Geographic Atrophy Risk-Associated Biomarkers Identified in Intermediate AMD

Credit: Scott Camazine/Medical Images.
OCT findings may predict disease development, a study shows.

Researchers have identified multiple novel optical coherence tomography (OCT) biomarkers associated with geographic atrophy (GA) development, including reduced ellipsoid zone (EZ) integrity and compartmental textural and shape-based features. The biomarkers might enable early detection of eyes with intermediate dry age-related macular degeneration (AMD) at greatest risk for GA development. The findings were presented at the American Society of Retina Specialists (ASRS) 40th Annual Meeting, held in New York July 13-16, 2022.

The researchers conducted a retrospective 5-year longitudinal study of patients with intermediate dry AMD without baseline GA. They aimed to identify novel spectral domain (SD)-OCT features using compartmental outer retinal segmentation and radiomics, such as shape and texture, that are predictive of GA development from intermediate dry AMD.

All patients had 5 years of SD-OCT imaging, allowing the eyes to be categorized as GA developers or nondevelopers based on development of GA within 5 years. The team evaluated machine-learning classification models for predicting GA development using data from the SD-OCT, including quantitative EZ-integrity and retinal pigment epithelium (RPE)-compartment assessment and radiomics-based feature extraction. 

A total of 114 eyes of 114 patients with intermediate dry AMD without baseline GA were included in the study. Of those, 29% developed GA anywhere within the macula within 5 years. These eyes had significantly decreased EZ integrity compared with eyes that did not develop GA, indicated by mean EZ-RPE central subfield thickness (28.5 vs 36.2 μm; P <.001) and EZ-RPE volume (1.23 vs 1.30 mm³; P =.002). Eyes that developed GA also had significantly increased partial (5.4% vs 0.9%; P <.001) and total EZ attenuation (2.6 vs 0.3%; P <.001) compared with eyes that did not develop GA.

Using Random Forest modeling with EZ-RPE identified radiomic features, the researchers observed a classifier performance for area under the curve (AUC) of 0.88±0.02 for GA development. When modeling with subRPE features, they observed an AUC of 0.92±0.02. The team also found that increased heterogeneity in the EZ-RPE compartment was associated with the development of GA.

“With further validation, the fractal and texture features could potentially be used for clinical trial enrichment and assessments for future therapeutic response in dry AMD patients,” researchers report.

Disclosure: Some study authors declared affiliations with biotech, pharmaceutical, and/or device companies. Please see the original reference for a full list of authors’ disclosures.  


Abraham J, Sil Kar S, Cetin H, et al. Geographic atrophy risk-associated OCT biomarkers identified with quantitative outer retinal integrity and compartmental radiomics in intermediate AMD. Poster presented at: The 40th annual meeting of the American Society of Retina Specialists; June 13-16, 2022; New York, NY.  Poster 117.