Cellular-Level Characterization of Cone Photoreceptors and Retinal Pigment Epithelium in Stargardt Disease with Adaptive Optics Imaging Authors: Masakazu Hiraoka, Leandre Fura, Marika Hirota, Kiyoko Gocho, Natsuko Nakamura, Satoshi Yokota, Yasuhiko Hirami, Masayo Takahashi, Shuhei Kimura, Michiko Mandai, Yasuo Kurimoto, Akiko Maeda Published: 2026 ABSTRACT PurposeThe purpose of this study was to simultaneously visualize cone photoreceptors (cone) and retinal pigment epithelium (RPE) cells at the cellular level in Stargardt disease (STGD) using adaptive optics (AO) imaging. In addition, the feasibility of quantitative analysis was explored.DesignObservational case series.SubjectsPatients with STGD.MethodsCone and RPE cells were simultaneously imaged at the cellular level using AO flood-illumination ophthalmoscopy (AO-FIO) and AO transscleral flood illumination (AO-TFI). Quantitative metrics were derived and normalized as z-scores to provide an exploratory assessment of the feasibility of cellular-level quantification.Main Outcome MeasuresCellular-level cone and RPE integrity and abnormalities assessed by simultaneous AO-FIO and AO-TFI imaging.ResultsUsing AO-FIO and AO-TFI, cone and RPE cells could be simultaneously visualized in eyes with STGD. In a subset of cases, cellular-level quantitative analysis was feasible. Imaging findings varied across patients and retinal regions, showing heterogeneous degrees of structural alteration in cone and RPE cells. In the quantitatively analyzed cases, cone density z-scores showed marked reductions, with the lowest values of –3.28 in case 1 and –2.90 in case 2, whereas RPE cell density z-scores showed only mild deviations and remained within the reference range in the analyzed regions.ConclusionsAdaptive optics flood-illumination ophthalmoscopy and AO-TFI enable simultaneous cellular-level visualization of cone and RPE cells in STGD. This study characterized cellular-level structural features in STGD and demonstrated the feasibility of quantitative analysis in a subset of cases. These findings could contribute to a deeper understanding of STGD pathology and support future investigations aimed at treatment optimization.Financial DisclosuresProprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article. Stargardt disease (STGD) is an autosomal recessive inherited macular dystrophy, characterized by progressive retinal degeneration centered on the macula, where cone photoreceptors (cone) are densely distributed, with an estimated prevalence of 1 in 8000 to 10 000 individuals.1, 2, 3 Most patients develop symptoms in adolescence or early adulthood, presenting with progressive visual decline, central scotoma, photophobia, and color vision abnormalities.4 ABCA4 is the major causative gene for STGD, accounting for 70%–90% of cases.5, 6, 7, 8 ATP-binding cassette subfamily A member 4 functions as a lipid flippase in cone and rod outer segments9,10 and plays a key role in the visual cycle.1,11 Abnormal ATP-binding cassette subfamily A member 4 function delays clearance of all-trans-retinal from photo-activated photoreceptors, leading to the accumulation of all-trans-retinal condensation products such as A2E and other bisretinoids. Accumulation of these products can be detected in retinal pigment epithelium (RPE) cells12 as a result of phagocytosis of outer segments of photoreceptor. Such accumulation is thought to contribute to functional impairment and structural alterations of both photoreceptors and RPE cells.13 Retinal pigment epithelium transplantation has been developed as one therapeutic strategy targeting RPE dysfunction in STGD, and clinical benefits have been reported.14 In recent years, therapeutic development for STGD has expanded to include pharmacological approaches targeting retinoid metabolism within photoreceptors, as well as gene therapies directed at the causative gene ABCA4.15, 16, 17, 18 As therapeutic strategies targeting different retinal cell types are being explored, cellular-level characterization of the structural features of cone and RPE cells in STGD may provide important insights into disease pathology and has the potential to contribute to the optimization of therapeutic approaches. Adaptive optics (AO) is an imaging technology that corrects ocular aberrations to achieve high-resolution retinal imaging.19 Adaptive optics flood-illumination ophthalmoscopy (AO-FIO) enables visualization of the cone mosaic.19 More recently, AO transscleral flood illumination (AO-TFI), which uses oblique illumination, has been developed to visualize the RPE mosaic, and this device allows simultaneous acquisition of AO-FIO for comparative analysis.20 In addition, AO-TFI allows imaging with a wider field of view and shorter acquisition time compared with AO scanning laser ophthalmoscope (AO-SLO), making it a promising modality for clinical application.20 In this study, cone and RPE cells