Photoreceptor Function and Structure in Retinal Areas With Intraretinal Hyperreflective Foci in Age-Related Macular Degeneration

Abstract

Purpose

To assess retinal light sensitivity in regions with intraretinal hyperreflective foci (IHRFs) in eyes with intermediate age-related macular degeneration (AMD) and examine the photoreceptor structure in these areas using adaptive optics scanning laser ophthalmoscopy (AOSLO) and optical coherence tomography (OCT).

Methods

A retrospective analysis was conducted on 82 eyes of 57 subjects (age: 76.4 ± 7.0 years) with intermediate AMD. IHRFs were identified in OCT B-scans. Drusen and subretinal drusenoid deposits (SDDs) were evaluated using multimodal imaging. Photoreceptor structure was assessed with AOSLO, and choroidal and retinal thicknesses were measured in areas with IHRFs. In 16 eyes, mesopic and scotopic light sensitivities were compared in regions with and without IHRFs but with similar SDD/drusen load in the same eye.

Results

Retinal areas with IHRFs had significantly reduced mesopic (17.19 ± 5.68 dB vs. 18.49 ± 5.35 dB, P = 0.0029) and scotopic (8.39 ± 5.67 dB vs. 9.72 ± 6.28 dB, P = 0.0096) light sensitivity compared to areas without IHRFs. AOSLO revealed disrupted cone photoreceptor structure in IHRF regions. Choroidal thickness beneath areas with IHRFs was thinner than in those without IHRFs (196.71 ± 73.31 µm vs. 202.37 ± 70.64 µm, P = 0.0211). Retinal thickness in regions with IHRFs was not significantly different from those without IHRFs (320.40 ± 31.16 µm vs. 316.92 ± 26.32 µm, P = 0.3537).

Conclusions

IHRF presence is associated with localized reduced visual function and photoreceptor degeneration in intermediate AMD. Prospective studies are warranted to further investigate the mechanisms of photoreceptor and sensitivity loss in the context of IHRF.

Major risk factors for age-related macular degeneration (AMD) progression include drusen volume, subretinal drusenoid deposits (SDDs), hyporeflective foci within drusenoid lesions, intraretinal hyperreflective foci (IHRFs), thin double-layer sign, and complete retinal pigment epithelial and outer retinal atrophy (cRORA) in the fellow eye. Among these factors, IHRFs are the strongest predictor for developing late AMD.^1,2 Understanding the pathogenesis and role of IHRFs in AMD pathophysiology is crucial for improving clinical management and preventing vision loss associated with late-stage disease progression.

Compared to classically recognized AMD lesions such as drusen and SDDs,^3–8 IHRFs are a more recently studied feature captured by optical coherence tomography (OCT).^9–20 These foci are discrete, well-circumscribed punctate lesions on OCT B-scans with reflectivity equal to or greater than that of the RPE band, appearing singly or in clusters in the outer retina above the ellipsoid zone (EZ) or in the inner nuclear layer (INL).^9,10,21,22 IHRFs on OCT colocalize with areas of RPE pigment clumping on color fundus photography.⁹

Clinicopathologic AMD studies indicate that IHRFs represent migrating RPE cells.^{11,12,16,17,23} These cells migrate toward the deep retinal capillary plexus and are referred to as “dissociated” and “sloughed” RPE cells. These migrating foci can be found in various retinal locations, including the subretinal space, outer nuclear layer (ONL), Henle's fiber layer, INL, inner plexiform layer, and even the ganglion cell layer.^12,16,19,24 Conversely, other studies suggest that IHRFs may represent activated microglia or macrophages. These reactive immune cells can ingest RPE cellular components and appear as melanosome/melanolipofuscin-laden mononuclear phagocytes.^19,25

IHRF accumulation and migration from the outer retina to the inner retina are risk factors for the development of incomplete and complete retinal pigment epithelial and outer retinal atrophy, as well as progression of AMD to the late stages.^{10,11,22,23,26} IHRFs are precursors for the development of type 3 macular neovascularization^18,27–29 through a 4-stage evolution.²⁸ The presence of IHRFs, either through a process of cellular transdifferentiation^{11,12,16,17,30} or as part of an immune response,^19,25 is an indicator of impairment of the photoreceptors, which impacts visual function. Eyes exhibiting IHRFs experience loss of visual function,^31–33 and the cellular mechanisms underlying this process are still under investigation.

