[Skip to Content]
[Skip to Content Landing]
Download PDF
Figure 1.
Left, Scan obtained 5 minutes after injection showing the 18 images. Top right, Fluorometric profile of the 18 images. Bottom right, Subtraction of 2 profiles (eg, 30 and 5 minutes). The shaded area represents fluorescein that penetrated into the vitreous.

Left, Scan obtained 5 minutes after injection showing the 18 images. Top right, Fluorometric profile of the 18 images. Bottom right, Subtraction of 2 profiles (eg, 30 and 5 minutes). The shaded area represents fluorescein that penetrated into the vitreous.

Figure 2.
Top row, Normal; middle row, minimal diabetic retinopathy; and bottom row, moderate diabetic retinopathy. Left, Fundus image; center, 30-minute RLMaps; and right, 60-minute RLMaps. Note progressive diffusion of fluorescein in the vitreous with time. The color coding can be deciphered with the aid of the color bar on the right. Units are 10−7 cm/s.

Top row, Normal; middle row, minimal diabetic retinopathy; and bottom row, moderate diabetic retinopathy. Left, Fundus image; center, 30-minute RLMaps; and right, 60-minute RLMaps. Note progressive diffusion of fluorescein in the vitreous with time. The color coding can be deciphered with the aid of the color bar on the right. Units are 10−7 cm/s.

Figure 3.
Left, Fundus image. Right, RLmaps. Top row, 3-day photocoagulated eye showing highly leaking regions at burn sites; middle row, vitrectomized eye showing uniform distribution of fluorescein; and bottom row, eye with multiple drusen showing little leakage, independent from background fluorescence. The color coding can be deciphered with the aid of the color bar on the right. Units are 10−7 cm/s.

Left, Fundus image. Right, RLmaps. Top row, 3-day photocoagulated eye showing highly leaking regions at burn sites; middle row, vitrectomized eye showing uniform distribution of fluorescein; and bottom row, eye with multiple drusen showing little leakage, independent from background fluorescence. The color coding can be deciphered with the aid of the color bar on the right. Units are 10−7 cm/s.

1.
Wise  GNDollery  CTHenkind  P The Retinal Circulation.  New York, NY Harper & Row1971;143- 165421- 455
2.
Cunha-Vaz  JGFaria Abreu  JRCampos  AJFigo  GM Early breakdown of the blood-retinal barrier in diabetes. Br J Ophthalmol. 1975;59649- 656Article
3.
Cunha-Vaz  JG Vitreous fluorophotometry. Osborne  NNGhader  GJeds.Progress in Retinal Research. Oxford, England Pergamon Press1985;90- 114
4.
Chahal  PSChowienczyk  PJKohner  EM Measurement of blood-retinal barrier permeability: a reproducibility study in normal eyes. Invest Ophthalmol Vis Sci. 1985;26977- 982
5.
Asrani  SZeimer  RGoldberg  MFZou  S Application of rapid scanning retinal thickness analysis in retinal diseases. Ophthalmology. 1997;1041145- 1151Article
6.
Puliafito  CAHee  MRLin  CP  et al.  Imaging of macular disease with optical coherence tomography. Ophthalmology. 1995;102217- 229Article
Clinical Sciences
May 1999

Mapping Retinal Fluorescein Leakage With Confocal Scanning Laser Fluorometry of the Human Vitreous

Author Affiliations

From the Center of Ophthalmology, University Hospital and Institute for Biomedical Research in Light and Image, Coimbra, Portugal.

Arch Ophthalmol. 1999;117(5):631-637. doi:10.1001/archopht.117.5.631
Abstract

Objective  To demonstrate an objective, quantitative, and sensitive method of mapping retinal fluorescein leakage into the vitreous while simultaneously imaging the retina.

