Email:
global@GCP-ClinPlus.com
The appropriate selection of endpoints in ophthalmic drug clinical trials directly determines the scientific validity of a study and its likelihood of regulatory approval. Endpoint selection strategies vary significantly across diseases due to differences in pathophysiology and progression patterns.
1. Glaucoma
Intraocular Pressure (IOP) : Current glaucoma medications are approved based on their IOP-lowering effects. However, the degree of visual function protection conferred by different agents may vary even at comparable IOP reductions, making IOP an imperfect surrogate marker [1].
Standard Automated Perimetry (SAP) Parameters: Parameters such as mean deviation (MD) correlate more directly with clinical outcomes, yet SAP may show no abnormalities in early-stage glaucoma [1].
Imaging Biomarkers: Retinal nerve fiber layer (RNFL) thickness and optic disc structural parameters measured by OCT have been shown to correlate with glaucoma diagnosis and progression, and are considered candidate surrogate endpoints [1].
Composite Endpoints: Combining structural and functional metrics into composite endpoints can address the limitations of single measures, which may lack sensitivity across different disease stages [1].
2. Diabetic Retinopathy (DR) and Diabetic Macular Edema (DME)
Best Corrected Visual Acuity (BCVA): BCVA is the most widely used primary endpoint in DME clinical trials [2].
Central Subfield Thickness (CST) and Total Macular Volume (TMV): CST correlates with baseline visual acuity, while TMV better captures the full extent of edema. Combined assessment is recommended [2].
Diabetic Retinopathy Severity Scale (DRSS) Score: Worsening of DRSS score predicts long-term vision loss. A ≥2-step improvement in DRSS has been used as a primary efficacy endpoint in trials such as ZETA-1 [3].
Disorganization of Retinal Inner Layers (DRIL) : A 2024 study confirmed that baseline DRIL area and maximal horizontal extent negatively correlate with BCVA, while post-treatment reduction in DRIL correlates positively with visual improvement [4].
OCTA Parameters: Changes in foveal avascular zone (FAZ) area and vascular density can reflect microvascular disease progression [2].
3. Neovascular Age-Related Macular Degeneration (nAMD)
BCVA: BCVA remains the core endpoint in nAMD clinical trials [2].
Fluorescein Angiography (FA) : FA enables dynamic monitoring of leakage, and the parameters it provides for macular neovascularization (MNV) measurement remain key biomarkers [2].
OCT and OCTA: OCT can assess subretinal fluid, intraretinal fluid, and pigment epithelial detachment. OCTA can delineate neovascular network morphology, though results may vary across devices; consistent use of a single device throughout a trial is recommended [2].
4. Dry Age-Related Macular Degeneration / Geographic Atrophy (GA)
Atrophy Area: Fundus autofluorescence (FAF) has largely replaced color fundus photography as the standard method for GA area assessment [2].
Disconnect Between Anatomic and Functional Endpoints: Recent trials of complement inhibitors have shown that while treatment modestly slows GA expansion, corresponding improvements in visual function metrics have not been observed. Resolving this disconnect remains a critical challenge [5].
cRORA and iRORA: A post hoc analysis of the GATHER1 trial showed that avacincaptad pegol delayed progression from iRORA to cRORA, with differences emerging as early as 6 months into treatment, suggesting that iRORA/cRORA may serve as endpoints for earlier intervention [6].
Other OCT Risk Biomarkers: Intraretinal hyperreflective foci (IHRF), hyporeflective core drusen (hCD), and subretinal drusenoid deposits (SDD) are all associated with a high risk of progression to advanced AMD [2].
5. Inherited Retinal Diseases (IRD)
Limitations of Traditional Endpoints: Schmetterer et al. noted that while visual acuity is a gold standard, additional endpoints are needed for indications such as GA and IRD [1]. The Chinese Medical Association has issued consensus recommendations on efficacy endpoints for IRD trials, providing systematic guidance that accounts for clinical and genetic heterogeneity [7].
Functional Endpoints:
The Freiburg Visual Acuity Test (FrACT) can quantify visual acuity down to hand motion levels, making it suitable for patients with advanced disease [2].
Low-luminance visual acuity (LLVA) and low-luminance deficit (LLD) detect vision loss earlier than high-contrast acuity. The MACUSTAR study confirmed good multicenter repeatability (ICC ≥ 0.7) [8].
Full-field stimulus threshold (FST) testing quantifies visual function in patients with severe vision loss and has been used as a surrogate endpoint in subretinal gene therapy trials [1].