in STGD were simultaneously visualized at the cellular level using AO-FIO and AO-TFI, and their structural features were characterized to explore potential relevance for disease understanding and future therapeutic considerations. Methods Study Design and Ethical Approval This exploratory observational case series was conducted at Kobe City Eye Hospital between November 15, 2024, and September 15, 2025. The study protocol was approved by the institutional review board of Kobe City Medical Center General Hospital (No. E23004). The study adhered to the tenets of the Declaration of Helsinki. Written informed consent was obtained from all participants. Patients Five patients clinically diagnosed with STGD were enrolled. The diagnosis was based on comprehensive ophthalmic examinations, including assessment of bilateral visual acuity, fundus evaluation for macular atrophy or pigmentary changes, and OCT imaging to detect ellipsoid zone disruption. Typical features of STGD, such as yellow-white flecks and relative peripapillary sparing, were also observed. Visual field was tested using Humphrey Field Analyzer (Carl Zeiss-Humphrey Systems) and Goldmann perimetry (Haag Streit). Retinal pigment epithelium cells and photoreceptors were evaluated by fundus autofluorescence (FAF) (Optos 200Tx; Optos) and OCT (Spectralis; Heidelberg Engineering). Adaptive optics flood-illumination ophthalmoscopy (rtx1, Imagine Eyes) was used for detecting cone density and visualizing RPE cell morphology. For comparison, 5 healthy volunteers were also examined using the same protocol. Image Acquisition and AO Analysis Adaptive optics flood-illumination ophthalmoscopy images for detecting cone density were acquired with rtx1. Adaptive optics transscleral flood illumination was used for RPE visualization. Image analysis was performed with AOdetect mosaic software (Imagine Eyes). All AO imaging was performed with consideration of the Stiles–Crawford effect, in which cones most efficiently capture incident light entering near the center of the pupil.21 This approach minimized signal attenuation and light scatter, thereby enabling acquisition of high-resolution mosaic images.22 Regions-of-interest (ROIs) were selected from atrophic, transition, and preserved areas. Montage and alignment were performed with i2k Retina (Dual Align LLC) by registering AO with SLO and OCT. The field of view was 1200 × 1200 μm at axial length (AL) of 24 mm. For each eye, the image scale was corrected for individual AL using the built-in scaling function of AOdetect before calculating cellular densities.23 For the analysis of cone and RPE cell mosaic patterns, Voronoi analysis was applied to extract the centroid of each cell, and Voronoi tessellation based on these points was performed to quantitatively determine cell density and regularity.24, 25, 26 Automated detections were manually reviewed by 2 retinal specialists (M.H. and M.H.). They reviewed the automatically detected cell centroids in each ROI and corrected obvious false-positive and false-negative detections when necessary. Low-quality ROIs were excluded when cellular mosaics could not be reliably identified because of severe motion artifacts, defocus, poor contrast, media opacity, or extensive lesion-related disruption. For cone density analysis, 24 of 27 ROIs were included and 3 ROIs were excluded (11.1%). For RPE cell density analysis, 18 of 27 ROIs were included and 9 ROIs were excluded (33.3%). Statistical Analysis Reference models from healthy controls were constructed according to eccentricity. Locally estimated scatterplot smoothing regression was applied; if limited data points were available, linear regression or mean values were used. At each eccentricity, mean and standard deviation served as reference. For patient data, %diff and z-scores were calculated. Z-scores indicated deviation from normal distribution, with out-of-range values treated as extrapolations. Statistical comparisons of baseline characteristics between STGD cases and healthy volunteers were performed using the Mann–Whitney U test for continuous variables and Fisher exact test for categorical variables. A 2-sided P value of <0.05 was considered statistically significant. Z-scores were mainly used to support interpretation and standardize comparisons between cases. All analyses and visualizations were performed in R version 4.4.1 (R Foundation for Statistical Computing) using lme4, lmerTest, emmeans, ggplot2, dplyr, and tidyr. Results Participant Characteristics This study included 5 eyes from 5 patients with STGD and 5 eyes from 5 healthy volunteers. Demographic characteristics of the STGD cases and healthy volunteers are summarized in Table 1. Compared with healthy volunteers, STGD cases had significantly worse logarithm of the minimum angle of resolution visual acuity, whereas age, sex, and AL did not differ significantly between groups. Individual