Methods  A prototype Zeiss confocal scanning laser ophthalmoscope was modified to obtain fluorometric measurements from 18 optical planes across the retina and cortical vitreous, separated from each other by 150 µm, and parallel to the retinal surface. After intravenous administration of fluorescein, an axial graphic of equivalent fluorescein concentration in the vitreous may be obtained from any region of interest. After correcting for fluorescence levels in the retina and choroid and plasma levels of free fluorescein, permeability values of the blood-retinal barrier to fluorescein were obtained from 1512 regions measuring 75 × 75 µm, from a total 3150 × 2700-µm area of the fundus, generating a detailed map of retinal fluorescein leakage. The method was assessed in vitro and in 7 healthy subjects who underwent scans during separate visits. Depth resolution and influence of chorioretinal fluorescence were further tested in 2 patients with multiple drusen and in 2 eyes after vitrectomy. Fourteen eyes from 7 patients with diabetes and nonproliferative retinopathy were also examined. Lateral resolution was tested in 3 diabetic eyes that underwent focal photocoagulation. Four eyes from 2 patients with diabetes and minimal retinopathy were examined at 3-month intervals. All eyes examined had less than 2 diopters of astigmatism.

Results  Characteristics of the modified confocal scanning laser fluorometer included a lower limit of detection equal to 0.40 Eq ng/mL and depth precision of ±15 µm. Values for the blood-retinal barrier permeability index in healthy subjects, measured 30 minutes after a single intravenous pulse of fluorescein (14 mg/kg), ranged from 1.3 ± 0.4 × 10−6 cm/s over the foveal avascular zone to 2.2 ± 0.6 × 10−6 cm/s over vessels in the retina. Diabetic eyes with retinopathy showed higher values, ranging from 1.4 to 15.0 × 10−6 cm/s. Vitrectomized eyes and eyes with multiple drusen showed the validity of the correction algorithm demonstrating that measurements of fluorescence in the vitreous are not influenced by the chorioretinal fluorescence level. Argon laser photocoagulation burns placed in the diabetic retina demonstrated a lateral resolution on the order of 75 to 100 µm. Intravisit and intervisit reproducibility was ±10.2% and ±13%, respectively.

Conclusions  This new method measures localized alterations of the blood-retinal barrier and allows for direct correlation with retinal anatomy. Its most interesting feature is the ability to map retinal fluorescein leakage while simultaneously imaging the retina. This capability is expected to improve our understanding and management of retinal disease.

SEVERAL RETINAL diseases are associated with altered permeability of the blood-retinal barrier (BRB). Widespread use of fluorescein angiography has shown that a breakdown of the BRB, as demonstrated by abnormal fluorescein leakage, is a prominent feature of retinal pathologic lesions.1 One example of such retinal pathology is diabetic macular edema, which is considered the most frequent cause of visual loss in diabetic retinopathy and is associated with increased BRB leakage. Fluorescein angiography shows the location of the leakage but does not quantify it.

Quantitative evaluation of the BRB became possible when vitreous fluorophotometry was introduced to clinical research.2 Fluorophotometric studies of BRB permeability have been performed in patients with diabetes mellitus, glaucoma, macular degeneration, retinitis pigmentosa, uveitis, central retinal vein occlusion, and acquired immunodeficiency syndrome.3 However, a major limitation of currently available instrumentation is associated with the fact that the permeability of the BRB is measured as an average over the macular area. Accurate mapping of localized changes in the permeability of the BRB would benefit early diagnosis, explanation of the natural history of retinal disease, and prediction of its effect on visual acuity.

We have developed a method of retinal fluorescein leakage mapping (RLMapping) that is capable of measuring localized changes in fluorescein leakage across the BRB while simultaneously imaging the retina. The instrument is based on a prototype Zeiss confocal scanning laser ophthalmoscope that was modified into a confocal scanning laser fluorometer. In this article, we describe the modifications introduced and the test results of the instrument. The studies were performed in vitro, in a model eye, and in vivo, in the eyes of healthy volunteers and of patients with and without alterations to the BRB.

