Drusen with Accompanying Fluidunderneath the Sensory Retina

Bartosz L. Sikorski, MD,1,2 Danuta Bukowska, MS,2 Jakub J. Kaluzny, MD, PhD,1 Maciej Szkulmowski, PhD,2

Andrzej Kowalczyk, PhD,2 Maciej Wojtkowski, PhD2

Purpose: To investigate whether confluent drusen may be accompanied by fluid accumulation underneaththe sensory retina and to determine if the detection of subretinal fluid on spectral-domain optical coherencetomography (OCT) in patients with coalescent drusen is indicative of choroidal neovascularization (CNV).

Design: Prospective, noncomparative case series.Participants: Seventy-four eyes of 57 patients with large, confluent drusen.Methods: The retinal structure of patients with coalescent drusen was studied by spectral-domain OCT.

Optical coherence tomography reflectivity and outer retina topography maps were created and compared withfluorescein angiography (FA) and indocyanine green angiography (ICGA) images as well as with microperimetry.

Main Outcome Measures: Optical coherence tomography-derived retinal morphologic features.Results: What appears to be fluid beneath the sensory retina was found on spectral-domain OCT in 8 eyes

of 7 patients. The outer retina topography maps demonstrated that fluid accumulates only in the concavitybetween clustering soft drusen, not on their outward slopes. The maps also revealed a reduced distance betweenthe retinal pigment epithelium (RPE) and the photoreceptor inner/outer segment (IS/OS) junction over largedrusen and tiny elevations of the IS/OS junction around drusen of all sizes. Microperimetry showed decreasedretinal light sensitivity at the site of diminished distance between the RPE and the IS/OS junction. Seven eyes of6 patients who were followed up were found to have no retinal changes other than confluent drusen along withsubretinal fluid during the entire observational period (12–27 months). There was no evidence of CNV on FA orICGA in any of the patients.

Conclusions: Large, confluent drusen may be accompanied by subretinal spaces that appear to be filledwith fluid. Specific distribution of the fluid limited to the depression between adjacent drusen may indicate thatthe cluster of coalescent drusen produces mechanical strain to the outer retinal layers that locally pulls thesensory retina away from its normal position. Consequently, the appearance of fluid within subretinal compart-ment between coalescent drusen in OCT cross-sectional images may not be a reliable marker for the presenceof CNV.

Financial Disclosure(s): Proprietary or commercial disclosure may be found after the references.

Ophthalmology 2011;118:82–92 © 2011 by the American Academy of Ophthalmology.

Age-related macular degeneration (AMD) occurs as 2 dis-tinct forms, nonexudative and exudative, each representinga different stage of the same disease process. An earlynonexudative form is associated with the presence of softdrusen, which are localized deposits of membranous debrissituated between the retinal pigment epithelium (RPE) basallamina and inner collagenous layer of Bruch’s membrane.As this material continues to accumulate, such drusen mayenlarge and coalesce. This process has been linked to higherrates of progression to an advanced form of nonexudativeAMD represented by geographic atrophy.1–3 The formationof confluent drusen also constitutes a significant risk factorfor the development of the exudative form of AMD. Thelatter condition is characterized by choroidal neovascular-ization (CNV) that invades the subretinal space, leading toexudation of subretinal fluid and hemorrhage.1–3 The loca-tion of the leakage that is producing the subretinal fluidaccompanying CNV can be determined with the aid offluorescein angiography (FA) and indocyanine green an-

giography (ICGA). A supplementary approach to the iden-

82 © 2011 by the American Academy of OphthalmologyPublished by Elsevier Inc.

tification of subretinal fluid is provided by optical coherencetomography (OCT). This method demonstrates collection ofthe transparent fluid as a weakly scattering area surroundedby strongly scattering tissue. The backscattered light intensitywithin this region usually is smaller than the noise level (Drex-ler W, Ko TH, Sattmann H, et al. Clinical feasibility of ultra-high resolution ophthalmic optical coherence tomography. In-vest Ophthalmol Vis Sci 2002;43:E-Abstract 264).4–6 Thecontrast in optical reflectivity between nonreflective fluidand the more reflective posterior boundary of the neurosen-sory retina allows detection of even a small amount ofsubretinal fluid.7 Because FA and ICGA examine the cir-culation of the retina and choroid, whereas OCT recon-structs the retinal morphologic features, the best imagingstrategy to detect the CNV is to perform all these teststogether.

There is still special interest paid to the earlier identifi-cation of CNV, since most treatment options have poorvisual outcomes after severe AMD development. In the

context of recent advances in basic sciences and the intro-

ISSN 0161-6420/11/$–see front matterdoi:10.1016/j.ophtha.2010.04.017

Sikorski et al � Drusen with Accompanying Fluid underneath the Sensory Retina

duction of new pharmacologic agents, the prompt detectionof CNV would be particularly beneficial for reducing visionloss resulting from AMD.8,9 One possible way of searchingfor markers of an early stage of CNV is imaging andmonitoring over time patients without CNV but who are athigher risk for developing CNV.