clinical characteristics of the STGD cases are shown in Table 2. The median age of the STGD cases was 41 years (range, 20–49 years); 4 were female and 1 was male. Axial length ranged from 23.37 to 25.21 mm, except for 1 patient in whom AL was not available. Visual acuity, converted to logarithm of the minimum angle of resolution, ranged from 0.07 to 1.20. The age at onset ranged from 8 to 35 years, with 1 case undetermined. The study eye was the right eye in 3 patients and the left eye in 2 patients. Humphrey Field mean deviation values were available in 3 patients and ranged from –13.06 dB to –34.92 dB. Foveal sensitivity ranged from <0 to 36 dB.Table 1Demographic and Clinical Characteristics of Stargardt Disease Cases and Healthy VolunteersParameterSTGD CasesHealthy VolunteersP ValueNumber of eyes55—Age, years38.4 ± 11.429.8 ± 2.680.142SexM = 1, F = 4M = 3, F = 20.524Axial length, mm24.16 ± 0.7924.91 ± 0.750.190LogMAR VA0.54 ± 0.52–0.14 ± 0.050.011F = female; LogMAR = logarithm of the minimum angle of resolution; M = male; STGD = Stargardt disease; VA = visual acuity.Data are presented as mean ± standard deviation unless otherwise indicated. Axial length in the STGD group was calculated from 4 eyes because axial length was unavailable in case 3. Comparisons between STGD cases and healthy volunteers were performed using the Mann–Whitney U test for continuous variables and Fisher exact test for sex.Table 2Clinical Characteristics of Individual Stargardt Disease CasesCaseAge, ySexAL, mmVA, LogMAROnset Age, yEyeHF MD, dBFoveal Sensitivity, dB120M23.370.1514OD–13.0626236F23.791.00NAOSNANA346FNA0.078OD–34.92<0449F25.211.2030OD–29.1936541F24.280.3035OSNANAAL = axial length; F = female; HF MD = Humphrey Field mean deviation; LogMAR = logarithm of the minimum angle of resolution; M = male; NA = not available; OD = right eye; OS = left eye; VA = visual acuity.Foveal sensitivity was measured using Humphrey Field Analyzer. A value of <0 indicates that foveal sensitivity was below the measurable range. Table 1Demographic and Clinical Characteristics of Stargardt Disease Cases and Healthy VolunteersParameterSTGD CasesHealthy VolunteersP ValueNumber of eyes55—Age, years38.4 ± 11.429.8 ± 2.680.142SexM = 1, F = 4M = 3, F = 20.524Axial length, mm24.16 ± 0.7924.91 ± 0.750.190LogMAR VA0.54 ± 0.52–0.14 ± 0.050.011F = female; LogMAR = logarithm of the minimum angle of resolution; M = male; STGD = Stargardt disease; VA = visual acuity.Data are presented as mean ± standard deviation unless otherwise indicated. Axial length in the STGD group was calculated from 4 eyes because axial length was unavailable in case 3. Comparisons between STGD cases and healthy volunteers were performed using the Mann–Whitney U test for continuous variables and Fisher exact test for sex. Demographic and Clinical Characteristics of Stargardt Disease Cases and Healthy Volunteers F = female; LogMAR = logarithm of the minimum angle of resolution; M = male; STGD = Stargardt disease; VA = visual acuity. Data are presented as mean ± standard deviation unless otherwise indicated. Axial length in the STGD group was calculated from 4 eyes because axial length was unavailable in case 3. Comparisons between STGD cases and healthy volunteers were performed using the Mann–Whitney U test for continuous variables and Fisher exact test for sex. Table 2Clinical Characteristics of Individual Stargardt Disease CasesCaseAge, ySexAL, mmVA, LogMAROnset Age, yEyeHF MD, dBFoveal Sensitivity, dB120M23.370.1514OD–13.0626236F23.791.00NAOSNANA346FNA0.078OD–34.92<0449F25.211.2030OD–29.1936541F24.280.3035OSNANAAL = axial length; F = female; HF MD = Humphrey Field mean deviation; LogMAR = logarithm of the minimum angle of resolution; M = male; NA = not available; OD = right eye; OS = left eye; VA = visual acuity.Foveal sensitivity was measured using Humphrey Field Analyzer. A value of <0 indicates that foveal sensitivity was below the measurable range. Clinical Characteristics of Individual Stargardt Disease Cases AL = axial length; F = female; HF MD = Humphrey Field mean deviation; LogMAR = logarithm of the minimum angle of resolution; M = male; NA = not available; OD = right eye; OS = left eye; VA = visual acuity. Foveal sensitivity was measured using Humphrey Field Analyzer. A value of <0 indicates that foveal sensitivity was below the measurable range. Multimodal AO Findings in the Transition Zone Representative findings in the transition zone are shown for Case 1 (Fig 1). OCT showed irregularity and loss of the external limiting membrane and ellipsoid zone (Fig 1E5, E3), with concomitant thinning of the RPE layer (Fig 1E1). Adaptive optics flood-illumination ophthalmoscopy images corresponding to the blue box in Figure 1B show disrupted cone mosaics with occasional clustering within the upper atrophic lesion. Toward the lower preserved areas, cone visibility gradually decreases (Fig 2A). Adaptive