Adaptive optics scanning laser ophthalmoscopy (AOSLO), operating with confocal^34–37 or non-confocal imaging modes,^38,39 is a high-resolution imaging modality to study cone photoreceptor structure at the cellular level, which can aid in defining the mechanisms of underlying visual function loss in AMD. Previous studies have examined the cone photoreceptors overlying drusen,^35,36,40,41 SDDs,^42–47 outer retinal tubulation,^48,49 and at the edge of geographic atrophy.^36,50 However, the status of cone photoreceptors in retinal regions with IHRFs remains to be characterized.

A precise assessment of photoreceptor function and structure as correlated with recognized risk factors for AMD progression is crucial for understanding the pathways of disease progression and improving the management strategies to prevent significant vision loss. In this study, we evaluated the light sensitivity in retinal regions with and without IHRFs in the same eye using microperimetry and examined the cone structure in these areas using multimodal imaging, including AOSLO and OCT.

Methods

The study followed the tenets of the Declaration of Helsinki, complied with the Health Insurance Portability and Accountability Act of 1996, and was approved by the Institutional Review Boards at the University of Alabama at Birmingham and the University of California–Los Angeles. Written informed consent was obtained from all participants.

This was a retrospective study of a group of subjects who participated in the adaptive optics ophthalmoscopy of the AMD project.^37,51 Subject enrollment, image acquisition and process, and microperimetry of cone- and rod-mediated light sensitivity have been reported elsewhere.^{37,43–47,51,52} For the readers’ convenience, we summarize the procedures here.

• •Subjects: The participants were recruited from the clinical research registry of the Department of Ophthalmology and Visual Sciences of the University of Alabama at Birmingham and through its Retina Service. All these subjects were diagnosed with AMD previously, with a best-corrected visual acuity (BCVA) of 20/100 or better and a refractive error within ±6 diopters spherical and ±3 diopters cylinder. Exclusion criteria included diabetes, history of retinal vascular occlusion, and any historical or examination features of hereditary retinal dystrophy. AMD severity was assessed by a masked, experienced grader, who graded the color digital 30° fundus photographs taken from the subjects’ eyes using the Age-Related Eye Disease Study 2 (AREDS2) severity scale for AMD. Disease severity ranged from early to advanced (AREDS steps 2 to 9).⁵³
• •Multimodal imaging: Color fundus photography (Zeiss 450 plus; Carl Zeiss Meditec, Dublin, CA, USA), en face near-infrared reflectance (NIR; λ = 830 nm), blue reflectance (BR; λ = 488 nm), fundus autofluorescence (FAF; excitation, λ = 488 nm; emission, λ > 600 nm), and spectral domain OCT (central macula 15° × 15° to 20° × 20°, 97 to 121 B-scans) (Spectralis; Heidelberg Engineering, Carlsbad, CA, USA) were acquired in all subjects.
• •AOSLO^42,54–56 was performed to image the macula with a low coherent near-infrared superluminescent diode (λ = 840 nm) through a dilated pupil (with 1.0% tropicamide and 2.5% phenylephrine hydrochloride). The AOSLO images were continuously recorded across the macula (15° × 15° to 20° × 20°). Postprocessing corrected the nonlinear distortion caused by the resonant scanner, registered successive frames to enhance the signal-to-noise ratio, and montaged all individual images with cell-to-cell precision using custom and commercial software (Photoshop; Adobe Systems, Mountain View, CA, USA). The AOSLO montage was registered to overlay on the color fundus photograph, NIR, FAF, BR images, and OCT en face image using custom software.
• •Mesopic and scotopic microperimetry: Microperimetry (MP1-S; Nidek USA, San Jose, CA, USA) was conducted to evaluate both cone- and rod-mediated light sensitivity in one eye of the subject under mesopic and scotopic conditions. To minimize ceiling and floor effects, the dynamic range of the microperimetry was extended from 20 to 30 dB. Mesopic testing utilized a white light stimulus with a maximum intensity of 2.7 cd/m² and a minimum intensity of 0.127 cd/m². Scotopic testing employed a blue light stimulus with a maximum intensity of 2.55 cd/m² and a minimum intensity of 0.003 cd/m². Light sensitivity was tested at 68 points across the macula using a 4-2 thresholding strategy and a Goldmann III stimulus (Fig. 1A). The normal sensitivity for mesopic vision is 23.59 dB (22.11 dB, 24.43 dB), and for scotopic vision, it is 17.81 dB (17.59 dB, 18.75 dB), expressed as median (lower quartile, upper quartile).⁵¹
• •AMD lesions: Drusen and SDDs were identified on OCT B-scans and confirmed on color fundus photograph, NIR, FAF, and BR images.⁴³

