HONU: A Multicenter, Prospective, Observational Study of the Progression of Intermediate Age-Related Macular Degeneration Authors: Robyn H. Guymer, Zhichao Wu, Simon S. Gao, Miao Zhang, Verena Steffen, Chen Chen, Jiameng Zhang, Katie M. Litts, Michael C. Chang, Justin Elstrott, Brian Yaspan, Seema Datta, Jeffrey R. Willis, Liang Liu, Hao Chen, HONU investigators Published: 2026 ABSTRACT PurposeLarge and lengthy clinical trials are currently required to study the efficacy of interventions aiming to slow the progression of the early stages of age-related macular degeneration (AMD). To enable the design of more feasible trials, an observational study HONU aims to comprehensively study eyes with intermediate AMD (iAMD) to explore novel clinical endpoints and patient enrichment strategies.DesignProspective, multicenter, international observational study.ParticipantsFour hundred patients, whose study eye is required to have large drusen and pigmentary change, but without evidence of any atrophic changes other than incomplete retinal pigment epithelium and outer retinal atrophy. The nonstudy eye is required to have some evidence of atrophy.MethodsPatients are assessed every 12 weeks for up to 36 months, with standard opthalmic examinations and multimodal retinal imaging as well as a range of novel visual functional tests. Patient-reported outcomes and biomarker samples are also collected. Structural changes on multimodal imaging are evaluated longitudinally and their relationship with visual function parameters analyzed. Exploratory interim analyses are planned.Main Outcome MeasuresThe rates of conversion from baseline high risk iAMD to more advanced atrophic AMD stages, and the rate of ellipsoid zone loss.ResultsRecruitment commenced in May 2022 and 400 patients were enrolled by August 2024, allowing for up to 3 years follow-up. Enrollment occurred at 72 sites across 6 countries. The median age of the patients at baseline was 75 years, ranging from 50 to 93 years. The cohort is predominantly female, making up 69% of patients, and 95% identified as White.ConclusionsHONU will provide critical and robust natural history data on AMD progression in patients with high-risk iAMD in the study eye. Potential novel earlier clinical trial endpoints and enrichment strategy for high-risk patients will be determined. Knowledge gained will inform design and feasibility of early interventional studies aiming to slow progression to vision-threatening late AMD.Financial Disclosure(s)Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article. Significant progress has been made in the past 2 decades in treating age-related macular degeneration (AMD). This includes the development of anti-VEGF treatments for neovascular AMD,1, 2, 3 and more recently, complement inhibitors for slowing the progression of geographic atrophy (GA).4, 5, 6 Despite these advances, there remains a significant unmet need to effectively slow or stop the progression of AMD from its earlier stages to the more advanced vision-threatening late stages. However, current treatment strategies for the early stages of AMD primarily focus on lifestyle modifications and supplement use.7, 8, 9 A major obstacle in developing early interventions for AMD is the difficulty of conducting feasible trials that can detect clinically meaningful treatment effects within a manageable sample size and study duration. Overcoming this challenge requires enriching the clinical trial cohort with patients who are most likely to progress, as well as the establishment of robust outcome measures that can serve as early endpoints or surrogates in interventional trials. In studies that have evaluated early interventions for AMD, the presence of large drusen and/or pigmentary abnormalities on color fundus photographs have been used to identify higher risk patients for enrollment.10, 11, 12, 13, 14 Within eyes with drusen and/or pigmentary abnormalities on clinical examination, OCT has allowed the visualization of additional anatomical features that signal a higher risk of progression to traditionally determined late-stage AMD. Key OCT findings include nascent GA (nGA),15,16 incomplete retinal pigment epithelium and outer retinal atrophy (iRORA),17, 18, 19 and complete retinal pigment epithelium and outer retinal atrophy (cRORA).17,18,20 OCT facilitates identification and quantification of other risk factors such as reticular pseudodrusen,21, 22, 23, 24, 25, 26, 27 also known as subretinal drusenoid deposits,28 drusen volume, hyperreflective foci, and ellipsoid zone (EZ) integrity.29, 30, 31, 32, 33, 34 Some of these OCT features could also potentially serve as earlier surrogate endpoints, but require further investigation to determine their robustness as an endpoint in an interventional clinical trial.11,34 To