TLDR:
- An AI screening tool trained to measure geographic atrophy agreed with retina specialists in a majority of cases.
- Agreement between the AI tool and specialists was 73%, highlighting the potential to expedite clinical trial enrollment and improve consistency in disease measurement.
Main Article:
An AI screening tool has shown promising results in screening participants for clinical trial enrollment by effectively detecting geographic atrophy (GA). The agreement between the AI tool and retina specialists was 73%, indicating potential benefits in expediting enrollment and increasing consistency in disease measurement. This tool could fill a critical need for accurate GA measurement, especially in the absence of current clinical tools for this purpose.
The study involved the training of the AI model using FAF images from the Age-Related Eye Disease Studies 2 (AREDS2). This training enabled the AI to effectively detect GA, showcasing its potential in real-world applications such as clinical trials. The AI model demonstrated consistent performance across diverse datasets, including clinical, clinical trial, and prospective and retrospective datasets, underscoring its generalizability and robustness.
The findings of this study were presented at the Association for Research in Vision and Ophthalmology (ARVO) 2024 Meeting and highlighted the importance of validation and implementation studies to bridge the gap between AI model creation and real-world application. By enhancing the accuracy of patient selection in clinical trials, AI tools like this could play a crucial role in improving screening processes and reducing screen failure rates.
Sources: