To investigate longitudinal changes in the retinal and choroidal microstructure of the macula in patients with retinitis pigmentosa (RP).
Retrospective, observational cohort study.
A total of 69 patients with RP and 69 age- and sex-matched controls who underwent optical coherence tomography (OCT) over a 4-year follow-up period were included. The severity of RP was classified into 3 stages according to the integrity of the inner segment ellipsoid zone. The retinal and choroidal layers were segmented manually from OCT images. The areas of retinal pigment epithelium (RPE) atrophy and choroidal vascular index (CVI) were also analyzed. Longitudinal changes in the OCT parameters were compared among the groups.
Significant decreases (median [interquartile range]) in the thickness of the ganglion cell inner plexiform layer (GCIPL; −1.04 [−2.41 to −0.17]), outer nuclear layer (ONL; −1.44 [−1.86 to −0.28]), and inner segment ellipsoid (ISE; −0.74 [−1.33 to −0.49]) at the moderate stage and retinal nerve fiber layer (RNFL; −1.49 [−2.08 to −0.66]) and GCIPL (0.58 [−1.79 to 0.06]) at the advanced stage were observed. Choroidal thickness decreased significantly from −7.62 to −9.40 μm per year at all stages. RPE atrophy and CVI reduction were observed at the advanced stage. There was no change in the control group.
ONL and GCIPL thicknesses decreased at the moderate and advanced stages of RP; RNFL thickness decreased only at the advanced stage; and choroidal thickness decreased continuously. In addition, RPE atrophy and CVI reduction were prominent at the advanced stage. These results indicate that there is a temporal variation in the damage of each retinal layer and the choroid in RP patients.
R etinitis pigmentosa (RP) is the most common and devastating inherited retinal disorder that eventually progresses to blindness. RP is caused by mutations in genes that play crucial roles in the photoreceptors and retinal pigment epithelium (RPE). A pronounced early pathologic finding of RP is photoreceptor cell count reduction. As human retinal specimens are difficult to obtain, histological studies are extremely limited. Alternatively, rapid advances in imaging modalities, such as optical coherence tomography (OCT), have expanded insights into retinal microstructures in vivo. Understanding and interpreting the changes in macular microstructure is becoming more important to assess the remaining retinal structure and function in patients with RP. The approach to restoration of vision should be tailored according to the level of structural damage. For example, the use of an artificial retina is feasible only if retinal ganglion cells are preserved. Likewise, other treatment options that are currently under development require the integrity of certain macular structures, including the inner retina, RPE, and choroid, to be intact.
Macular structures including the retinal nerve fiber layer (RNFL) and ganglion cell inner-plexiform layer (GCIPL) are reported to be preserved or even thickened in patients with RP. , Previous studies using OCT found that RNFL thickness was increased and GCIPL was preserved in patients with RP compared to heathy controls. In addition, the difference in choroidal thickness between patients with RP and healthy controls has been debated in several studies. , However, the actual changes in these structures in patients with RP could not be revealed robustly in the above-mentioned cross-sectional studies. Cross-sectional studies can be confounded by innate interindividual variances rather than the pathologic process. Considerable variation in choroid thickness has been reported even among healthy individuals. Therefore, longitudinal observation is required to assess the degeneration of the retina and choroid in patients with RP. OCT is very useful in observing slowly progressing RP, as the OCT measurement is reliable and reproducible. Thus, the present study aimed to investigate the longitudinal macular changes in patients with RP using OCT. The thickness and structural changes in each layer of the retina and choroid were analyzed.
This study adhered to the tenets of the Declaration of Helsinki, and the study protocol was approved by the institutional review board of Seoul National University Hospital (approval no.: 1404-136-574).
