the differences between emmetropic non-glaucomatous controls and emmetropic glaucomatous eyes using confocal scanning laser ophthalmoscopy. Further, the ability of the apparatus in glaucoma diagnosis in highly myopic eyes was compared with that in emmetropic eyes.

Methods

Healthy subjects and age-matched patients with early-stage open-angle glaucoma were divided into two groups: emmetropic eyes (−1.0 to +1.0 diopters) and highly myopic eyes (−12.0 to −5.0 diopters).The participants were comprised of 65 emmetropic normal eyes, 59 emmetropic glaucomatous eyes, 62 highly myopic normal eyes, and 68 highly myopic glaucomatous eyes and eyes with pathologic myopia were carefully excluded. Confocal scanning laser tomographic parameters were compared among all subjects after adjustment for age and disc area. The ROC curves and sensitivity and specificity for glaucoma detection using several clinical methods were then compared between the emmetropic and highly myopic eyes.

Results

Rim area, cup/disc area ratio, mean cup depth, and cup shape measure of glaucoma eyes are significantly different from those of normal eyes in both highly myopic eyes and emmetropic eyes. Methodological overestimation of retinal nerve fiber layer cross sectional area due to optic disc tilting was suggested in the highly myopic eyes. The diagnostic performance of glaucoma using several discriminant methods significantly deteriorated in the highly myopic eyes.

Conclusions

In the highly myopic glaucomatous eyes, confocal scanning laser tomographic parameters were significantly different from that of non-glaucomatous highly myopic eyes but diagnostic performance of glaucoma was deteriorated than that in emmetropic eyes. These findings demonstrate the utility and limitations of the apparatus in diagnosing glaucoma in highly myopic patients.     

Optic Disc Change during Childhood Myopic Shift: Comparison between Eyes with an Enlarged Cup-To-Disc Ratio and Childhood Glaucoma Compared to Normal Myopic Eyes

BackgroundProgressive disc tilting and the development or enlargement of peripapillary atrophy (PPA) are observed during a myopic shift in children. This could be related to the changes around the optic nerve head during eyeball elongation. If the biomechanical properties at or around the optic nerve head are changed after exposure to elevated intraocular pressure (IOP) in glaucoma eyes, different response of the disc tilting and PPA changes could take place during eyeball elongation by myopic shift. On the basis of this background, the aim of this study was to compare the morphological changes in the optic disc induced by a myopic shift during childhood between normal control eyes, eyes from disc suspects with an enlarged cup-to-disc ratio (CDR), and eyes with childhood glaucoma.MethodsTotal of 82 eyes from 82 subjects younger than 14 years of age were included in the study. Serial disc photographs were classified into one of two groups: eyes with an optic nerve head (ONH) or peripapillary atrophy (PPA) change or without an ONH/PPA change. Using ImageJ software, the outlines of the optic disc and PPA were plotted, and the vertical disc diameter (VDD), horizontal disc diameter (HDD), and maximum PPA width (PPW) were measured. The changes in the ratios of these parameters and the relationships between the degree of myopic shift or the ONH/PPA change were analyzed.ResultsTwenty-five eyes with normal optic disc appearance, 36 eyes with enlarged cup-to-disc ratio, and 21 eyes of glaucoma patients were analyzed. The initial intraocular pressure (IOP) at diagnosis was significantly different among the groups (P<0.001). The degree of myopic shift during follow-up period was not significantly different among the groups (P=0.612). However, the changes in the HDD/VDD and PPW/VDD ratios were significantly greater in the disc suspect group and significantly smaller in the glaucoma group. Among the 42 eyes with an ONH/PPA change, 16 (38.1%) were from the normal control group, 24 (57.1%) were from the disc suspect group, and 2 (4.8%) were from the glaucoma group (P < 0.001).

Conclusions and Relevance

The optic disc change during childhood myopic shift was different in eyes with various conditions. Eyes of childhood glaucoma showed less change in the disc morphology during myopic shift compared to eyes with normal disc or enlarged cup-to-disc ratio.     

