Analysis of Progression of Reticular Pseudodrusen by Spectral Domain–Optical Coherence Tomography

Abstract

Purpose.: To analyze reticular pseudodrusen progression using spectral domain–optical coherence tomography (SD-OCT).

Methods.: Thirty-three consecutive patients (48 eyes) underwent SD-OCT using the eye-tracked follow-up protocol 24 ± 2 months after baseline examination. Each pair of B-scans (only one per eye was evaluated among those showing pseudodrusen progression) was compared with respect to pseudodrusen appearance and retinal layer structure. Stage 1 pseudodrusen was defined as granular material between the RPE and the inner segment/outer segment (IS/OS), stage 2 as mounds of material sufficient to alter the contour of the IS/OS, stage 3 as thicker material adopting a conical appearance and breaking through the IS/OS, and stage 4 as fading of the material because of reabsorption and migration within the inner retinal layers.

Results.: A total of 78 pseudodrusen (detected on the 48 analyzed B-scans, and counting for a mean of 2.3 pseudodrusen per scan) showed progression over a mean of 23.9 ± 1.2 months. All 58 pseudodrusen (100%) graded as stage 1 at baseline examination progressed to stage 2. Thirteen of 16 pseudodrusen (81.3%) graded as stage 2 at baseline examination progressed to stage 3, and three (18.7%) progressed to stage 4. All four pseudodrusen (100%) graded as stage 3 at baseline examination progressed to stage 4. Among pseudodrusen that were stage 3 or 4 at follow-up (n = 20), 100% had IS/OS disruption whereas 12.1% (n = 7) had IS/OS disruption at stage 1 or 2 (n = 58) (OR, 1.736; 95% CI, 1.02–2.43).

Conclusions.: The frequency of stage changes over time suggest that reticular pseudodrusen are dynamic pathologic structures.

In 1990 we described reticular pseudodrusen as a peculiar yellowish pattern in the macula of patients with age-related macular degeneration (AMD). ¹ In the original description we called this peculiar form of drusen les pseudodrusen bleus because of their enhanced visibility using blue light. ¹ More recently, using spectral domain–optical coherence tomography (SD-OCT), we and other researchers ^{2 –4} demonstrated that discrete collections of hyperreflective material located not under (as typical of drusen in AMD) but above the retinal pigment epithelium (RPE) were associated with reticular pseudodrusen. 

Zweifel et al. ³ suggested that the hyperreflective material as visualized by SD-OCT could be graded by the thickness of the accumulation above the RPE and proposed a defined grading system of three stages. According to their grading system, stage 1 is characterized by diffuse deposition of granular hyperreflective material between the RPE and the boundary between the inner segments (IS) and the outer segments (OS) of the photoreceptors (the IS/OS boundary). In stage 2, mounds of accumulated material are sufficient to alter the contour of the IS/OS boundary. In stage 3, the material is thicker, adopts a conical appearance, and breaks through the IS/OS boundary. Despite its being well detailed and logical, this proposed grading system is limited because of its derivation from a cross-sectional analysis. Most important, the possible evolution of this material accumulated above the RPE, once it has reached the stage 3, remains to be elucidated. 

High-resolution SD-OCT (Spectralis; Heidelberg Engineering, Heidelberg, Germany) is a high-speed OCT system (up to 40,000 axial scans per second) using spectral/Fourier domain detection, with an axial image resolution of 7 μm. Real-time eye-tracking technology in this SD-OCT system allows not only precise orientation of the scan toward the region of interest but also rescanning of the region during follow-up. 

In this study, our purpose was to analyze the progression of reticular pseudodrusen (morphologic changes over time of the hyperreflective material located above the RPE, associated with reticular pseudodrusen) using the eye-tracker follow-up protocol of Spectralis SD-OCT. 