MATERIALS AND METHODS
INSTRUMENTATION

A prototype Zeiss confocal scanning laser ophthalmoscope (Carl Zeiss, Oberkochen, Germany) was used in this work. The data and images obtained were of a 20° field of view that corresponds to a scanned area 3150 µm wide by 2700 µm tall. In the fluorometry mode, a set of 9 images was sequentially captured. The first image was in the posterior plane (located at the retina), while the last image represented the most anterior scanned region (in the vitreous), ie, the movement was posterior to anterior, with a distance of approximately 300 µm between planes. Continuous movement of the confocal plane of the system led to a tilting effect. An exponential fit performed took this into consideration, and no differences were observed between the top and bottom lines of the maps of permeability. The instrument has an argon laser source (488 nm) with a light power of 400 µW at the eye; confocal stop No. 3 was used.

To perform vitreous fluorometry, the first modification was substituting the original automatic gain control with a fixed one that is manually adjustable by a precision potentiometer. This modification allowed us to obtain a linear correlation between collected light from the eye and image intensity. We also needed to separate the fluorescent light, so we placed a barrier filter in the light optical path before starting each scan. The filter is a bandpass filter, with minimal attenuation from 520 to 630 nm (fluorescent light) and high attenuation for reflected light (488 nm). The second modification was to introduce an automatic electronic on/off switch for the barrier filter control. When the user presses a knob to start the acquisition, the system automatically puts the filter in the optical path of the light. To use the instrument for obtaining RLMaps, we created a computer program that (1) controls all image acquisition and visualization, (2) stores information on computer disk, and (3) processes the data. Using the program, we can select a region of interest (ROI) among the images and calculate the mean value and SD of the pixels inside that ROI. By representing those mean values as a function of the corresponding image's axial position, we obtained a fluorescence axial graphic in arbitrary units that could be converted to equivalent fluorescein concentrations after calibration. This software allowed for separating the odd and even lines of each image, thus creating 2 images that, in fact, represented 2 consecutive planes with half of the original distance between them. This design is based on the fact that the video signal is interlaced, ie, the generation of 1 image is performed in 2 steps: in the first step the system generates all the odd lines, and in the second step it generates all the even lines. The focal plane is continuously moved during image acquisition so that the odd and even lines represent 2 physically distinct planes separated by half of the distance between 2 consecutive full-image planes. As a result, we obtain 18 images and thus more detailed axial information with 18 points in the fluorescence axial graphic.

Another modification involves introducing a personal computer (PC) board, a digital-to-analog converter, some analog support hardware, and an IBM-compatible PC bus interface logic circuitry. This board controls the manual focusing knob, thus enhancing the reproducibility of in vivo scans. The process of image acquisition has been simplified because the manual focus on the patient eye is always made on the retinal vessels, the most visible and high-contrast regions, and because the location before the scan is automatically set by the developed PC board.

The instrument is, therefore, able to obtain fluorescence axial graphics of the vitreous for 1512 ROIs simultaneously, each of which allows for calculation of individual values of permeability of the BRB after administration of fluorescein, considering plasma fluorescence levels and correcting for vascular fluorescence and background scatter. The simultaneous representation of such individual values of permeability of the BRB form an RLMap. The instrument described herein is identified as a retinal leakage analyzer (RLA).

CHARACTERISTICS OF RLA PERFORMANCE
Calibration in Equivalent Fluorescein Concentration

We used a set of cuvettes filled with different known fluorescein concentrations to perform the calibration. Each cuvette is placed in the focal plane of a 66.6-dioptries convex lens, and a scan is taken after careful alignment. After image acquisition, we selected the most relevant ROI and, using the specifically developed software program, obtained the corresponding fluorescence axial graphic of the cuvette. The maximum value of each scan is plotted by function of the fluorescein concentration in the cuvette. If saturation is not reached, a linear relation should be obtained between RLA values of the cuvettes and their fluorescein concentrations, using the minimum square deviations method to adjust a straight line to the points. This calibration procedure gives a conversion factor for RLA values into equivalent fluorescein concentrations.