As the activity of CNV was found to be associated withintraretinal and subretinal fluid accumulation on OCT, itseems a reasonable assumption that the detection of sub-retinal fluid in patients diagnosed with confluent drusen maybe a potential marker of CNV presence.10,11 Therefore, thisstudy investigates whether nonexudative AMD in the formof large confluent drusen—that constitute a significant riskfactor for CNV development—can be accompanied by thefluid underneath the sensory retina and whether the detec-tion of this is indicative of CNV. Recently, Stopa et al12

briefly mentioned the presence of very small foci of re-duced-light backscattered intensity on OCT in a patient withcoalescent drusen that may represent subretinal fluid. How-ever, they did not discuss this issue and did not present afollow-up study. They also did not perform ICGA. To thebest of the authors’ knowledge, there have been no reportsgiving the evidence for the possibility of fluid accumulationunderneath the sensory retina in patients with confluentdrusen.

Because high-speed OCT instruments such as spectral-domain OCT enable visualizing 3-dimensional retinal mor-phologic features without motion artifacts, they can be usedvery effectively for detecting and tracing subretinal fluiddistribution.7,13 Thus, this study applied laboratory high-speed spectral-domain OCT device to search for fluid-filledintraretinal and subretinal spaces in patients with confluentdrusen. To investigate the spatial distribution of potentialfluid spaces with respect to drusen topography, a newmethod of 3-dimensional OCT data analysis and visualiza-tion was used. This was performed with the aid of a custom-made automated segmentation software. Additionally, acommercially available manual segmentation tool was usedto create straightforward 3-dimensional visualization of theouter retina morphology and to verify findings based on thenew processing software. Results were compared with thoseof FA, ICGA, and microperimetry.

Patients and Methods

Tomographic images presented in this article were obtained by theprototype high-resolution spectral-domain OCT system developedand constructed at Nicolaus Copernicus University, Poland.14,15

The instrument provides 3-dimensional images with 4.5-�m axial,15-�m transverse resolution, and acquisition speed of 25 000 linesper second. The device uses Broadlighter D830 light source (Su-perlum, Moscow, Russia) emitting light of central wavelength 830nm and 70 nm of full width at half maximum. The optical powerincident on the eye is 750 �W.

To unveil subtle disturbances in the retina, 2 measurementprotocols were used. To collect general information on the struc-tural changes, 35 cross-sectional images, each consisting of 3000A-scans, covering the area of 6 � 3 mm, were obtained. Anotherprotocol provided data for creating contour and reflectivity mapsof individual OCT layers as well as for a 3-dimensional recon-

struction of the outer retina architecture. The latter one measured

200 cross-sectional images each consisting of 400 A-scans, cov-ering the area of 6 � 6 mm. The scanning was performed in rasterpattern, enabling uniform reconstruction of the fundus image. Rawdata were stored on the hard drive and were used for the furtherprocessing.

The reconstruction of the outer retinal structure was performedin the following steps. First, the outline of the complex of the junctionbetween the photoreceptor inner segment and outer segment (IS/OS)and the retinal pigment epithelium (RPE) in all 200 spectral-domain OCT cross-sectional images was marked automatically.In the next step, each 2-dimensional cross-sectional image wasrealigned to flatten the outer contour of the retina (ORC) whilepreserving the original radial dimensions of the tomogram. Thisprocedure was performed by custom-designed software, whichhas been described in details elsewhere.14 –16 All flattened 2-dimensional cross-sectional images then were imported into theAmira 4.1 (Visage Imaging GmbH, Berlin, Germany) visualiza-tion and modeling system to perform a straightforward reconstruc-tion of the 3-dimensional structure of the retina. The inner retinallayers were removed for better visualization of the outer retinadisturbances. The boundaries of hyporeflective dark space underthe neural retina in all cross-sectional images were marked man-ually. The volume of delimited space was displayed in false bluecolor.

For additional detailed investigation of observed structuralchanges, we independently performed analysis of the outer retinausing the automated segmentation software. The detailed descrip-tion of the method of reflectivity and topographic contour mapconstruction has been published elsewhere.15–17 The most atten-tion was paid to 3 contour maps displaying the distances betweenthe ORC and RPE (RPE topography), between the RPE and IS/OSjunction (the IS/OS-RPE thickness), and between the ORC andIS/OS junction (IS/OS topography). The RPE topography mapreveals zones of increased distance between the ORC and RPEresulting from sub-RPE deposits. By assigning a different color toeach height range, this map can depict the 3-dimensional structureof drusen. Areas where the RPE is not raised and overlays the ORCare registered in black. The IS/OS-RPE thickness map demon-strates regions where the sensory retina is separated from under-lying RPE. If fluid accumulates in the subretinal space, it elevatesthe sensory retina, and thus can be showed in this map as a thickerregion indicated in red. The normal distance between the RPE andIS/OS junction is displayed in green. The IS/OS topography mapcombines information from both previous maps, facilitating eval-uation of RPE elevations with respect to areas of increased dis-tance between the RPE and IS/OS junction. This article alsopresents so-called reflectivity maps displaying summed reflectivityof the IS/OS junction, outer photoreceptor segments, RPE, and asmall portion of the choroid to show precisely the horizontal extentof external retinal layers disturbances.

The 3-dimensional reconstruction of the outer retina along withreflectivity and contour maps were compared with FA, ICGA(Topcon TRC 50DX/Type IA; Topcon, Tokyo, Japan), and micro-perimetry (MP-1; Nidek Technologies, Padova, Italy) images. Forthe microperimetric study, the following parameters were used: ared cross as a fixation object, a white monochromatic backgroundat 4 asb, stimulus size Goldman III with 200 ms projection time,and a 4-2-1 double-staircase strategy.

