Eye Fundus Images Research Articles

New developments in OCT evaluations

AbstractDiabetic macular edema (DME) is one of the most common causes for visual loss (Ciulla et al., Diabetes Care, 2003). DME is characterized by an abnormal increase of the retinal thickness (RT) due to the accumulation of liquids in the retinal tissue, consequence of the first response to the inflammatory process occurring in diabetic retinopathy (DR). It is therefore considered one of the most important indicators of DR progression. Other indicators, such as microaneurysms and retinal leakage, were also observed and quantified in a prospective longitudinal 3‐year follow‐up study involving type 2 diabetic patients with nonproliferative DR. Patterns of progression were established based on these ophthalmic indicators (Lobo et al., Arch Opthalmol., 2004). OCT plays an important role in the assessment and quantification of DME. Stratus OCT version established a 9 area map for RT, being the central area of 1000µm in diameter. However, there is a lack of detailed information when studying DME. This problem was addressed by the use of an atlas of the human macula in a recent work from our group, where the merging of radial and circular scans was made possible (Bernardes et al., IOVS, 2008). Recently, Carl Zeiss introduced Cirrus HD‐OCT, a Fourier domain OCT, allowing to achieve more detailed information of the macular area, performing either 512×128 or 200×200 B‐scans, and allowing to export an eye fundus image for reference. It is now possible to assess local RT and correlate it with any imaging modalities available.

Automatic estimation of the arteriolar‐to‐venular diameter ratio (AVR) in retinal images

Abstract Purpose The Arteriolar‐to‐Venular diameter Ratio (AVR), a parameter derived from vessel caliber measurements in a specific region of retinal images, is used as a descriptor of generalized arteriolar narrowing. We developed a computerized system to compute AVR in a totally automatic way. Methods Images are at first enhanced to highlight the vessel network, which is then traced by a vessel tracking algorithm. From the detected vessel structure, the position of the optic disc is derived and the region inside which the AVR data are to be measured is determined. Vessels within this region are classified as either arteries or veins, their caliber estimated and the AVR parameter is eventually computed. Results provided by the system have been compared with manually derived AVR values on 20 eye fundus images. Results Mean and SD values from the two sets of measurements are the same, and automatic/manual ratios have an average value of 1 and 95% confidence interval of (0.98‐1.02). The correlation coefficient between the two methods is 0.88. In only two images the ratio is appreciably different from unity, 1.14 and 0.89 respectively. A detailed analysis of these cases revealed that in each image the wrong classification of one vessel only was the cause for these unsatisfactory results. When these misclassifications are manually corrected, e.g. with a quick editing tool that can be easily made available to the user, the ratios become 0.95 and 0.99, respectively, and the overall correlation coefficient becomes 0.97. Conclusion Additional evaluation on a larger set of images, acquired from subjects exhibiting wide variations of AVR, will be performed in order to fully assess the reliability and clinical applicability of this automatic procedure.

Identification of the optic nerve head with genetic algorithms

This work proposes creating an automatic system to locate and segment the optic nerve head (ONH) in eye fundus photographic images using genetic algorithms. Domain knowledge is used to create a set of heuristics that guide the various steps involved in the process. Initially, using an eye fundus colour image as input, a set of hypothesis points was obtained that exhibited geometric properties and intensity levels similar to the ONH contour pixels. Next, a genetic algorithm was used to find an ellipse containing the maximum number of hypothesis points in an offset of its perimeter, considering some constraints. The ellipse thus obtained is the approximation to the ONH. The segmentation method is tested in a sample of 110 eye fundus images, belonging to 55 patients with glaucoma (23.1%) and eye hypertension (76.9%) and random selected from an eye fundus image base belonging to the Ophthalmology Service at Miguel Servet Hospital, Saragossa (Spain). The results obtained are competitive with those in the literature. The method's generalization capability is reinforced when it is applied to a different image base from the one used in our study and a discrepancy curve is obtained very similar to the one obtained in our image base. In addition, the robustness of the method proposed can be seen in the high percentage of images obtained with a discrepancy delta<5 (96% and 99% in our and a different image base, respectively). The results also confirm the hypothesis that the ONH contour can be properly approached with a non-deformable ellipse. Another important aspect of the method is that it directly provides the parameters characterising the shape of the papilla: lengths of its major and minor axes, its centre of location and its orientation with regard to the horizontal position.