Lower Limit of Detection

The lower limit of detection (LLOD) represents the lowest amount of fluorescein that can be reliably detected by the system. In vitro, noise performance of the system is the factor that determines the LLOD. We defined the LLOD as 3 times the SD of the background, which means that there is a 0.27% probability of originating such a value from a background random artifact.

Accuracy

Accuracy represents deviations of measured values from the true value. In fluorometry, it is common to define accuracy as the percentage deviation of an in vitro fluorometry reading from its known concentration.

Reproducibility

Reproducibility describes the ability of a system to reproduce the output value when the same value to be measured is applied to it consecutively under the same conditions. It is expressed as the ratio between SD and the mean of repetitive measurements performed under similar conditions.

RLMAPPING IN HUMAN EYES

Fluorometric measurements were performed with the RLA after dilation of the pupil with tropicamide (1%). Retinal and vitreous fluorescence was measured within the first 5 minutes after intravenous injection of 14 mg/kg of 20% sodium fluorescein and at 15, 30, and 60 minutes after the injection. Blood was sampled forplasma fluorescein level at 10, 15, and 50 minutes after fluorescein injection. As described previously, the RLA acquires an image of the entire retina with real 3-dimensional information. With use of the developed software, it is possible to select an arbitrary-sized ROI over an image as small as 75 × 75 µm or the total area (Figure 1) and to build the corresponding fluorescence axial graphic. Using this facility, the ROI can be placed over a localized area, a vessel, the foveal avascular zone, or wherever desired, and thus the leakage from these areas can be quantitatively compared.

Developed software also allows us to have an image of the retina constructed from the 18 planes using image-processing algorithms and to obtain profiles for different points chosen on the constructed image of the retina.

Calculations of BRB Permeability to Fluorescein

We chose to calculate BRB permeability using a simple index of permeability4 that involves measuring the penetration of fluorescein into the posterior vitreous by estimating the corrected area under the vitreous fluorescence curve and dividing by the total "exposure" to plasma-free fluorescein. Thus, at time t = t1 the permeability index can be derived from our data by dividing the corrected area under the vitreous fluorescence curve by the area under the plasma-free fluorescence curve. Measurement of the permeability index assumes 1-dimensional diffusion of fluorescein across a retina plane and includes measurements of fluorescence near the retina. These calculations may be made with the instrumentation described herein.

To calculate the corrected area under the vitreous fluorescence curve at the time of the measurement and to include values near the retina, we proceed as follows. We makea scan less than 5 minutes after the fluorescein administration, followed by a measurement scan (eg, 30 minutes). These 2 images are then shown simultaneously so the user can select a common reference point in both. With this information, the system is able to align the images and extract the common portion of each one. Both scan profiles are filtered and interpolated. The profile of the first scan is then scaled down until its maximum equals the maximum of the second profile. Profiles are aligned and the difference is computed. The difference profile starts at zero, reaches a peak, and decays as an exponential. Using the decay portion of the curve, the exponential is adjusted by a best-fit algorithm. The area under the curve is the corrected area under the vitreous fluorescence curve (Figure 1). The permeability index is calculated and the maps of BRB permeability built.

Lateral Resolution

To evaluate lateral resolution of the RLA for RLMapping, we applied 100-µm photocoagulation spots to 3 eyes with diabetic macular edema. These burns were applied contiguously or 100 µm apart. Measurements were performed before and at 3 days and 3 weeks after photocoagulation to measure retinal leakage before, at its height, and after healing has occurred.

Reproducibility

Reproducibility was assessed in human subjects with no known history of ocular or systemic disease. Seven subjects, 2 women and 5 men aged 20 to 40 years were recruited. Tenets of the Declaration of Helsinki were followed, and approval of the institutional review board was obtained. Before measurement, informed consent was received from each subject, and the pupil was dilated with 1% tropicamide.