date, the onset of an anatomic endpoint such as OCT-defined atrophic AMD changes (i.e. nGA) has not been accepted as an incident endpoint for regulatory authorities to approve an intervention. However, a recent Delphi study of international AMD experts reported that there was consensus from this group that the onset of atrophic AMD changes, which have previously been shown in clinical studies to be associated with marked functional loss, would be an appropriate clinical endpoint.35 However, to date, a progression endpoint such as the mean rate of change in GA lesion size over a 12-month or longer period has been the acceptable primary efficacy endpoint by the US Food and Drug Administration (FDA) when approving treatments for GA.4,5 Currently, any AMD interventional trial that starts with patients presenting with earlier disease must wait for the onset and growth of GA to assess the effectiveness of treatments in slowing the rate of growth in these areas over time. These endpoints are impractical for early intervention, given the slow progression of the disease and the need to show a change in the slope of growth over time. Therefore, there is an unmet need for an early anatomical endpoint that can act as a robust surrogate for the development of GA. Establishing such an endpoint requires additional evidence to demonstrate a strong correlation between a potential early biomarker and late AMD. Demonstrating their association with functional loss is also critically important when considering regulatory approval of interventions. To this end, we are conducting HONU (NCT05300724),36 a well-powered, prospective, multicenter, international, observational study of patients with intermediate AMD (iAMD) in the study eye. The objective is to gain a comprehensive understanding of the clinical features, imaging characteristics, genetic risks, aqueous humor and plasma biomarkers, visual function, and the structure–function relationships relevant to disease progression in this cohort. This study utilizes innovative modalities and approaches, such as individualized assessment of visual function in topographically relevant areas of the macula37, 38, 39, 40 and employs artificial intelligence (AI) algorithms to facilitate the screening and assessment processes. Additionally, HONU specifically targets a population with significant AMD disease burden to increase likelihood of progression to the late AMD stages in the study eye, within a 36-month trial period. The design enables accurate tracking of OCT-defined disease progression over time and will allow longitudinal assessment of novel anatomical features such as the rate of EZ loss which was recently used as an endpoint to allow the registration of Encelto by FDA for macular telangiectasia type 2. Despite growing interest in early disease progression, no natural history study to date has specifically targeted this higher-risk subgroup or employed a design that allows for individualized functional testing of retinal regions at highest risk for progression. This strategic focus will generate crucial novel data enabling identification of prognostic markers of rapid progression as well as establishing robust early endpoints. Critical new insights from HONU will help inform the design of feasible clinical trials that assess interventions aimed at slowing disease progression before sight-threatening late-stage disease develops in high-risk patients. Methods Study Design and Objectives HONU is a prospective, multicenter, observational study designed to enroll 400 patients with iAMD in their study eye. The study is being conducted across 72 investigational sites globally. The study protocols were approved by the Institutional Review Board/Independent Ethics Committee (Supplementary Table 2, available at www.ophthalmologyscience.org) prior to patient recruitment and conducted in accordance with ethical principles of the Declaration of Helsinki, the International Conference on Harmonization guidelines, and all applicable local laws and regulations. Each subject provided written informed consent for participation before enrollment. The study is a 36-month study with visits every 12 weeks to evaluate multimodal imaging biomarkers, visual function, fluid, and genetic biomarkers (Fig 1). The screening period was up to 28 days and the observation period is 144 weeks, unless both eyes receive treatment for late AMD (e.g. exudative macular neovascularization or GA).Figure 1Study schema. AMD = age-related macular degeneration; BCVA = best-corrected visual acuity; CFP = color fundus photography; FA = fluorescein angiography; FAF = fundus autofluorescence; FPF = flavoprotein fluorescence; iAMD = intermediate age-related macular degeneration; LLBCVA = low-luminance best-corrected visual acuity; MNV = macular