CONSISTENCY OF OCULAR COHERENCE TOMOGRAPHY FAST MACULAR THICKNESS MAPPING IN DIABETIC DIFFUSE MACULAR EDEMA

OBJECTIVES: To investigate optical coherence tomography consistency on foveal thickness, foveal volume, and macular volume measurements in patients with and without diffuse diabetic macular edema. INTRODUCTION: Optical coherence tomography represents an objective technique that provides cross-sectional tomographs of retinal structure in vivo. However, it is expected that poor fixation ability, as seen in diabetic macular edema, could alter its results. Several authors have discussed the reproducibility of optical coherence tomography, but only a few have addressed the topic with respect to diabetic maculopathy. METHODS: The study recruited diabetic patients without clinically evident retinopathy (control group) and with diffuse macular edema (case group). Only one eye of each patient was evaluated. Five consecutive fast macular scans were taken using Ocular Coherence Tomography 3; the 6 mm macular map was chosen. The consistency in measurements of foveal thickness, foveal volume, and total macular volume for both groups was evaluated using the Pearson’s coefficient of variation. The T-test for independent samples was used in order to compare measurements of both groups. RESULTS: Each group consisted of 20 patients. All measurements had a coefficient of variation less than 10%. The most consistent parameter for both groups was the total macular volume. DISCUSSION: Consistency in measurement is a mainstay of any test. A test is unreliable if its measurements can not be correctly repeated. We found a good index of consistency, even considering patients with an unstable gaze. CONCLUSIONS: Optical coherence tomography is a consistent method for diabetic subjects with diffuse macular edema


INTRODUCTION
Diabetic retinopathy is an important cause of visual impairment. 1 Slit-lamp biomicroscopy has a low sensitivity and can misdiagnose macular edema smaller than 100 µm. 2 Another well known technique is fluorescein angiography; however, it provides a more qualitative rather than quantitative analysis of the retina.Therefore, it can be difficult to evalu-ate the progression of macular edema or treatment efficacy.
New imaging modalities have emerged in the past few years.A promising technology is the one described in detail by Hee et al. in 1995 named optical coherence tomography (OCT). 2,3It represents an objective and noninvasive technique that provides cross-sectional tomographs of retinal structure in vivo with a resolution of 10 µm. 2,3Other available instruments such as ultrasound and retinal thickness analyzers cannot provide this data. 3,4,5everal authors have already demonstrated that OCT can play an important role in the study of macular and optic disk diseases. 2,5,6][9][10][11] Polito et al. 10 reported a reproducibility of 4-11% in diabetic macular edema.However, focal and diffuse macular edema were not distinguished.
Since ocular motion can affect image resolution, as seen in subjects with diffuse diabetic macular edema (DME), it is expected that poor fixation ability can alter the outcome of the text. 2 The aim of this study was to investigate optical coherence tomography consistency on foveal thickness, foveal volume, and macular volume measurements in patients with and without diabetic diffuse macular edema (DME).

MATERIALS AND METHODS
The protocol was approved by the ethical review committee of the hospital.All patients understood and signed an informed consent.The study recruited 40 type II diabetic patients from the Hospital das Clínicas of the University of São Paulo Medical School.Twenty patients did not have clinically evident retinopathy, and the other 20 had diffuse macular edema with no proliferative changes and no macular ischemia.The diffuse DME was characterized by generalized leakage from a dilated macular capillary bed; it was assessed by slit-lamp biomicroscopy with a Volk 78D lens and fluorescein angiography.Patients were consecutively enrolled over 2 months.We evaluated only one eye of each patient to avoid bias by dependent samples.If both eyes of a subject met the criteria, we randomly selected one eye.Patients who had a severe eye disease other than DME (such as glaucoma or severe hypertensive retinopathy) were excluded from the study.
Patients were divided into two different groups: (1) Case group -eyes with DME; (2) Control group -eyes without retinopathy.Five consecutive scans were taken by a trained ophthalmologist using OCT-3 (Stratus OCT ™Model 3000; Version 4.0.1;Carl Zeiss Meditec Inc. Dublin; CA, USA) after mydriasis.The fast macular scan protocol was preferred, attempting to reduce errors caused by scan positioning misalignment; this problem occurs most often when separate line scans are used.This protocol consists of six consecutive 6-mm radial line scans centered on the macula, each containing 128 A-scans, performed in 1.92 seconds.Therefore, retinal thickness is measured at a total of 768 points along these six intersecting lines.A signal strength higher than five was obtained in all scans.
The measurement analyses were done using a retinal thickness/volume tabular program.The 6 mm macular map was chosen.Consistency in measurements of foveal thickness, foveal volume, and total macular volume for both groups was evaluated using the Pearson's coefficient of variation (CV).The T-test for independent samples was used in order to compare measurements of both groups.A p-value smaller than 5% was considered statistically significant.