Two different reproducibilities were assessed. The first—intravisit reproducibility—evaluated the reproducibility of multiple measurements obtained in a single session. Six scans were performed at the fovea in each of the 7 subjects. The instrument was realigned for each scan. The coefficient of variation (SD/mean) of BRB permeability values in the 6 scans of each subject was calculated, based on the 30-minute scan. An average was calculated, and in vivo intrasession reproducibility was taken as the mean of values of the 7 subjects.

The second reproducibility assessed variability between 2 visits in 5 subjects. Two sessions were performed 2 weeks apart. In each session, 6 scans were obtained at the fovea, and an average permeability was calculated. The coefficient of variation of BRB permeability values from both visits of each subject was calculated, and the mean of the values of 5 subjects was taken as the in vivo intervisit reproducibility.

RESULTS
CHARACTERIZATION OF RLA PERFORMANCE
Calibration in Equivalent Fluorescein Concentration

A curve of calibration in equivalent fluorescein concentrations was obtained using cuvettes filled with fluorescein solutions having concentrations of 1, 5, 10, 50, 100, 150, and 200 ng/mL. A linear correlation was found between fluorescein concentration and the measured values in a range greater than 2 orders of magnitude.

Lower Limit of Detection

The LLOD was experimentally determined by performing 10 consecutive scans of a cuvette filled with distilled water. The average and SD were calculated for the 18 values measured for each scan. The final average and SD of the 10 scans were 0.313 and 0.0286 Eq ng/mL, respectively, which represents an LLOD of 0.31 + 3 × 0.0286 equal to 0.40 Eq ng/mL.

Accuracy

Accuracy was assessed by performing 3 sets of 5 consecutive measurements of cuvettes with known concentrations (10, 50, 100, 150, and 200 ng/mL). The results showed an error of measurement lower than 10% for concentrations of 10 to 50 ng/mL, and lower than 1.5% for concentrations of 100 to 200 ng/mL.

Reproducibility

Using a set of cuvettes with 1, 5, 10, and 20 ng/mL, a reproducibility better than 4.2% was obtained for concentrations of 1 to 5 ng/mL. This figure became better than 1.5% for concentrations of 100 to 200 ng/mL.

ASSESSMENT OF RLMAPPING IN HUMAN EYES
RLMapping in Healthy Human Eyes

A large 3150 × 2700-µm field covering the macular area was assessed in 7 healthy subjects. Each series of 18 images (1 scan) was obtained within the first 5 minutes after injection and at 15, 30, and 60 minutes after intravenous administration of fluorescein. In an analogy to computed tomographic imaging, 3-dimensional information is obtained by using the optical sectioning of the RLA to acquire multiple images of 18 consecutive slices throughout the retina and vitreous, each 150 µm apart. Two types of information are obtained simultaneously: one for optical imaging of the retinal and vitreous structures, and the other representing fluorescence measurements from the areas being scanned.

Retinal anatomic features such as the fovea and retinal vessels were identifiable by their characteristic morphological features, the retinal vessels being readily detected immediately after fluorescein injection.

Axial graphics of the fluorescence measurements obtained from the vitreous representing a volume of 75 × 75 × 2550 µm are converted into RLMaps (Figure 2). The BRB permeability index in normal retina ranged from 0.9 to 2.8 × 10−6 cm/s. The values were lower in the region of the foveal avascular zone and higher in vascularized areas of the macula, with a mean value of 1.9 ± 0.4 × 10−6 cm/s for the total area of the 7 healthy subjects examined. Localized differences were better identified in the 15- and 30-minute examinations, becoming less apparent in the 60-minute examination because of progressive diffusion of fluorescein in the vitreous.

