Heidelberg Publishes Eye2Gene Study Evaluating Potential of AI-Powered Analysis
Heidelberg Engineering has publised the Eye2Gene study in Nature Machine Intelligence. Titled “Next-generation phenotyping of inherited retinal diseases from multimodal imaging with Eye2Gene,” the study highlights the potential of AI-powered analysis of multimodal Spectralis imaging to accelerate genetic diagnosis in patients with inherited retinal diseases (IRDs).
According to Heidelberg, Eye2Gene leverages fundus autofluorescence (FAF), infrared reflectance (IR), and spectral-domain OCT (SD-OCT) to predict the likely causative gene in IRD cases—providing a non-nvasive decision-support tool for clinicians.
The AI system was trained on 58,030 multimodal retinal scans from 2,451 patients with confirmed genetic diagnoses and further externally validated on 775 patients from five sites. Covering 63 disease-associated genes, Eye2Gene captured more than 90% of IRD cases in Europe, demonstrating a broad clinical relevance.
“I am very excited to announce the publication of our long anticipated Eye2Gene paper,” said Associate Professor Nikolas Pontikos (UCL), lead author of the study. “We demonstrate a top 5 prediction accuracy of 83% compared to world-leading experts.”
In interpreting only FAF images, the modal reached an accuracy of 76%, compared to 36% or less by experienced clinicians who took part in the study, according to Heidelberg. These results were consistently reproduced across five independent clinical centers—including institutions in Tokyo, Bonn, São Paulo, Oxford, and Liverpool. In more than 75% of tested cases, it outperformed popular phenotyping-only tools in prioritizing disease-causing genetic variants, thereby increasing the likelihood of achieving a definitive diagnosis.
Eye2Gene is powered by an ensemble of 15 convolutional neural networks—five per imaging modality—which together generate patient-level predictions by averaging across scans and modalities. This architecture is designed to improve accuracy and ensure that the system can adapt to variations in imaging conditions across different sites.
“By combining the power of Heidelberg image quality and AI, we empower eye care professionals with new insights into the genetic landscape of IRDs—enhancing diagnosis and ultimately the development of new treatments,” said Arianna Schoess Vargas, Managing Director of Heidelberg Engineering.
Heidelberg Engineering demonstrated Eye2Gene live at ARVO 2024 and 2025, where participants experienced its integration within the HEYEX 2 platform via Heidelberg AppWay. This setup enables real-time gene prediction directly from multimodal Spectralis scans—bringing AI-assisted diagnosis one step closer to clinical routine.