Stereoscopic Analysis of Optic Nerve Head Parameters in Primary Open Angle Glaucoma: The Glaucoma Stereo Analysis Study

PurposeThe Glaucoma Stereo Analysis Study (GSAS), a cross sectional multicenter collaborative study, used a stereo fundus camera to assess various morphological parameters of the optic nerve head (ONH) in glaucoma patients and investigated the relationships between these parameters and patient characteristics.ResultsPatient characteristics included refractive error of −3.38±3.75 diopters, intraocular pressure (IOP) of 13.6±2.6 mmHg, and visual field mean deviation (MD) of −4.71±3.26 dB. Representative ONH parameters included a horizontal disc width of 1.66±0.28 mm, vertical disc width of 1.86±0.23 mm, disc area of 2.42±0.63 mm², cup area of 1.45±0.57 mm², and cup volume of 0.31±0.22 mm³. Correlation analysis revealed significant negative associations between vertical cup-to-disc ratio (0.82±0.08) and MD (r = −0.40, P<0.01) and between disc tilt angle (10.5±12.5 degrees) and refractive error (r = −0.36, P<0.01). Seventy-five percent of the eyes had a positive value for rim decentering (0.30±0.42), indicating that rim thinning manifested more often as an inferior lesion than a superior lesion.ConclusionWe used stereoscopic analysis to establish a database of ONH parameters, which may facilitate future studies of glaucomatous changes in ONH morphology.     

Peripapillary Intrachoroidal Cavitations. The Beijing Eye Study

Purpose

To assess prevalence, size and location of peripapillary intrachoroidal cavitations (PICCs) and their associations in a population-based sample. .

Methods

The population-based Beijing Eye Study 2011 included 3468 individuals with a mean age of 64.6±9.8 years (range:50-93 years). A detailed ophthalmic examination included enhanced depth imaging of the choroid by spectral-domain optical coherence tomography and fundus photography. PICCs were defined as triangular thickening of the choroid with the base at the optic disc border and a distance between Bruch´s membrane and sclera of ≥200μm. Parapapillary large choroidal vessels were excluded.

Results

Out of 94 subjects with high myopia (refractive error <-6.0 diopters or axial length >26.5mm in right eyes), OCT images were available for 89 (94.7%) participants. A PICC was detected in 15 out of these 89 highly myopic subjects (prevalence:16.9±4.0%) and in none of hyperopic, emmetropic or medium myopic subgroups each consisting of 100 randomly selected subjects. Mean PICC width was 4.2±2.3 hours (30°) of disc circumference and mean length was 1363±384μm. PICCs were located most frequently (40%) at the inferior disc border. On fundus photos, a typical yellow-orange lesion was found in 8 (53%) eyes with PICCs. In binary regression analysis, presence of PICCs was significantly associated with optic disc tilting (P=0.04) and presence of posterior staphylomata (P=0.046).

Conclusions

Prevalence of PICCs in the adult Chinese population was 16.9±4.0% in the highly myopic group, with no PICCs detected in non-highly myopic eyes. PICCs were located most frequently at the inferior optic disc border. Only half of the PICCs detected on OCT images showed a yellow-orange lesion on fundus photos. Presence of PICC was significantly associated only with an increased optic disc tilting and presence of posterior staphylomata, while it was not associated with axial length, refractive error or other ocular or systemic parameters.     

兔先天性青光眼的临床研究

目的 总结兔先天性青光眼的临床特点。方法 对先天性青光眼兔和正常兔进行临床观察,研究其在眼压、眼球结构、视功能方面的变化。结果 青光眼兔的眼压明显升高,角膜直径变大,前房变深,眼轴变长,房角变宽,眼底视乳头损害明显,视觉诱发电位明显异常。结论 兔先天性青光眼的房水排泄障碍部位可能在小梁,兔眼球壁对高眼压的耐受力弱,在高眼压下容易出现眼球扩张,视功能损害。     

A Novel Image Recuperation Approach for Diagnosing and Ranking Retinopathy Disease Level Using Diabetic Fundus Image

