Graefe's Archive for Clinical and Experimental Ophthalmology

, Volume 247, Issue 6, pp 729–734

Macular autofluorescence in eyes with cystoid macula edema, detected with 488 nm-excitation but not with 580 nm-excitation

Authors

  • Kenichiro Bessho
    • Department of OphthalmologyOsaka University Medical School
    • Department of OphthalmologyOsaka University Medical School
    • Department of ophthalmologyOsaka University Medical School
  • Seiyo Harino
    • OphthalmologyYodogawa Christian Hospital
  • Miki Sawa
    • Department of OphthalmologyOsaka University Medical School
  • Kaori Sayanagi
    • Department of OphthalmologyOsaka University Medical School
  • Motokazu Tsujikawa
    • Department of OphthalmologyOsaka University Medical School
  • Yasuo Tano
    • Department of OphthalmologyOsaka University Medical School
Retinal Disorders

DOI: 10.1007/s00417-008-1033-y

Cite this article as:
Bessho, K., Gomi, F., Harino, S. et al. Graefes Arch Clin Exp Ophthalmol (2009) 247: 729. doi:10.1007/s00417-008-1033-y

Abstract

Background

Fundus autofluorescence (AF) derives from lipofuscin in the retinal pigment epithelium (RPE). Because lipofuscin is a by-product of phagocytosis of photoreceptors by RPE, AF imaging is expected to describe some functional aspect of the retina. In this study we report distribution of AF in patients showing macular edema.

Methods

Three eyes with diabetic macular edema (DME) and 11 with retinal vein occlusion (RVO), associated with macular edema (ME) were examined. ME was determined by standard fundus examination, fluorescein angiography (FA) and optical coherence tomography (OCT). AF was recorded using a Heidelberg confocal scanning laser ophthalmoscope (cSLO) with 488 nm laser exciter (488 nm-AF), and a conventional Topcon fundus camera with halogen lamp exciter and 580 nm band-pass filter (580 nm-AF). Color fundus picture, FA image and these two AF images were analyzed by superimposing all images.

Results

All subjects presented cystoid macular edema (CME) with petaloid pattern hyperfluorescence in FA. In 488 nm-AF, all eyes (100%) showed macular autofluorescence of a similar shape to that of the CME in FA. In contrast, in 580 nm-AF only one eye (7%) presented this corresponding petaloid-shaped autofluorescence. In all cases, peripheral retinal edemas did not show autofluorescence corresponding to the leakage in FA.

Conclusions

In eyes with CME, analogous hyperautofluorescence to the CME was always observed in 488 nm-AF, while it was rarely observed in 580 nm-AF. Moreover, this CME hyperautofluorescence was only seen in the macular area. We hypothesize that autofluorescence from CME may be considered as a “pseudo” or “relative” autofluorescence, due to macular stretching following CME that may result in lateral displacement of macular pigments (MPs) and subsequent reduction of MPs density, as MPs block 488 nm-AF more intensely than 580 nm-AF. Although this phenomenon may not directly indicate change of RPE function, it may be used as a method to assess or track CME non-invasively.

Keywords

Cystoid macular edemaFundus autofluorescenceHRA2580 nm excitation488 nm excitation

Introduction

Fundus autofluorescence (AF) has been studied clinically and histologically [112]. Autofluorescence of the fundus is thought to be mainly a property of the retinal pigment epithelium (RPE), particularly reflecting the existence of lipofuscin, a pigment that is a by-product of phagocytosis of photoreceptor outer segments [11, 1315]. Therefore, AF has been considered to reflect some aspect of RPE function and integrity. Interpretation of the AF image seems to have not been fully understood yet. In reports studying properties of the diseased retina using AF, for example in eyes with age-related macular degeneration (AMD), more intense AF was observed, even in the fellow eye of the patient with exudative AMD [5], in eyes with acute central serous retinopathy, increased AF was seen at the site of leakage and in the area of retinal detachment [6]. Classification of AF pattern has been attempted in early age-related maculopathy [16].