RLMapping of Eyes With Retinal Pathologic Lesions

We examined a convenience sample of 14 eyes with diabetic retinopathy, 2 eyes after complete vitrectomy, and 2 eyes with multiple drusen. Three of the diabetic eyes were subjected to focal argon laser photocoagulation and examined at different intervals before and after treatment. The other 4 diabetic eyes were examined twice at 3-month intervals. This sample of eyes with varied chorioretinal pathology was chosen to represent (1) localized retinal pathologic lesions, eyes with nonproliferative diabetic retinopathy; (2) newly induced and well-defined localized alterations of the BRB, diabetic eyes subjected to focal photocoagulation; and (3) clinical situations appropriate to test the effect of chorioretinal fluorescence levels on vitreous fluorescence measurements and the efficiency of the correction algorithm, vitrectomized eyes and eyes with multiple drusen (Figure 3).

RLMapping of Eyes With Focal Retinal Disease

Fourteen eyes with diabetic retinopathy that showed localized vascular lesions in the posterior pole were examined with the RLA. Vascular lesions included dilated retinal capillaries and microaneurysms that showed different degrees of leakage on both fluorescein angiography and RLA imaging. Increased BRB permeability was measured in sites of morphological vascular abnormalities and in areas where no retinal pathologic lesions could be identified (Figure 2). The BRB permeability ranged from 1.4 to 15.0 × 10−6 cm/s. Four eyes were reexamined 3 months later, showing varying degrees of increased BRB permeability. The pattern of distribution of increased BRB permeability, represented by "hot spots" in the false-color RLMaps, showed good correlation between the 2 examinations.

RLMapping of Diabetic Eyes Submitted to Focal Photocoagulation to Estimate Lateral Resolution

To document lateral resolution of RLA measurements, we applied 100-µm argon laser photocoagulation spots to 3 eyes with diabetic macular edema. We used different patterns and different spacing between the photocoagulation burns. The eyes were measured before photocoagulation, 3 days after photocoagulation at the expected height of fluorescein leakage, and 21 days after, when healing would have occurred. In every case, the RLMaps obtained 3 days after photocoagulation clearly identified central zones of the photocoagulation burns as sites of increased fluorescein leakage. This finding was true even in the eye where four 100-µm photocoagulation spots were placed next to each other in the center of a circinate ring, demonstrating a lateral resolution on the order of 75 to 100 µm (Figure 3). The RLMaps obtained at 15, 30, and 60 minutes after fluorescein injection, at the height of increased BRB permeability, ie, 3 days after photocoagulation, correlated well with the expected distribution of fluorescein in the vitreous after a localized breakdown of the BRB. Comparison between the 30-minute RLMaps obtained before photocoagulation and 3 days and 21 days after photocoagulation illustrated well the known evolution of retinal fluorescein leakage resulting from photocoagulation. The BRB permeability values before and 21 days after photocoagulation were similar and in the order of 5 × 10−6 cm/s in the photocoagulated area. At 3 days after photocoagulation, the BRB permeability index reached a peak about 10 times higher than the previous value.

RLMapping of 2 Vitrectomized Eyes and Eyes With Multiple Drusen to Test the Effect of Chorioretinal Fluorescence Levels on Vitreous Fluorescence Measurements
Vitrectomized Eyes.

RLMaps were obtained from 2 vitrectomized eyes (Figure 3). The distribution of fluorescence in the vitreous appeared uniform—independent of the interval after fluorescein injection—demonstrating rapid mixing of fluorescein in the vitreous cavity in the absence of vitreous. Fluorescence levels of the choroid and retina, which were not uniform, did not influence fluorescence levels in the vitreous, demonstrating that the correction algorithm leaves only real fluorescence levels in the vitreous.

Eyes With Multiple Drusen.

These eyes were chosen because they are characterized by high levels of fluorescence in the choroid and retina in the presence of minimally affected BRB. The BRB permeability index values in regions with high levels of chorioretinal fluorescence were similar to those obtained from regions facing lower levels of chorioretinal fluorescence. This showed the efficacy of the correction algorithm in eliminating the influence of higher fluorescence levels in the choroid and retina on fluorescence measurements in the vitreous (Figure 3).