Seventy-four eyes of 57 patients with large confluent drusenwere included in the study. Examination took place at the Depart-ment of Ophthalmology, Nicolaus Copernicus University, betweenOctober 2005 and December 2008. Clinical diagnosis of confluentdrusen was determined by slit-lamp examination. Patients with anyother concomitant macular pathology were excluded from thestudy. In particular, none of the enrolled patients had clinical

evidence of retinal hemorrhage, lipid exudates, fibrosis, and

83

Ophthalmology Volume 118, Number 1, January 2011

subretinal fluid. All patients were subsequently evaluated withspectral-domain OCT. The identification of subretinal fluid wasmade based on backscattering.18–23 As optical scattering is aproperty of a heterogeneous medium and occurs because of mi-croscopic spatial variations in the refractive index within tissue,18

the poorly reflective (hyporeflective) dark spaces located betweenthe RPE and sensory retina that were composed of opticallyhomogenous material were identified as containing gel or fluid.Eight eyes of 7 patients (3 men and 4 women; age range, 65–80years; mean age�standard deviation, 72.2�4.5 years) with unmis-takable poorly reflective (hyporeflective) dark space under the sensoryretina on spectral-domain OCT examination were selected for furtheranalysis and underwent FA and ICGA. Seven eyes of 6 patients werefollowed up for a mean of 16.5 months (range, 12–27 months). Theyadditionally underwent microperimetric examination.

As the term CNV is essential for the study, it requires a morespecific description. Choroidal neovascularization was defined asbeing present based on certain patterns of leaking on FA accordingto the modified Macular Photocoagulation Study grading protocolused in the Photodynamic Therapy Investigation and Verteporfinin Photodynamic Therapy Trial.24,25

The examination was performed under the tenets of the Dec-laration of Helsinki, the protocol was approved by the EthicsCommittee of Nicolaus Copernicus University, and signed in-formed consent was obtained from all the study participants.

Results

Based on spectral-domain OCT cross-sectional images, the regulardome-shaped elevations of the RPE were found, along with theoptically clear space beneath the sensory retina that had the ap-pearance of fluid, in 8 eyes of 7 subjects (Figs 1D, 2D, 3D). Seveneyes of 6 patients who were followed up were found to havenothing but confluent drusen along with the fluid during the entireobservational period (Figs 1G, 2G, 3G). Selected representativecases are presented below.

Case Reports

Case 1

A 71-year-old woman with coalescent drusen in both eyes soughttreatment at the ophthalmology department. Her visual acuity inthe right eye was 0.9. A spectral-domain OCT examination re-vealed multiple regular confluent elevations of the RPE. In thedepression between them, the hyporeflective material consistentwith subretinal fluid accumulation was observed. Fluorescein an-giography showed staining of drusen without any trace of a leak.The ICGA images displayed hypofluorescent areas at the site ofdrusen. Twelve months later, drusen increased in size. However,the visual acuity remained 0.9 and spectral-domain OCT showed astable amount of subretinal fluid. There was no evidence of CNVon FA or ICGA (Figs 1, 4A, 5).

Case 2

A 77-year-old man had coalescent drusen in the left eye. His visualacuity was 0.7. Spectral-domain OCT revealed fused elevations ofthe RPE with fluid-filled space between them. On FA, drusenexhibited hyperfluorescence. The ICGA images displayed hy-pofluorescent areas at the site of drusen. After 15 months, visualacuity improved to 0.8 and spectral-domain OCT demonstrated

disappearance of the prominent druse in the temporal macula. The

84

amount of the fluid also decreased. There was no evidence of CNVon FA and ICGA (Figs 2, 4B, 6).

Case 3A 71-year-old woman with coalescent drusen in both eyes wasreferred to the ophthalmology department. Her visual acuity in theleft eye was 0.9. On spectral-domain OCT examination, drusenwith fluid accumulation under the sensory retina was shown.Fluorescein angiography revealed nonleaking hyperfluorescent le-sions corresponding to drusen. The ICGA images demonstratedhypofluorescence of drusen. Twelve months later, her visual acuitywas 0.8. Spectral-domain OCT showed the presence of a newfocus of subretinal fluid resulting from growth of drusen. Therewas no evidence of CNV on FA or ICGA (Figs 3, 4C, 7).