RESULTS
The study consisted of 40 eyes of 40 diabetic patients.Twenty patients (11 females, 9 males) had a normal fundus appearance and the other 20 (9 females, 11 males) were diagnosed with diffuse macular edema (DME).Mean age was 56.9 ± 8.9 years in the control group and 62. 6  All measurements had a CV smaller than 10%.The most consistent parameter for both groups was the total macular volume.This was the only parameter that was statistically different between groups (table 2).All others values were smaller in the control group but not statistically significant.

DISCUSSION
Consistency is a mainstay of any test and can be assessed by Pearson's coefficient of variation, which is the standard deviation of the measurements divided by its mean times 100.A test is unreliable if its measurements can not be correctly repeated.Several authors have discussed OCT reproducibil- ity but only a few studies addressed its use with diabetic maculopathy. 5,9,10It is already known that OCT measurements change little in people with normal eye fixation. 2,5,7owever, greater variation is likely to occur in patients with macular disease due to their reduced ability to fixate during examination; 2,10 this can be visualized in figures 1 and 2, which represent two consecutive fast macular scans of two different patients with diabetic DME. Figure 1 shows that small eye movements are able to produce completely different outcomes, especially in the 1000µm inner ring where we found a difference as great as 135 µm.Gürses-Özden et al. 7 studied 10 normal subjects and demonstrated a CV less than 10% in mean foveal thickness and total macular volume using OCT-3.They also observed better reproducibility during macular scans than retinal nerve fiber layer thickness scans.The latter requires manual placement of the scan by the operator, and the former only depends upon the patient focusing on the instrument's internal fixation light.Goebel and Kretzchmar-Gross 5 reported a reproducibility of 2.3% in diabetic retinopathy measurements.However, it is important to note that they included all stages of diabetic retinopathy, and they excluded patients that were not able to maintain stable fixation.Since we cannot exclude these patients from our routine ophthalmologic practice, it is also essential to have data on OCT reproducibility in all diabetic subjects independent of fixation ability.In our study, we found a good index of consistency even considering patients with an unstable gaze.
Polito and coworkers 10 found similar retinal thickness results comparing undilated and dilated pupils in normal eyes and those with clinically significant macular edema.Moreover, they observed greater data variability in the outer macular areas than in the foveal region.Reduced numbers of measured points and greater distance between tomograms done in the periphery explain this variability.Schuman et al. 12 studied patients in different OCT machines and found similar measurements in the studied parameters.
The mean foveal thickness of our diabetic patients without retinopathy was larger than published data in normal patients, similar to the results obtained by Massin et al. 13 This probably happened because slit-lamp biomicroscopy is worse than OCT at detecting subtle macular edema. 2 Yang et al. 14 found swelling or cystoid macular edema in 58% (11/19) of diabetic eyes without ophthalmoscopic evi-  dence of clinically significant macular edema.

Figure 1 -
Figure 1 -Fast macular scans of a patient with diffuse diabetic macular edema obtained consecutively.Note the differences in the central macular values between scans.

Figure 2 -
Figure 2 -Fast macular scans of a patient with diffuse diabetic macular edema obtained consecutively.Despite the smaller difference than observed in figure 1, there is still a reasonable amount of variation in central values.
± 9.85 years in the case group.The results of the measurements are shown in table 1; the mean value of the coefficient of variation (CV) is summarized in table 2. Examples of greater variation due to lack of fixation are shown in figures 1 and 2.

Table 2 -
Mean coefficient of variation of the foveal thickness, the foveal volume and the total macular volume in the case and the control group.

Table 1 -
Mean foveal thickness, foveal volume and total macular volume in the case group and the control group.