Patients and Methods

We performed Spectralis SD-OCT follow-up examination on all consecutive patients with reticular pseudodrusen who presented at the University Eye Clinic of Creteil between January 2011 and March 2011 and who had undergone Spectralis SD-OCT examination 24 ± 2 months earlier (baseline examination). All patients were recruited from the AMD outpatient clinic at the University Eye Clinic of Creteil and underwent complete ophthalmologic examination, including confocal scanning laser ophthalmoscope (cSLO) imaging (Heidelberg Retina Angiograph; Heidelberg Engineering) and tracked Spectralis SD-OCT, as part of their routine clinical workup. Inclusion criteria were reticular pseudodrusen defined by the peculiar yellowish reticular pattern at the macula, whose visibility was enhanced by infrared (IR) reflectance (as groups of hyporeflective lesions interspersed against a background of mild hyperreflectance). ^{2 –5} Exclusion criteria were choroidal neovascularization (CNV) in the study eye and signs of any other active retinal disease in the study eye such as retinal vascular (e.g., diabetic retinopathy, retinal vein occlusion) or vitreoretinal (e.g., vitreomacular traction syndrome, epiretinal membrane) disease. Informed consent was obtained, as required by the French bioethical legislation, in agreement with the Declaration of Helsinki for research involving human subjects. The University Paris Est Creteil Institutional Review Board approved this study. 

cSLO and SD-OCT Image Acquisition

Spectralis SD-OCT allows reliable detection and assessment of small changes over time with the use of cSLO technology to track the eye and guide the OCT device to the selected location. Using a previously selected reference scan, the Spectralis SD-OCT aligns the reference fundus image with the live patient fundus image at follow-up. The eye tracker recognizes the retina and then directs the SD-OCT scan to the same location. This eliminates the potential bias of subjective placement of the scan by the operator. 

All cSLO images and SD-OCT scans were acquired by three of the authors (GQ, FC, EHS) using the follow-up protocol of Spectralis SD-OCT 24 ± 2 months after baseline examination. At baseline examination, a minimum standardized imaging protocol was performed in all patients that included raster sets of high-resolution SD-OCT B-scans (19 scans, each composed of nine averaged EDI OCT B-scans at 240-μm intervals) used to image the macula and reticular pseudodrusen. Examination field size was 30° × 30°, and frequency of image acquisition was up to nine images per second. In addition, by using cSLO technology to track the eye in a subset of eyes at the baseline examination, further high-resolution single B-scans (each composed of up to 100 averaged EDI OCT B-scans) were guided from the pseudodrusen, as visualized on IR frames; thus, cSLO images were used as a real-time reference for locating the SD-OCT scan. 

SD-OCT Image Analysis

The Spectralis SD-OCT system provides in vivo details of the anatomy of the retina that nearly resemble histologic specimens (the optical resolution of the Spectralis SD-OCT is 7 μm in axial direction [z-axis]), but no strict correlations with histology have been demonstrated thus far. However, to describe the SD-OCT images, according to Pircher et al., ⁶ the following correspondence has been applied to the outer retinal layers: the innermost band reflects the external limiting membrane; a second band corresponds to photoreceptors IS/OS interface; a third band represents the RPE/OS interface; and the most external band, corresponds to RPE/Bruch's membrane complex. 

SD-OCT images consecutively collected were proportionally magnified for better visualization of intraretinal changes and evaluated with regard to microstructural changes over time. For SD-OCT scan selection, we looked at the 19 scans used to image the macula and any further single B-scan acquired. Given that this study was designed as a qualitative analysis of the reticular pseudodrusen progression, two expert retinal physicians (GQ, EHS) selected only one pair (both baseline and follow-up examination) of high-quality, reliably tracked SD-OCT scans per eye (the most representative of pseudodrusen progression). When available, high-resolution single B-scans (each composed of up to 100 averaged EDI OCT B-scans), guided from the pseudodrusen as visualized on the IR frame (cSLO images were used as a real-time reference for locating the SD-OCT scan), were chosen. 

Only lesions judged to show progression were selected. The pseudodrusen appearance and retinal layer structure observed in each pair (baseline and follow-up examination) of the selected SD-OCT scans were described, analyzed, and interpreted by two expert retinal physicians (GQ, EHS). Disagreement between readers regarding the detection of features was resolved by open adjudication. 