A total of 69 patients diagnosed with RP at the hereditary retinal disease clinic of Seoul National University Hospital between 2009 and 2019 were enrolled in the present study. Only patients who were followed up for more than 42 months were included. Patients with a history of vitreoretinal surgery, glaucoma, and macular diseases, such as epiretinal membrane and cystoid macular edema, were excluded. Patients with cataract were not excluded unless the OCT signal strength was less than 6, and those with a history of cataract surgery before or after study enrollment were also not excluded. Patients with an epiretinal membrane were excluded if the foveal pit was absent. We included patients in whom the retina was not stretched to make the foveal dimple evident. Cystoid macular edema was defined as present if more than 5 intraretinal cysts of any size existed, regardless of definite retinal thickening. Because the central retinal thickness generally is decreased in RP, the macular edema can be underestimated. We also excluded patients if CME developed during the follow-up period. In addition, data of 69 age-and sex-matched healthy controls were also collected. The patients were divided into early, moderate, and advanced RP groups based on the initial OCT findings. The advanced and moderate groups were distinguished based on the visibility of the inner segment ellipsoid zone (ISE) on OCT images within 2500 µm from the fovea; ISE was visible in the moderate RP group, but not in the advanced group. The early RP group showed preserved ISE of more than 2500 µm from the fovea, whereas the moderate RP group showed constricted ISE within 2500 µm from the fovea on OCT scans ( Figure 1 ).
Vertical cross-sectional high-definition macular OCT scans were obtained using spectral-domain OCT (Cirrus 4000 HD OCT, Carl Zeiss Meditec, Inc). The right eye was selected for the analysis; however, if the right eye had severe media opacity or other diseases, the left eye was selected. The retinal and choroidal layers were segmented manually because the integrated segmentation algorithm of the OCT device is prone to errors under pathologic conditions. , The borders of the following were manually delineated: (a) vitreous/RNFL, (b) RNFL/ganglion cell layer, (c) inner plexiform layer (IPL)/inner nuclear layer (INL), (d) INL/outer plexiform layer, (e) upper margin of the ISE, (f) upper margin of the RPE, and (g) choroid/sclera. The choroidal layer was subdivided into the inner and outer choroid, which contained mainly small/medium-sized and large choroidal vessels, respectively. , Segmentation was performed by 2 masked graders (D.I.K., K.W.B.) ( Figure 1 ).
After manual segmentation, the average thickness of each layer within 2500 µm from the fovea was calculated using a custom program developed in Python (Python Language Reference, version 3.6; The Python Software Foundation). This is the same region where the severity of RP was graded. RNFL was between (a) and (b), GCIPL was between (b) and (c), INL was between (c) and (d), outer nuclear layer (ONL) was between (d) and (e), ISE was between (e) and (f), and the choroid was between (f) and (g). The thickness was calculated by dividing the area with the horizontal length. The OCT image at the last follow-up visit was superimposed on the initial OCT image based on the RPE contour and fovea location using Photoshop CC (Adobe). Thereafter, layer segmentation and thickness measurements were performed in the same manner as for the initial image.
Segmentation data constructed by 1 grader (K.W.B.) were used for the analysis, after the reliability of manual segmentation was confirmed by intraclass correlation coefficient (ICC) Comparisons of retinal thickness and visual acuity between the initial and last follow-up visits were performed using the Wilcoxon signed-rank test. Kruskal–Wallis 1-way analysis of variance was performed to compare the baseline RPE and CVI difference among the 3 RP stages, and Dunn post test was used for multiple comparisons. Statistical calculations were performed using individual logarithm of minimal angle of resolution (logMAR) acuity data, not decimal values, and the results were converted back to Snellen vision. Preserved ISE width is defined as the horizontal length where the ISE thickness is measurable. Pearson correlation analysis was performed to evaluate the ISE width and retinal layer thickness. Statistical analyses were performed using SPSS version 25.0 (IBM Corporation).
The atrophic area of the RPE was measured using the Cirrus OCT software. The advanced RPE analysis program provides a sub-RPE illumination area where the light transmitted through the RPE is increased. RPE atrophy area was assessed through the sub-RPE illumination area in a 5-mm circle. Niblack auto local thresholding was used to calculate the choroidal vascular index (CVI). The segmented choroidal layer was converted into an 8-bit grayscale format and binarized using the auto local threshold. The Scikit-image library of the Python language was used to perform this binarization. White pixels and dark pixels were considered the stromal and luminal areas, respectively. CVI was defined as the proportion of the luminal area to the total choroidal area.
