Small Dense Particles in the Retina Observable by Spectral-Domain Optical Coherence Tomography in Age-Related Macular Degeneration

Abstract

Purpose.: To observe detailed changes in neurosensory retinal structure after anti-VEGF upload in age-related macular degeneration (AMD), by using spectral domain optical coherence tomography (SD-OCT).

Methods.: The retinal structure was observed by using SD-OCT in 61 patients, before and 1 month after the third ranibizumab injection (upload phase). The main focus of attention was a subjective determination of the amount and behavior of the numerous small, dense particles (SDPs) frequently observed within the outer and inner neurosensory layers in eyes with neovascular AMD. The Spearman rho correlation was used for statistical analysis.

Results.: In all eyes, various amounts of SDPs were seen within the neurosensory layer of the foveal and parafoveal area. In 54%, the amount of SDPs became significantly less after ranibizumab therapy (stable, 41%; higher, 5%). SDP reduction correlated positively with the reduction of retinal disease according to OCT (P = 0.000), with central foveal thickness (P = 0.040), and with the improvement in best corrected visual acuity (BCVA; P = 0.006). The baseline amount of SDPs also correlated positively with the increase in BCVA (P = 0.005).

Conclusions.: The origin of the SDPs observable in SD-OCT is unknown, but they may represent migrating RPE cells or leukocytes, indicating a certain status of retinal inflammation. The amount of SDPs is substantially reduced after ranibizumab upload therapy and correlates positively with BCVA. Moreover, an initial large number of SDPs may indicate a higher grade of inflammation, but the presence of a high number enhances the effect of ranibizumab therapy. Thus, the amount of SDPs before treatment may be a predictive factor for the therapy's outcome.

Age-related macular degeneration (AMD), especially its wet form characterized by an abnormal growth of choroidal neovascularization (CNV) into the subretinal space, is still the leading cause of blindness in developed countries in people older than 50 years. ^1,2  

Since the advent of ranibizumab (Lucentis; Novartis, Basel, Switzerland), it has been possible to treat neovascular AMD and to achieve an improvement in visual acuity. Ranibizumab is a recombinant monoclonal antibody fragment that neutralizes all active forms of vascular endothelial growth factor (VEGF) A and has been shown in large clinical trials to inhibit the CNV in AMD significantly. ^{3 –5} Ranibizumab is FDA approved for all subtypes of choroidal neovascularizations (CNVs) due to AMD and is the most commonly used treatment method. It is initially intravitreally injected three times during a period of 2 months (upload phase) followed by a monitoring phase starting 4 weeks after the third injection. 

As a noninvasive and fast diagnostic tool for adequate follow-up of treated AMD, the technique of optical coherence tomography (OCT) is widely approved, whereas the center of attention in clinical studies is mainly the regression of retinal edema after therapy. ^{6 –9}  

With the advent of the high-resolution spectral-domain OCT technique (SD-OCT), a much-improved differentiation of the retinal structures and also of CNV became possible. ^{9 –11}  

To our knowledge, small dense particles (SDPs) have not been examined yet in detail, but they are frequently observed by SD-OCT in various amounts and distributions within the outer and inner neurosensory layers of the foveal and parafoveal regions in neovascular AMD (Fig. 1). The origin of the SDPs is unknown but may represent macrophages, indicating a possible inflammatory condition in AMD ¹² or migrating RPE cells. The purpose of this retrospective chart study was to examine the impact of SDP changes after upload therapy with ranibizumab in neovascular AMD and to correlate these changes with retinal disease derived from the SD-OCT findings and best corrected visual acuity (BCVA). 

Figure 1.

 

In both SD-OCT images, SDPs can be observed throughout the inner and outer neurosensory retinal layers. They are more condensed in the first image to the left of the large area of retinal edema (A) and are linearly distributed on the inner retinal layers toward the cystic space (B).

Material and Methods

Spectral-Domain OCT

OCT images were acquired with a combined SD-OCT and scanning laser ophthalmoscope (SLO) imaging system (Spectralis HRA+OCT; Heidelberg Engineering, Heidelberg, Germany). The system acquires en face SLO images in angiographic, autofluorescence (AF), and reflectance imaging modes, as well as cross-sectional SD-OCT images. An 870-nm superluminescent diode (SLD) is used for OCT imaging. In SD-OCT mode, the retina is scanned at 40,000 A-scans per second, presenting highly detailed images of the retinal structure. The OCT depth resolution (full width at half-maximum [FWHM]) is 7 μm. The images of the SLO and OCT modes can be overlaid and automatically spatially co-registered. Images can also be co-registered over time to visualize changes as the patient's treatment progresses. With this setup, similar OCT sections before and after therapy were obtainable throughout the study. 