Figure 1.

Comparison of retinal light sensitivities in regions with and without IHRFs. (A) Retinal light sensitivity measured in the macula using microperimetry (under mesopic conditions). The magenta box delimits a retinal region with IHRFs, while the blue box indicates a near area without IHRFs. (B) Infrared (IR) reflectance retinal image showing the Early Treatment Diabetic Retinopathy Study (ETDRS) grid; adjacent retinal areas are highlighted in magenta and blue boxes where retinal light sensitivities were compared. (C) An OCT B-scan indicated by the green line C in panel B shows the retinal region contains a cluster of IHRFs (yellow arrows). The red arrow points to a SDD. (D) An OCT B-scan indicated by the green line D in panel B shows the retinal region without IHRFs but has a similar lesion load to the area with IHRFs. The white arrow points to a SDD. The light sensitivities in these two regions were averaged and compared. Scale bar: 200 µm.

IHRF Identification

IHRFs were identified based on previously published criteria.^9,10,21,22 On OCT B-scan, they were discrete, well-circumscribed punctate lesions, with a minimum size of 3 pixels and a reflectivity equal to or greater than that of the RPE band. The foci were identified singly or in clusters, located within the retina from the EZ to the INL. To better appreciate the appearance of the retina with IHRFs on various clinical imaging modalities and to understand their impact, retinal regions with IHRFs were further examined on color fundus photograph, NIR, FAF, and BR images.

Assessing the Local Retinal Light Sensitivity in Regions With and Without IHRF

To evaluate the spatial relationship between IHRFs and regional photoreceptor function, we compared mesopic and scotopic light sensitivities in retinal regions containing IHRFs with those in control regions without IHRF in the same eye. The control region was selected to match the region with IHRFs based on a similar number of drusen and/or SDDs, assessed using cross-sectional and en face OCT images. To ensure a fair comparison of light sensitivities, both regions were located within the same subfield of the Early Treatment Diabetic Retinopathy Study grid (Fig. 1).

Since IHRFs presented in singular or in clusters of various sizes, the study region was selected to include a singular or a cluster of closely distributed IHRFs, spanning 2 to 7 microperimetry testing points, with areas ranging from approximately 2.2° × 2.2° to 4.4° × 4.4°, depending on the IHRF size. Control regions contained an equivalent number of microperimetry points. Local retinal sensitivity was calculated by averaging the data from the included testing points.

Measuring Local Chorioretinal Structure in Regions With and Without IHRFs

To determine whether retinal regions with IHRFs exhibited more severe chorioretinal degeneration compared to regions without IHRFs in the same eye, retinal thickness and choroidal thickness were measured using OCT with the manufacturer integrated software (Heidelberg Eye Explorer software, version 1.7.0.0; Heidelberg Engineering). These measurements were taken in the same areas where cone- and rod-mediated light sensitivities were to be compared. Following the method of Spaide,⁵⁷ choroidal thickness was measured from the outer border of the RPE to the choroidal–scleral junction, while retinal thickness was measured from the internal limiting membrane to Bruch's membrane band. To minimize spatial variability, we measured the choroidal and retinal thicknesses near both edges of the IHRFs and used the mean values for comparison.

Statistical Analysis

The BCVA was converted to the logMAR for summary calculations. Subject age, mesopic sensitivity, scotopic sensitivity, choroidal thickness, and retinal thickness were reported as means ± standard deviations. Mesopic sensitivity, scotopic sensitivity, choroidal thickness, and retinal thickness in the selected retinal regions with and without IHRFs were compared using a paired t-test (MATLAB R2023a; The MathWorks, Natick, MA, USA). A P value <0.05 was considered significant. The differences in mesopic sensitivity, scotopic sensitivity, choroidal thickness, and retinal thickness between the paired regions were illustrated using boxplots to display statistical significance.