A total of 69 patients diagnosed with RP were included in the study. One case from the moderate and 2 cases from the advanced group were excluded as the follow-up OCT did not align well with the initial examination. The early, moderate, and advanced groups comprised 13, 17, and 39 patients, respectively. Visual acuity decreased significantly during the follow-up period only in the advanced group. Age, sex, and follow-up duration were not significantly different between the groups ( Table 1 ). ICC analysis between the independent graders was performed using the retinal layer thickness measurement from 15 samples. Five cases were randomly selected from the mild, moderate, and advanced groups each. The thickness of all 6 retinal layers of each sample were used. ICC was excellent (0.929, 95% CI = 0.912-0.940, P < .001). In the early and control groups, the retinal layer thickness did not change significantly between the initial and final examinations. In the moderate group, GCIPL, ONL, and ISE thickness (median [interquartile range]) decreased significantly over the follow-up period (79.09 [75.59-82.37] to 71.91 [67.00-78.88] μm, 84.44 [72.44-97.21] to 82.84 [64.88-93.65] μm, and 18.81 [9.25-27.91] to 13.80 [6.37-23.43] μm, respectively; P = .001, P = .005, and P < .001, respectively; Wilcoxon signed-rank test). In the advanced group, the RNFL and GCIPL thickness (median [interquartile range]) decreased significantly over the follow-up period (52.13 [46.12-59.80] to 45.87 [41.22-51.09] μm and 69.05 [63.19-76.03] to 67.11 [56.82-72.85] μm, respectively; P < .001 and P < .001, respectively; Wilcoxon signed-rank test). The thicknesses between the IPL and RPE in the advanced group remained unchanged. The changes in the retinal layer thicknesses are summarized in Table 2 and Figure 2 .
|Group||M:F||Age, y||Follow-up, y||VA_ini||VA_fu||P Value a|
|Early||4:9||36 [30-49]||4.13 [3.64-4.93]||1.00 [0.50-1.20]||0.80 [0.60-1.20]||1|
|Moderate||9:8||34 [32-42]||4.05 [3.73-5.52]||0.70 [0.40-0.90]||0.60 [0.40-0.70]||.555|
|Advanced||21:18||47 [34-53]||4.43 [3.82-5.04]||0.02 [0.01-0.10]||0.01 [0.01-0.02]||.005 b|
|Control||33.36||42 [30-48]||4.19 [3.79-4.99]||1.00 [0.90-1.00]||1.00 [0.80-1.20]||.468|
|P Value c||.372||.145||.324||<.001 b||<.001 b|
a Wilcoxon signed-rank test between initial and final visual acuity values.
c χ 2 Test for comparing sex distribution, and Kruskal–Wallis 1-way analysis of variance for comparing age and visual acuity among groups. Values are expressed as median [interquartile range].
|Early||Change||–0.33 [–0.45 to 0.07]||0.21 [–0.52 to 1.06]||–0.08 [–0.73 to 0.70]||–0.48 [–0.72 to 0.05]||0.79 [–0.26 to 0.98]||—|
|Moderate||Change||0.29 [–0.76 to 0.65]||–1.04 [–2.41 to –0.17]||0.11 [–0.19 to 0.63]||–1.44 [–1.86 to –0.28]||–0.74 [–1.33 to –0.49]||—|
|P value||.782||.001 a||.378||.005 a||<.001 a|
|Advanced||Change||–1.49 [–2.08 to –0.66]||–0.58 [–1.79 to 0.06]||–0.27 [–1.65 to 2.33]|
|P value||<.001 a||<.001 a||.874|
|Control||Change||0.12 [–0.86 to 0.88]||–0.16 [–1.41 to 0.77]||–0.28 [–0.78 to 1.31]||0.36 [–1.47 to 1.20]||–0.14 [–1.94 to 1.27]||0.20 [–2.09 to 2.05]|