Multi-Luminance Mobility Testing (MLMT) : MLMT quantifies functional vision by assessing navigation of an obstacle course under varying illuminance levels. It has become a global standard in gene therapy clinical trials [1]. In 2025, China issued an expert consensus to promote standardized application of MLMT [2].
Structural Endpoints: Ellipsoid zone (EZ) width is a key anatomical marker in diseases such as retinitis pigmentosa; progressive narrowing of EZ width correlates with disease advancement and has been widely adopted as a surrogate endpoint [2].
Centralized reading, through independent and standardized interpretation, ensures data objectivity. According to the reading center framework recommendations published in *Ophthalmology Science*, standard procedures include the following components [9].
1. Protocol Design and Standardization
The reading center engages at the trial initiation stage to develop disease-specific imaging evaluation criteria and acquisition manuals. The Bern Photographic Reading Center (BPRC), for example, adheres to EVI.CT.SE and ICH-GCP standards, coordinates acquisition workflows with participating sites, and maintains independent evaluation to ensure unbiased image assessment [10].
2. Personnel Training and Certification
All graders must complete structured training including at least 200 supervised image interpretations. Certification requires reading a set of 100 standardized images, with intergrader agreement of κ ≥ 0.80 for primary diagnoses and κ ≥ 0.70 for secondary features [9]. International reading centers such as BPRC provide certification and real-time instruction for both investigators and photographers, ensuring high-quality imaging standards across sites [10].
3. Double-Blind Cross-Review and Adjudication
Independent dual grading with adjudication is the core quality control measure. When intergrader agreement falls below the preset threshold, a senior expert independently arbitrates. In cases of persistent disagreement, a consensus meeting is convened, with the final ruling serving as the analytical “gold standard” [9].
4. Quality Traceability and Data Security
Quality control includes duplicate grading (secondary grading of 1/20 to 1/5 of images) and intra-grader reliability assessment (re-reading of ≥5% of prior evaluations after an interval of ≥4 weeks). Recommended consistency thresholds are κ ≥ 0.80 for primary outcomes and intraclass correlation coefficient ≥ 0.90 for quantitative measures [9]. Reading systems must comply with international data security standards such as HIPAA and support full-process data traceability [10].
5. Data Delivery and Regulatory Support
Final data are exported in CDISC-compliant formats to support regulatory review. During project execution, the centralized reading center monitors image quality in real time, provides technical guidance, and regularly reports evaluation progress [9, 10].
Multicenter imaging data consistency faces three major sources of variability: differences in devices, operators, and interpretation standards. Key mitigation strategies are outlined below.
1. Standardized Imaging Data Model
The Radiology Common Data Model (R-CDM) , developed as an extension of the OMOP-CDM, is specifically designed to standardize medical imaging data such as OCT. A Korean proof-of-concept study successfully standardized 737,500 OCT images from two hospitals into the R-CDM format, enabling efficient cross-center analysis of central macular thickness and RNFL thickness [11].
2. End-to-End Quality Control by Reading Centers
Centralized reading centers ensure multicenter data consistency through unified acquisition parameters, duplicate grading, standardized image quality criteria, and interand intra-grader reliability assessments [9, 10].
3. Imaging Transfer and Management Platforms
Dedicated imaging management platforms enable closed-loop management encompassing image acquisition, de-identified upload, quality checks, and grading assignment. Institutions such as BPRC utilize standardized databases for full-lifecycle imaging management, applying the highest security standards from acquisition to long-term storage [10].
4. Emerging Technology Applications
AIand blockchain-based data management frameworks, along with deep learning algorithms for detecting and quantifying cRORA/iRORA, offer new pathways for standardized multicenter imaging analysis and tamper-proof data integrity [2].
Ophthalmic drug clinical trial endpoints are evolving from a singular focus on BCVA toward a multidimensional framework encompassing functional, structural, and patient-reported outcomes. In glaucoma, the limitations of IOP as a surrogate endpoint are increasingly recognized, and structure-function composite endpoints represent a future direction. Across DR/DME, AMD, and IRD, various imaging biomarkers (DRIL, cRORA/iRORA, EZ width) and functional metrics (MLMT, FST) have been validated as surrogate endpoints. Centralized imaging evaluation requires establishing standardized processes covering protocol design, personnel certification, dual grading, adjudication, and full-process traceability. Multicenter data consistency depends on the synergistic application of the R-CDM standardization model, end-to-end reading center quality control, and dedicated imaging management platforms.