In the qualitative analysis of pseudodrusen appearance on SD-OCT (collections of hyperreflective material located above the RPE), we adopted the grading system proposed by Zweifel et al., ³ which, for the purpose of the present study, was modified from three stages to four stages. Briefly, stage 1 was defined as diffuse deposition of granular hyperreflective material between the RPE and the IS/OS boundary. Stage 2 was defined as mounds of accumulated material sufficient to alter the contour of the IS/OS boundary. Stage 3 was defined by thicker material that adopted a conical appearance and broke through the IS/OS boundary. Stage 4, the last stage of evolution, was defined by fading of the material because of reabsorption and, eventually, migration within the inner retinal layers. 

For each pair of scans (baseline and follow-up examination) included in the qualitative analysis, the stages of progression of each pseudodrusen (as visualized by SD-OCT) over the 24-month study period were recorded. In addition, for each pseudodrusen showing progression, changes in the IS/OS boundary (from intact [no discontinuity of IS/OS boundary] to disrupted [reduced intensity ± focal discontinuity of IS/OS boundary] or absent [definitive discontinuity of IS/OS boundary]) were investigated. 

Statistical analysis was performed using data analysis and statistical software (Stata, version 11; StataCorp, College Station, TX). Continuous data were expressed as mean (SD), and qualitative data were expressed as n (percentages). Best-corrected visual acuity (BCVA) was transformed to the logarithm of the minimum angle of resolution [LogMAR]). The concordance correlation coefficient (Pearson correlation) was calculated for interobserver correlations and considered strong if the correlation coefficient was >0.8. The Stuart-Maxwell test of marginal homogeneity for paired data was used to compare pseudodrusen stages from baseline to follow-up visit. Wilcoxon matched-pair signed-rank test was used to assess changes in BVCA (from baseline to follow-up visit). Exact logistic regression was performed to assess the association between progression to stage 3 or 4 and changes in IS/OS boundary. All tests were two sided. The chosen level of statistical significance was P < 0.05. 

Results

Patients Demographics and Clinical Characteristics

A total of 48 eyes of 33 consecutive patients (21 women [63.6%]; mean age, 80.5 ± 6.1 years; range, 66–95 years) were included in the analysis (Table 1). Eighteen eyes were excluded because of the presence of CNV. In all included eyes, reticular pseudodrusen were defined by the peculiar yellowish reticular pattern at the macula, whose visibility was enhanced using IR reflectance. Mean follow-up was 23.9 ± 1.2 months (range, 22–26 months). Mean BCVA changed from 0.58 (±0.22; range, 0.16–1) at the baseline visit to 0.54 (±0.21; range, 0.16–1) at the last visit (P = 0.19) (Table 1). Seven of 48 eyes of 5 of 33 patients with reticular pseudodrusen showed geographic atrophy within the macular area at baseline; 10 of 48 eyes of 8 of 33 patients showed geographic atrophy at the last visit. Five eyes of four patients with reticular pseudodrusen showed coincident pseudovitelliform material within the macular area at both the baseline and the last visit. 

Table 1.

 

Demographics of AMD Patients with Pseudodrusen and Evolution of BVCA and Pseudodrusen Characteristics

	Baseline	24-Month Follow-up	P
Patients (n = 33)
Age, mean (SD)	80.5 (6.1)	—	—
Female, n (%)	21 (63.6)	—	—
Eyes (n = 48)
BVCA	0.58 (0.22)	0.54 (0.21)	0.19*
Pseudodrusen (n = 78)
Stage, n (%)
1	58 (74.4)	0 (0)
2	16 (20.5)	58 (74.4)	<0.001†
3	4 (5.1)	13 (16.7)
4	0 (0)	7 (8.9)
IS/OS boundary disruption	—	27 (34.6)

*  Wilcoxon matched-pair signed-rank test.

†  Test of marginal homogeneity (Stuart-Maxwell).

Spectral Domain–Optical Coherence Tomography Analysis of Pseudodrusen Changes from Baseline to Last Visit

Overall, 48 pairs of high-quality SD-OCT B-scans (both baseline and follow-up examinations) of 48 eyes (33 patients) were included in the analysis. By using the eye-tracking feature of Spectralis SD-OCT that aligns the reference fundus image (baseline) with the live patient fundus image (follow-up examination), a total of 78 pseudodrusen (detected on the 48 analyzed B-scans, for a mean of 2.3 pseudodrusen per scan) were judged to show progression over a mean of 23.9 ± 1.2 months (Table 1) (P < 0.001). 