Macular epiretinal membrane formation after low-dose radiation exposure in a child with papilledema from a craniopharyngioma.

To describe a child with a craniopharyngioma and papilledema who developed a unilateral epiretinal membrane (ERM) after completing radiation therapy for his tumor. A clinical case is presented including fundus images of the patient's left eye and radiotherapy dosimetry curves. The patient presented with unilateral formation of an ERM after resolution of papilledema from a craniopharyngioma. Dosimetry curves from radiotherapy showed low-dose exposure in the region of ERM formation in the left eye at the level of 500 cGy to 1,000 cGy. The concordance of the low-dose radiation with the site of ERM formation suggests a possible contribution of radiotherapy treatment to the formation of the macular epiretinal tissue in this patient.

Choroidal Metastasis from Submandibular Salivary Gland Adenoid Cystic Carcinoma

A case of choroidal metastasis from an adenoid cystic carcinoma ofsubmandibular salivary gland is described. A 50-year-old woman with loss of vision in her left eye for 1 week was evaluated clinically, radiologically, and pathologically. A fundus image of the left eye showed an amelanotic, plateau-shaped choroidal mass measuring 17 x 12 X 4 mm with overlying exudative retinal detachment. A choroidal biopsy was consistent with metastatic adenoid cystic carcinoma from her submandibular salivary gland treated 5 years previously. She was treated with external beam radiotherapy but developed liver and lung metastases 3 months later and died within 1 month. All of the metastases demonstrated a basaloid histologic variant of adenoid cystic carcinoma that is known for its aggressive potential. Salivary gland carcinoma rarely metastasizes to the choroid; however, evaluation of patients with this carcinoma should include ophthalmic examination.

Análise fractal da vasculatura retínica: métodos de segmentação e de cálculo dimensional

Although it has been proposed that retinal vasculature is fractal, no method of standardization has been performed for vascular segmentation or for dimension calculation, thus resulting in great variability among values of fractal dimensions. The present study was designed to determine if estimation of retinal vessel fractal dimensions is dependent on vascular segmentation and dimensional calculation methods. Ten eye fundus images were segmented to extract their vascular trees by four computational methods ("multi-threshold", "scale-space", "pixel classification" and "ridge based detection"). Their information, mass-radius and box counting fractal dimensions were calculated and compared with those of the same vascular trees manually segmented (gold standard). The mean vascular tree dimension varied among the groups of different segmentation methods, from 1.39 to 1.47 for box counting, from 1.47 to 1.52 for information and from 1.48 to 1.57 for mass-radius dimensions. The utilization of different vascular segmentation methods and different dimension calculation methods introduced significant difference among fractal dimension of vessels. Estimation of retinal vessel fractal dimensions was dependent on both vascular segmentation and dimension calculation methods.

Evaluation of Retinal Digital Images by a General Practitioner

Diabetes and systemic hypertension are among the most prevalent diseases in developed countries. Since both can produce retinal involvement, they often cause long waiting lists for ophthalmologist consultation. The inspection of digital eye fundus images by a general practitioner would help to divert only those cases with retinal involvement. To evaluate this possibility we studied the agreement between the assessment of digital fundus images made by a general practitioner and an ophthalmologist. A total of 776 fundus images of 194 patients were inspected by both observers looking for retinal lesions consistent with diabetic or hypertensive retinopathy. Contingency tables were made, and the agreement between both observers was assessed by using the kappa index. We found good agreement between the findings reported by the general practitioner and those reported by the ophthalmologist (kappa = 0.80 for patients with only diabetes, 1.00 for patients with only hypertension, and 0.79 for patients with both diseases). Our results suggest that the inspection of digital retinal images by a general practitioner could reduce the number of referred patients for specialist consultation.