Image Acquisition

All consecutive patients who had received a first-time diagnosis of neovascular AMD and were assigned to standard intravitreal ranibizumab therapy ¹³ underwent pre- and posttreatment SD-OCT examinations in a regular clinical setting. The protocol adhered to the guidelines set forth in the Declaration of Helsinki. The posttreatment examination took place usually 4 weeks after the third injection. OCTs of six different sections were performed with a cross-sectional technique. The same sections were scanned during the follow-up to ensure matching sections for evaluation. The images were displayed with the Heidelberg Eye Explorer software (Heidelberg Engineering). For SDP evaluation, similar sections from both images (pre- and posttreatment examination) were retrospectively evaluated. The amount of SDPs within the parafoveal area was subjectively graded in all patients by dividing them into three categories: 1, few, representing approximately 2 to 10 SDPs; 2, moderate, representing approximately 11 to 20 SDPs; and 3, many, representing approximately 21 or more SDPs (Fig. 2). 

Figure 2.

 

The amount and density of the SDPs was subjectively graded as few particles (A), a moderate amount of particles (B), and many particles (C).

Additional measurements included the best corrected visual acuity (BCVA) before and after treatment, the central retinal thickness (CRT) determined with the standard protocols of the Heidelberg SD-OCT software and the qualitative posttreatment judgment of retinal disease regarding the status of the edema: 1, stable, no change in edema; 2, improved, but still present; and 3, dry retina. 

Statistical analysis for correlations was performed with Spearman's rho and for paired samples with the t-test (SPSS for Windows 16.0; SPSS, Chicago, IL). 

Results

The overall BCVA improved from logMAR 0.60 (range, 0.1–1.3) to 0.50 (range, 0.1–1.7). The equivalent decimal BCVA change was 0.25 to 0.32, representing a significant ETDRS change (5 letters; P = 0.004). SD-OCT showed that retinal thickness improved from 416.7 ± 15.6 to 294.1 ± 11.8 μm (P = 0.000). The qualitative judgment of sub- and intraretinal fluid revealed dry retinal conditions in 49.2% (n = 30), improvement of fluid in 37.7% (n = 23), and no change in 13.1% (n = 8). 

At baseline in 18 eyes, the number of SDPs was judged to be few (grade 1), in 32 eyes to be moderate (grade 2), and in 11 eyes to be many (grade 3). Regarding the retinal changes after ranibizumab therapy, the amount of SDPs became fewer in 54% (n = 33), was stable in 41% (n = 25), and showed enhancement in 5% (n = 3). In cases of OCT-determined fluid improvement, the number of SDPs was clearly reduced (Table 1; 0.500; P = 0.000). There was also a significant correlation between SDP reduction and the reduction in OCT-measured CRT (0.268; P = 0.040) and BCVA improvement (0.349; P = 0.006). Four representative SD-OCT courses are displayed in Figures 3, through 6. 

Table 1.

 

Dependence of SDP Reduction on the Course of Fluid Change after Ranibizumab Upload, as Determined by SD-OCT

OCT Fluid	SDP
	Amount	%	n
Stable	Less	0	0
	Stable	87.5	7
	Enhanced	12.5	1
Subtotal			8
Improved	Less	43.5	10
	Stable	52.5	12
	Enhanced	4.0	1
Subtotal			23
Dry	Less	73.3	22
	Stable	23.3	7
	Enhanced	3.4	1
Subtotal			30
Total			61

 A clear correlation was observed between the amount of SDPs and fluid reduction in OCT (P = 0.000).

Figure 3.

 

(A) On both sides of the baseline image, SDPs can be observed lying lateral to the fluid formation. (B) After ranibizumab upload, the retina appears to be dry and SDP formation significantly reduced. BCVA improved from decimal 0.25 to 0.5.

Figure 4.

 

(A) Central edema of retinal structures accompanied with a moderate amount of SDPs (decimal BCVA, 0.5) can be observed in this case of occult CNV (note the CNV formation above Bruch's membrane in the sub-RPE space). (B) The amount of SDPs significantly decreased after ranibizumab upload, while the retinal appeared to be dry. BCVA did not improve but stayed stable.

Figure 5.

 

(A) Another case of occult CNV (BCVA, 0.5) revealed SDPs predominantly in the outer retinal layers. (B) After drug upload, the SDPs became fewer and BCVA improved to 0.6. However, despite the improvement in BCVA, subretinal fluid decreased only slightly.

Figure 6.

 

(A) In this case, OCT revealed significant subretinal fluid and a moderate amount of SDPs before treatment (BCVA, 0.5). (B) Despite the fact that subretinal fluid was still present after the upload therapy, BCVA improved to 0.8 and, accordingly, SDP formation appeared to be less.

A significant positive correlation was also found between the baseline amount of SDPs and the change in BCVA. Thus, if many SDPs (grade 3) were present before treatment, the improvement in BCVA was more enhanced (0.359; P = 0.005); however, no positive correlation was found between the number of SDPs before treatment and retinal thickness (0.050; P = 0.709) or the (0.077; P = 0.554) resolution of retinal edema, as shown by OCT. The correlation between SDPs and ETDRS visual acuity before treatment was 0.228 (P = 0.077). 