neovascularization; NI = near-infrared imaging; Q24W = every 24 weeks; Q48W = every 48 weeks; qCSF = quantitative contrast sensitivity function; SD-OCT = spectral-domain OCT; SS-OCTA = swept-source OCT angiography; VILL = vision impairment in low luminance; W = week. Study schema. AMD = age-related macular degeneration; BCVA = best-corrected visual acuity; CFP = color fundus photography; FA = fluorescein angiography; FAF = fundus autofluorescence; FPF = flavoprotein fluorescence; iAMD = intermediate age-related macular degeneration; LLBCVA = low-luminance best-corrected visual acuity; MNV = macular neovascularization; NI = near-infrared imaging; Q24W = every 24 weeks; Q48W = every 48 weeks; qCSF = quantitative contrast sensitivity function; SD-OCT = spectral-domain OCT; SS-OCTA = swept-source OCT angiography; VILL = vision impairment in low luminance; W = week. Standard demographic and key systemic data are being collected via questionnaires. Ocular assessments include slit-lamp examination and dilated binocular indirect ophthalmoscopy. Functional testing consists of best-corrected visual acuity (BCVA), low-luminance BCVA, quantitative contrast sensitivity function (qCSF), and mesopic microperimetry. Imaging assessments include spectral-domain OCT (SD-OCT), fundus autofluorescence, near-infrared imaging, color fundus photography, flavoprotein fluorescence (FPF), and optional swept-source OCT angiography (SS-OCTA) (Fig 1). Microperimetry, qCSF, FPF, and SS-OCTA are undertaken at selected sites. In addition to these evaluations, plasma samples and optional aqueous humor samples are being collected for biomarker analysis. Patient-reported outcomes are also assessed annually using the Vision Impairment in Low Luminance (VILL) questionnaire. Although only 1 eye is chosen as the study eye, both eyes are assessed throughout the course of the study. HONU has 2 primary objectives. To determine the rate of progression from baseline to more progressed atrophic stages of AMD, assessed by multimodal imaging defined anatomical changes (see definitions in “Ocular imaging grading” in Supplementary Materials) such as nGA or cRORA on SD-OCT, or GA as defined on fundus autofluorescence or color fundus photography, in the study eye. Assessment of the rate of photoreceptor loss, based upon loss of EZ will also be done to determine its potential utility as an endpoint for early intervention studies. Exploratory objectives of this study are outlined in Supplementary Table 1, available at www.ophthalmologyscience.org, investigating imaging, molecular, and genetic biomarkers, as well as visual functions. Participants Key Inclusion Criteria Patients are required to be ≥50 and <95 years of age with BCVA letter score of 72 letters (Snellen equivalent of 20/40) or better using ETDRS chart in the study eye. Age-related macular degeneration status in the study eye, for eligibility, is iAMD with a high risk of progression (high-risk subset of iAMD), defined by the presence of >1 large drusen ≥125 μm and pigmentary abnormalities within a circle with a radius of 3600 μm centered on the fovea. The nonstudy eye is required to have the presence of any atrophic changes, determined on the entire volume scan of the SD-OCT to increase the risk of disease progression (Table 1).Table 1Key Ocular Inclusion and Exclusion CriteriaStudy EyeNonstudy EyeInclusion criteria•High-risk intermediate AMD with >1 large drusen ≥125 μm, AND•AMD pigmentary abnormalities within a circle with a radius of 3600 μm centered on the fovea•Any atrophic signs on the entire volume scan of the SD-OCTExclusion criteria•Any atrophic signs on the entire volume scan of the SD-OCT, except for iRORA•Prior or active exudative MNV•Prior or active exudative MNVAMD = age-related macular degeneration; iRORA = incomplete retinal pigment epithelium and outer retinal atrophy; MNV = macular neovascularization; SD-OCT = spectral-domain OCT. Table 1Key Ocular Inclusion and Exclusion CriteriaStudy EyeNonstudy EyeInclusion criteria•High-risk intermediate AMD with >1 large drusen ≥125 μm, AND•AMD pigmentary abnormalities within a circle with a radius of 3600 μm centered on the fovea•Any atrophic signs on the entire volume scan of the SD-OCTExclusion criteria•Any atrophic signs on the entire volume scan of the SD-OCT, except for iRORA•Prior or active exudative MNV•Prior or active exudative MNVAMD = age-related macular degeneration; iRORA = incomplete retinal pigment epithelium and outer retinal atrophy; MNV = macular neovascularization; SD-OCT = spectral-domain OCT. Key Ocular Inclusion and Exclusion Criteria •High-risk intermediate AMD with >1 large drusen ≥125 μm, AND High-risk intermediate AMD with >1 large drusen ≥125 μm, AND •AMD pigmentary abnormalities within a circle with a radius of 3600 μm centered on the fovea AMD