Detailed Description of Imaging ResultsVolumetric rendering of outer retinal layers plainly demonstratedthat fluid-filled spaces were situated in the depression betweencontiguous drusen (Figs 1E, 2E, 3E, 1H, 2H, 3H). The superiorside of each fluid-filled compartment was formed by the sensoryretina extending between the peaks of adjacent drusen, which isclearly visible on spectral-domain OCT cross-sectional images(Figs 1D, 2D, 3D, 1G, 2G, 3G). Laterally, the spaces werebounded by drusen edges, which can be delineated by looking atthe reflectivity maps (Figs 1F, 2F, 3F, 1I, 2I, 3I). The depressionsare visible either as bright small islands or peninsulas surroundedby the black area related to the basal part of the drusen. The outerretina contour maps, presented in Figures 1F, 2F, and 3F and inFigures 1I, 2I, and 3I, also showed that fluid accumulated only inthe concavity between clustering soft drusen, not on their outwardslopes. The RPE topography maps demonstrated the localization ofeach depression in the center of the cluster of coalescent largedrusen (white arrows). The IS/OS-RPE thickness maps revealedelevation of the IS/OS junction right over the concavity betweendrusen consistent with fluid accumulation (black arrows) and thereduced distance between the RPE and IS/OS junction at the site ofthe large drusen (blue areas). These maps also showed tiny eleva-tions of the IS/OS junction around drusen of all sizes (blackarrowheads). The IS/OS topography maps depicted the mutuallocation of drusen and fluid (black arrows). Microperimetry dis-played a decrease in retinal light sensitivity in the areas corre-sponding to the position of coalescent drusen (Fig 4). There was noevidence of CNV on FA and ICGA in any of the patients (Figs 1B,2B, 3B, 1C, 2C, 3C, 5–7).

Discussion

Soft drusen appear in spectral-domain OCT cross-sectionalimages as regular, dome-shaped focal elevations of theRPE.26 Choroidal neovascularization is demonstrated bymore irregular detachment of the RPE and indirect signs andfindings related to its abnormal exudation and the accumu-lation of fluid within or underneath the sensory retina, orboth.9,10,27–29 Thus, elevation of the RPE with spaces hav-ing the appearance of subretinal fluid that were observed inspectral-domain OCT images from the patients suggestedthe presence of CNV.

All 7 patients with regular merging dome-shaped eleva-tions of the RPE and hyporeflective dark space under thesensory retina in spectral-domain OCT images had good

visual acuity and showed no leakage on FA. There also was

Sikorski et al � Drusen with Accompanying Fluid underneath the Sensory Retina

Figure 1. Case 1: right eye of a 71-year-old woman with nonexudative age-related macular degeneration. A, Color fundus photograph. B, Late-phasefluorescein angiogram showing a staining of drusen. C, Indocyanine green angiography (ICG) with hypofluorescent spots corresponding to drusen.D, Spectral-domain optical coherence tomography (SD-OCT) cross-sectional image. Gray arrow points to the subretinal fluid. The thickened detachedphotoreceptor outer segments are marked with curved arrow. E, Volumetric reconstruction of retinal pigment epithelium (RPE) surface showing drusentopography; subretinal fluid is marked in blue. F, Group of SD-OCT contour maps. Case 1: at the 12-month follow-up. G, SD-OCT cross-sectional image.Gray arrow points to the subretinal fluid. The thickened detached photoreceptor outer segments are marked with curved arrow. H, Volumetricreconstruction of RPE surface showing the progression of drusen coalescence (white arrows). I, Group of SD-OCT contour maps. The reflectivity and RPEtopography maps demonstrate increased horizontal and vertical extent of drusen. The inner segment/outer segment junction (IS/OS)-RPE thickness and

IS/OS topography maps show no apparent change in fluid volume that is located in the depression between clustering drusen.

85

Ophthalmology Volume 118, Number 1, January 2011

Figure 2. Case 2: left eye of a 77-year-old man with nonexudative age-related macular degeneration. A, Color fundus photograph. B, Late-phasefluorescein angiogram showing a staining of drusen. C, Indocyanine green angiogram (ICG) with hypofluorescent spots corresponding to drusen. D,Spectral-domain optical coherence tomography (SD-OCT) cross-sectional image. Gray arrow points to the subretinal fluid. The thickened detachedphotoreceptor outer segments are marked with curved arrow. E, Volumetric reconstruction of retinal pigment epithelium (RPE) surface showing drusentopography; subretinal fluid is marked in blue. F, Group of SD-OCT contour maps. Case 2: at the 15-month follow-up. G, Spectral-domain OCTcross-sectional image demonstrating disappearance of the prominent druse in the temporal macula (white arrow). Gray arrow points to the subretinal fluid.The thickened detached photoreceptor outer segments are marked with curved arrow. H, Volumetric reconstruction of RPE surface showing drusentopography. White arrow points to the site where the druse has disappeared. The amount of subretinal fluid is decreased significantly. I, Group of(SD-OCT) contour maps. The reflectivity and RPE topography maps demonstrate the reduced horizontal and vertical extent of drusen. The innersegment/outer segment junction (IS/OS)-RPE thickness map reveals normalization of the previously reduced distance between the RPE and IS/OS

junction (asterisk).

86

Sikorski et al � Drusen with Accompanying Fluid underneath the Sensory Retina

Figure 3. Case 3, left eye of a 71-year-old woman with nonexudative age-related macular degeneration. A, Color fundus photograph. B, Late-phasefluorescein angiogram showing a staining of drusen. C, Indocyanine green angiography (ICG) with hypofluorescent spots corresponding to drusen.D, Spectral-domain optical coherence tomography (SD-OCT) cross-sectional image. Gray arrow points to the subretinal fluid. The thickened detachedphotoreceptor outer segments are marked with curved arrow. E, Volumetric reconstruction of retinal pigment epithelium (RPE) surface showing drusentopography; subretinal fluid is marked in blue. F, Group of SD-OCT contour maps. Case 3: at 12-month follow-up. G, SD-OCT cross-sectional image.Gray arrow points to the subretinal fluid. The thickened detached photoreceptor outer segments are marked with curved arrow. H, Volumetricreconstruction of RPE surface showing slight confluence of drusen (white arrows) with new focus of subretinal fluid.I, Group of SD-OCT contour maps.The reflectivity and RPE topography maps demonstrate the small increase in horizontal extent of drusen. The inner segment/outer segment junction

(IS/OS)-RPE thickness and IS/OS topography maps demonstrate that fluid is located in the concavities between clustering drusen.

