Related Experiment Video
Updated: Aug 31, 2026
Ex Vivo OCT-Based Multimodal Imaging of Human Donor Eyes for Research into Age-Related Macular Degeneration
Published on: May 26, 2023
Structure-Function Correlations Using Artificial Intelligence-Extracted Retinal Layer Attenuation Measurements in OCT
Kenta Yoshida1, Miao Zhang2, Adam Pely2
1Clinical Pharmacology, Genentech, Inc., South San Francisco, California.
Purpose:
To investigate the structure-function relationships between retinal layer attenuation on OCT and a range of functional assessments in patients with geographic atrophy (GA).
Design:
Retrospective analysis of 2 phase III clinical trials.
Subjects:
One thousand five hundred and nineteen subjects with GA who were enrolled in phase III studies investigating Lampalizumab: Chroma (NCT02247479) and Spectri (NCT02247531).
Methods:
Structural retinal features were assessed by artificial intelligence (AI)-based automated segmentation algorithms on OCT images to calculate metrics for outer retinal layer attenuations: loss for external limiting membrane (ELM), ellipsoid zone (EZ), and retinal pigment epithelium (RPE). Functional measures included best-corrected visual acuity (BCVA), low-luminance visual acuity (LLVA), and reading speed collected in all patients. Microperimetry was collected in a subset of patients.
Main Outcome Measures:
Correlations (Pearson r) between the OCT layer loss metrics with visual function measures.
Results:
Cross-sectional analysis did not show any notable correlations between loss areas of each layer for the full OCT scan range and BCVA but did show correlations with LLVA (EZ loss area r = -0.39). After spatially defining regions using the ETDRS circles, notable correlations were observed between BCVA and ELM loss area within the central 1 mm diameter circle area (r = -0.46) and for LLVA with the central 3 mm diameter circle even after exclusion of the central 1 mm diameter region. Microperimetry outcomes correlated with OCT layer losses; the number of scotomatous points correlated equally well with total macular ELM and EZ losses (r = 0.67), mean macular sensitivity had the highest correlations with EZ loss areas, and perilesional sensitivity or responding sensitivity showed the highest correlation with the difference in EZ and RPE loss area (r = -0.50, r = -0.48).
Conclusions:
The current analyses using AI-based OCT layer losses provide new insights into the spatial dependency of the relationship between anatomical and functional changes in GA. The differential associations noted between BCVA and LLVA with central and paracentral macular pathology, respectively, suggest alternative pathways for diverse aspects of visual function and help inform the design of anatomical and functional assessments for clinical trials in GA.
Financial Disclosures:
Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.