Retinal fundus images are widely used in diagnosing and providing treatment for several eye diseases. Prior works using retinal fundus images detected the presence of exudation with the aid of publicly available dataset using extensive segmentation process. Though it was proved to be computationally efficient, it failed to create a diabetic retinopathy feature selection system for transparently diagnosing the disease state. Also the diagnosis of diseases did not employ machine learning methods to categorize candidate fundus images into true positive and true negative ratio. Several candidate fundus images did not include more detailed feature selection technique for diabetic retinopathy. To apply machine learning methods and classify the candidate fundus images on the basis of sliding window a method called, Diabetic Fundus Image Recuperation (DFIR) is designed in this paper. The initial phase of DFIR method select the feature of optic cup in digital retinal fundus images based on Sliding Window Approach. With this, the disease state for diabetic retinopathy is assessed. The feature selection in DFIR method uses collection of sliding windows to obtain the features based on the histogram value. The histogram based feature selection with the aid of Group Sparsity Non-overlapping function provides more detailed information of features. Using Support Vector Model in the second phase, the DFIR method based on Spiral Basis Function effectively ranks the diabetic retinopathy diseases. The ranking of disease level for each candidate set provides a much promising result for developing practically automated diabetic retinopathy diagnosis system. Experimental work on digital fundus images using the DFIR method performs research on the factors such as sensitivity, specificity rate, ranking efficiency and feature selection time.     

Comparison of Optic Disc Morphology of Optic Nerve Atrophy between Compressive Optic Neuropathy and Glaucomatous Optic Neuropathy

Objectives

To compare the optic nerve head (ONH) structure between compressive optic neuropathy (CON) and glaucomatous optic neuropathy (GON), and to determine whether selected ONH quantitative parameters effectively discriminate between GON and CON, especially CON cases presenting with a glaucoma-like disc.

Methods

We prospectively assessed 34 patients with CON, 34 age-matched patients with moderate or severe GON, and 34 age-matched healthy control subjects. The quantitative parameters of ONH structure were compared using the Heidelberg Retina Tomograph 2 (HRT2) and Spectralis optical coherence tomography with an enhanced depth imaging method.

Results

The mean and maximum cup depths of CON were significantly smaller than those with GON (P<0.001 and P<0.001, respectively). The distance between Bruch''s membrane opening and anterior surface of the lamina cribrosa (BMO-anterior LC) of CON was also significantly smaller than that of glaucoma but was similar to that of the healthy group (P<0.001 and P = 0.47, respectively). Based on Moorfields regression analysis of the glaucoma classification of HRT2, 15 eyes with CON were classified with a glaucoma-like disc. The cup/disc area ratio did not differ between cases of CON with a glaucoma-like disc and cases of GON (P = 0.16), but the BMO-anterior LC and mean and maximum cup depths of CON cases with a glaucoma-like disc were smaller than those in GON (P = 0.005, P = 0.003, and P = 0.001, respectively).

Conclusions

Measurements of the cup depths and the LC depth had good ability to differentiate between CON with a glaucoma-like disc and glaucoma. There was no laminar remodeling detected by laminar surface position in the patients with CON compared to those with GON.     

High-resolution imaging of the retinal nerve fiber layer in normal eyes using adaptive optics scanning laser ophthalmoscopy

Purpose

To conduct high-resolution imaging of the retinal nerve fiber layer (RNFL) in normal eyes using adaptive optics scanning laser ophthalmoscopy (AO-SLO).

Methods

AO-SLO images were obtained in 20 normal eyes at multiple locations in the posterior polar area and a circular path with a 3–4-mm diameter around the optic disc. For each eye, images focused on the RNFL were recorded and a montage of AO-SLO images was created.

Results

AO-SLO images for all eyes showed many hyperreflective bundles in the RNFL. Hyperreflective bundles above or below the fovea were seen in an arch from the temporal periphery on either side of a horizontal dividing line to the optic disc. The dark lines among the hyperreflective bundles were narrower around the optic disc compared with those in the temporal raphe. The hyperreflective bundles corresponded with the direction of the striations on SLO red-free images. The resolution and contrast of the bundles were much higher in AO-SLO images than in red-free fundus photography or SLO red-free images. The mean hyperreflective bundle width around the optic disc had a double-humped shape; the bundles at the temporal and nasal sides of the optic disc were narrower than those above and below the optic disc (P<0.001). RNFL thickness obtained by optical coherence tomography correlated with the hyperreflective bundle widths on AO-SLO (P<0.001)

Conclusions

AO-SLO revealed hyperreflective bundles and dark lines in the RNFL, believed to be retinal nerve fiber bundles and Müller cell septa. The widths of the nerve fiber bundles appear to be proportional to the RNFL thickness at equivalent distances from the optic disc.     

