Evaluation of retinal pigment epithelium changes in serous pigment epithelial detachment in age-related macular degeneration

The purpose of this study was to quantitatively evaluate retinal pigment epithelium (RPE) changes in serous pigment epithelial detachment (PED) among patients with age-related macular degeneration by means of prototype multi-contrast optical coherence tomography (OCT), which is capable of simultaneous collection of OCT angiography, polarization-sensitive OCT, and standard OCT images. We evaluated 26 eyes of 21 patients with serous PED. RPE-melanin OCT images were calculated from the multi-contrast OCT dataset and compared with near-infrared autofluorescence images. An active RPE lesion was defined as an area of thickened RPE-melanin (≥ 70 μm; RPE70) on RPE-melanin OCT. Each PED area was divided into peak and slope regions. RPE70 area ratios were compared with the maximum PED height, PED area, PED volume, and slope area ratio (area of slope region/area of whole PED). RPE-melanin OCT images were consistent with near-infrared autofluorescence images. The RPE70 area ratio in the slope region was significantly negatively correlated with the slope area ratio. Development of active RPE lesions in the slope region was correlated with the PED configuration. Multi-contrast OCT is useful for objective evaluation of changes in the RPE in patients with age-related macular degeneration.


Scientific Reports
| (2021) 11:2764 | https://doi.org/10.1038/s41598-021-82563-z www.nature.com/scientificreports/ Polarization-sensitive OCT is a functional extension of OCT technology that permits advanced tissue differentiation in retinal disease 20,21 . Multiple scattered lights from melanin in the RPE induce depolarization or scrambled polarization 22 . Comprehensive evaluation of RPE changes in AMD can be performed using a combination of polarization-sensitive OCT and AF imaging 8,23,24 . However, the evaluation of RPE changes by polarization-sensitive OCT has some limitations, including a lack of discriminability between melanin in the RPE (RPE-melanin) and melanin in the choroid. In projection images captured by polarization-sensitive OCT, RPE changes are obscured by depolarization involving choroidal melanin 25 . To overcome this limitation, we developed an automatic highlighting algorithm for RPE-melanin using multi-contrast OCT (MC-OCT) 26 , which enables simultaneous collection of OCT angiography, polarization-sensitive OCT, and standard OCT images [27][28][29] . From the MC-OCT dataset, we derived RPE-melanin-specific contrast OCT (RPE-melanin OCT) images, which we used to evaluate RPE changes 26 . In our previous studies of patients with AMD 26 and patients with Vogt-Koyanagi-Harada disease 25 , we confirmed similarities between NIR-AF images and RPE-melanin OCT projection images. Thus, the combination of RPE-melanin OCT and AF imaging might be useful for evaluation of RPE changes in PED. In this study, we evaluated RPE changes in serous PEDs by means of AF and MC-OCT imaging. We also analyzed the relationships of RPE changes with PED parameters.