With 23 years of industry expertise, GCP ClinPlus leverages its proprietary ClinX intelligent platform—deeply integrated with AI technologies—to deliver efficient, predictable, end-to-end clinical research solutions for global pharmaceutical companies.
Served 500+ pharmaceutical companies and executed 2,300+ clinical research projects
Contributed to 220+ product approvals, including China’s first stem cell therapy
Executed 200+ international multicenter clinical trials (MRCTs), with 4 products receiving FDA/EMA approval
40+ ophthalmic clinical research projects, building capabilities across the full spectrum of ophthalmic diseases
Disease areas spanning anterior segment conditions (dry eye disease, glaucoma, cataract) and posterior segment conditions (age-related macular degeneration, diabetic retinopathy, retinal vein occlusion)
Extensive real-world study experience supporting evidence generation across the full lifecycle of ophthalmic products
Long-term partnerships with leading ophthalmic companies, including Chengdu Kanghong Pharmaceutical Group Co., Ltd., supporting clinical development of innovative products such as Lifitegrast Eye Drops (Langyueming®)
Driven by its “AI + Globalization” dual strategy, GCP ClinPlus remains committed to advancing intelligent clinical research and serving as a trusted partner in global clinical development.
[1] Schmetterer L, Scholl HPN, Garhöfer G, et al. Endpoints for clinical trials in ophthalmology. *Prog Retin Eye Res*, 2023, 97: 101160. DOI: 10.1016/j.preteyeres.2022.101160.
[2] Chinese Medical Association Ophthalmology Branch Fundus Disease Group, Chinese Medical Doctor Association Ophthalmology Branch Fundus Disease Group. Chinese expert consensus: Standardized application of multi-luminance mobility testing in clinical trials for inherited retinal dystrophies. *Chinese Journal of Ocular Fundus Diseases*, 2025, 41(3): 169-177. DOI: 10.3760/cma.j.cn511434-20241231-00513.
[3] Lally D, et al. Efficacy of oral APX3330 for diabetic retinopathy: results from the phase 2 ZETA-1 trial. Presented at: American Society of Retina Specialists (ASRS) Annual Meeting; July 2023; Seattle, WA.
[4] Velaga SB, et al. OCT outcomes as biomarkers for disease status, visual function, and prognosis in diabetic macular edema. *Can J Ophthalmol*, 2024, 59(2): 109-118. DOI: 10.1016/j.jcjo.2023.01.012.
[5] Lad EM, Fleckenstein M, Holz FG, et al. Informing endpoints for clinical trials of geographic atrophy. *Annu Rev Vis Sci*, 2024, 10: 455-476. DOI: 10.1146/annurev-vision-101623-101232.
[6] Patel SS, Lally DR, et al. Avacincaptad pegol for geographic atrophy secondary to age-related macular degeneration: 18-month findings from the GATHER1 trial. *Eye (Lond)*, 2023, 37(17): 3558-3565. DOI: 10.1038/s41433-023-02447-6.
[7] Chinese Medical Association Ophthalmology Branch Fundus Disease Group, Chinese Medical Doctor Association Ophthalmology Branch Fundus Disease Committee. Chinese expert consensus on visual function assessment and endpoint recommendations for clinical trials in inherited retinal diseases. *Chinese Journal of Ocular Fundus Diseases*, 2022, 38(8): 626-635. DOI: 10.3760/cma.j.cn511434-20220808-00443.
[8] Dunbar HMP, Behning C, Lüning A, et al. Repeatability and Discriminatory Power of Chart-Based Visual Function Tests in Individuals With Age-Related Macular Degeneration: A MACUSTAR Study Report. *JAMA Ophthalmol*, 2022, 140(8): 780-787. DOI: 10.1001/jamaophthalmol.2022.2113.
[9] Suggested Framework for Reading Centers Evaluating Fibrosis. *Ophthalmology Science*, 2025. DOI: 10.1016/j.xops.2025.100670.
[10] Bern Photographic Reading Center (BPRC). About the Bern Photographic Reading Center. Available at: https://augenheilkunde.insel.ch/en/teaching-and-research/bern-photographic-reading-center-bprc/about.
[11] Park CH, Park SJ, Lee DY, et al. Multi-Institutional Collaborative Research Using Ophthalmic Medical Image Data Standardized by Radiology Common Data Model (R-CDM). *Stud Health Technol Inform*, 2024, 310: 48-52. DOI: 10.3233/SHTI230925.