optics transscleral flood illumination images corresponding to the blue box in Figure 1B show large hyperreflective dots aligned with disrupted cones, surrounded by sparsely distributed pigmented cells. The RPE mosaic becomes more apparent farther from the atrophic lesion, corresponding to areas with preserved RPE on OCT (Fig 2B). Consistent findings regarding the relationship between cone and RPE cells were observed in other cases as well.Figure 1Transition zone of Stargardt disease in a 20-year-old male (right eye, ABCA4 compound heterozygous pathogenic variants: c.1760+2T>G and c.4195G>A). (A) Infrared image. Adaptive optics flood-illumination ophthalmoscopy and AO-TFI imaging were performed in the orange box. Green dotted lines indicate OCT scan locations. The central atrophic lesion shows high reflectivity. (B) Adaptive optics flood-illumination ophthalmoscopy image showing the transition from degenerative to preserved retina. The orange double arrow indicates the atrophic lesion. (C) Fundus autofluorescence image. Adaptive optics imaging was performed in the orange box. Central hypofluorescence corresponds to the atrophic lesion. (D) Adaptive optics transscleral flood illumination image demonstrating the same transition zone. The orange double arrow indicates the atrophic lesion. (E) OCT B-scans obtained along the green dotted lines (1–6). Black arrows indicate AO-FIO sites and white arrows AO-TFI sites. Photoreceptor layers are preserved at peripheral area (6), become irregular (5), and disappear (3, 4). Retinal pigment epithelium changes follow photoreceptor loss, showing irregularity (4), thinning, and eventual disappearance (1, 2). AO-FIO = adaptive optics flood-illumination ophthalmoscopy; AO-TFI = adaptive optics transscleral flood illumination; FAF = fundus autofluorescence; IR = infrared.Figure 2Magnified images from Figure 1 and Voronoi analysis of RPE cells. (A) Adaptive optics flood-illumination ophthalmoscopy image of the blue box in Figure 1B. Within the upper atrophic lesion, cone mosaics are disrupted with occasional clustering. Toward lower preserved areas, cone visibility decreases, but the RPE mosaic remains identifiable. Green dotted lines at the left edge of the image indicate OCT scan location. (B) Adaptive optics transscleral flood illumination image corresponding to the blue box in Figure 1B. Large hyperreflective dots align with disrupted cones, surrounded by sparsely distributed pigmented cells. The RPE mosaic becomes evident further from the atrophic lesion. Green dotted line indicates OCT scan location. (C) Regions-of-interest corresponding to the black box in Figure 1D, showing the region used for Voronoi analysis of RPE cells. (D) Retinal pigment epithelium cell images corresponding to yellow boxes 1, 2, and 3 in (C), with the corresponding Voronoi diagrams. AO-FIO = adaptive optics flood-illumination ophthalmoscopy; AO-TFI = adaptive optics transscleral flood illumination; ROI = region of interest; RPE = retinal pigment epithelium. Transition zone of Stargardt disease in a 20-year-old male (right eye, ABCA4 compound heterozygous pathogenic variants: c.1760+2T>G and c.4195G>A). (A) Infrared image. Adaptive optics flood-illumination ophthalmoscopy and AO-TFI imaging were performed in the orange box. Green dotted lines indicate OCT scan locations. The central atrophic lesion shows high reflectivity. (B) Adaptive optics flood-illumination ophthalmoscopy image showing the transition from degenerative to preserved retina. The orange double arrow indicates the atrophic lesion. (C) Fundus autofluorescence image. Adaptive optics imaging was performed in the orange box. Central hypofluorescence corresponds to the atrophic lesion. (D) Adaptive optics transscleral flood illumination image demonstrating the same transition zone. The orange double arrow indicates the atrophic lesion. (E) OCT B-scans obtained along the green dotted lines (1–6). Black arrows indicate AO-FIO sites and white arrows AO-TFI sites. Photoreceptor layers are preserved at peripheral area (6), become irregular (5), and disappear (3, 4). Retinal pigment epithelium changes follow photoreceptor loss, showing irregularity (4), thinning, and eventual disappearance (1, 2). AO-FIO = adaptive optics flood-illumination ophthalmoscopy; AO-TFI = adaptive optics transscleral flood illumination; FAF = fundus autofluorescence; IR = infrared. Magnified images from Figure 1 and Voronoi analysis of RPE cells. (A) Adaptive optics flood-illumination ophthalmoscopy image of the blue box in Figure 1B. Within the upper atrophic lesion, cone mosaics are disrupted with occasional clustering. Toward lower preserved areas, cone visibility decreases, but the RPE mosaic remains identifiable. Green dotted lines at the left edge of the image indicate OCT scan location. (B) Adaptive optics transscleral flood illumination image corresponding to the blue box