pigmentary abnormalities within a circle with a radius of 3600 μm centered on the fovea •Any atrophic signs on the entire volume scan of the SD-OCT Any atrophic signs on the entire volume scan of the SD-OCT •Any atrophic signs on the entire volume scan of the SD-OCT, except for iRORA Any atrophic signs on the entire volume scan of the SD-OCT, except for iRORA •Prior or active exudative MNV Prior or active exudative MNV •Prior or active exudative MNV Prior or active exudative MNV AMD = age-related macular degeneration; iRORA = incomplete retinal pigment epithelium and outer retinal atrophy; MNV = macular neovascularization; SD-OCT = spectral-domain OCT. Key Exclusion Criteria Study eye exclusion criteria require that there be no evidence of any atrophic changes except iRORA in the study eye. Evidence of prior or active exudative macular neovascularization is an exclusion criterion in both eyes. Other usual exclusionary criteria applied (Supplementary materials). Ocular Functional Testing At each study visit, ocular assessments are performed on both eyes except for microperimetry, which is conducted at alternate visits on the study eye and the nonstudy eye. Best-corrected visual acuity is measured using the Precision Vision Distance Visual Acuity Test chart (Precision Vision) starting at 4 m before pupil dilation. Low-luminance BCVA is measured by placing a 2.0-log unit neutral density filter (Kodak Wratten 2.0 neutral density filter) over the best correction for the eye and having the participant read the standard illuminated ETDRS chart. Contrast sensitivity is assessed at selected sites by the qCSF testing on the Manifold Contrast Vision Meter (Adaptive Sensory Technology). The test is performed prior to pupil dilation and is undertaken only at selected sites. The qCSF uses an adaptive testing algorithm to rapidly measure contrast sensitivity. Mesopic microperimetry is performed at selected sites using a Macular Integrity Assessment microperimeter (CenterVue) which presents a Goldmann III stimulus with a dynamic range of 0-36 dB and a maximum luminance of 318.3 cd/m2 on a background of 1.27 cd/m2. Microperimetry testing includes using a fixed stimulus pattern (approximately 3.5 mm or 12° in diameter) centered on the fovea with 37 stimulus points, and an individualized targeted pattern (approximately 1.2 mm or 4° in diameter) centered on 1 high-risk lesion, also with 37 stimuli points. (See the Supplementary Materials section for details on targeted microperimetry testing). Intraocular pressure measurements, slit lamp examination with dilated binocular indirect high-magnification ophthalmoscopy, and all imaging are conducted after all functional tests are performed. Ocular Imaging Assessments These assessments include:•SD-OCT, fundus autofluorescence, and near-infrared imaging: Performed using the Spectralis HRA + OCT or Spectralis OCT with BluePeak module (Heidelberg Engineering GmbH/Inc). Details of the OCT imaging protocols are provided in the Supplementary Materials.•Color fundus photography: Captured using any digital fundus camera with a 45–60° field of view and a minimum output resolution of 1536 × 1536 pixels, or the Zeiss Clarus 500/700 (Carl Zeiss Meditec). Required imaging fields include field 1M (optic disc-centered view), field 2 (fovea-centered view), and field 3M (temporal-to-macula view).•FPF: Conducted at selected sites using the Ocumet Beacon system (OcuSciences). Flavoprotein fluorescence is an imaging modality used to assess retinal mitochondrial stress in the retinal pigment epithelium and photoreceptors.•SS-OCTA: Performed at selected sites using the ZEISS PLEX Elite 9000 system (Carl Zeiss Meditec), acquiring 3 × 3 mm and 6 × 6 mm scans to assess the changes in choriocapillaris perfusion.•Keratometry: Performed to obtain corneal curvature values, which are necessary for accurate calibration and scaling of retinal imaging data.•Fluorescein angiography: When exudative macular neovascularization is suspected. For fluorescein angiography, field 2 is captured at the start and end of the injection and throughout the early phase in the study eye. During the mid-phase, fields 1M, 2, and 3M are obtained in both eyes. In the late phase, field 2 is captured again in both eyes. •SD-OCT, fundus autofluorescence, and near-infrared imaging: Performed using the Spectralis HRA + OCT or Spectralis OCT with BluePeak module (Heidelberg Engineering GmbH/Inc). Details of the OCT imaging protocols are provided in the Supplementary Materials. SD-OCT, fundus autofluorescence, and near-infrared imaging: Performed using the Spectralis HRA + OCT or Spectralis OCT with BluePeak module (Heidelberg Engineering GmbH/Inc). Details of the OCT imaging protocols are provided in the Supplementary Materials. •Color fundus photography: Captured