87

to the fixation target. dB � decibels.

Ophthalmology Volume 118, Number 1, January 2011

88

no evidence of hot spots or plaques on ICGA. Six of them,all of whom were followed up, demonstrated almost un-changed vision and no significant progression on FA andICGA after 12 to 27 months. This indicates a diagnosis ofsoft confluent drusen rather than CNV. Because drusen areconsidered to be a nonexudative form of AMD, the questionarises as to the source of fluid detected under the sensoryretina in these patients. The FA examination did not revealany disturbances of the retinal circulation, so fluid does notleak from blood vessels of the retina. Another potentialmechanism by which subretinal fluid may occur would bethe presence of a retinal break that could allow liquefiedvitreous to gain access under the sensory retina, separatingit from the RPE. As there was no evidence of retinal break,this mechanism can be ruled out as well. Hence, the mostreasonable explanation for fluid accumulation underneaththe neural retina in these patients is that drusen may producemechanical strain to the outer retinal layers that pulls awaythe sensory retina from its normal position.

Analysis of spectral-domain OCT cross-sectional imagesshowed that the neural retina is originally stretched outbetween 3 noncollinear points. The upper 2 points areconstituted by drusen peaks, whereas the lower one is lo-cated at the site of contact between adjacent drusen. Asdrusen grow, the RPE gradually lifts off the Bruch’s mem-brane. Thus, the distances between upper points and thelower point become longer. This can entail a tensile stresswithin the outer sensory retina. When the magnitude of theresultant upward force acting at the lower point overcomesthe RPE-sensory retina adhesiveness, retinal detachmentmay occur. The sequence of these events can be calledtenting of the overlying retina by large adjacent drusen.Because solitary soft drusen do not touch each other, theydo not possess a lower point, and therefore they cannot pullthe sensory retina away in such a strong fashion from itsnormal position. For the same reason, fluid is not present onthe outward slopes of wreathlike configuration of clusteringsoft drusen.

To perform detailed topographic analysis of all cross-sectional images in the entire set of OCT data, contour mapsof the outer retina were created. The RPE topography mapsshowed the same spatial distribution of drusen across themacula as seen on fundus photographs, proving the capa-bility of automated outer retina structure analysis to depictthese lesions adequately. The IS/OS-RPE thickness andIS/OS topography maps confirmed that the sensory retinaremains attached to drusen, except in the area within thedepression between them. Additionally, the IS/OS-RPEthickness map showed that drusen of all sizes are sur-rounded by tiny elevations of the IS/OS junction. This isprobably caused by drusen stretching and elevating thesensory retina at their bases. Interestingly, the IS/OS-RPEthickness map also revealed the reduced distance betweenthe RPE and IS/OS junction over the large drusen that couldcorrespond to photoreceptor layer thinning. The latter find-ing was likewise documented by the spectral-domain OCTstudy carried out by Schuman et al.30 They found thatdrusen height had a much stronger correlation than drusenwidth, with the extent of photoreceptor thinning over drusen

Figure 4. Microperimetric examination in patients presented in Figures 1through 3 superimposed on color fundus photographs: (A) case 1, (B) case2, and (C) case 3. Retinal sensitivity significantly decreases at the site oflarge confluent drusen. In cases 2 and 3 (B, C), microperimetry also revealsarea of small dense scotoma (empty red squares). A red cross corresponds

locations. This result seems to support the assumption that

resen

Sikorski et al � Drusen with Accompanying Fluid underneath the Sensory Retina

drusen can produce mechanical strain on the outer retinallayers. The higher drusen would include the greater stresswithin the outer sensory retina. In this context, the areas ofreduced distance between the RPE and IS/OS junction dem-onstrated in this study over the large drusen could be re-lated, at least partially, to the strain causing photoreceptorsdeflection. Moreover, it was also observed that after disap-pearance of drusen (Fig 2I), the previously reduced distancebetween the RPE and IS/OS junction normalized what couldbe explained by stress relief within the sensory retina thatrecovered its original configuration. This explanation also isconsistent with the findings demonstrated by Johnson et al31

on histologic sections of human eyes. They identified short-ening of photoreceptor outer segments overlying drusen andconcluded that both inner and outer photoreceptor segmentswere physically deflected. Because the reduced distancebetween the RPE and IS/OS junction should affect macularfunction, microperimetric examination also was performedand its results were compared with the outer retinal struc-ture. In all cases, microperimetry revealed decreased retinalsensitivity at the sites of reduced distance between the RPEand IS/OS junction seen on the IS/OS-RPE thickness maps.Similar findings of significant diminution in macular sensi-

Figure 5. Early, middle, and late frames from the fluorescein angiogram p

Figure 6. Early, middle, and late frames from the fluorescein angiogram presen

tivity over large soft drusen despite normal visual acuitywere also reported by Midena et al.32

Interestingly, spectral-domain OCT cross-sectional im-ages also demonstrated that detached photoreceptor outersegments at the site of subretinal fluid accumulation exhib-ited structural features analogous to those observed withinthe serous retinal detachment in central serous chorioreti-nopathy.33,34 They appeared thicker and had increasedbackscattering as well as an irregular outer profile (Fig 1D,2D, 3D, 1G, 2G, 3G), possibly indicative of a disruption innormal photoreceptors metabolism resulting from their sep-aration from the underlying RPE.