Results

In 225 eyes of 141 subjects, IHRFs were identified in 82 eyes of 57 subjects with intermediate AMD (AREDS grade from 5 to 8) aged 76.4 ± 7.0 years. Among the subjects with IHRFs, 16 eyes from 16 subjects were tested with microperimetry for mesopic and scotopic light sensitivity. This subgroup of subjects was aged 73.8 ± 6.1 years (range, 61 to 85 years), including 10 females and 6 males, and all were white non-Hispanic individuals. BCVA was –0.08 ± 0.16 logMAR.

In the 16 subjects who underwent microperimetry testing, retinal areas with IHRFs exhibited significantly reduced mesopic (17.19 ± 5.68 dB vs. 18.49 ± 5.35 dB, P = 0.0029) and scotopic (8.39 ± 5.67 dB vs. 9.72 ± 6.28 dB, P = 0.0096) light sensitivity compared to areas without IHRFs within the same eye (Fig. 2).

Figure 2.

Comparison of mesopic and scotopic retinal light sensitivities in regions with and without IHRFs. (A) Mesopic retinal light sensitivities in regions with and without IHRFs. (B) Scotopic retinal light sensitivities in regions with and without IHRFs. (C) Differences in mesopic (red boxplot) and in scotopic (blue boxplot) sensitivity between retinal regions with and without IHRFs. A paired t-test reveals significant differences in mesopic light sensitivity (P = 0.0029) and scotopic light sensitivity (P = 0.0096) between regions with and without IHRFs. Both boxplots of light sensitivity difference show that most (>75%) values of the differences clustered above zero, indicating consistent and significant reductions in mesopic and scotopic sensitivities in retinal areas with IHRFs compared to those without.

Although the retinal thickness in the region with IHRFs was not significantly different from the region without IHRFs (320.40 ± 31.16 µm vs. 316.92 ± 26.32 µm, P = 0.3537), the choroidal thickness beneath the area with IHRFs was significantly thinner than that beneath the area without IHRFs (196.71 ± 73.31 µm vs. 202.37 ± 70.64 µm, P = 0.0211) (Fig. 3).

Figure 3.

Comparison of retinal and choroidal thickness in regions with and without IHRFs. The blue boxplot illustrates the difference in retinal thickness between regions with and without IHRFs. A paired t-test between these two groups produced a P value of 0.3537. The orange boxplot represents the difference in choroidal thickness between regions with and without IHRFs, where a paired t-test yielded a P value of 0.0211. While the distribution of the retinal thickness differences is spread around zero, indicating no significant difference, the choroidal thickness difference shows a distribution predominantly above zero, reflecting a significant difference in choroidal thickness between areas with and without IHRFs.

On OCT, the IHRFs were classified as either single foci (Figs. 4⁵–6) or clusters (Figs. 7 and 8), within the ONL and INL, in both the fovea and parafoveal regions (Figs. 4⁵⁶⁷–8). IHRFs were associated with other AMD lesions, including drusen, and SDDs. Notably, large IHRFs were linked with advanced drusen (Fig. 5) or tall drusen (Fig. 6) or thick basal laminar deposit (BLamD) (Fig. 8). IHRFs near the foveal avascular zone were more anteriorly located in the inner retina (Figs. 6 and 7).

Figure 4.

Multimodal imaging of the retina with single small IHRFs. The subject is a 76-year-old woman with non-neovascular AMD (Age-Related Eye Disease Study grade: 7, best-corrected visual acuity: 20/25). (A) Color fundus photograph. (B) Infrared reflectance image. (C) Autofluorescence image. Retinal areas with IHRFs, indicated by small yellow boxes, are enlarged to highlight the IHRFs pointed by the yellow arrows. (D) The OCT B-scan along the green line in panel B. The yellow arrow marks the IHRFs. The red arrow points to a subretinal drusenoid deposit, while the cyan arrow indicates a druse. (E) AOSLO image of photoreceptors. Yellow arrowheads delimit the areas containing IHRFs. RPE-BL-BrM, retinal pigment epithelium–basal lamina–Bruch's membrane. Scale bars: 200 µm for OCT, 50 µm for AOSLO.