using any digital fundus camera with a 45–60° field of view and a minimum output resolution of 1536 × 1536 pixels, or the Zeiss Clarus 500/700 (Carl Zeiss Meditec). Required imaging fields include field 1M (optic disc-centered view), field 2 (fovea-centered view), and field 3M (temporal-to-macula view). Color fundus photography: Captured using any digital fundus camera with a 45–60° field of view and a minimum output resolution of 1536 × 1536 pixels, or the Zeiss Clarus 500/700 (Carl Zeiss Meditec). Required imaging fields include field 1M (optic disc-centered view), field 2 (fovea-centered view), and field 3M (temporal-to-macula view). •FPF: Conducted at selected sites using the Ocumet Beacon system (OcuSciences). Flavoprotein fluorescence is an imaging modality used to assess retinal mitochondrial stress in the retinal pigment epithelium and photoreceptors. FPF: Conducted at selected sites using the Ocumet Beacon system (OcuSciences). Flavoprotein fluorescence is an imaging modality used to assess retinal mitochondrial stress in the retinal pigment epithelium and photoreceptors. •SS-OCTA: Performed at selected sites using the ZEISS PLEX Elite 9000 system (Carl Zeiss Meditec), acquiring 3 × 3 mm and 6 × 6 mm scans to assess the changes in choriocapillaris perfusion. SS-OCTA: Performed at selected sites using the ZEISS PLEX Elite 9000 system (Carl Zeiss Meditec), acquiring 3 × 3 mm and 6 × 6 mm scans to assess the changes in choriocapillaris perfusion. •Keratometry: Performed to obtain corneal curvature values, which are necessary for accurate calibration and scaling of retinal imaging data. Keratometry: Performed to obtain corneal curvature values, which are necessary for accurate calibration and scaling of retinal imaging data. •Fluorescein angiography: When exudative macular neovascularization is suspected. For fluorescein angiography, field 2 is captured at the start and end of the injection and throughout the early phase in the study eye. During the mid-phase, fields 1M, 2, and 3M are obtained in both eyes. In the late phase, field 2 is captured again in both eyes. Fluorescein angiography: When exudative macular neovascularization is suspected. For fluorescein angiography, field 2 is captured at the start and end of the injection and throughout the early phase in the study eye. During the mid-phase, fields 1M, 2, and 3M are obtained in both eyes. In the late phase, field 2 is captured again in both eyes. AI Algorithm Development Several AI algorithms, including one referred to as EyeNotate, have been, or are being developed, to assist HONU imaging biomarker extraction. Some of these algorithms are also undergoing validation on HONU images. Applications include nGA lesion detection on OCT,41 drusen volume calculation, retinal OCT layers (e.g. EZ), reticular pseudodrusen/subretinal drusenoid deposits, and hyperreflective foci segmentation.42,43 Of these, the EyeNotate retina segmentation algorithm was trained on annotated OCT B-scans from participants enrolled in iAMD and GA clinical trials (NCT01790802 and NCT02399072). Preliminary validation of its clinical robustness, particularly for detecting outer retinal disruption, has been described previously.42 EyeNotate has demonstrated performance comparable to that of human experts in quantifying areas of outer retinal loss (unpublished data). Further validation of EZ loss measurement is ongoing with additional human-in-the-loop annotation by the Utah Retinal Reading Center on HONU data. Poor quality scans determined by the reading center are excluded from the downstream analysis. Biomarker Sample Collection and Analysis The study includes a comprehensive biomarker collection to investigate genetic, ocular, and systemic factors associated with iAMD progression. Whole blood samples are being collected for DNA extraction, to identify novel genetic factors predictive of iAMD progression. Nonfasting plasma samples were collected at baseline and then will continue annually, while fasting samples will be obtained at the final study visit. Aqueous humor samples were collected from the study eye at baseline (day 1), with the nonstudy eye aqueous collected at week 12. Thereafter, samples from each eye will be collected annually, with the collections for the 2 eyes occurring at different visits relative to their initial collection dates, enabling within-patient comparisons across different disease states. The procedures for aqueous humor and plasma sample collection are detailed in the supplementary material. Both aqueous humor and plasma samples will be stored at –70°C until analysis. Proteomic analysis will be conducted using the Olink® Explore HT platform (Olink Proteomics) on plasma and aqueous humor samples. Future additional analyses may be conducted at the conclusion of the study