in Figure 1B. Large hyperreflective dots align with disrupted cones, surrounded by sparsely distributed pigmented cells. The RPE mosaic becomes evident further from the atrophic lesion. Green dotted line indicates OCT scan location. (C) Regions-of-interest corresponding to the black box in Figure 1D, showing the region used for Voronoi analysis of RPE cells. (D) Retinal pigment epithelium cell images corresponding to yellow boxes 1, 2, and 3 in (C), with the corresponding Voronoi diagrams. AO-FIO = adaptive optics flood-illumination ophthalmoscopy; AO-TFI = adaptive optics transscleral flood illumination; ROI = region of interest; RPE = retinal pigment epithelium. Voronoi-Based Quantification of Cone and RPE Cell Density The normative cone density and RPE cell density values obtained from healthy volunteers are summarized in Supplementary Table 1 (available at www.ophthalmologyscience.org). Cone density decreased with increasing eccentricity, whereas RPE cell density remained relatively stable across the measured eccentricities. These values were within the ranges reported in previous histologic and AO-based studies.27, 28, 29, 30, 31 Voronoi analysis was performed for both cone and RPE cells in cases 1 and 2. The quantitative analysis for case 1 is shown in Figure 2C, D. Z-scores were calculated (Fig 3 and Table 3). In case 1, cone density showed the lowest z-score at 5° (–3.28), whereas RPE cell density remained within the reference range at corresponding eccentricities. In case 2, cone density was reduced at 3° and 4°, whereas RPE cell density remained within the reference range across the analyzed eccentricities.Figure 3Quantitative z-score analysis of cone and RPE cell density in cases 1 and 2. Z-scores for cone density and RPE cell density are plotted according to retinal eccentricity. The plotted ROIs correspond to extralesional or transition-zone regions evaluated on FAF and OCT. Cone density showed reduced z-scores in several regions, whereas RPE cell density was relatively preserved in some regions. FAF = fundus autofluorescence; ROI = region-of-interest; RPE = retinal pigment epithelium.Table 3Z-Scores of Cone and RPE Density by Eccentricity in Cases 1 and 2CaseEccentricity, DegreesCone Density Z-ScoreRPE Cell Density Z-Score14–1.33–0.2115–3.28–0.6118–1.21NA1100.01NA112NA2.0723–2.90–1.2224–2.43–1.0127–1.91–0.8928–0.74NARPE = retinal pigment epithelium.Z-scores were calculated using normative reference data. The reference range was defined as –1.96 to +1.96. Values outside this range were considered to indicate deviation from the normative reference. NA indicates that the corresponding cell type was not quantified at that eccentricity. Quantitative z-score analysis of cone and RPE cell density in cases 1 and 2. Z-scores for cone density and RPE cell density are plotted according to retinal eccentricity. The plotted ROIs correspond to extralesional or transition-zone regions evaluated on FAF and OCT. Cone density showed reduced z-scores in several regions, whereas RPE cell density was relatively preserved in some regions. FAF = fundus autofluorescence; ROI = region-of-interest; RPE = retinal pigment epithelium. Table 3Z-Scores of Cone and RPE Density by Eccentricity in Cases 1 and 2CaseEccentricity, DegreesCone Density Z-ScoreRPE Cell Density Z-Score14–1.33–0.2115–3.28–0.6118–1.21NA1100.01NA112NA2.0723–2.90–1.2224–2.43–1.0127–1.91–0.8928–0.74NARPE = retinal pigment epithelium.Z-scores were calculated using normative reference data. The reference range was defined as –1.96 to +1.96. Values outside this range were considered to indicate deviation from the normative reference. NA indicates that the corresponding cell type was not quantified at that eccentricity. Z-Scores of Cone and RPE Density by Eccentricity in Cases 1 and 2 RPE = retinal pigment epithelium. Z-scores were calculated using normative reference data. The reference range was defined as –1.96 to +1.96. Values outside this range were considered to indicate deviation from the normative reference. NA indicates that the corresponding cell type was not quantified at that eccentricity. AO Findings around Hyperfluorescent Rings Regions corresponding to hyperfluorescent rings on FAF were also evaluated in Case 2 (Fig 4). In these areas, AO-FIO demonstrated indistinct cone mosaics (Fig 4B), with cone density showing z-scores below –2.0 compared with normative references (Fig 3 and Table 3). Cone arrangement was more disorganized inside the transition zone (Fig 4C) but relatively preserved outside (Fig 4D). Adaptive optics transscleral flood illumination showed irregular, large pigmented cells within the atrophic lesion, with the RPE mosaic becoming apparent toward the hyperfluorescent ring, while it became indistinct within the ring (Fig 4G). Outside the ring, the RPE mosaic was clearly visible (Fig 4H). Retinal pigment epithelium cell density in these regions