Could AI Improve Available Treatments for Geographic Atrophy?

Why It Matters

AI‑enhanced imaging provides objective metrics that can accelerate diagnosis, personalize therapy, and meet regulatory demands, reshaping the GA treatment landscape in both Europe and the United States.

Key Takeaways

• AI identifies ellipsoid zone loss with micron precision
• EZ‑RPE ratio predicts rapid geographic atrophy progression
• RetinAI can screen OCT databases for trial eligibility
• EU has GA imaging tools; US lacks AI regulatory approval
• AI‑driven functional OCT may replace time‑consuming microperimetry

Pulse Analysis

The integration of artificial intelligence into optical coherence tomography (OCT) is redefining how clinicians monitor geographic atrophy. By automating the detection of subtle ellipsoid zone (EZ) changes, AI platforms achieve a measurement accuracy of 1.5 µm, far surpassing manual interpretation. This precision enables the calculation of novel biomarkers such as the EZ‑RPE loss ratio and EZ thickness, which have emerged as strong predictors of disease velocity and visual function. Consequently, ophthalmologists can differentiate fast‑progressing lesions from indolent ones, tailoring follow‑up intervals and therapeutic decisions with unprecedented confidence.

Regulatory environments on opposite sides of the Atlantic shape the deployment of these tools. Europe already permits AI‑enabled imaging devices like the Topcon GA Monitor, allowing clinicians to leverage real‑time analytics in routine care. In contrast, the United States must navigate the FDA’s software‑as‑a‑medical‑device pathway, limiting AI use to research settings until robust outcome data demonstrate clinical benefit. Nonetheless, precedents such as IDx‑DR illustrate viable routes for autonomous AI approval, suggesting that GA‑focused algorithms could follow a similar trajectory once they prove to improve visual and anatomical endpoints.

From a business perspective, AI promises operational efficiencies and new revenue streams. Automated OCT screening can dramatically shorten patient selection timelines for clinical trials, reducing costly manual chart reviews. Moreover, visualizing disease trajectories for patients enhances adherence to complement inhibitor therapies, potentially extending market uptake. As payers increasingly demand evidence of value, AI‑derived metrics that correlate structural preservation with functional outcomes may become essential for reimbursement negotiations, positioning AI as a strategic asset in the evolving GA treatment ecosystem.