Clinical Detection of Macular Edema in Diabetic Patients Using the Automated Retinal Imaging System (ARISTM)
R. S. Doherty, A. P. Ciardella, J. Olsen, N. Mandava; Ophthalmology, Denver Health Hospital Authority, Denver,CO, Rocky Mountain Lions Eye Institute, Aurora, CO

BACKGROUND

Diabetic Retinopathy is a well known complication of diabetes mellitus. Macular edema or swelling of the retina is one of the retinal changes associated with diabetic retinopathy and worsening of visual acuity. Patients with clinically significant macular edema may benefit from additional treatments to prevent the progression of diabetic retinopathy. Proper screening and early detection of macular edema can lead to early treatment and better long term visual outcomes.
The Automated Retinal Imaging System (ARIS) is a fully automated, digital, fundus camera, which collects high quality constant- base stereo fundus photographs which demonstrate the retinal topography. Each image covers a thirty degree field. The images of the fundus reveal the topography of the retina, when viewed stereoscopically. These images are taken with minimal discomfort to the patient. This new technology potentially detects clinically significant macular edema in patients with diabetic retinopathy.

Optical Coherence Tomography (OCT) is a well established noninvasive method for measuring the thickness of the central retina and is a standard tool in detecting macular edema in diabetic patients.

PURPOSE

To compare the stereo photographic images produced by the Automated Retinal Imaging System (ARIS) and the retinal image maps created using Optical Coherence Tomography to determine if the ARIS is an effective screening tool to detect the appearance of macular edema in diabetic retinopathy. Also to examine the OCT thickness threshold values as measures of macular edema.

MATERIALS AND METHODS


Patients who presented to the clinic with a history of diabetes qualified for the study. After a complete ophthalmic exam patients had OCT and photographs taken. OCT images were first taken using the Zeiss OCT3 system. OCT images maps were obtained by taking six 5 mms scans centered on the macula wich generated a thickness map of the macula divided into nine standard sections. Fundus photographs of retinal fields one, two, and three were collected by an operator with minimal fundus imaging experience. 90 eyes from 45 patients were evaluated using the ARIS. A retinal specialist who had no prior knowledge of the patients nor had access to the OCT readings reviewed the photos. The photographs were reviewed stereoscopically for the presence or absence of macular edema. The macula was divided into nine standard sections, which correspond to the nine areas that the OCT images depict (Figure 1). The presence of edema was quantified as not present (0), slight edema or possible edema (1), and apparent macular edema (3). The results were then compared to the OCT scans results. Macular edema was defined as greater than 250 microns in the central macula or greater than 300 microns in any of the eight adjacent sectors on the macula map.

RESULTS

Fundus photographs were obtained for all patients. The presence of macular edema as recognized viewing stereo pairs of field 2 (centered on the macula) correlated with a central macular thickness greater than 250 microns by OCT and outer macular thickness greater than 300 microns. Of 90 eyes measured we obtained results from 72 eyes. 4 eyes (2 patients) could not be dilated sufficiently for ARIS imaging (the ARIS requires at least a 4mm pupil), 3 eyes (2 patients) could not be viewed in stereo, and 5 eyes (three patients) yielded images that were not gradable due to possible staphylomas or other pathological processes, 1 eye did not have a corresponding OCT, and 5 eyes were not able to fixate on the target properly to obtain an image of the macula. The ARIS sensitivity was 97% and specificity was 43%. Assuming that the OCT results are the gold standard, ARIS calls 3% of eyes negative that are edematous, and calls 56% of eyes positive when they are negative (Tables 1, 2).

Historically, the treatment threshold for macular edema has been a macular thickness of 250 microns in the center and 300 microns in the outer sectors. In our study, these were the parameters for macular edema that we chose. Clinically, the patients may need to be treated at a lower thickness or a much greater thickness depending on anatomical variance of normal retinal thickness and the overall health of the entire retina. As part of the statistical analysis it was found that these numbers are good estimates of treatment thresholds and that the ARIS is a good screening tool at these thresholds. In a screening test, false negatives are much more critical than false positives in order to effectively diagnose people who have the disease. At about 240 microns and 300 microns there were the minimum number of false negatives and the fewest number of false positives (See graphs 1,2 below).

Table 1

Table 2

Graph 1

Graph 2

CONCLUSIONS

In preliminary results, the ARIS appears to be an effective screening tool for early diagnosis of macular edema. Few cases of macular edema were missed, however there were a substantial number of false positives. This would make it an effective screening tool, especially since an operator with minimal experience in fundus photography can use the ARIS system. It is superior to OCT in that it provides not only information on macular thickness, but also vascular abnormalities such as microaneurysms, neovascularization and exudates. The ARIS uses new technology and creates high quality stereo photographs, which can provide a vast array of information. The ARIS camera compared to a traditional fundus camera provides the potential advantages of creating automated stereo pair photographs of the macula and optic disc and a mosaic of the fundus in 30 degree fields that can cover up to nearly a total 110 degrees (Figure 2). There is a short learning curve in operating the ARIS, after which both the quality of the images and the length of the photographic study were greatly improved.

AUTHOR DISCLOSURES

R.S. Doherty, Visual Pathways paid for some travel expenses; A.P. Ciardella, None; J. Olsen, None; N. Mandava, None