Discussion

OCT provides an indirect image of the retina and RPE, based on the reflective properties of the various cellular layers. The various intensities seen on the OCT scan correlate well with the different levels of the retina and RPE. ^14,15  

SD-OCT can be used to observe details of the disease in the central retinal structure. Thus, it was striking to observe SDPs within the inner and outer layers of the neurosensory retina in patients with neovascular AMD. The origin is unclear; however, these particles may reflect the leukocytes that invade the extracellular spaces in inflamed regions. An inflammatory component in neovascular AMD has been postulated and discussed in recent years, ¹² and the SDPs may be the in vivo expression of this condition, which can be observed by SD-OCT. On the other hand, these particles may also reflect migrating RPE cells. The RPE provides the most highly reflective surface and usually appears as a dense hyperreflective (color-coded bright orange-red) layer on the OCT scan. Thus, the SDPs may also be consistent with RPE proliferation and migration through the retina into the macular region. These findings suggest that the transretinal migration of proliferating RPE cells plays a role in the formation of intraretinal membranes in these patients. 

Similar SDPs are described in epiretinal membranes or after retinal detachment. RPE cells play an important role in the formation of epiretinal membranes after retinal detachment. ¹⁴ The RPE cells may access the inner retinal surface through the retinal break that caused the detachment. Subsequent proliferation of these cells on the retinal surface then contributes to the formation of the epiretinal membranes, which typically consist of a variety of cell types. The second mechanism by which the RPE cells may access the inner retinal surface is migration through the retina. The ability of proliferating RPE cells to migrate is well documented and can be altered by the cells' microenvironment. ^{15 –18}  

The proliferation of RPE cells occurs not infrequently after trauma and after repair of a retinal detachment. ^19,20 Experimental models of retinal detachment show that RPE cell proliferation begins early in the course of the detachment and is confined to the region of the detachment. ^21,22 Translocation of RPE pigment into the retina occurs in some diseases such as retinitis pigmentosa (bone spicula), as well as in experimental models of subretinal pigment clumping. ^23,24  

In our group of patients treated with ranibizumab loading-dose injections, it was interesting that SDPs were present in nearly all cases. This pathologic feature is usually not observable in healthy retinas. However, the amount and size of the SDPs and the exact location within the neurosensory retinal layers appears to be highly variable. The SDPs were observed within the neurosensory retinal structure itself, but also in the inner boundaries of the subretinal space in cases of enhanced subretinal fluid. Such cases of SDPs overlying drusen formation have also been described by the SD-OCT technique in dry forms of drusen maculopathy due to AMD. ^25,26 We also speculate that this entity may represent the progression of RPE cell migration into the retina. ^25,26  

With the Heidelberg Eye Explorer Software it was possible to examine exactly the same sections before and after therapy. It became obvious that the amount of SDPs was significantly reduced after treatment. This positive correlation could also have been observed for CRT reduction and BCVA improvement that could lead to the impression that besides the standard judgment of OCT-imaged subretinal fluid after ranibizumab intravitreal therapy, ^{9,27 –29} the amount of SDPs may be an additional indicator of therapeutic success. 

However, it is obviously easier to judge the amount of subretinal and intraretinal fluid, which is mainly performed clinically in a subjective manner, than it is to grade and eventually count SDPs. Thus, the question arose as to whether the amount of SDPs before treatment would be predictive of the success of therapy in terms of subretinal fluid resolution. A high number of SDPs at baseline signaled an improvement in BCVA of 5 ETDRS letters but not an improvement in edema, according to OCT. Also, no positive correlation was found in this case series between BCVA and OCT-determined fluid improvement (0.118; P = 0.365) or CRT (0.048; P = 0.718). This result shows that changes observed by OCT do not necessarily correlate with BCVA. However, since the grading by OCT of fluid reduction is subjective and the fluid's location (e.g., intraretinal, subretinal, central, or parafoveal) is highly variable, a clear correlation with BCVA seems to be unnecessary. 

The subjective judgment and grading of the appearance on OCT of edema and SDPs was a drawback in this observational study. Special software could be developed to objectively calculate the exact area of sub- and intraretinal fluid and to count SDPs in the SD-OCT images. 

In summary, with the SD-OCT technique, SDPs of unknown origin are observable in neovascular AMD and may represent the status of the accompanying retinal inflammation. The amount of SDPs is predominantly reduced after ranibizumab upload therapy, and this reduction correlates positively with improved BCVA. Moreover, an initial larger amount of SDPs may indicate a higher grade of inflammation, but it shows clearly the benefits of ranibizumab therapy. Thus, the amount of pretreatment SDPs may be a predictive factor for therapy outcome. 

Footnotes

 Disclosure: C. Framme, None; S. Wolf, None; U. Wolf-Schnurrbusch, None