Figure 5.

Multimodal imaging of the retina with IHRFs associated with an advanced druse. The subject is a 77-year-old woman with non-neovascular AMD (Age-Related Eye Disease Study grade: 7). (A) Color fundus photograph. (B) Infrared reflectance image. (C) Autofluorescence image. Retinal areas with IHRFs, indicated by small yellow boxes, are enlarged to highlight the IHRFs. (D, E) OCT B-scans along the green lines in panel B. (F) AOSLO image of photoreceptors. Yellow arrows point to the areas containing IHRFs. Scale bars: 200 µm for OCT, 100 µm for AOSLO.

Figure 6.

Multimodal imaging of the retina with IHRFs atop a large and tall druse. Images were obtained from a 67-year-old woman diagnosed with AREDS grade 7. (A) Color fundus photograph. (B) Infrared reflectance image. (C) Autofluorescence image. Retinal areas containing IHRFs, indicated by small yellow boxes, are enlarged within the larger yellow box to show the patterns of the foci. (D) The OCT B-scan taken along the green line in panel B. (E) AOSLO image. Yellow arrowheads indicate the foci area. OPL, outer plexiform layer. Scale bar: 100 µm for both the OCT and AOSLO images.

Figure 7.

Multimodal imaging of the retina with multiple IHRFs. Images were obtained from a 67-year-old woman diagnosed with AREDS grade 5. (A) Color fundus photograph. (B) Infrared reflectance image. (C) Autofluorescence image. Retinal areas containing IHRFs, marked by small yellow boxes, are enlarged within a larger yellow box. The IHRFs are indicated by yellow arrows. (D, E) OCT B-scans taken along the green lines in panel B. (F) AOSLO image. The yellow arrowheads mark the areas with IHRFs. IPL, inner plexiform layer. Scale bar: 100 µm for both the OCT and AOSLO images.

Figure 8.

IHRFs in an eye with intermediate-stage AMD. Images were obtained from a 77-year-old woman diagnosed with AREDS grade 7. (A) Color fundus photograph. (B) Infrared reflectance image. (C) Autofluorescence image. Enlarged yellow boxes display five IHRFs, which are indicated by yellow arrows. (D–G) OCT B-scans taken along the green lines in panel B. Yellow arrows point to the IHRFs. The red arrow indicates a subretinal drusenoid deposit, while the paired magenta arrowheads mark the basal laminar deposit. Scale bar: 200 µm for OCT.

Multimodal imaging revealed characteristic alterations in retinal reflectivity and structure associated with IHRFs, within the context of AMD. As shown in Figure 4, in an eye with intermediate-stage AMD (AREDS step 7), as evidenced by large drusen (cyan arrows) and SDDs (red arrows), the retina region with a small singular IHRF (<100 µm in diameter) exhibited minimal disparity from the adjacent region (Figs. 4A–C). OCT showed only subtle RPE discontinuity (Fig. 4D). However, AOSLO disclosed distinctively reduced reflectivity and perturbed structures of cone photoreceptors in the region where the IHRFs were present. In contrast, the surrounding area maintained normal waveguiding properties and an intact photoreceptor mosaic (Fig. 4E).

Figure 5 shows multimodal imaging of a region with a large cluster of IHRFs at the top of a collapsing large druse. The glistening appearance and a punctate reflectivity of the RPE atop the druse, the pronounced hypertransmissions beneath the lesion, and the descent of the external limiting membrane (ELM) indicate the calcific stage of the druse and suggest impending collapse and atrophy.^26,58 Color fundus photo displays a distinct hypopigmentation (Fig. 5A), and other imaging modalities show slightly increased infrared reflectance (Fig. 5B), increased fundus autofluorescence (Fig. 5C), and an attenuated photoreceptor EZ band (Fig. 5D). The evident hypertransmissions and the ELM descent near the lesion suggest impending atrophy development.⁵⁹ AOSLO discloses a complete loss of the photoreceptor mosaic in the affected regions, indicating severe structural disruption (Fig. 5E). Similar observations are noted over a large druse in Figure 6 and in areas with clusters of IHRFs (Figs. 7⁸–9), underscoring the significant impact of such features on retinal structure and function.