Development of the fundus imaging device has enabled us to obtain AF image easier in a clinical setting. Currently, the Heidelberg retina angiograph 2 (HRA2) and Topcon fundus camera are commercially available for obtaining AF images. The HRA2 uses 488 nm excitation and the Topcon camera uses excitation around 580 nm using a band-pass filter. There is a difference of excitation wavelength and resulting fluorescence between the two devices, and images by these devices do not necessarily show similar results. But both are still considered to be reflecting mostly fluorescence from lipofuscin, thanks to its wide emission spectrum.

In this report, we describe AF in retinas presenting with cystoid macular edema(CME), and differences in the images obtained on both instruments. This may indicate that hyperautofluorescence in CME may not always directly reflect RPE function.

Methods and subjects

Subjects

The study was conducted at Osaka University Hospital following the tenets of the Declaration of Helsinki. Fourteen eyes [three eyes with diabetic macular edema (DME) and 11 eyes with retinal vein occlusion (RVO)] with clinically significant macular edema (CSME) were retrospectively reviewed in this study. CSME was determined by standard fundus examination, fluorescein angiography (FA: Topcon TRC50IX, Topcon, Tokyo, Japan), and optical coherence tomography (OCT: Stratus or Cirrus, Carl Zeiss Meditec, Dublin, CA, USA).

Methods

Written consent was obtained before all examinations. Autofluorescence was recorded using two devices: a confocal scanning laser ophthalmoscope (cSLO: HRA2, Heidelberg Engineering, Vista, CA, USA) which was equipped with a 488 nm laser exciter and 500 nm barrier filter (488 nm-AF), and a conventional fundus camera (Topcon TRC50LX, Topcon, Tokyo, Japan) with a halogen lamp exciter with a 580 nm band-pass filter (bandwidth 500–610 nm) and a 695 nm barrier filter (bandwidth 675-715 nm) (580 nm-AF) [5]. In the HRA2, fundus images were taken as a movie sequence, and subsequently frame averaging was performed using internal software. Generally more than eight frames were averaged. In the Topcon camera, fundus image was taken as a single image.

To avoid contaminating fluorescence from residual dye of previous FA examination, AF examination was performed before FA when both examinations were planned at the same day, and at least 10 days after previous FA examination. Using computer software (Photoshop 7.0, Adobe, San Jose, CA, USA), FA image, color fundus image and these two AF images per subject were overlaid as separate layer. By changing transparency of layer, images were aligned with each other. By switching on and off each layer, location and size of lesion were compared to that of other layers. Analysis was performed by a single observer (KB) without knowing background condition of subjects.

Results

All fourteen subject eyes presented cystoid macular edema (CME); FA showed a petaloid-shaped leakage, and OCT showed cystic space within the macula in all eyes. The profile of all subjects and results are shown in Table 1.
Table 1

Profile of patients

 

Gender

Age

Eye

Diagnosis#

FT(micron)$

AF(488)*

AF(580)**

1

female

63

OD

BRVO

506

positive

negative

2

male

80

OS

BRVO

498

positive

negative

3

male

74

OD

BRVO

514

positive

negative

4

female

58

OS

BRVO

531

positive

negative

5

male

50

OS

BRVO

457

positive

negative

6

male

80

OD

BRVO

713

positive

negative

7

female

77

OS

BRVO

531

positive

negative

8

male

72

OS

BRVO

573

positive

negative

9

female

74

OS

BRVO

659

positive

negative

10

female

69

OD

CRVO

938

positive

negative

11

male

68

OD

CRVO

431

positive

positive

12

male

66

OS

DME

539

positive

negative

13

male

67

OD

DME

439

positive

negative

14

male

72

OS

DME

514

positive

negative

# BRVO: branch retinal vein occlusion, CRVO: central retinal vein occlusion, DME: diabetic macular edema

$ FT: foveal thickness measured using OCT 3000 (Zeiss)

* AF (488): macular autofluorescence corresponding to the cystoid macular edema detected by 488 nm excitation (HRA2)