Glaucomatous-Type Optic Discs in High Myopia

Purpose

To assess the prevalence of glaucoma in patients with high myopia defined as myopic refractive error of >-8 diopters or axial length ≥26.5 mm.

Methods

The hospital-based observational study included 172 patients (336 eyes) with a mean age of 61.9±12.3 years and mean axial length of 30.1±2.3 mm (range: 24.7–39.1mm). Glaucomatous-type optic discs were defined by glaucomatous optic disc appearance. Glaucoma was defined by glaucomatous optic disc appearance and glaucomatous Goldmann visual field defects not corresponding with myopic macular changes.

Results

Larger disc area (mean: 3.18±1.94 mm²) was associated with longer axial length (P<0.001; standardized correlation coefficient: 0.45). Glaucoma was detected in 94 (28%; 95% Confidence intervals: 23%, 33%) eyes. In multivariate analysis, glaucoma prevalence was 3.2 times higher (P<0.001) in megalodiscs (>3.79 mm²) than in normal-sized discs or small discs (<1.51 mm²) after adjusting for older age. Axial length was not significantly (P = 0.38) associated with glaucoma prevalence in that model. Glaucoma prevalence increased by a factor of 1.39 for each increase in optic disc area by one mm². Again, axial length was not significantly (P = 0.38) associated with glaucoma prevalence when added to this multivariate model.

Conclusion

Within highly myopic individuals, glaucoma prevalence increased with larger optic disc size beyond a disc area of 3.8 mm². Highly myopic megalodiscs as compared to normal sized discs or small discs had a 3.2 times higher risk for glaucomatous optic nerve neuropathy. The increased glaucoma prevalence in axial high myopia was primarily associated with axial myopia associated disc enlargement and not with axial elongation itself.     

Automatic non-proliferative diabetic retinopathy screening system based on color fundus image

Background

Non-proliferative diabetic retinopathy is the early stage of diabetic retinopathy. Automatic detection of non-proliferative diabetic retinopathy is significant for clinical diagnosis, early screening and course progression of patients.

Methods

This paper introduces the design and implementation of an automatic system for screening non-proliferative diabetic retinopathy based on color fundus images. Firstly, the fundus structures, including blood vessels, optic disc and macula, are extracted and located, respectively. In particular, a new optic disc localization method using parabolic fitting is proposed based on the physiological structure characteristics of optic disc and blood vessels. Then, early lesions, such as microaneurysms, hemorrhages and hard exudates, are detected based on their respective characteristics. An equivalent optical model simulating human eyes is designed based on the anatomical structure of retina. Main structures and early lesions are reconstructed in the 3D space for better visualization. Finally, the severity of each image is evaluated based on the international criteria of diabetic retinopathy.

Results

The system has been tested on public databases and images from hospitals. Experimental results demonstrate that the proposed system achieves high accuracy for main structures and early lesions detection. The results of severity classification for non-proliferative diabetic retinopathy are also accurate and suitable.

Conclusions

Our system can assist ophthalmologists for clinical diagnosis, automatic screening and course progression of patients.

     

一种新的可用于干细胞移植的新西兰白兔青光眼视神经损伤模型

青光眼是一种致盲且不可逆神经退行性疾病,目前还没有能够彻底解决其视野缺失的治疗方法。神经干细胞研究的兴起为该病的治疗带来了可能,且迫切需要一种适用于干细胞移植研究的青光眼造模方法。该实验对新西兰白兔球结膜下注射地塞米松给药,2.5mg/次,一周3次,持续8周。眼底照相显示造模眼眼球屈光间质保持清晰,视乳头凹陷明显扩大、血管呈屈膝状;病理切片显示造模眼视神经明显损伤;海德堡视网膜断层扫描仪(Heidelberg retina tomography,HRT)定量分析显示造模眼盘沿面积减小(1.10±0.88)mm2、杯/盘比增大0.17±0.13,视网膜神经纤维层平均厚度降低(0.44±0.31)mm等青光眼性质病理改变,且均达到极显著水平(P<0.001)。建模成功后,于角膜缘后4mm向造模眼玻璃体注入猕猴神经干细胞,植入5个月后摘取眼球切片,通过荧光显微镜观察到存活的移植细胞。该实验建立了一种简单可靠、重复性强且适用于神经干细胞移植研究的慢性青光眼造模方法。     