In contrast to the soft confluent drusen, an exudativeform of AMD, which can cause the rapid and profoundvisual impairment, requires an urgent therapy. Early diag-nosis of an active CNV is crucial for its effective treatment.Therefore, much effort has been expended on finding hall-marks of an active neovascularization process. The activityof CNV was found to be associated with intraretinal andsubretinal fluid accumulation related to exudation frompathologic vessels.10,11 As large, confluent drusen also canbe accompanied by subretinal spaces that appear to be filledwith fluid, as shown in this study, they can be misdiagnosed

ted in Figure 1B showing staining of drusen without any trace of a leak.

ted in Figure 2B revealing staining of drusen with no leakage.

89

Ophthalmology Volume 118, Number 1, January 2011

by OCT as active CNV. Although no one has reliably testedwhether the treat-if-there-is-fluid paradigm is correct, manydoctors treat if they see fluid. Thus, it is important toidentify the characteristics that will distinguish these 2entities. The following features of OCT cross-sectional im-ages may help to recognize large confluent drusen accom-panied by fluid accumulation under the neural retina: (1)merging regular dome-shaped elevations of the RPE, (2)hyporeflective dark space beneath the sensory retina in thedepression between RPE elevations that does not exceedtheir peaks, (3) an intact sensory retina, (4) relatively goodvisual acuity unless other conditions causing diminishedvision are present (e.g., cataract, optic nerve atrophy), and(5) no leakage on FA.

When interpreting OCT images, one should rememberthat the evolution of drusen is a dynamic process and apatient may be encountered who seeks treatment at the stageof drusen shrinking. In this case, fluid distribution can beatypical, because the height and morphologic features ofdrusen peaks are changed significantly (Fig 2G).

The results that we presented herein also show that atleast in some cases, 3-dimensional spectral-domain OCTimaging is more sensitive than FA for the detection of thesubretinal fluid collection. If the passage of the dye betweencirculation and the pathologic region is very slow or there isno active leakage, fluid-filled space can be visualized as astructural rather than functional change. Therefore, it isidentified more easily by topographic imaging with OCTthan by FA. From this point of view, it could be theorizedthat there may be CNV in these cases but that it does notappear on FA or ICGA. If it is not leaking, its appearance onOCT need not correlate with angiographic CNV. More than30 years ago, Green and Key35 demonstrated histologicallythat subjects in whom FA does not reveal the presence ofCNV nevertheless can have CNV that has not yet invadedthe retina. In another study recently conducted in mice,Takeda et al36 showed that in regions of the fundus that aresilent on FA, a novel in vivo imaging technique with CCR3(chemokine [C-C motif] receptor-3)-targeting quantum dots

Figure 7. Early, middle, and late frames from the fluorescein angiogramcorresponding to drusen.

is capable of detecting the CNV that is still limited to the

90

choroid. However, such an indolent CNV would not belikely to produce the subretinal fluid that was seen in thecurrent patients. It is also worth noting that the pocket ofsubretinal fluid that accompanied the cluster of soft drusenis located within the depression between them and seems tobe dependent only on their size and spatial configuration,whereas fluid overlying an active CNV can be present at anysite around it in an unpredictable manner.

Interestingly, unlike previous research on confluentdrusen,37,38 this study did not identify any sign of plaqueson ICGA in any of the studied eyes. This apparent discrep-ancy most likely can be explained by the fact that the felloweyes of these patients showed no exudative AMD, whereasthe other authors studied the ICGA findings in fellow eyeswith drusen of patients with unilateral exudative AMD ordisciform scars.

Finally, further research on the reflectivity of both con-fluent drusen and subretinal fluid is necessary. As differentspectral-domain OCT drusen ultrastructures may be precur-sors of various local changes, the investigation of the rela-tionship between the morphologic patterns of drusen thatare accompanied by subretinal fluid and a potential risk forprogression to an advanced form of AMD may have aconsiderable impact on monitoring such patients. In thisstudy, these particular drusen presented relatively uniform,although not perfectly homogeneous, internal reflectivity.However, due to the small sample size, it remains an openquestion whether this appearance is typical of these kinds ofdrusen. Another interesting issue that should be addressed infuture research is whether the reflectivity pattern of thesubretinal fluid accompanying coalescent drusen could beused as a feature to distinguish it from the typical exudativefluid seen in association with CNV lesions. As this is along-standing subretinal fluid, its reflectivity pattern maynot necessarily be the same in all patients. For instance, it isnoteworthy that in the case shown in Figure 2D, the signalof the fluid is slightly stronger than that of the vitreous,which could suggest that the fluid has some density.