showed mildly reduced z-scores but only minor deviations from normal (Fig 3 and Table 3).Figure 4Transition zone at the hyperfluorescent ring in a 36-year-old woman (left eye). (A) Fundus autofluorescence image. Adaptive optics imaging was performed in the yellow box, and the black box indicates the magnified area. The blue and orange boxes in (A) correspond to the same retinal regions indicated by the matching blue and orange boxes in (B) and (F), respectively. (B) Adaptive optics flood-illumination ophthalmoscopy image showing the transition zone including the hyperfluorescent ring. Cone visibility is markedly reduced. (C) Magnified AO-FIO image of the orange box (inside the ring), showing sparse cones with clustering. (D) Magnified AO-FIO image of the blue box (outside the ring), showing more evenly distributed cones compared with (C). (E) Horizontal OCT scan demonstrating progressive loss of photoreceptor layers toward the lesion. (F) Adaptive optics transscleral flood illumination image showing the transition zone with reduced visibility of the RPE mosaic at the ring. (G, H) Magnified AO-TFI images from the orange box (inside the ring) and blue box (outside the ring), respectively; (G) shows irregular large pigmented cells, with the RPE mosaic becoming apparent toward the periphery, while (H) shows a reappearing RPE mosaic structure. AO-FIO = adaptive optics flood-illumination ophthalmoscopy; AO-TFI = adaptive optics transscleral flood illumination; RPE = retinal pigment epithelium. Transition zone at the hyperfluorescent ring in a 36-year-old woman (left eye). (A) Fundus autofluorescence image. Adaptive optics imaging was performed in the yellow box, and the black box indicates the magnified area. The blue and orange boxes in (A) correspond to the same retinal regions indicated by the matching blue and orange boxes in (B) and (F), respectively. (B) Adaptive optics flood-illumination ophthalmoscopy image showing the transition zone including the hyperfluorescent ring. Cone visibility is markedly reduced. (C) Magnified AO-FIO image of the orange box (inside the ring), showing sparse cones with clustering. (D) Magnified AO-FIO image of the blue box (outside the ring), showing more evenly distributed cones compared with (C). (E) Horizontal OCT scan demonstrating progressive loss of photoreceptor layers toward the lesion. (F) Adaptive optics transscleral flood illumination image showing the transition zone with reduced visibility of the RPE mosaic at the ring. (G, H) Magnified AO-TFI images from the orange box (inside the ring) and blue box (outside the ring), respectively; (G) shows irregular large pigmented cells, with the RPE mosaic becoming apparent toward the periphery, while (H) shows a reappearing RPE mosaic structure. AO-FIO = adaptive optics flood-illumination ophthalmoscopy; AO-TFI = adaptive optics transscleral flood illumination; RPE = retinal pigment epithelium. Cellular Findings in Atrophic Lesions In atrophic lesion, OCT and FAF demonstrated preserved retinal structures in some regions (Fig 5). In such regions, AO-FIO revealed generally unclear visualization of cones but with possible focal cone clustering (Fig 5C, D). These clusters, indicated by white arrows, appeared hyperreflective, and OCT showed deposits extending from the preserved external limiting membrane to the ellipsoid zone and RPE at the corresponding sites. Adaptive optics transscleral flood illumination showed large white dots spatially corresponding to these clusters and widespread irregular pigment-containing cells in the surrounding area (Fig 5E,F). In other atrophic lesion, AO-TFI showed no regular RPE mosaic and demonstrated widespread pigment-containing cells, sometimes aggregated in a coiled-like pattern (Supplementary Figure 1, E, F, available at www.ophthalmologyscience.org).Figure 5Preserved retinal area within an atrophic lesion in a 49-year-old woman (right eye). (A) Fundus autofluorescence image showing central atrophy with preserved retinal islands. Adaptive optics imaging was performed in the yellow box. (B) Infrared image with OCT scan lines (green = horizontal, blue = vertical). Enlarged AO images from the green and blue boxes are shown in panels C–F. (C, D) Adaptive optics flood-illumination ophthalmoscopy images of the preserved retinal islands. Residual cones are visible, with hyperreflective cone clusters indicated by white (C) and black (D) arrows. (E, F) Corresponding AO-TFI images. The hyperreflective clusters indicated by the white arrow (E) and black arrow (F) appear as large white dots, and no clear RPE mosaic is observed. Pigment-containing cells were observed along the preserved retinal area. (G) OCT images. The upper panel shows the horizontal scan along the green line, and the lower panel shows the vertical scan along the blue line. Partial preservation of the ELM and EZ is noted. AO-FIO = adaptive optics flood-illumination