Figure 9.

AOSLO of the retina with IHRFs. AOSLO images depict retinal regions exhibiting IHRFs, as indicated in Figure 8. (A) The retinal area containing IHRFs 1 and 2. The green arrow line indicates the location where the OCT B-scan was taken, as shown in Figure 8D. (B) The retinal area containing IHRF 3. The green arrow line indicates the OCT B-scan, as shown in Figure 8E. (C) The retinal area containing IHRF 4. The green arrow line indicates the OCT B-scan, as shown in Figure 8F. (D) The retinal area containing IHRF 5. The green arrow line indicates the OCT B-scan, as shown in Figure 8G. (E) AOSLO montage overlaid on color fundus photograph displaying the areas shown in panels A to D. AOSLO reveals numerous hyperreflective polygonal structures (yellow arrows). The average diameter of these structures is 17.35 ± 3.5 µm. Red arrowheads indicate preserved cone photoreceptors in the areas affected by IHRFs. Scale bar: 100 µm for all AOSLO images.

In Figure 7, a cluster of IHRFs is noted anterior to the outer nuclear layer in the outer plexiform layer, the inner nuclear layer, and the inner plexiform layer. Notably, these lesions developed within and on the border of the foveal avascular zone.

In regions where a large cluster of IHRFs is present (Figs. 8 and 9), we observed numerous hyperreflective polygonal structures. As illustrated in Figure 9D, the diameter of these structures is 17.35 ± 3.5 µm. These IHRFs are located above the BLamD⁵⁸ and share notable similarities with those shown by Borella and colleagues⁶⁰ (figs. 3b2 and 4F in Borella et al.⁶⁰). The photoreceptor mosaic is not visible in the area affected by these lesions.

Discussion

In this study, we assessed cone- and rod-mediated light sensitivity in retinal regions exhibiting IHRFs in patients with intermediate AMD using mesopic and scotopic microperimetry. We also examined the retinal and choroidal structures with multimodal imaging, including high-resolution AOSLO and spectral-domain OCT (SD-OCT). This study shows that retinal areas with IHRFs exhibit reduced light sensitivity compared to those without IHRFs. This diminished sensitivity is associated with structural changes in the photoreceptors, as disclosed by AOSLO. Furthermore, we observed that retinal regions with IHRFs exhibit a thinner underlying choroid versus regions without IHRFs.

This study represents a meaningful advancement in understanding IHRFs’ impact on photoreceptor function and structure, providing new insights into the factors contributing to IHRF development. Eyes with intermediate-stage AMD and IHRFs can have reduced visual function. Echols and colleagues³¹ assessed visual function through various metrics, including BCVA, contrast sensitivity, low-luminance acuity, low-luminance deficit, mesopic light sensitivity, scotopic light sensitivity, and rod-mediated dark adaptation. Their findings indicated that IHRFs were associated with visual dysfunction. Wu et al.³² demonstrated that changes in mesopic light sensitivity corresponded with the number of IHRFs present. Similarly, Fragiotta et al.,³³ using microperimetry under mesopic conditions, found a strong inverse relationship between retinal sensitivity and the presence of IHRFs. All these studies were conducted on a whole-eye basis.

In contrast, this study focuses on the differences in light sensitivity between regions with and without IHRFs in the same eye. We observed significantly reduced light sensitivity in regions with IHRFs under both mesopic (17.19 ± 5.68 dB vs. 18.49 ± 5.35 dB, P = 0.0029) and scotopic (8.39 ± 5.67 dB vs. 9.72 ± 6.28 dB, P = 0.0096) conditions compared to areas without IHRFs. These findings highlight a clear spatial association between the presence of IHRFs and decreased retinal light sensitivity.