** AF (580): macular autofluorescence corresponding to the cystoid macular edema detected by 580 nm excitation (Topcon camera)

In 488 nm-AF, all eyes (100%) showed macular autofluorescence in a petaloid pattern, similar to that of the CME in FA as shown in Figs. 1 and 2. In 580 nm-AF, this corresponding autofluorescence was not seen except for one case (6.6%). However, even in this case, contrast of petaloid pattern AF to background was much lower than that in 488 nm-AF (Fig. 2). In all cases, peripheral retinal edemas did not show increased AF. Lesions or retinal structures such as vessels or retinal hemorrhage were likely to block fluorescence from the RPE and thus appeared as hypofluorescent. Hard exudates appeared as both hyperautofluorescent lesion and hypoautofluorescent lesion and those AF changes were correspondent on both AF modalities.
https://static-content.springer.com/image/art%3A10.1007%2Fs00417-008-1033-y/MediaObjects/417_2008_1033_Fig1_HTML.gif
Fig. 1

Typical sample of the CME autofluorescence image of patient #13 (DME). a Color fundus picture. In FA (b), petaloid-shape macular cysts are visible in the foveal area. Macular edema can be confirmed in OCT (c) as well. N: nasal, T: temporal. In 488 nm-AF, macular autofluorescence analogous to that in FA is seen (d). Note other retinal area showing leakage in FA does not present this hyperautofluorescent change. In 580 nm-AF (e), this hyperautofluorescent change is not observed, while photocoagulation scars can be seen as hypoautofluorescent lesions. Hard exudates appear mostly as hypoautofluorescent lesions. FA was performed after the other examination so that residual fluorescein dye could not contaminate AF result

https://static-content.springer.com/image/art%3A10.1007%2Fs00417-008-1033-y/MediaObjects/417_2008_1033_Fig2_HTML.gif
Fig. 2

Case of both 488 nm- and 580 nm- positive AF image of patient #11 (CRVO). In FA (b) and 488 nm-AF, cystic compartment structure is obvious. In 580 nm-AF (e), the same lesion shows very mild hyperautofluorescence. FA was performed after the other examination so that residual fluorescein dye may not contaminate AF result

Discussion

Fundus autofluorescence (AF) has been studied, and is currently considered to be mainly derived from RPE lipofuscin, which is the by-product of phagocytosis of photoreceptors by the RPE [1, 2, 5, 8, 11], and thus AF is expected to reflect the function of RPE. Recent findings have shown that the precursors of lipofuscin located in the outersegment of photoreceptors such as A2PE, A2PE-H2 and A2-rhodopsin might induce the increased AF [17, 18].

Currently 488 nm and 580 nm excitation are available for obtaining AF. The emission spectrum of the RPE lipofuscin ranges between 500 nm and 750 nm with a peak of approximately 630 nm [11]. Optimal excitation is obtained when excited at 510 nm [11]. The HRA confocal device uses 488 nm laser excitation. 488 nm excitation has a potential problem that autofluorescence from crystalline lens may contaminate and degrade contrast of the resulting image. The confocal optics of the HRA has an advantage for avoiding this problem, and its internal frame-averaging software also help to improve the image. The conventional Topcon camera uses 580 nm excitation, which is naturally advantageous for avoiding lens autofluorescence [13]; thus, it doesn’t need the frame-averaging procedure. In addition, 580 nm excitation light may be less affected by macular pigments (MPs). MPs which mainly consist of lutein and zeaxanthine are located in the outer plexiform layer and partly in the inner plexiform layer, especially in highest density in the foveal area. MPs show broad range absorption of short wavelength light with a peak of 460 nm, with an absorption spectra ranging between 400 to 540 nm [19, 20]; therefore, excitation light at 488 nm is also blocked by MPs, resulting in insufficient excitation, which causes macular hypoautofluorescence, while 580 nm excitation is not likely to be influenced. It seems as yet unclear whether the responsible fundus fluorophore for these two wavelengths is exactly the same, although wide fluorescence spectrum of lipofuscin covers both wavelengths. A recent study investigating the spectrum of autofluorescence separately in four emission ranges suggested that lipofuscin was still a major fluorophore throughout all studied emission ranges, and there might be additional fluorophores with a short emission wavelength, which might be advanced glycation end products (AGEs), collagen or elastin [21]. In eyes with age-related macular degeneration, Bruch’s membrane and sub-RPE deposits also seem to act as additional fluorophores [22].