Reproducibility of RLMapping.

Intravisit Reproducibility. The mean reproducibility of points on RLMaps obtained on the same examination was ±10.2%.

Intervisit Reproducibility.

The reproducibility of points on the average maps per visit was ±13%. This value was the average across all 30-minute RLMaps from the 5 healthy subjects and included all points. To evaluate local reproducibility, namely, the reproducibility at each point on maps obtained in 2 different visits, the coefficient of variation was plotted as a frequency distribution for 1 healthy individual, selected because the mean reproducibility was 18%—closest to the overall average.

COMMENT

We have developed a new method to perform localized measurements of retinal fluorescein leakage using a confocal scanning laser ophthalmoscope modified to obtain fluorescence measurements in the vitreous. The instrument described herein (the RLA) allows localized measurements of fluorescein leakage into the vitreous in the immediate vicinity of the retina and simultaneous high-resolution imaging of the retina. We believe that simultaneous retinal imaging and localized vitreous fluorometry measurements offer the first opportunity to correlate changes in the BRB with retinal pathologic lesions. Localized vitreous fluorometry performed with this instrument offers a major advantage over previously available vitreous fluorometry instrumentation that could only give averaged information without precise knowledge of its location. Multiple measurements of retinal leakage using the RLA can be graphically assembled in a false-color RLMap that represents the distribution of localized alterations in the BRB in any chosen area of the total 3150 × 2700 µm of the posterior pole under examination.

RLMaps of the central macular region of a series of healthy eyes showed the topographic distribution of the inward permeability of the BRB and its dynamics. The maps show, in a healthy eye, fluorescein penetration into the vitreous over the entire area under examination, with more fluorescein leakage over the vascularized retina and less over the foveal avascularized zone. This distribution is progressively less clear when comparing measurements performed 15, 30, and 60 minutes after fluorescein administration, demonstrating the diffusion of fluorescein in the vitreous and allowing an estimation of the fluorescein diffusion coefficient of the vitreous under examination.

The permeability index of the BRB at 30 minutes in a series of 7 healthy eyes showed values ranging from a mean of 1.3 ± 0.4 × 10−6 cm/s over the foveal avascular zone to a mean of 2.2 ± 0.6 × 10−6 cm/s over vessels in the retina, with a mean of 1.9 ± 0.4 × 10−6 cm/s for the central macular area (3150 × 2700 µm).

These values are higher than those reported for the permeability index using the Fluorotron Master,4 but this may be explained by taking into consideration the fact that no deconvolution has been performed and our values are obtained from the posterior vitreous, which is in the immediate vicinity of the retina.

The RLMaps and permeability index on which these values are based were tested in vivo in healthy eyes for intravisit and intervisit reproducibility and the correction procedures in a series of extreme clinical situations.

Axial resolution was tested in diabetic eyes with localized pathologic lesions, in vitrectomized eyes, and in eyes with highly fluorescent drusen. Lateral resolution was tested in diabetic eyes by measuring retinal fluorescein leakage after performing localized 100-µm green argon laser photocoagulation burns of the diseased retina at different distances from each other.

Our observations showed that the correction procedure used to eliminate the influence of high levels of fluorescence in the choroid and retina on vitreous measurements was effective. The correction procedures involve subtraction for each measurement scan of an initial scan—obtained within the first 5 minutes after the fluorescein administration—in each individual eye after automatic alignment.

To verify that fluorescence measurements made in the vitreous are not influenced by fluorescence levels registered in the choroid and retina, 3 specific clinical situations were chosen: (1) multiple drusen, which have characteristically high fluorescence levels in the choroid and retina and are associated with minimal alteration of the BRB; (2) vitrectomized eyes, which are characterized by rapid mixing of fluorescein in the vitreous cavity in the presence of a highly abnormal BRB; and (3) diabetic eyes with nonproliferative retinopathy and localized retinal pathologic lesions with varying background levels of fluorescence in the choroid and retina.