Elastic Registration for Retinal Images Based on Reconstructed Vascular Trees

The vascular tree of the retina is likely the most representative and stable feature for eye fundus images in registration. Based on the reconstructed vascular tree, we propose an elastic matching algorithm to register pairs of fundus images. The identified vessels are thinned and approximated using short line segments of equal length that results a set of elements. The set of elements corresponding to one vascular tree are elastically deformed to optimally match the set of elements of another vascular tree, with the guide of an energy function to finally establish pixel relationship between both vascular trees. The mapped positions of pixels in the transformed retinal image are computed to be the sum of their original locations and corresponding displacement vectors. For the purpose of performance comparison, a weak affine model based fast chamfer matching technique is proposed and implemented. Experiment results validated the effectiveness of the elastic matching algorithm and its advantage over the weak affine model for registration of retinal fundus images.

Computerized measurement of retinal blood vessel calibre: description, validation and use to determine the influence of ageing and hypertension

To validate a computer-based method for measuring the calibre of retinal blood vessels, and use it to determine the effects of ageing and arterial hypertension on the calibres of these vessels and on their ratio. Digital eye fundus images covering a 50 degrees field and centred on the optic disc were obtained using a 540 nm filter. The boundaries of blood vessels crossing a series of circles concentric with the optic disc were located by an image analysis program; the calibres of vessels crossing the circles perpendicularly were determined automatically, the average arteriole and average vein calibres were calculated, and the arteriovenous ratio (AVR) was calculated as the ratio of these averages. The within-operator, between-operator and within-eye reliability of this method was investigated using images of 30 or 40 eyes; the effects of ageing on average vessel calibres and AVR using both eyes of each of 120 normotensive volunteers aged 10-69 years (60 males, 60 females); and the effects of arterial hypertension using a group of 54 hypertensive patients aged 50.9 +/- 13.9 years. Within-operator, between-operator and within-eye correlation coefficients for AVR were all better than 0.95, and the corresponding coefficients of variation were all better than 3%. The results of Bland Altman approach show a very good agreement. There were no significant differences between right and left eyes or between the sexes in either the normotensive or the hypertensive group. In the normotensive group, vein calibre was almost constant (111 +/- 6 microm), but arteriole calibre and AVR fell significantly from 96 +/- 6 microm and 0.870 +/- 0.046, respectively, in the second decade of life to 85 +/- 4 microm and 0.761 +/- 0.044 in the seventh decade. Arterial hypertension was not associated with changes in vein calibre, but was associated with decreases in arteriole calibre (from 91 +/- 7 microm among normotensive individuals to 84 +/- 2 microm) and AVR (from 0.816 +/- 0.056 to 0.755 +/- 0.027). Age significantly modulated the AVR-reducing influence of hypertension (P = 0.003). The proposed method of measuring retinal blood vessel calibres is reliable and precise (especially for the AVR). In this study, its results confirmed that increasing age and arterial hypertension are both associated with reductions in retinal arteriole calibre and AVR.

Knowledge modeling for the image understanding task as a design task

This paper formally analyzes the image understanding (IU) task at the knowledge level and in the observer domain. The analysis is done at three levels: task, method and domain knowledge, distinguishing the generic components in most of the IU tasks, thereby enabling the components to be reused. We model the IU problem as a design task and define a generic problem solving method (PSM) that allows us to tackle the task in a hierarchical and recursive way. The main advantage of this generic PSM is the possibility of instantiating specific PSMs through parameter space configuration, which enables the structure to be reused both across the task decomposition at different hierarchical levels and across different application domains. This work has been done following the well-established foundations of knowledge engineering that prescribe the maintenance of the conceptual structure from the modeling stage at the knowledge level down to the implementation. Finally, we apply the proposed framework to the problem of identifying the papilla in eye fundus images in order to exemplify the successive stages in the modeling process and system design, and accordingly justify the framework's validity.

Nonproliferative retinopathy in diabetes type 2. Initial stages and characterization of phenotypes

This review addresses the initial stages of nonproliferative diabetic retinopathy in diabetes type 2. The natural history of the initial lesions occurring in the diabetic retina has particular relevance for our understanding and management of diabetic retinal disease, one of the major causes of vision loss in the western world. Diabetic retinal lesions are still reversible at this stage opening entirely new opportunities for effective intervention. Four main alterations characterize these early stages of diabetic retinopathy: microaneurysms/hemorrhages, alteration of the blood-retinal barrier, capillary closure and alterations in the neuronal and glial cells of the retina. These alterations may be monitored by red-dot counting on eye fundus images and by fluorescein leakage and retinal thickness measurements. A combination of these methods through multimodal macula mapping has contributed by identifying three different phenotypes of diabetic retinopathy. They show different types and rates of progression which suggest the involvement of different susceptibility genes. The identification of different phenotypes opens the door for genotype characterization, different management strategies targeted treatments.

