Project 8 - Novel AI-based biomarkers for retinal gene therapy outcomes by integrating functional neuroimaging and retinal imaging

The development of gene therapies is more than a glimmer of hope for many people with rare inherited retinal diseases. Unfortunately, phase III clinical trials have failed recently because the endpoints of measuring success in terms of vision were not met. This is partly because of the small numbers of cases available, but mainly because each inherited retinal disease affects vision differently and common clinical endpoints such as visual acuity are far too one-dimensional to assess the impact of therapy on vision.

However, vision is not only a performance of the retina; The visual impression rather arises as a consequence of information processing of primary and higher cortical areas in the brain. So far, this has received little attention in ophthalmological studies. We have therefore set ourselves the goal of combining the expertise of three principal investigators from the fields of ophthalmology, neuroscience, and artificial intelligence (AI), to collaboratively develop novel biomarkers for visual performance.

Here, ophthalmological examination with all modalities important according to the current state of practice is combined with neuroimaging including visually-induced functional and task-dependent activity assessments of the affected brain areas. Using image analysis methods based on AI, features relevant to vision are extracted from retinal examinations and fused with signatures and trajectories of cortical representations of vision to form multi-layered models of visual performance. These models are generated using data from a cross-sectional study with a single examination of a deliberately heterogeneous group of patients in terms of diseases and visual deficits. Using data from a simultaneously followed longitudinal group with repeated examinations, these models can then be sharpened and internally cross-validated. The longitudinal group will also include patients with the only approved gene therapy to date (Voretigene Neparvovec-rzyl), to prospectively assess the recording of improvements in visual performance by the models. 

At the end of the funding period, integrative biomarkers of vision and thus a novel way of assessing the therapeutic success of retinal gene therapy will be available. As an outlook, these methods will be further developed into predictive tools for individually predicting therapy success with regard to visual performance in a possible second funding period.