The eyes with drusen showed clearly that retinal leakage was minimal and not dependent of chorioretinal fluorescence levels. This dissociation between vitreous fluorescence values and chorioretinal fluorescence was also well demonstrated in vitrectomized eyes. These eyes showed a near uniform distribution of fluorescence in the vitreous as early as 15 minutes after fluorescein administration—independent of the variations in background fluorescence due to chorioretinal pathologic lesions.

The RLMaps of diabetic eyes showed also that fluorescence levels measured in the vitreous are not influenced by increased chorioretinal fluorescence, when present.

Lateral resolution of the RLMaps was found to be in the order of 75 to 100 µm, with the RLA being able to discriminate the central leaking areas from the 100-µm laser photocoagulation burns placed next to each other. RLMaps obtained before and 3 days and 3 weeks after photocoagulation correlated well with present knowledge of the course of alterations of the BRB after laser photocoagulation. Maximal alteration of the BRB was recorded 3 days after photocoagulation, with good recovery after 3 weeks.

Intravisit and intervisit reproducibility of the RLA measurements were 10.2% and 13%, respectively, indicating that, in a clinical setting, the RLA is capable of measuring fluorescence levels in the vitreous near the retina, allowing for accurate and reproducible mapping of BRB to fluorescein in the human eye.

The method described allows localized measurements of BRB permeability that discriminate leaking sites of approximately 75 to 100 µm in size. This capability offers unique opportunities to examine the focal alteration of the BRB that occurs in a variety of retinal diseases. Simultaneous imaging of the retina is another important feature of the RLA. It is now possible to measure BRB permeability in specific locations of the fundus where there are retinal abnormalities and to follow both changes in retinal pathologic lesions and modifications in the BRB. An area of retinal edema may be characterized as cytotoxic or vasogenic, depending on its association with a localized breakdown of the BRB identified by the RLA. Information obtained with the RLA, together with information from examining the same retinal location with the Retinal Thickness Analyzer5 or Optical Coherence Tomography6 is, therefore, extremely promising.

Diabetic retinopathy, which has been shown to be associated with an early alteration of the BRB, appears immediately as a natural candidate for the application of this method. RLMapping may contribute to our understanding of the natural history of BRB alterations in diabetes and its correlation with microaneurysm formation, hard exudates, and other retinal lesions.

The RLA, by performing quantitative RLMapping and simultaneous imaging of the retina, offers a novel approach to examine the role of BRB breakdown in the development and progression of retinal pathologic lesions.

Back to top
Article Information

Accepted for publication November 6, 1998.

Reprints: Conceição L. Lobo, MD, MSc, Center of Ophthalmology, University Hospital and Institute for Biomedical Research in Light and Image, 3000 Coimbra, Portugal (e-mail: clobo@imagem.ibili.uc.pt).

References
1.
Wise  GNDollery  CTHenkind  P The Retinal Circulation.  New York, NY Harper & Row1971;143- 165421- 455
2.
Cunha-Vaz  JGFaria Abreu  JRCampos  AJFigo  GM Early breakdown of the blood-retinal barrier in diabetes. Br J Ophthalmol. 1975;59649- 656Article
3.
Cunha-Vaz  JG Vitreous fluorophotometry. Osborne  NNGhader  GJeds.Progress in Retinal Research. Oxford, England Pergamon Press1985;90- 114
4.
Chahal  PSChowienczyk  PJKohner  EM Measurement of blood-retinal barrier permeability: a reproducibility study in normal eyes. Invest Ophthalmol Vis Sci. 1985;26977- 982
5.
Asrani  SZeimer  RGoldberg  MFZou  S Application of rapid scanning retinal thickness analysis in retinal diseases. Ophthalmology. 1997;1041145- 1151Article
6.
Puliafito  CAHee  MRLin  CP  et al.  Imaging of macular disease with optical coherence tomography. Ophthalmology. 1995;102217- 229Article
×