Retinal Nerve Fiber and Optic Disc Morphology in Patients with Human Immunodeficiency Virus Using the Heidelberg Retina Tomography 3

Purpose

To use novel confocal scanning ophthalmoscopy technology to test hypothesis that HIV-seropositive patients without history of retinitis with a history of a low CD4 count are more likely to have damage to their retinal nerve fiber layer (RNFL) when compared to patients with high CD4 count. In addition, we compared optic disc morphologic changes with glaucoma.

Design

Cross-sectional study.

Participants and Controls

171 patients were divided into four groups. The control group consisted of 40 eyes of 20 HIV-seronegative patients. The second group consisted of 80 eyes of 41 HIV-positive patients whose CD4 cell count never dropped below 100 (1.0 x 10⁹/L). The third group consisted of 44 eyes of 26 HIV-positive patients with a history of low CD4 counts <100. Fourth group consisted of 79 eyes of 79 patients with confirmed glaucoma who served as positive controls.

Testing

Confocal scanning laser ophthalmoscopy was performed with the Heidelberg Retina Tomograph (HRT3) and data were analyzed with HRT3, software (Heyex version 1.5.10.0).

Main Outcome Measures

Disc area, cup area, cup volume, rim volume, mean cup depth, maximum cup depth, cup-to-disc ration, mean RNFL thickness, and RNFL cross-sectional area.

Results

Analysis of the global optic nerve and cup parameters showed no difference in disk area among the four groups. There was also no difference in cup, rim volume, mean cup depth, or maximum cup depth among the first three groups but they were all different from glaucoma group. The RNFL was thinner in glaucoma and both HIV-positive groups compared to HIV-seronegative subjects. The cross sectional RNFL area was thinner in both high and low CD4 HIV-positive groups compared to HIV-seronegative group in the nasal and temporal/inferior sectors, respectively. Glaucoma group showed thinning in all sectors.

Conclusions

HIV retinopathy results in retinal nerve fiber layer loss without structural optic nerve supportive tissue change. RNFL damage may occur early in HIV disease by mechanism different than in glaucoma.     

The formation and distribution of intercellular junctions in the rhesus monkey optic cup: The early development of the cilio-iridic and sensory retinas

The eyes of prenatal monkeys from 30 to 102 ± 2 days old were examined by light microscopy, conventional electron microscopy, and the freeze-fracturing technique. At 30 days, invagination of the optic vesicle has begun, and the inner and outer walls of the forming optic cup are closely apposed anteriorly; invagination is complete at 45 days. By 58 days, the rudiment of the ciliary body and iris has appeared; at 71 days, primitive ciliary processes are present and retinal photoreceptors begin to differentiate. The distribution of intercellular junctions varies both in different regions of the optic cup and at different stages of development. At 30 days, adjacent ventricular and adjacent pigmented cells are joined throughout the optic cup by zonulae adhaerentes and gap junctions. The anterior region of the cup, however, contains two additional junctional specializations: (1) fasciae occludentes between ventricular cells and (2) intermediate and gap junctions between the apposing luminal surfaces of ventricular and pigmented cells. By 36 days the fasciae occludentes between ventricular cells in the anterior optic cup become zonular, signaling the morphological development of the blood-aqueous barrier. In the posterior optic cup, zonulae occludentes appear between adjacent pigmented cells at 36 days; furthermore, with the continuing obliteration of the optic ventricle, luminal junctions spread toward the optic stalk but do not reach the optic disc until 45 days, when invagination is complete. Between 58 and 102 days there are no further changes in the distribution of the junctions anteriorly between the primitive cilio-iridial epithelial cells, whereas in the posterior optic cup the luminal gap and intermediate junctions between pigmented cells and differentiating photoreceptors decrease in number and finally disappear. Two main conclusions can be drawn from this study. (1) In the optic cup, intermediate junctions are consistently present in regions of the plasma membrane which later contain junctional complexes. The temporal and spatial pattern of junctional development suggests that intermediate junctions are necessary for the establishment of tight and gap junctions. (2) Twenty days before the ciliary body-iris anlage becomes visible in the light microscope, the distribution of junctions in the anterior part of the optic cup is identical to that in the adult cilio-iridic retina. The time-honored view that the cilio-iridic retina appears late in development is, therefore, no longer tenable. In the monkey, the optic cup is divided into a cilio-iridic and a sensory region soon after the onset of invagination.     