Adaptive system for eye-fundus imaging

A compact adaptive system capable of imaging a human-eye retina with a spatial resolution as high as 6 μm and a field of view of 15° is developed. It is shown that a modal bimorph corrector with nonlocalised response functions provides the efficient suppression of dynamic aberrations of a human eye. The residual root-mean-square error in correction of aberrations of a real eye with nonparalysed accommodation lies in the range of 0.1 — 0.15 μm.

Digital Retinal Images and Teleophthalmology for Detecting and Grading Diabetic Retinopathy

Detecting and grading of diabetic retinopathy (DR) by means of digital retinal images sent via the Internet. Four nonstereoscopic digital retinal images (45 degrees field each) of 126 eye fundus images from 70 diabetic patients were obtained with a nonmydriatic camera at two peripheral units. The images were sent via the Internet using a web-based system to a reference center, where they were diagnosed and graded by one ophthalmologist. These results were compared with those obtained by two other ophthalmologists, one at each peripheral unit, after direct examination of the patients. A modified severity scale of Airlie House was used for grading DR in all cases. Agreement between observers was assessed using unweighted kappa for categorical data and the intraclass correlation coefficient (ICC) for continuous data. Presence of DR was detected in 69 eyes (55%). All eyes with DR (69 of 69, 100%) were correctly identified (kappa = 1) by inspecting the digital images. In 118 eyes (118 of 126, 94%), 57 with no DR and 61 with DR, there was an agreement between the gradation made after the direct examination and the gradation made after the inspection of the images (ICC = 0.92). In eight eyes with DR (8 of 126, 6%), there was disagreement in the grading made with both techniques. Inspection of digital retinal images sent via the Internet allowed diagnosis and grading of DR. The presence or absence of DR was correctly assessed by inspection of the images in all instances. We also found agreement, in most cases, between retinopathy gradation made from the images and the gradation made by direct examination of the eyes. These findings suggest that this technique is suitable for screening procedures.

A multimodal registration algorithm of eye fundus images using vessels detection and Hough transform.

Image registration is a real challenge because physicians handle many images. Temporal registration is necessary in order to follow the various steps of a disease, whereas multimodal registration allows us to improve the identification of some lesions or to compare pieces of information gathered from different sources. This paper presents an algorithm for temporal and/or multimodal registration of retinal images based on point correspondence. As an example, the algorithm has been applied to the registration of fluorescein images (obtained after a fluorescein dye injection) with green images (green filter of a color image). The vascular tree is first detected in each type of images and bifurcation points are labeled with surrounding vessel orientations. An angle-based invariant is then computed in order to give a probability for two points to match. Then a Bayesian Hough transform is used to sort the transformations with their respective likelihoods. A precise affine estimate is finally computed for most likely transformations. The best transformation is chosen for registration.

Reliability of photometric measurements in the Zeiss fundus camera.

The authors test the hypothesis of systematic errors in photometric and/or densitometric measurements performed by Fundus camera equipment. By means of 2 different methods (reflection and transmission) the non-uniformity of light-distribution on sample surfaces placed in front of the fundus camera is shown. The consequent point by point variability in retinal illumination can be defined as the 'retinal shading problem'. This conclusion must be taken into account when equidensitometric measurements are carried out on eye fundus images with fundus camera, by photographic, cinematographic or TV techniques.

Recent developments in eye fundus imaging for clinical application: television fluoroangiography and new technologies.

New technical developments are reported in television fluoroangiography. The use of “equidensitometry” and fluorescein dilution curves for the determination of retinal blood flow is discussed. The authors present a digital system for TV fundus image processing, which appears to be the most promising field for obtaining quantitative information, i.e. quantifying clinical data from the eye fundus.