This study has demonstrated that large confluent drusen

sented in Figure 3B demonstrating nonleaking hyperfluorescent lesions

pre

can be accompanied by hyporeflective spaces under the

Sikorski et al � Drusen with Accompanying Fluid underneath the Sensory Retina

sensory retina that appear to represent fluid. As this fluid isnot the result of exudation from new blood vessels, the termnonexudative form of AMD is still adequate in this case.However, these findings imply that the presence of fluidbeneath the neural retina is not a specific attribute of exu-dative form of AMD and may also be associated with thenonexudative form of this disease. In this context, it is veryimportant to realize that the detection of fluid below thesensory retina in conjunction with prominent regular RPEelevations in OCT cross-sectional images does not automat-ically indicate an active neovascularization process. Theprecise mechanism of fluid accumulation that accompaniessoft coalescent drusen requires further detailed studies. Fu-ture work also is needed to identify whether the presence ofthis fluid increases the risk of developing geographic atro-phy, CNV, or both.

References

1. van Leeuwen R, Klaver CC, Vingerling JR, et al. The risk andnatural course of age-related maculopathy: follow-up at 6 1/2years in the Rotterdam study. Arch Ophthalmol 2003;121:519–26.

2. Wang JJ, Foran S, Smith W, Mitchell P. Risk of age-relatedmacular degeneration in eyes with macular drusen or hyper-pigmentation: the Blue Mountains Eye Study cohort. ArchOphthalmol 2003;121:658–63.

3. Klein R, Klein BE, Knudtson MD, et al. Fifteen-year cumu-lative incidence of age-related macular degeneration: the Bea-ver Dam Eye Study. Ophthalmology 2007;114:253–62.

4. Hee MR, Puliafito CA, Wong C, et al. Optical coherencetomography of central serous chorioretinopathy. Am J Oph-thalmol 1995;120:65–74.

5. Srinivasan VJ, Wojtkowski M, Witkin AJ, et al. High-definitionand 3-dimensional imaging of macular pathologies with high-speed ultrahigh-resolution optical coherence tomography. Oph-thalmology 2006;113:2054–65.

6. Srinivasan VJ, Monson BK, Wojtkowski M, et al. Character-ization of outer retinal morphology with high-speed, ultra-high-resolution optical coherence tomography. Invest Oph-thalmol Vis Sci 2008;49:1571–9.

7. Rivellese MJ, Puliafito CA. Optical coherence tomography inthe diagnosis and management of posterior segment disorders.In: Bouma BE, Tearney GJ, eds. Handbook of Optical Coher-ence Tomography. New York: Marcel Dekker; 2002:471–86.

8. Rohrer B, Guo Y, Kunchithapautham K, Gilkeson GS. Elim-inating complement factor D reduces photoreceptor suscepti-bility to light-induced damage. Invest Ophthalmol Vis Sci2007;48:5282–9.

9. Wong RW, Richa DC, Hahn P, et al. Iron toxicity as apotential factor in AMD. Retina 2007;27:997–1003.

10. Salinas-Alaman A, Garcia-Layana A, Maldonado MJ, et al.Using optical coherence tomography to monitor photody-namic therapy in age related macular degeneration. Am JOphthalmol 2005;140:23–8.

11. Sayanagi K, Sharma S, Yamamoto T, Kaiser PK. Comparisonof spectral-domain versus time-domain optical coherence to-mography in management of age-related macular degenerationwith ranibizumab. Ophthalmology 2009;116:947–55.

12. Stopa M, Bower BA, Davies E, et al. Correlation of patho-logic features in spectral domain optical coherence tomog-raphy with conventional retinal studies. Retina 2008;28:

298 –308.

13. Wojtkowski M, Srinivasan V, Fujimoto JG, et al. Three-dimensional retinal imaging with high-speed ultrahigh-resolution optical coherence tomography. Ophthalmology2005;112:1734 – 46.

14. Szkulmowski M, Wojtkowski M, Sikorski B, et al. Analysis ofposterior retinal layers in spectral optical coherence tomogra-phy images of the normal retina and retinal pathologies.J Biomed Opt 2007;12:041207.

15. Wojtkowski M, Sikorski BL, Gorczynska I, et al. Comparisonof reflectivity maps and outer retinal topography in retinaldisease by 3-D Fourier domain optical coherence tomography.Opt Express 2009;17:4189–207.

16. Kaluzny JJ, Wojtkowski M, Sikorski BL, et al. Analysis of theouter retina reconstructed by high-resolution, three-dimensionalspectral domain optical coherence tomography. Ophthalmic SurgLasers Imaging 2009;40:102–8.

17. Sikorski BL, Wojtkowski M, Kaluzny JJ, et al. Correlation ofspectral optical coherence tomography with fluorescein andindocyanine green angiography in multiple evanescent whitedot syndrome. Br J Ophthalmol 2008;92:1552–7.

18. Schuman JS, Puliafito CA, Fujimoto JG. Optical CoherenceTomography of Ocular Diseases. 2nd ed. Thorofare, NJ:SLACK; 2004:21–43.

19. Eter N, Spaide RF. Comparison of fluorescein angiographyand optical coherence tomography for patients with choroidalneovascularization after photodynamic therapy. Retina 2005;25:691–6.

20. Jaffe GJ, Caprioli J. Optical coherence tomography to detectand manage retinal disease and glaucoma. Am J Ophthalmol2004;137:156–69.