to generate additional disease-relevant insights and analyzed using advanced technologies, such as proteomics. The objectives include identifying baseline and longitudinal biomarkers associated with iAMD progression, assessing their variability over time, and exploring their relationships with imaging features, genetic factors, and demographic or environmental variables. Patient-Reported Outcomes The VILL questionnaire44 was administered at baseline and then is repeated every 48 weeks until the end of the study. The questionnaire captures vision-related quality of life for all stages of AMD. The VILL includes 33 items that measure 3 subscales: Reading and accessing information, mobility and safety, and emotional wellbeing. Higher scores on the subscales indicate better function and vision-related quality of life. The VILL is self-administered or interviewer-administered by the site staff at the scheduled visit prior to the performance of any other assessments. Safety and Tolerability Assessments The World Health Organization toxicity grading scale is used to assess adverse event severity for ocular and systemic adverse events. If a patient experiences an adverse event directly resulting from a per-protocol procedure, this event is documented and monitored per investigator judgment. Statistical Analyses A primary objective of this study is to estimate conversion rates and timelines from baseline high-risk iAMD to signs of atrophy and subsequently to GA. Precision-based sample size calculations were conducted using the Wald normal approximation method, implemented via the prec_prop function in the R package presize. Assuming a 3-year event rate of 30% and a desired 95% confidence interval width of ±5% (total width of 10%), the required number of evaluable patients was estimated at approximately 320. To account for an anticipated 20% dropout rate, the total planned sample size was increased to 400. All patients who are enrolled and have ≥1 postbaseline assessment will be included in the analyses. The conversion event rates to more advanced atrophic AMD stages (nGA, cRORA, or GA) at each scheduled visit will be determined, including 95% confidence intervals. Kaplan–Meier plots will illustrate time-to-event data. Cox proportional hazards models will be used to evaluate the association between various baseline covariates and the development of conversion events. For longitudinal continuous outcomes such as EZ loss, descriptive statistics will be provided at each visit. Linear mixed models will be used to estimate rates of change in EZ loss over time. Baseline values will be included as covariates where appropriate. Visual function assessment will include evaluating changes from baseline BCVA under both normal and low luminance conditions, as well as the low luminance deficit; descriptive summaries of these visual function outcomes and their changes over time will be provided. Mesopic microperimetry will be used to assess changes in mean sensitivity, the percentage of test locations with a ≥7 dB loss from baseline, and the proportion of locations exhibiting deep sensitivity defects (≤10 dB) over time, using a high-density, targeted approach.45 These longitudinal changes will be analyzed in relation to baseline atrophic AMD severity and its progression over time. The study will also assess changes in contrast sensitivity as measured by qCSF, along with the reliability of this metric. All analyses will be adjusted for relevant baseline confounders, as appropriate. Final analysis will be performed once all enrolled patients who did not discontinue early have completed the 144-week study period. Results Four-hundred patients (Table 2) were enrolled from 72 investigational sites from 6 countries: the United States, Germany, France, Italy, Israel, and Australia (Fig 2). Recruitment started in May 2022 and was completed in August 2024. Over 1000 patients were screened, with a screen failure rate of approximately 60%. Among the enrolled patients, 101 (25%) opted into the microperimetry study, 163 (41%) into the qCSF testing, 86 (22%) into the FPF imaging, and 84 (21%) into the SS-OCTA. Aqueous humor samples were collected from 123 patients (31%) at the baseline visit.Table 2Baseline Demographics of PatientsCharacteristicAll Patients (N = 400)Age in years, median (range)75 (50-93)Sex Male124 (31%) Female276 (69%)Race White379 (95%) Unknown20 (5%) Asian1 (0%)Smoking history Never used245 (61%) Previous user142 (36%) Current user13 (3%)Weight in kg, median (range)77 (41-168)BMI in kg/m2, median (range)27 (15-66)AREDS/AREDS2 supplements (yes)216 (54%)AREDS = Age-related Eye Disease Study; BMI = body mass index.Figure 2Study enrollment sites. Table 2Baseline Demographics of PatientsCharacteristicAll Patients (N = 400)Age in years, median (range)75 (50-93)Sex Male124 (31%) Female276 (69%)Race White379 (95%) Unknown20 (5%) Asian1 (0%)Smoking history Never used245 (61%) Previous user142 (36%) Current user13 (3%)Weight in kg, median (range)77 (41-168)BMI in kg/m2, median (range)27 (15-66)AREDS/AREDS2 supplements (yes)216 (54%)AREDS = Age-related Eye Disease Study; BMI = body mass index. Baseline Demographics of Patients AREDS = Age-related Eye Disease Study; BMI = body mass index. Study enrollment sites. The median age of the HONU cohort at baseline is 75 years, ranging from 50 to 93 years. The cohort predominantly consists of women, making up 69% of the patients. The racial composition is largely White (95%), with 1 Asian patient, and 20 patients whose race was indicated as unknown. The median weight of the group is 77 kg, with a range from 41 to 168 kg. The median body mass index is 27 kg/m2, spanning from 15 to 66 kg/m2. At baseline, the majority of patients had never smoked (61%), and 54% were on Age-related Eye Disease Study (AREDS)/AREDS2 supplements. Discussion Currently, most AMD interventions target eyes that have already progressed to vision-threatening late-stage disease. However, intervening earlier—prior to significant vision loss—with the goal of preventing or slowing progression to late AMD is a highly desired objective. Apart from dietary supplementation46, 47, 48 and lifestyle modifications,49,50 there are no widely adopted, well-validated treatments proven to effectively slow or prevent the development of late-stage AMD. This highlights a significant unmet medical need. A major challenge in advancing early interventions is the requirement for large, lengthy, and costly clinical trials to demonstrate efficacy. HONU will create a robust and unprecedented longitudinal dataset on the natural history of individuals with iAMD in the study eye and atrophic AMD changes in the nonstudy eye, aiming to provide data to enable the design of more feasible future interventional trials. This will be achieved by assessing operational feasibility, identifying clinical enrichment strategies, helping identify appropriate study outcome measures, and providing more accurate estimates of sample size and study timelines. Together, these elements are critical for enhancing the design and execution of future clinical trials, ensuring they are both efficient and tailored to address specific therapeutic goals effectively. Utilizing a novel enrichment strategy, HONU applies heterogeneous eye eligibility criteria to assemble a cohort at high risk for disease progression. The study eye requires large drusen and pigmentary abnormalities, while the nonstudy eye is required to have additional features of atrophy. The presence of signs of advancing atrophic changes in the nonstudy eye increases the likelihood that the study eye, with nonlate AMD changes, will progress to advanced disease.51 To balance eligibility with the ability to longitudinally assess disease progression, the study permitted only the earliest OCT signs of atrophy (iRORA)16,18,19 in the study eye at baseline, while excluding more advanced atrophic changes. This approach avoided excluding a substantial proportion of iAMD eyes while still enabling accurate tracking of progression to more advanced OCT-defined atrophic endpoints over time. However, complex imaging-based eligibility criteria are associated with high screen failure rates. The rate observed in this study is consistent with a review of 87 retina clinical trials reporting an overall screen failure rate of 51.8%, with imaging-related criteria being the most common cause.52 Consistent with these findings, approximately 83% of screen failures in the present study were due to failure to meet imaging eligibility criteria (Table 1; Supplementary Figure 1, available at www.ophthalmologyscience.org). Endpoints will also focus on the rate of photoreceptor loss, which could be used to directly assess the impact of interventions in future interventional trials. The EZ represents the photoreceptor inner segments that are packed with mitochondria.53,54 The rate of change in EZ loss was used in studies for macular telangiectasia as a primary outcome measure55, 56, 57 and has been recently accepted by the FDA as a potentially approvable endpoint for a phase III trial in GA (clinicaltrials.gov identifier, NCT06373731). HONU will allow the assessment of loss of the EZ in progression of iAMD. HONU will enable a thorough analysis of progression rates from high-risk iAMD in the study eye at baseline to more advanced atrophic AMD detected on OCT, such as the onset of iRORA, nGA, cRORA or GA. Nascent GA and iRORA represent evidence of photoreceptor loss,11,20,58,59 and they are associated with a higher risk of progression to GA or cRORA,34,60 and most of their features on OCT can