ophthalmoscopy; AO-TFI = adaptive optics transscleral flood illumination; ELM = external limiting membrane; EZ = ellipsoid zone; FAF = fundus autofluorescence; IR = infrared; RPE = retinal pigment epithelium. Preserved retinal area within an atrophic lesion in a 49-year-old woman (right eye). (A) Fundus autofluorescence image showing central atrophy with preserved retinal islands. Adaptive optics imaging was performed in the yellow box. (B) Infrared image with OCT scan lines (green = horizontal, blue = vertical). Enlarged AO images from the green and blue boxes are shown in panels C–F. (C, D) Adaptive optics flood-illumination ophthalmoscopy images of the preserved retinal islands. Residual cones are visible, with hyperreflective cone clusters indicated by white (C) and black (D) arrows. (E, F) Corresponding AO-TFI images. The hyperreflective clusters indicated by the white arrow (E) and black arrow (F) appear as large white dots, and no clear RPE mosaic is observed. Pigment-containing cells were observed along the preserved retinal area. (G) OCT images. The upper panel shows the horizontal scan along the green line, and the lower panel shows the vertical scan along the blue line. Partial preservation of the ELM and EZ is noted. AO-FIO = adaptive optics flood-illumination ophthalmoscopy; AO-TFI = adaptive optics transscleral flood illumination; ELM = external limiting membrane; EZ = ellipsoid zone; FAF = fundus autofluorescence; IR = infrared; RPE = retinal pigment epithelium. Clinically Normal-Appearing Regions In clinically normal-appearing regions (Fig 6), FAF was isofluorescent and OCT revealed no abnormalities in the outer layers. In these areas, AO-FIO showed cone irregularity (Fig 6C), but AO-TFI demonstrated slightly irregular RPE mosaics (Fig 6F). Quantitative analysis showed that cone density at 5° exhibited reduced z-scores, whereas RPE cell density remained within the normal range (–1.96 to +1.96) (Fig 3 and Table 3).Figure 6Clinically normal region in case 1. A: Fundus autofluorescence image. The region distant from the atrophic lesion is indicated by the yellow box, where AO imaging was performed. No hypofluorescence or abnormal hyperfluorescence is observed. B: Adaptive optics flood-illumination ophthalmoscopy image. In the clinically normal region, the cone mosaic appeared irregular, with disorganization observed in some areas. C: Magnified AO-FIO image. Cones appear irregular, with partial disorganization observed in the mosaic pattern. D: Infrared image. The yellow-marked region is distant from the atrophic lesion, with no abnormal reflectance observed. E: Adaptive optics transscleral flood illumination image. In the clinically normal region, the RPE mosaic is slightly irregular. F: Magnified AO-TFI image. Retinal pigment epithelium cells are arranged slightly irregularly, showing mild variation in the mosaic pattern. G: OCT image. In the clinically normal region, the outer retinal layers are well preserved, and no abnormalities are detected. AO-FIO = adaptive optics flood-illumination ophthalmoscopy; AO-TFI = adaptive optics transscleral flood illumination; FAF = fundus autofluorescence; IR = infrared; RPE = retinal pigment epithelium. Clinically normal region in case 1. A: Fundus autofluorescence image. The region distant from the atrophic lesion is indicated by the yellow box, where AO imaging was performed. No hypofluorescence or abnormal hyperfluorescence is observed. B: Adaptive optics flood-illumination ophthalmoscopy image. In the clinically normal region, the cone mosaic appeared irregular, with disorganization observed in some areas. C: Magnified AO-FIO image. Cones appear irregular, with partial disorganization observed in the mosaic pattern. D: Infrared image. The yellow-marked region is distant from the atrophic lesion, with no abnormal reflectance observed. E: Adaptive optics transscleral flood illumination image. In the clinically normal region, the RPE mosaic is slightly irregular. F: Magnified AO-TFI image. Retinal pigment epithelium cells are arranged slightly irregularly, showing mild variation in the mosaic pattern. G: OCT image. In the clinically normal region, the outer retinal layers are well preserved, and no abnormalities are detected. AO-FIO = adaptive optics flood-illumination ophthalmoscopy; AO-TFI = adaptive optics transscleral flood illumination; FAF = fundus autofluorescence; IR = infrared; RPE = retinal pigment epithelium. Discussion In this study, we demonstrated that the combined use of AO-TFI and AO-FIO enables simultaneous cellular-level visualization of cones and RPE cells in STGD, thereby revealing structural features that are difficult to capture using other clinical imaging modalities. Reports using AO-TFI remain limited, and to the best of our knowledge, this study represents the first application of AO-TFI in STGD.32,33 As a characteristic structural feature