21. Rogers AH, Martidis A, Greenberg PB, Puliafito CA. Opticalcoherence tomography findings following photodynamic ther-apy of choroidal neovascularization. Am J Ophthalmol 2002;134:566–76.

22. Sahni J, Stanga P, Wong D, Harding S. Optical coherencetomography in photodynamic therapy for subfoveal choroi-dal neovascularisation secondary to age related maculardegeneration: a cross sectional study. Br J Ophthalmol2005;89:316 –20.

23. Keane PA, Liakopoulos S, Ongchin SC, et al. Quantitativesubanalysis of optical coherence tomography after treatmentwith ranibizumab for neovascular age-related macular degen-eration. Invest Ophthalmol Vis Sci 2008;49:3115–20.

24. Macular Photocoagulation Study Group. Subfoveal neovascu-lar lesions in age-related macular degeneration: guidelines forevaluation and treatment in the Macular PhotocoagulationStudy. Arch Ophthalmol 1991;109:1242–57.

25. Treatment of Age-Related Macular Degeneration with Photo-dynamic Therapy Study Group, Verteporfin in PhotodynamicTherapy Study Group. Photodynamic therapy of subfovealchoroidal neovascularization with verteporfin: fluorescein an-giographic guidelines for evaluation and treatment—TAP andVIP report no. 2. Arch Ophthalmol 2003;121:1253–68.

26. Yi K, Mujat M, Park BH, et al. Spectral domain optical coher-ence tomography for quantitative evaluation of drusen and asso-ciated structural changes in non-neovascular age-related maculardegeneration. Br J Ophthalmol 2009;93:176–81.

27. Hee MR, Baumal CR, Puliafito CA, et al. Optical coherencetomography of age-related macular degeneration and choroi-dal neovascularization. Ophthalmology 1996;103:1260–70.

28. Coscas F, Coscas G, Souied E, et al. Optical coherence to-mography identification of occult choroidal neovasculariza-tion in age-related macular degeneration. Am J Ophthalmol2007;144:592–9.

29. Sandhu SS, Talks SJ. Correlation of optical coherence tomog-

raphy, with or without additional colour fundus photography,

91

Ophthalmology Volume 118, Number 1, January 2011

with stereo fundus fluorescein angiography in diagnosingchoroidal neovascular membranes. Br J Ophthalmol 2005;89:967–70.

30. Schuman SG, Koreishi AF, Farsiu S, et al. Photoreceptor layerthinning over drusen in eyes with age-related macular degen-eration imaged in vivo with spectral-domain optical coherencetomography. Ophthalmology 2009;116:488–96.

31. Johnson PT, Lewis GP, Talaga KC, et al. Drusen-associateddegeneration in the retina. Invest Ophthalmol Vis Sci 2003;44:4481–8.

32. Midena E, Vujosevic S, Convento E, et al. Microperimetry andfundus autofluorescence in patients with early age-related mac-ular degeneration. Br J Ophthalmol 2007;91:1499–503.

33. Matsumoto H, Kishi S, Otani T, Sato T. Elongation of pho-toreceptor outer segment in central serous chorioretinopathy.

Am J Ophthalmol 2008;145:162–8.

ogy Annual Meeting, May 2009, Fort Lauderdale, Florida.

92

34. Piccolino FC, de la Longrais RR, Ravera G, et al. The fovealphotoreceptor layer and visual acuity loss in central serouschorioretinopathy. Am J Ophthalmol 2005;139:87–99.

35. Green WR, Key SN III. Senile macular degeneration: a his-topathologic study. Trans Am Ophthalmol Soc 1977;75:180–254.

36. Takeda A, Baffi JZ, Kleinman ME, et al. CCR3 is a target forage-related macular degeneration diagnosis and therapy. Na-ture 2009;460:225–30.

37. Brancato R, Trabucchi G, Introini U, Avanza P. Hyperfluo-rescent plaque lesions in the late phases of indocyanine greenangiography: a possible contraindication to the laser treatmentof drusen. Am J Ophthalmol 1997;124:554–7.

38. Hanutsaha P, Guyer DR, Yannuzzi LA, et al. Indocyanine-greenvideoangiography of drusen as a possible predictive indicator of

exudative maculopathy. Ophthalmology 1998;105:1632–6.

Footnotes and Financial Disclosures

Originally received: May 13, 2009.Final revision: April 8, 2010.Accepted: April 9, 2010.Available online: August 11, 2010. Manuscript no. 2009-640.

1 Department of Ophthalmology, Nicolaus Copernicus University, Byd-goszcz, Poland.

2 Institute of Physics, Nicolaus Copernicus University, Torun, Poland.

Presented in part at: Association for Research in Vision and Ophthalmol-

Financial Disclosure(s):The author(s) have made the following disclosure(s):Maciej Szkulmowski - Consultant - Optopol SAAndrzej Kowalczyk - Consultant - Optopol SA

Supported by EuroHORCs-European Science Foundation EURYI AwardEURYI-01/2008-PL (M.W.).

Correspondence:Bartosz L. Sikorski, MD, Department of Ophthalmology, Nicolaus Coper-nicus University, 9 M. Sklodowskiej-Curie St., Bydgoszcz 85-094, Poland.

E-mail: sikorski@doctors.org.uk.