be graded reliably.33 Although the onset of atrophic AMD has not yet been accepted by regulatory authorities as an endpoint for approving interventions, its use could significantly enhance the feasibility of early intervention trials aimed at preventing or slowing progression to vision-threatening late AMD. Incorporating such endpoints could also support earlier decision-making before advancing to pivotal trials. The study also supports the development of algorithms for the segmentation of OCT features and retinal layers on individual B-scans. Artificial intelligence-driven tools allow for efficient, time-saving assessment of multiple anatomical changes and offer adaptability to evolving regulatory guidance on endpoints. For example, AI enabled measurements of the rate of change in EZ loss, an anatomic marker recognized by the FDA as an acceptable for assessing photoreceptor loss over time, can be implemented consistently across large datasets. Artificial intelligence-assisted labeling facilitates grading by our reading center, enhancing the study's feasibility. Leveraging these capabilities, we plan to create and/or validate diagnostic and prognostic models across different stages of AMD. Importantly, the collection of imaging and functional data in HONU will allow correlations between atrophic lesions and corresponding functional changes. The study employs several novel modalities, including qCSF, FPF, and individualized targeted microperimetry. The qCSF, a measure of contrast sensitivity, has been shown to decrease with AMD progression61 and may thus be a more sensitive measure of vision loss compared with BCVA.62 Flavoprotein fluorescence is an imaging modality that assesses retinal metabolic stress and understanding its prognostic significance in non-neovascular AMD will give novel insights into disease progression. The study introduces a high-density, targeted microperimetry approach, individualized for each eye, which allows for detailed assessment of visual function in and around small areas of atrophy. This individualized testing enables high-spatial resolution sampling and more precise characterization of functional loss,37, 38, 39 potentially offering a more sensitive means of detecting progressive visual functional decline in early atrophic AMD. Implementing individualized microperimetry required the adoption of a novel workflow, supported by an integrated reading center. Uniquely, the central reading center determines the grid placement based on feature grading and provides each site with custom-designed patterns specific to each patient, a methodology not previously used in a multicenter clinical study setting.40 Collection of longitudinal biological samples, as well as DNA for genetic analysis, will allow HONU to contribute to the growing interest in molecular biomarkers probed through Omics analysis, and to interpret the outcomes in the setting of genetic risk.63,64 The HONU data set will be a major addition to the current longitudinal datasets of patients with eyes that have drusen, followed meticulously over time. It will complement other ongoing studies, such as ALSTAR1 and ALSTAR2,65, 66, 67 MACUSTAR,68 and PINNACLE,69 and to the completed AREDS1 and AREDS2 and LEAD cohorts.11,46 The information gathered will be critical as we move toward earlier intervention in this common sight-threatening disease. In summary, HONU is a novel, global prospective observational study in iAMD, designed to inform the development of feasible early intervention trials and strategies for identifying high-risk patients for recruitment. The study aims to support the identification of earlier AMD endpoints that could enable shorter, more efficient trials, while still detecting meaningful treatment effects. Additionally, HONU has the potential to uncover novel imaging, genetic, ocular, and serological fluid biomarkers that are prognostic for disease progression. By providing a more comprehensive understanding of AMD progression in this population, HONU will deliver valuable insights that can help advance clinical care for individuals affected by this vision-threatening disease. Data Availability For eligible studies, qualified researchers may request access to individual patient level clinical data through a data request platform. At the time of writing, this request platform is Vivli: https://vivli.org/ourmember/roche/. For up to date details on Roche’s Global Policy on the Sharing of Clinical Information and how to request access to related clinical study documents, see here: https://www.roche.com/innovation/process/clinical-trials/data-sharing. Anonymized records for individual patients across more than 1 data source external to Roche cannot, and should not, be linked due to a potential increase in risk of patient reidentification.