Results
NIR-AF imaging showed hyper NIR-AF lesions in the PED (Figs. 1a and 2a) in 23 of 26 eyes examined (88%). SW-AF images showed hyper SW-AF at corresponding locations (Figs. 1b and 2b). However, hyper SW-AF was somewhat obscured in the foveal region, possibly by the presence of macular pigment, in the patient shown in Fig. 1b. MC-OCT provides standard OCT, OCT angiography, and degree of polarization uniformity (DOPU) 30 data from a single measurement. Standard OCT B-scan images showed a thickened RPE band and intraretinal hyper-reflective foci in hyper NIR-AF lesions (Figs. 1c and 2c). The presence of low DOPU is presumed to indicate depolarization by multiple scattered lights from melanin 22 . A chorioretinal melanin thickness map was computed by counting the number of pixels with low DOPU (< 0.8) on each A-line in the three-dimensional dataset. The resulting chorioretinal melanin thickness map represents the overall thickness of melanin in the choroid and RPE. DOPU B-scan images in polarization-sensitive OCT clearly showed melanin accumulation in the RPE band, as well as intraretinal melanin migration (Figs. 1d and 2d). However, these melanin changes were obscured by choroidal melanin in the chorioretinal melanin thickness maps (Figs. 1e and 2e). For automatic discrimination of RPE-melanin from choroidal melanin based on DOPU images, we computed a new index (F RPE ) based on the concept of feature engineering by the presence or absence of blood flows 26 . Melanin in both the RPE and choroid showed low DOPU due to depolarization. However, the OCT angiography signal in the RPE cells was low because vascularization was absent, whereas the choroid showed a high OCT angiography signal because of its dense vasculature. The presence of high F RPE is presumed to indicate the presence of RPE-melanin. The RPE-melanin B-scan images were used to determine the distribution of F RPE in the B-scan images, which enabled evaluation of the depth-resolved distribution of RPE-melanin. RPE-melanin thickness maps were created by counting the number of pixels with high F RPE (≥ 0.15) on each A-line in the volume dataset. RPE-melanin thickness maps clearly showed the en face distributions of these melanin changes. Greyscale RPE-melanin thickness maps were similar to both NIR-AF and SW-AF images in all patients (Figs. 1f, 2f). Hyper NIR-AF lesions coincided with areas of thickened RPE melanin. This similarity facilitated quantitative evaluation of the RPE-melanin changes in hyper NIR-AF lesions through RPE-melanin thickness maps. The thickness of RPE-melanin, according to location within the PED, could readily be visualized with color-coded RPE-melanin thickness maps (Figs. 1g and 2g). The RPE-melanin B-scan images clearly showed that thickened RPE-melanin lesions were composed of RPE-melanin accumulation at the RPE band and intraretinal RPE-melanin migration (Figs. 1h and 2h). To evaluate the distributions of thickened RPE lesions, the PED area was divided into peak and slope regions. The inner area (70% of the maximum PED height) was defined as the peak region, while the outer residual area of the PED was defined as the slope region ( Supplementary Fig. S1 online). Areas of thickened RPE-melanin (RPE 70 : ≥ 70 μm) were distributed in both the peak and slope regions in 19 eyes (Fig. 1i), while they were solely in the slope region in four eyes (Fig. 2i).
For quantitative evaluations, we compared morphometric PED parameters and the parameters associated with RPE 70 (Table 1). Tomographic volumes in standard OCT images, obtained from MC-OCT, were used to evaluate morphometric parameters (maximum PED height, PED area, PED volume, and slope area ratio). Obliqueness of the slope area was evaluated as the slope area ratio (area of slope region/area of whole PED). A small slope area ratio indicated the presence of a steep slope.
We calculated the RPE 70 areas for each whole PED, peak region, and slope region. RPE 70 areas were significantly greater in the slope region than in the peak region (P = 0.002, Wilcoxon signed-rank test; Table 1). Simple linear regression analysis was performed between RPE 70 areas and morphometric PED parameters ( Table 2, Fig. 3, and Supplementary Fig. S2 online). Figure 3 shows the scatter plots of combinations with statistically significant correlations. The RPE 70 areas in whole PED, peak region, and slope region showed significant positive correlations with PED area and PED volume. The RPE 70 area in whole PED showed a significant negative correlation with slope area ratio. The RPE 70 area in the peak region showed a significant positive correlation with maximum PED height. Subsequently, multiple linear regression analysis was conducted to consider interactions among morphometric parameters. Stepwise multiple linear regression analysis of RPE 70 area with four morphometric PED parameters (maximum PED height, PED area, PED volume, and slope area ratio) revealed that the RPE 70 areas in both whole PED and the slope region showed significant positive correlations with PED volume (Table 3).