At baseline examination, 58 of 78 pseudodrusen (74.4%) were graded as stage 1, 16 of 78 (20.5%) were graded as stage 2, and 4 of 78 (5.1%) were graded as stage 3 (Table 1). At follow-up examination, 58 of 78 pseudodrusen (74.4%) were graded as stage 2, 13 of 78 (16.7%) were graded as stage 3, and 7 of 78 (8.9%) were graded as stage 4 (Table 1). Pseudodrusen grading in the selected SD-OCT B-scans was significantly correlated between the two readers for each stage (Pearson's coefficient, r = 0.83, P = 0.01). 

All 58 pseudodrusen (100%) graded as stage 1 at baseline examination showed progression to stage 2 at the follow-up examination (Figs. 1, 2); disruption of IS/OS was present in seven cases (P = 0.2), whereas no pseudodrusen graded as stage 1 at the baseline examination showed progression to stage 3 or 4 at the follow-up examination. Thirteen of 16 pseudodrusen (81.3%) graded as stage 2 at the baseline examination showed progression to stage 3 at the follow-up examination (Figs. 2, 3), with disruption of IS/OS in each case (P < 0.001), and 3 of 16 (18.7%) showed progression to stage 4 at the follow-up examination, with disruption of IS/OS in each case (P < 0.001); no pseudodrusen graded as stage 2 at baseline examination showed regression to stage 1. All (4 of 4) pseudodrusen (100%) graded as stage 3 at the baseline examination showed progression to stage 4 at the follow-up examination (Figs. 4 5–6), with disruption or absence of IS/OS in each case (P < 0.001); no pseudodrusen graded as stage 3 at baseline examination showed regression to either stage 2 or stage 1. 

Figure 1.

 

Left eye color fundus photography (top left), blue light (top middle), and IR reflectance (top right) frames of a 78-year-old woman (case 6) with reticular pseudodrusen showing the area investigated by SD-OCT scans at baseline examination and 25 months later. Enlarged views show stage 1 pseudodrusen (arrowheads, middle) that progressed to stage 2 over 25 months (arrowheads, bottom). Note the absence of progression for a stage 2 pseudodrusen (arrows, middle and bottom) over 25 months. Also note the intact IS/OS boundary overlying stage 1 and stage 2 pseudodrusen.

Figure 2.

 

SD-OCT scans of the right eye of an 87-year-old woman (case 8) with reticular pseudodrusen at baseline examination and 24 months later. Enlarged views show a stage 1 pseudodrusen (arrowhead, top) that progressed to stage 2 (arrowhead, bottom) over 24 months and a stage 2 pseudodrusen that progressed to stage 3 (arrowheads, top and bottom) over 24 months. Note the absence of progression for a stage 2 pseudodrusen (single arrows, top and bottom) and two stage 3 pseudodrusen (double arrows, top and bottom) over 24 months. Also note the disrupted IS/OS boundary overlying stage 3 pseudodrusen.

Figure 3.

 

SD-OCT scans of the left eye of a 77-year-old man (case 12) with reticular pseudodrusen and coincident pseudovitelliform material within the macula at baseline examination and 24 months later. Enlarged views show a stage 2 pseudodrusen (arrowhead, top) that progressed to stage 3 (arrowhead, bottom) over 24 months, accompanied by IS/OS boundary disruption. Note the thickening of the pseudovitelliform material within the macula (top and bottom) over 24 months.

Figure 4.

 

SD-OCT scans of the right eye of an 81-year-old woman (case 16) with reticular pseudodrusen at baseline examination and 26 months later. Enlarged views show a stage 3 pseudodrusen (asterisk, top) that progressed to stage 4 (asterisk, bottom) over 26 months. Note the absence of progression for three stage 3 pseudodrusen (arrows, top and bottom) over 26 months. Also note the disrupted IS/OS boundary overlying stage 3 and stage 4 pseudodrusen.