The intraretinal location of IHRFs implies a direct impact on photoreceptors. While the most prominent pathologic damage in AMD involves the RPE, Bruch's membrane (BrM), and choriocapillaris, the degeneration, dysfunction, and loss of photoreceptors ultimately cause vision loss. Normal photoreceptor function relies on both the integrity of individual cells and the surrounding extracellular matrix. The presence of IHRFs suggests a compromised retina–blood barrier and a disrupted supporting system for photoreceptors. AOSLO reveals distinctively reduced reflectivity and disrupted cone photoreceptor structure in areas affected by IHRFs. In contrast, the surrounding regions demonstrate normal waveguiding properties and exhibit an intact photoreceptor mosaic (Figs. 4⁵⁶–7, 9). Even beneath a small IHRF, photoreceptors lose their typical cell mosaic (Fig. 4), which is formed by the cells’ waveguiding property. This loss of waveguiding function reduces the cells’ ability for photon capture, leading to decreased light sensitivity.⁴⁸ Thus, our work supports previous research by investigating the direct impact of IHRFs on photoreceptor function and structure. This finding aligns with a recent clinicopathologic study, which suggests that IHRFs not only predict or signify a higher risk of AMD progression but correlate directly with retinal dysfunction at their topographic location.¹⁶

The examination of the spatial correlation between IHRFs and retinal function and chorioretinal structures offers insight into the pathogenesis of IHRFs. While IHRFs may represent phagocytes (macrophages or microglia) that have ingested RPE cells,²⁵ clinicopathologic studies strongly support the concept that IHRFs in AMD eyes are RPE cells migrating toward the deep retinal capillary plexus.^12,16,19,24 However, the underlying reasons for RPE migration toward the inner retina remain incompletely understood. Evidence for RPE transdifferentiation, possibly driven by ischemia, suggests an epithelial–mesenchymal transition process. In our study, we have observed that IHRFs are associated with various AMD lesions, including drusen and SDDs (Figs. 4⁵⁶⁷⁸–9). Notably, large IHRFs are linked to advanced-stage drusen (Fig. 5), tall drusen (Fig. 6), or thick BLamD (Fig. 8). The accumulation of drusen and BLamD can result in a significant drop of oxygen tension across the chorioretinal complex.⁶¹ In eyes with AMD, attenuation of the choriocapillaris has been observed under areas of drusen and SDDs.⁶² OCT angiography (OCTA) reveals increased flow deficits in the choriocapillaris of eyes with SDDs and drusen.^63,64 Consequently, reduced or impaired metabolic supply due to choriocapillaris dysfunction and the accumulation of extracellular lesions on both sides of the RPE likely trigger RPE cells to leave their monolayer and migrate toward the vascularized retina.

Our results support this hypothesis, as we found that the choroid beneath areas with IHRFs was significantly thinner than the choroid beneath areas without IHRFs (196.71 ± 73.31 µm vs. 202.37 ± 70.64 µm, P = 0.0211). Choroidal thinning is a recognized feature of chorioretinal degeneration associated with AMD.^45,52,57,62 Thus, the presence of IHRFs may reflect the overall chorioretinal degenerative process. However, our measurement indicates no significant difference in retinal thickness between the areas with IHRFs and the areas without IHRFs (320.40 ± 31.16 µm vs. 316.92 ± 26.32 µm, P = 0.3537), suggesting that while IHRFs are present in the neuroretina, they do not substantially affect retinal thickness. Future studies should investigate the correlation between choroidal thickness changes and IHRFs in larger populations over time.

A key strength of this study is the use of multimodal imaging, including high-resolution AOSLO and SD-OCT to examine the retinal and choroidal structures in areas containing IHRFs. Additionally, microperimetry was employed to measure cone- and rod-mediated light sensitivity. Classical color fundus photography offers a broad overview of AMD presence and disease severity. OCT provides high-resolution cross-sectional imaging of retinal structures. However, both imaging modalities have inherent limitations as they lack en face plane resolution sufficient to visualize photoreceptors, which are critical components in AMD's degenerative processes and most directly correspond to visual function. Consequently, degenerative changes at the photoreceptor level are largely invisible. Furthermore, the functional characteristics of AMD, such as visual acuity and light sensitivity, often show poor correlation with disease severity, complicating the accurate assessment and interpretation of disease progression. Confocal AOSLO overcomes some of these limitations by providing cellular-level resolution, allowing for a detailed, subjective assessment of photoreceptor directionality, thereby offering deep insights into retinal health and structure that are not accessible with traditional imaging methods.