of STGD, cones appeared indistinct within relatively preserved retinal tissue inside atrophic lesions, as well as in transition zones such as hyperfluorescent rings observed on FAF images. These findings may reflect alterations in the waveguiding properties of cones caused by inflammatory changes, misalignment of cell orientation, and density alterations in degenerating tissue. The fact that cones become visible when approaching degenerating regions suggests that, although cones are partly preserved in transition zones, disarray and other changes render them indistinct. Histological studies have also reported cone morphological abnormalities without density loss, supporting our findings.34 Adaptive optics transscleral flood illumination revealed large white dots corresponding to cone clusters observed with AO-FIO. These white dots may reflect combined hyperreflectance and strong scattering arising from multiple cones with irregular orientation and density under AO-TFI illumination, although other hyperreflective structures within the diseased retina cannot be excluded. Within atrophic lesions, RPE cells were absent, consistent with previous histological reports.35,36 In transition zones, RPE mosaics could be visualized in AO-TFI images; inside the transition zone, they appeared as pigmented cells, but toward the periphery, we observed well-formed mosaics of RPE cells (Fig 4G). In the transition zone, the RPE mosaic became indistinct, but still preserved, in contrast to the changes observed in cone with AO-FIO. These findings may also result from the same cone alterations described earlier, which hinder visualization in degenerating regions. The observation that RPE structure is relatively preserved whereas cone structure appears disrupted within transition zones is consistent with the possibility that toxic byproducts of impaired visual cycle activity, such as A2E and other bisretinoids, may accumulate within photoreceptors and affect cone structure and function, particularly under conditions of high light exposure and metabolic demand. However, the causal relationship and temporal sequence underlying these changes cannot be determined from the present cross-sectional data and warrant further investigation. Previous OCT studies in early STGD have reported external limiting membrane thickening and increased outer nuclear layer signal intensity as among the earliest detectable structural alterations, supporting the relevance of our observations.37 The findings of this study are consistent with the reports that used AO-SLO to examine microstructural changes in STGD.38,39 In line with these reports, our cellular-level analysis using AO imaging suggests that structural changes in photoreceptors and RPE do not necessarily progress in an identical pattern or to the same extent within the same retinal regions, supporting the importance of parallel evaluation of both cell types for understanding structural abnormalities in STGD. This study has several methodological advantages. First, AO-TFI provides a wider field of view than conventional AO-SLO, enabling detailed visualization of the transition zone and simultaneous imaging of photoreceptor and RPE changes within the same retinal field.20 Second, in a subset of cases with preserved mosaic structure, Voronoi-based quantitative parameters were applied and normalized as z-scores using healthy control data. This analysis was performed to explore the feasibility of more objective and comparable assessment of cellular-level structural features.40 Through these methodological refinements, our study not only supports prior observations,38,39 but also provides a methodological framework for more quantitative and comprehensive evaluation of cone and RPE cell status in STGD. This study includes several limitations. First, confocal and nonconfocal AO-SLO and AO-OCT imaging were not included. Second, the sample size was limited, and analyses were cross-sectional in design. Larger longitudinal studies are needed to validate the reproducibility and generalizability of our findings. Third, detailed functional testing such as microperimetry was not performed, limiting our ability to discuss the structural–functional relationship. Future studies should combine functional evaluation with structural parameters such as cone density and RPE cell density to better elucidate structure–function correlation and enhance the clinical relevance of early detection and treatment assessment. This study demonstrated that AO-FIO and AO-TFI enable cellular-level assessment of cone and RPE cells in STGD. By descriptively characterizing cellular-level structural features in STGD, our findings suggest that AO imaging may contribute to a deeper understanding of disease pathology. These observations may provide a foundation for future studies investigating treatment strategies and evaluation of therapeutic effects in STGD.