Figure 5.

 

SD-OCT scans of the left eye of an 81-year-old man (case 29) with reticular pseudodrusen at baseline examination and 24 months later. Enlarged views show two stage 2 pseudodrusen (arrowheads, top) that progressed to stage 3 (arrowheads, bottom) over 24 months and a stage 3 pseudodrusen (asterisk, top) that progressed to stage 4 (asterisk, bottom) over 24 months. Note the intact IS/OS boundary overlying stage 2 pseudodrusen and the disrupted IS/OS boundary overlying stages 3 and 4 pseudodrusen.

Figure 6.

 

SD-OCT scans of the right eye of an 85-year-old woman (case 24) with reticular pseudodrusen at baseline examination and 22 months later. Enlarged views show a stage 3 pseudodrusen (asterisk, top) that progressed to stage 4 (asterisk, bottom) over 22 months. Note the disrupted IS/OS boundary overlying stages 3 and 4 pseudodrusen.

Overall, a total of 27 pseudodrusen (34.6%) showing progression also showed changes in IS/OS boundary, from intact to disrupted or absent. Among pseudodrusen presenting stage 3 or 4 at the last visit (n = 20), 100% showed changes in IS/OS (Figs. 3 4 5–6), whereas among those not presenting stage 3 or 4 (n = 58), 12.1% (n = 7) showed changes in IS/OS (odds ratio = 1.736; 95% confidence interval, 1.02–2.43). All stage 4 pseudodrusen presented a disrupted (five pseudodrusen) or an absent (two pseudodrusen) IS/OS boundary (Figs. 4 5 6–7). 

Figure 7.

 

SD-OCT scans of the right eye of an 88-year-old man (case 20) with reticular pseudodrusen at baseline examination and 23 months later. Enlarged views show a stage 3 pseudodrusen (asterisk, top) that progressed to stage 4 (asterisk, bottom) over 23 months. Note the disrupted IS/OS boundary overlying the stage 3 pseudodrusen that became absent once it progressed to stage 4 (dotted arrow). Also note two stage 2 pseudodrusen (arrowheads, top) that progressed to stage 3 (arrowheads, bottom) over 23 months (arrows), accompanied by IS/OS boundary disruption.

Discussion

Recently, using SD-OCT, Zweifel et al. ³ described reticular pseudodrusen as discrete collections of hyperreflective material not under but above the RPE and suggested that the accumulated material could be graded by the thickness of the accumulation. The authors proposed a three-stage grading system for reticular pseudodrusen and showed that the thicker aggregates (stage 3) corresponded with the white punctate inclusions in the color photograph. In a previous study, ² we described a “target” aspect for reticular pseudodrusen and suggested that, for stage 3 (SD-OCT), the visualization of a core on integrated cSLO imaging (IR frames, fundus autofluorescence, and fluorescein angiography) could be explained by the build-up of pseudodrusen passing through different stages. However, the actual build-up and progression over time of pseudodrusen (as visualized by SD-OCT) has not been definitively evaluated in previous studies. Moreover, the possible evolution of stage 3 reticular pseudodrusen on SD-OCT remains to be elucidated. This issue is of particular interest given that it may help in the understanding of why reticular pseudodrusen represent a risk factor for late AMD. ^{7 –9}  

In this study, our aim was to analyze the progression, not the prevalence, of reticular pseudodrusen. We investigated changes in pseudodrusen appearance and retinal layer structure over time using the eye-tracker follow-up protocol of Spectralis SD-OCT, which allows reliable detection and assessment of small changes over time. In fact, by using a selected previous reference scan, the Spectralis SD-OCT aligns the reference fundus image with the live patient fundus image at follow-up. The eye tracker recognizes the retina and then directs the SD-OCT scan to the same location, thus eliminating subjective placement of the scan by the operator. The same scanning location during follow-up is not related to the location of the fixation light but to the eye-tracking system and its reference points. The reference points for the eye tracker are anatomic structures at the confocal scanning laser ophthalmoscope fundus image. As long as there is an overlap of the fundus image for the baseline and the follow-up examinations that allows the identification of enough reference points to align the two images automatically, automatic rescan is feasible. There might be a slight variation in the position of the follow-up scans compared with the baseline examination. Therefore, structural changes might also be due to differences in the scanning position. However, because of the active real-time eye tracker of Spectralis SD-OCT, in most cases, this variation in the position of the follow-up scans should be smaller than the size of the pseudodrusen lesions. 