However, the study has limitations, including the absence of imaging modalities that could provide more detailed insights into the photoreceptor status beneath IHRFs, choroidal circulation, and RPE health and function. Additionally, our work is limited by a small sample size and the retrospective design.

While confocal AOSLO revealed disrupted photoreceptor structure and loss of waveguiding ability in retinal regions containing IHRFs, indicative of disruptions in their supporting retinal cellular matrix, it cannot capture light scattered non-ballistically by retinal structures. The reduction in photoreceptor reflectivity observed on AOSLO may result from shadowing by the IHRFs⁶⁵ or perturbations in photoreceptor alignment in areas with small, singular IHRFs (Figs. 4 and 7), as indicated by the intact EZ and ELM. In contrast, the absence of the typical photoreceptor mosaic under large IHRFs suggests that photoreceptors may have been damaged (Figs. 5, 6, 8, and 9), as both the EZ and ELM exhibit significant disruption. Although we did not observe specific cell structures, such as macrophages, this absence does not necessarily imply that these cells are not present. Advanced non-confocal AOSLO imaging modes^38,39,60 may offer the potential to detect these structures, providing further insights into the cellular environment surrounding IHRFs.

Although our study identified mean differences of 1.30 dB in mesopic and 1.33 dB in scotopic retinal sensitivities for areas with versus without IHRFs, these values fall below the commonly accepted threshold for clinical significance (>3 dB⁶⁶ or even >4 dB^67,68). As such, a ∼1.3 dB difference may not translate into perceptible functional impairment under standard clinical assessments. Nevertheless, these findings provide quantitative insights into the role of IHRFs in the pathophysiology of AMD, suggesting that IHRFs may directly impact visual performance, in addition to being a recognized risk factor for AMD progression. This perspective aligns with clinicopathologic findings¹⁶ and highlights the significance of IHRFs in understanding the functional consequences of AMD before the disease progresses to advanced stages.

The limited number of subjects who underwent microperimetry testing and the sparse distribution of test points further reduced the statistical power for subgroup analyses, preventing a direct correlation between functional sensitivity reductions and specific IHRF features, such as size, clustering, or retinal layer location. Adaptive optics (AO) enhanced microperimetry has enabled the assessment of light sensitivity at the scale of single cone photoreceptors by leveraging real-time eye tracking and precise targeting of specific retinal regions.^69–71 This technique enables a more definitive assessment of localized functional deficits with finer spatial resolution and higher precision. Future studies utilizing AO microperimetry could significantly advance our understanding of the functional consequences of IHRFs.

Given the incompletely characterized composition, phenotype, and dynamic nature of IHRFs, further investigation into the pathogenesis of this important pathologic entity is essential. The varying dimensions of IHRFs suggest that they may be dynamic structures with a life cycle, raising critical questions about when or under what conditions IHRFs develop. These questions could be explored using multimodal longitudinal imaging that tracks the development of IHRFs over time and longitudinally assesses the health of the RPE and choriocapillaris in relation to different stages of SDDs and drusen.

Therefore, the assessment of RPE function is vital. Fluorescence lifetime imaging ophthalmoscopy holds promise for objectively measuring the metabolic function of the RPE. Dysli et al.⁷² reported that IHRFs are associated with prolonged fluorescence lifetime measurements (650 to 700 ps), even in the presence of surrounding short fluorescence lifetime drusen (300 to 400 ps). Similarly, Hammer et al.⁷³ demonstrated that hyperpigmentation, including IHRFs captured with OCT, exhibit significantly longer fluorescence lifetimes compared to fundus areas without hyperpigmentation or drusen. The observed prolongation may stem from changes in the intracellular environment or a rebalancing of fluorophore content in transdifferentiating and migrating RPE cells.^74–77

In conclusion, IHRF presence indicates localized photoreceptor degeneration and reduced visual function in intermediate AMD. Given the high risk of advanced AMD associated with IHRFs, awareness of the clinical significance of the regional IHRF load is important. If replicated with longitudinal studies in larger cohorts, these findings could enhance our understanding of the pathophysiology of IHRFs and its role in AMD progression.