Acquired pits of the optic nerve in glaucoma

An acquired pit of the optic nerve (APON) is a discrete, focal area of depression within the optic cup at the level of the lamina cribrosa. It is an under-diagnosed sign of glaucoma damage due to its subtle appearance. APONs occur more frequently unilaterally and in patients with normal-tension glaucoma (NTG). They often correspond to a deep, sharp-margined scotoma approaching or involving fixation. Given the location and progressive nature of the associated visual field defects, glaucoma patients and glaucoma suspects should be evaluated for this sign of localized optic nerve damage.     

THE OCULAR FUNDUS OF TWO CETACEANS

We provided a water-cornea interface to correct the usual aerial myopia encountered in fundus photography of marine mammal eyes. The 12 Tursiops eyes were consistent for vascular structure, optic papillae, dimensional components and tapetal coloration. Multiple photographs were assembled to produce one ocular fundus typical of Tursiops truncatus and one for Grampus griseus. The eyes have a vestigial hyaloid vessel and an optic disc („blind spot”) that occupies the center of the fundus. The disc is bounded in both species by a vascular structure that is continuous with a (retrobulbar) perineural organ called the ophthalmic rete. The vascular trees of the fundus do not suggest an area of specialization for high resolution. Both species exhibit total tapetalization. Tapetal spectral reflectance did not vary between Tursiops. There was, however, a difference between species in the short wavelength regions of the spectrum.     

Proper patterning of the optic fissure requires the sequential activity of BMP7 and SHH

The optic disc develops at the interface between optic stalk and retina, and enables both the exit of visual fibres and the entrance of mesenchymal cells that will form the hyaloid artery. In spite of the importance of the optic disc for eye function, little is known about the mechanisms that control its development. Here, we show that in mouse embryos, retinal fissure precursors can be recognised by the expression of netrin 1 and the overlapping distribution of both optic stalk (Pax2, Vax1) and ventral neural retina markers (Vax2, Raldh3). We also show that in the absence of Bmp7, fissure formation is not initiated. This absence is associated with a reduced cell proliferation and apoptosis in the proximoventral quadrant of the optic cup, lack of the hyaloid artery, optic nerve aplasia, and intra-retinal misrouting of RGC axons. BMP7 addition to organotypic cultures of optic vesicles from Bmp7-/- embryos rescues Pax2 expression in the ventral region, while follistatin, a BMP7 antagonist, prevents it in early, but not in late, optic vesicle cultures from wild-type embryos. The presence of Pax2-positive cells in late optic cup is instead abolished by interfering with Shh signalling. Furthermore, SHH addition re-establishes Pax2 expression in late optic cups derived from ocular retardation (or) embryos, where optic disc development is impaired owing to the near absence of SHH-producing RGC. Collectively, these data indicate that BMP7 is required for retinal fissure formation and that its activity is needed, before SHH signalling, for the generation of PAX2-positive cells at the optic disc.     

Neuroretinal rim area and body mass index

Purpose

To examine associations between neuroretinal rim area, pressure related factors and anthropometric parameters in a population-based setting.

Methods

The population-based cross-sectional Beijing Eye Study 2006 included 3251 subjects with an age of 45+ years. The participants underwent a detailed ophthalmic examination. Exclusion criteria for our study were high myopia of more than -8 diopters and angle-closure glaucoma.