In the current series, 48 pairs of SD-OCT B-scans (at both baseline and follow-up examinations) of 33 consecutive patients (48 eyes) showing reticular pseudodrusen on fundus biomicroscopy, confirmed by IR reflectance, were analyzed in the search for morphologic changes over a 24-month period of the hyperreflective material located above the RPE. Overall, we found that this material progressively thickened and adopted a conical appearance and that, later in its progression, the material faded. This allowed confirmation of the SD-OCT grading system proposed by Zweifel et al. ³ and proposal of a fourth stage in pseudodrusen progression, characterized by material reabsorption and, eventually, migration within the inner retinal layers. Interestingly, progression to stage 3 and stage 4 was always accompanied by significant changes in IS/OS boundary, and all stage 4 pseudodrusen showed a disrupted or absent IS/OS boundary. 

Our findings suggest that reticular pseudodrusen are dynamic pathologic structures whose progression on SD-OCT is characterized first by a continuous accumulation of focal material and later by reabsorption and, eventually, migration of the material within the inner retinal layers. We recently reported a similar progression for retinal flecks in Stargardt disease/fundus flavimaculatus. ^{10 –12} Of note, like Stargardt disease/fundus flavimaculatus, reticular pseudodrusen have been reported to be associated with the development of early-onset macular atrophy. ¹³ The IS/OS disruption and loss that characterize the advanced stages in pseudodrusen progression may explain why reticular pseudodrusen are associated not only with early onset macular atrophy but with geographic atrophy in AMD ¹⁴ (and thus represent a risk factor for late AMD ^{7 –9} ). 

Our study has several limitations. The series here presented is relatively small, and the study period was short. However, one must consider that SD-OCT is a relatively new technology, and, thus, it is not yet possible to have much information without longer follow-up. Moreover, only one high-quality scan per eye was selected and included in the analysis; therefore, only selected deposits were evaluated by SD-OCT. It is possible that other tomographic features were missed in our analysis. In addition, only pseudodrusen showing progression have been selected and included in the analysis, resulting in an opportunity for bias in unmasked analysis. Finally, we investigated reticular pseudodrusen changes over time using the eye-tracker follow-up protocol of Spectralis SD-OCT, which allows reliable detection and assessment of small changes over time, but we cannot be sure whether there was always point-to-point correlation because artifacts might have occurred that led to classification bias. However, we analyzed only high-quality tracked scans that were selected with regard to dominant anatomic reference points, such as blood vessels, to make sure the correlation was perfect and that eye movements were properly compensated for. In turn, correct placement of the OCT cross-sectional images during the follow-up visit was verified by the reader, who searched for the reference points; these are not supposed to change from visit to visit (the invariant reference points have been used to subjectively verify the accuracy of the follow-up images). In such conditions, because of the active real-time eye tracker of the Spectralis SD-OCT, this uncertainty is usually smaller than the size of each individual pseudodrusen. 

In summary, using the eye-tracker follow-up protocol of Spectralis SD-OCT, we demonstrated that reticular pseudodrusen are dynamic pathologic structures. Based on the frequency of stage changes over time, we suggest that the hyperreflective material located above the RPE and associated with reticular pseudodrusen progressively thickens. In the advanced stages, the material fades, accompanied by IS/OS disruption and loss. Such morphologic features and peculiar progression may give insight into the pathogenesis of reticular pseudodrusen and may help in the understanding of why it represents a risk factor for late AMD. 

Footnotes

 Disclosure: G. Querques, None; F. Canouï-Poitrine, None; F. Coscas, None; N. Massamba, None; L. Querques, None; G. Mimoun, None; F. Bandello, None; E.H. Souied, None