To eliminate the contributing factor of area, the RPE 70 area ratios (RPE 70 area in whole PED/area of whole PED, RPE 70 area in peak region/area of peak region, RPE 70 area in slope region/area of slope region) were  Fig. 3). Stepwise multiple linear regression analysis confirmed this significant negative correlation (Table 5).   www.nature.com/scientificreports/ cells, and RPE dysmorphia [6][7][8] . Notably, these RPE changes have been suggested to represent RPE activity 7 ; our multimodal imaging technique might be useful in the evaluation of the RPE cellular activity in PEDs. The present study showed positive correlations of PED volume with the area of active RPE lesions (RPE 70 ) in both the whole PED and the slope region (Table 3). This finding is consistent with the results of our previous polarization-sensitive OCT study regarding serous PEDs 8 . A study with drusenoid PED also confirmed the correlation with PED volume 7 . Hypoxia may contribute to RPE activation, due to the reduction of oxygen distribution from the choriocapillaris 31,32 . The diffusion of oxygen to the RPE could be impaired by the increased distance from the choroid in large PEDs 7 . If hypoxia is a predominant contributing factor with respect to RPE activation, RPE changes should occur mainly in the peak regions of PEDs, at a specific distance from the choroid.
However, RPE changes frequently occurred in the slope regions of serous PEDs. In the present study, RPE 70 existed in the slope region in 23 of 26 eyes (88%); the RPE 70 area was significantly greater in the slope region than in the peak region. This frequent occurrence of RPE changes in the slope region was also reported in drusenoid PEDs 6 , despite a possibly larger contribution of hypoxia with the presence of drusenoid material. The development of RPE activation in the slope region suggests the presence of induction factors other than hypoxia. One possible explanation involves size: there was a close relationship between the RPE 70 area and the PED area, according to simple linear regression analysis ( Table 2). The RPE 70 area might increase along with the underlying PED area. To eliminate the contributing factor of area, we evaluated the RPE 70 area ratio. We found a negative correlation between the slope area ratio and the RPE 70 area ratio in the slope region. A small slope area ratio indicated the presence of a steep slope; thus, active RPE lesions in the slope region tended to occur with steep slopes. A possible predisposing factor for a steep slope involves mechanical tension due to architectural changes in the RPE band. Mechanical forces may contribute to remodeling of the extracellular matrix around RPE cells and may facilitate the migration of RPE cells 33 . Although the exact mechanism by which RPE migration occurs in serous PEDs is not yet fully understood, hypoxia and mechanical stress are possible causative factors in RPE changes.
There were several limitations to this study. First, because it involved a relatively small number of patients without follow-up assessments, this preliminary study evaluated only some aspects of RPE changes in serous PEDs. Further studies, involving many patients with follow-up data, are needed to evaluate RPE changes in greater detail. Second, RPE-melanin measurement with MC-OCT in this study was based on depolarization by    www.nature.com/scientificreports/ although a previous study showed a monotonic relationship between DOPU and melanin, measurement of the RPE-melanin thickness could be affected by the DOPU kernel size or melanin packing density in RPE cells. Therefore, RPE-melanin thickness maps do not represent the actual thickness of RPE-melanin, although they are proportional to the thickness of the RPE-melanin. Seventh, a threshold of F RPE (≥ 0.15) in MC-OCT imaging was used to calculate the area of RPE melanin defects, in accordance with the method used in our previous studies 9, 25,26 . With this threshold, the RPE-melanin line at the normal RPE band occasionally became unclear (Fig. 1h). Finally, we used a cutoff value of ≥ 70 μm for statistical analysis of the areas of thickened RPE-melanin. Further analysis with a larger number of patients is necessary to determine the appropriate evaluation threshold.
In conclusion, this study demonstrated the clinical utility of multimodal imaging with RPE-melanin OCT imaging and MC-OCT for evaluation of RPE changes in serous PEDs. MC-OCT imaging enables three-dimensional evaluation of AF images; the combination of MC-OCT and AF imaging could be useful for assessment of various macular diseases. MC-OCT is a promising tool as a next-generation imaging system for the clinical evaluation of macular diseases.