Cystoid macular edema (CME) is one of the common important causes related to visual loss in the retinal diseases such as diabetic retinopathy, posterior uveitis, etc. Retinal cysts are mainly located in the outer plexiform layer of Henle fiber and inner nuclear layer. While the precise mechanism of developing CME is still uncertain, currently it is assumed to be the result of polycystic expansion of the extracellular matrix by the breakdown of the blood–retinal barrier, or swelling of Müller cells, due to various pathologic mechanisms [23].

In this study, we observed increased AF by 488 nm excitation in eyes with CME. The configuration of increased AF was exactly the same, with the petaloid shape of hyperfluorescence on FA, and they also seemed correspondent with cysts on OCT. However, interestingly, the increased AF was not apparent by 580 nm excitation except in one eye; even in this eye, increased AF was not prominent, as shown in Fig. 2.

One of the possible explanations for these inconsistent AF findings between 488 nm and 580 nm excitation is that the increased AF on 488 nm is “pseudo” increased AF resulting from the reduction of the blockage of AF by MPs, similar to the “window defect” in fluorescein angiography. The intraretinal cysts may cause lateral displacement of the MPs, and the consequent decrease in focal density of MP alters absorption of AF. Since 580 nm-AF is known to have the advantage of describing macular autofluorescence without blockage by the MPs [5], reduction of the MPs density is less likely to influence it, and this may result in low contrast of cystoid lesions in 580 nm-AF. This was supported by the findings that retinal cysts located apart from fovea with low MP density did not show hyperautofluorescence, even in 488 nm-AF. Intensity of fluorescence was not analyzed in the current study; however, there was no CME case showing autofluorescence obviously stronger than that in any other extramacular retinal area. This may also support the idea that the CME autofluorescence may be analogous with the window defect.

In addition, macular edema with other types such as retinal swelling or subretinal fluid [24] may not show the apparent increased AF.

As another possibility, there may be some cyst-specific fluorophore inside the cysts which is excited by 488 nm more intensely than 580 nm. In Fig. 2, CME was visible even in 580 nm-AF. From the absorption spectra of MPs, natural blockage of 580 nm excitation light by MPs is not likely to be caused, thus this may be a case supporting the second hypothesis. This hyperautofluorescence was very mild, and histochemical investigation may be necessary to address this issue; however, it would still not be easy to explain why only cysts in macular area show hyperautofluorescence. Intensity analysis may also help to distinguish whether it is “true” autofluorescence.

Although further study such as the measurement of MP density is necessary to confirm our hypothesis, currently we consider this “relative” or “pseudo” autofluorescence as a possible mechanism for explaining this phenomenon. From these findings, AF analysis with 488 nm has the potential to be a non-invasive diagnostic tool of CME. Recently, McBain et al. reported the sensitivity of AF imaging for CME detection [25]. They used the HRA2 (488 nm-AF), and CME was detected with 81% sensitivity and 69% specificity compared with standard FA. They concluded that AF imaging might be a non-invasive diagnostic tool for detecting CME. For the diagnosis and management of CME, we agree that OCT already plays a major role, with the widespread prevalence and improving technology showing even precise three-dimensional retinal morphology non-invasively. Thus, AF is not likely to replace OCT, but might be another option for diagnosing or tracking CME. For developing hyperautofluorescence in CME, McBain et al. also assumed a similar mechanism to us.

Our result using both 488 nm and 580 nm excitation is consistent with their result, and may reinforce expanding understanding about the reading of AF images.

Finally, we should notice that the AF imaging may show different results, depending on the excitation wave length currently available.

Copyright information

© Springer-Verlag 2009