Results

The study included 2917 subjects with a mean age of 59.8±9.8 years (range: 45–89 years). Mean neuroretinal rim area was 1.97±0.38 mm², mean intraocular pressure 15.6±3.0 mmHg, mean diastolic blood pressure 79.0±5.9 mm Hg, mean systolic blood pressure 133.5±11.1 mmHg, and mean body mass index was 25.5±3.7. In univariate analysis, neuroretinal rim area was significantly associated with optic disc size, open-angle glaucoma, refractive error, age and gender. After adjustment for these parameters in a multivariate analysis, a larger neuroretinal rim area was significantly correlated with a higher body mass index (P<0.001), in addition to be associated with a lower intraocular pressure (P = 0.004), lower mean blood pressure (P = 0.02), and higher ocular perfusion pressure.

Conclusions

In a general population, neuroretinal rim as equivalent of the optic nerve fibers is related to a higher body mass index, after adjustment for disc area, refractive error, age, gender, open-angle glaucoma, intraocular pressure, blood pressure and ocular perfusion pressure. Since body mass index is associated with cerebrospinal fluid pressure, the latter may be associated with neuroretinal rim area. It may serve as an indirect hint for an association between cerebrospinal fluid pressure and glaucoma.     

Walsh Hadamard Transform for Simple Linear Iterative Clustering (SLIC) Superpixel Based Spectral Clustering of Multimodal MRI Brain Tumor Segmentation

The automated brain tumor segmentation methods are challenging due to the diverse nature of tumors. Recently, the graph based spectral clustering method is utilized for brain tumor segmentation to make high-quality segmentation output. In this paper, a new Walsh Hadamard Transform (WHT) texture for superpixel based spectral clustering is proposed for segmentation of a brain tumor from multimodal MRI images. First, the selected kernels of WHT are utilized for creating texture saliency maps and it becomes the input for the Simple Linear Iterative Clustering (SLIC) algorithm, to generate more precise texture based superpixels. Then the texture superpixels become nodes in the graph of spectral clustering for segmenting brain tumors of MRI images. Finally, the original members of superpixels are recovered to represent Complete Tumor (CT), Tumor Core (TC) and Enhancing Tumor (ET) tissues. The observational results are taken out on BRATS 2015 datasets and evaluated using the Dice Score (DS), Hausdorff Distance (HD) and Volumetric Difference (VD) metrics. The proposed method produces competitive results than other existing clustering methods.     

Hierarchical clustering applied to automatic atlas based segmentation of 25 cardiac sub-structures

PurposeSegmentation of cardiac sub-structures for dosimetric analyses is usually performed manually in time-consuming procedure. Automatic segmentation may facilitate large-scale retrospective analysis and adaptive radiotherapy. Various approaches, among them Hierarchical Clustering, were applied to improve performance of atlas-based segmentation (ABS).MethodsTraining dataset of ABS consisted of 36 manually contoured CT-scans. Twenty-five cardiac sub-structures were contoured as regions of interest (ROIs). Five auto-segmentation methods were compared: simultaneous automatic contouring of all 25 ROIs (Method-1); automatic contouring of all 25 ROIs using lungs as anatomical barriers (Method-2); automatic contouring of a single ROI for each contouring cycle (Method-3); hierarchical cluster-based automatic contouring (Method-4); simultaneous truth and performance level estimation (STAPLE). Results were evaluated on 10 patients. Dice similarity coefficient (DSC), average Hausdorff distance (AHD), volume comparison and physician score were used as validation metrics.ResultsAtlas performance improved increasing number of atlases. Among the five ABS methods, Hierarchical Clustering workflow showed a significant improvement maintaining a clinically acceptable time for contouring. Physician scoring was acceptable for 70% of the ROI automatically contoured. Inter-observer evaluation showed that contours obtained by Hierarchical Clustering method are statistically comparable with them obtained by a second, independent, expert contourer considering DSC. Considering AHD, distance from the gold standard is lower for ROIs segmented by ABS.ConclusionsHierarchical clustering resulted in best ABS results for the primarily investigated platforms and compared favorably to a second benchmark system. Auto-contouring of smaller structures, being in range of variation between manual contourers, may be ideal for large-scale retrospective dosimetric analysis.