Methods
Patients. We prospectively examined 26 eyes of 21 Japanese patients with serous PEDs due to AMD (13 men, 8 women; age range, 55-83 years; mean age, 72.1 years). Among them, three patients were treated for hypertension, one patient was treated for diabetes, and one patient was treated for both hypertension and diabetes. To exclude the influence of these systemic diseases, eyes with retinal diseases other than AMD were excluded. Patients were considered to have "serous PED" when they exhibited serous PED without retinal hemorrhage, hard exudates, or choroidal neovascularization. OCT angiography images obtained from commercial OCT (DRI-OCT Triton, Topcon, Tokyo, Japan) and MC-OCT were used to exclude eyes with choroidal neovascularization. To avoid the possible influence of treatments for AMD, eyes with prior intravitreal injection or laser treatment were excluded. Eyes with severe cataract or other eye diseases that could significantly compromise the image quality were also excluded from this study.
This prospective, observational, cross-sectional study followed the tenets of the Declaration of Helsinki, and Institutional Review Board approval was obtained from the Tokyo Medical University Institutional Review Board (IB-1615 and T2019-0072). The study was registered in a public database (UMIN000026307 and 000039648; http://www.umin.ac.jp/ctr/index -j.htm). Written informed consent was obtained from each participant before any study procedures or examinations were performed. MC-OCT system. A detailed description of the prototype Jones-matrix MC-OCT has been published 28 .
MC-OCT provides standard OCT, OCT angiography, and DOPU data from a single measurement. Standard OCT images were obtained by coherent composition of four repetitive scans 28 . OCT angiography was calculated using the complex Jones matrix correlation method with noise correction 38 . DOPU was calculated with Makita's noise correction using a 3 × 3 pixel kernel 39 . The light source was a swept-source laser with a central wavelength of 1050 nm. The axial scan speed was 100,000 A-scans/s and the depth resolution in tissue was 6.0 μm. A horizontal-fast raster scanning protocol with 512 A-lines × 256 B-scans was used for volumetric scans; the approximate measurement area on the retina was 6.0 × 6.0 mm. For quantitative measurements, the transverse magnification of MC-OCT images was calibrated using a modified Littman's method 40 . B-scan measurements were repeated four times at a single location; the acquisition duration for volumetric measurement was 6.6 s. MC-OCT volumes without significant motion artifacts were used for this study.
Morphometric PED parameters. Tomographic volumes in standard OCT images, obtained from MC-OCT, were used for morphometric evaluation of serous PEDs using FIJI image-processing software 41 .
1. PED area: The locations of the PED margins in en face OCT images were manually determined using series of B-scan images. PED area was calculated as the transverse inner area within the margin of the PED. 2. Maximum PED height was measured as the greatest distance between Bruch's membrane and the outer boundary of the RPE band. 3. PED volume: The inner boundary of the PED was manually segmented in each B-scan image. The total PED volume was determined by summing the volumes of individual segments using the Cavalieri principle of stereological analysis 42 . 4. Peak and slope regions: Transverse locations with a height equal to 70% of the maximum PED height were manually determined using series of B-scan images ( Supplementary Fig. S1a online). In en face OCT images, the inner area (70% of the maximum PED height) was defined as the peak region, while the outer residual area of the PED was defined as the slope region ( Supplementary Fig. S1b online).

RPE-melanin-specific contrast OCT imaging.
For automatic discrimination of RPE-melanin from choroidal melanin based on DOPU images, a new index (F RPE ) was computed using an attenuation coefficient 43 , the DOPU, and the blood flow signal in OCT angiography, as follows 26 : where OCTA b is the binarized OCT angiography signal. This index was computed to selectively represent RPEmelanin by the absence of blood flow in the RPE. The RPE-melanin B-scan images were computed as the distribution of F RPE in the B-scan images; this enabled evaluation of the depth-resolved distribution of RPE-melanin. RPE-melanin thickness maps were created by counting the number of pixels with high F RPE (≥ 0.15) on each