SOIN - MI Data Lab: Personalized Ophthalmology Through Collaborative Data Collection and Dynamic Patient Consent

Area of Science:

• Ophthalmology
• Medical Informatics
• Machine Learning

Background:

• The Swiss Ophthalmic Image Network (SOIN) is part of the Swiss Personalized Health Network (SPHN).
• Chronic ocular diseases require personalized care, which can be enhanced by Machine Learning (ML) and medical imaging.
• Existing healthcare systems face burdens that can be alleviated through improved disease management.

Purpose of the Study:

• To establish a collaborative, privacy-preserving research environment (MI Data Lab) for data-driven research in ophthalmology.
• To develop novel ML algorithms for the early detection and prediction of chronic ocular diseases using medical imaging.
• To facilitate cooperation between clinicians and data scientists for advancing personalized eye care.

Main Methods:

• Utilizing the MI Data Lab for consolidating and curating datasets from research partners.
• Developing and applying ML algorithms for automated detection of ocular biomarkers.
• Analyzing over 100,000 retinal images to train and validate algorithms.

Main Results:

• Creation of several ML algorithms for automatic ocular biomarker detection.
• Successful application of these tools to a large dataset of over 100,000 retinal images.
• Demonstrated capability of MI Data Lab in enabling the development of predictive models.

Conclusions:

• MI Data Lab provides a robust platform for privacy-preserving, collaborative research in ophthalmic imaging.
• ML-driven analysis of retinal images holds significant potential for personalized care of chronic ocular diseases.
• The developed tools can extract novel traits for exploring -omic associations, treatment outcomes, and disease progression priors.