By Clinical Research News Staff 

July 14, 2026 | An experimental artificial intelligence (AI) model called OCTCube-M could help improve patient selection, predict disease progression, and streamline clinical trials by extracting new insights from three-dimensional optical coherence tomography (OCT) eye scans. In a recent study published in Nature Biomedical Engineering (DOI: 10.1038/s41551-026-01662-2), the model outperformed older 2D AI approaches in detecting multiple retinal diseases, including age-related macular degeneration and diabetic retinopathy. It also improved predictions of disease progression in geographic atrophy, an advanced form of dry macular degeneration. 

Developed using more than 1.62 million retinal images from 26,000 OCT scans, OCTCube-M is designed to extract information from the full 3D structure of the eye rather than relying on individual image slices. This approach provides an advantage for identifying patients, measuring disease progression, and improving clinical study design. 

One of the most promising applications is clinical trial optimization. AI-based analysis could help identify eligible participants more efficiently, create more uniform study populations, and reduce variability that can affect trial outcomes. Current enrollment strategies often depend on clinical assessments and restrictive inclusion criteria, which can slow recruitment and limit access to eligible patients. 

Aaron Lee, M.D., head of the Hardesty Department of Ophthalmology and Visual Sciences at Washington University School of Medicine in St. Louis, said the technology could also help reduce the sample size and time it typically takes to run a study. In collaboration with Genentech, researchers evaluated OCTCube-M using randomized controlled trial data and found that the model could be used to significantly reduce the number of patients needed for trials while achieving equal results. 

Researchers are also exploring whether AI models like OCTCube-M could support the development of digital twins—virtual patient models that predict how an individual would likely progress without receiving an experimental therapy. These models could eventually provide additional evidence in clinical trials by comparing a patient’s actual outcome with a personalized prediction of disease progression under standard care. 

The FDA has shown increasing interest in digital twin approaches, although their use in clinical trials will likely require additional validation. Lee suggested that early applications may be most appropriate in phase 2 studies, where sponsors make decisions about whether to advance therapies into larger and more expensive trials. 

Beyond ophthalmology, the technology supports the emerging field of “oculomics,” which uses eye imaging to identify signs of systemic disease. Because retinal blood vessels reflect changes occurring throughout the body, AI analysis of OCT scans could potentially help predict conditions such as cardiovascular disease, diabetes, and kidney dysfunction. 

To read the full article written by Deborah Borftiz, visit Diagnostics World News.