Using Artificial Intelligence to Investigate the Real-world effect of Anti-complement Medications on Retinal Photoreceptor Survival in Patients with Geographic Atrophy- Research project funded by the Macular Degeneration Association
Principal Investigator: Jeremiah Brown, Jr., MS, MD
Address: 17017 Interstate 35 North, Schertz, TX. 78154
IRB Project Title: Using Artificial Intelligence to Investigate the Real-world effect of Anti-complement Medications on Retinal Photoreceptor Survival in Patients with Geographic Atrophy
Protocol Version Date: 23 January 2025
- Introduction-Background, Rationale
AMD is a leading cause of blindness in the United States. Geographic atrophy is the advanced form of dry AMD. Patients with geographic atrophy develop sharply demarcated areas of retina thinning, characterized by loss of the photoreceptors, retinal pigmented epithelium (RPE) and choriocapillaris. Geographic atrophy affects over 5 million Americans and its incidence increases with increasing age. Photoreceptor degeneration results in diminished vision in the areas of atrophy. Despite extensive atrophy in the macula, patients may retain foveal function until very late in the disease. Thus, visual acuity is a poor measure for disease stage or progression.
A key mechanism underlying photoreceptor degeneration is dysregulation of the complement system, a part of the immune response that can contribute to tissue damage through inflammation and cellular destruction. Anti-complement therapies have emerged as promising strategies to slow the progression of retinal degeneration. Early clinical trials have shown that inhibiting select complement proteins can reduce inflammation and protect retinal cells in human subjects. However, the precise mechanisms by which complement inhibition preserves photoreceptor survival and the potential for optimizing these therapies remain areas of active investigation.
It is notable that the rate of progression of geographic atrophy varies amongst patients, amongst lesion types and amongst lesion patterns. Some patients are slow progressors and others progress more quickly. Patients with multifocal lesions progress faster than patients with single lesions. Patients with subfoveal lesions progress more slowly than patients with lesions outside the fovea. Large lesions tend to progress more quickly than slow lesions. The current anticomplement therapies appear to slow progression in all patients, but to varying degrees depending upon the lesion types. Patients with a larger rim of photoreceptor degeneration determined by OCT analysis of the ellipsoid zone compared to the size of the RPE defect tend to progress faster than patients where the photoreceptor degeneration size more closely matches the size of the RPE cell defect.1
In clinical practice, it has become interesting to note that some patients report visual improvement once initiating therapy with the current anticomplement medications. Some of these patients may demonstrate improved visual acuity, others report that the “quality” of their vision is better or that their contrast sensitivity has improved after starting therapy. Our goal is to better understand what is occurring at the level of the photoreceptors and RPE in patients treated with anticomplement therapies and determine if we can predict which patients are more likely to experience improvement in visual acuity.
Advances in artificial intelligence (AI) and machine learning (ML) offer new opportunities to deepen our understanding of these complex biological processes. In the context of anti-complement drugs, AI can be particularly useful for modeling the dynamic interactions between complement activation and photoreceptor survival, helping to identify patterns involved in disease progression and drug efficacy.
2.0 Objectives
This research protocol aims to utilize AI-driven approaches to investigate the impact of anti-complement therapies pegcetacoplan and avacincaptad pegol on photoreceptor (ellipsoid zone) and RPE degeneration through segmentation of the outer retinal layer images on OCT b scans. The findings from this research could not only enhance our understanding of the biological mechanisms driving retinal degeneration but also aid in the development research protocols taking into consideration baseline patient characteristics. This will aid in the development of personalized anti-complement therapies.
3.0 This study will be conducted at all participating Retina Consultants of America locations
4.0 Selection of patients
4.1 Inclusion Criteria
4.1.1. Patients with a diagnosis of Atrophic AMD involving the foveal center or Atrophic AMD not involving the foveal center.
4.1.2. Patients must have received a minimum of with pegcetacoplan or avacincaptad pegol.
4.1.3 Patients must have OCT b scan images performed on the Heidelberg Spectralis at +/- 1 month at each time point). Each OCT scan must have at least 45 scan lines.
4.1.4 Study eye must have Snellen visual acuity of 20/400 or better. Both eyes can be evaluated if both eyes qualify.
4.2. Exclusion criteria
4.2.1. Eyes being actively treated for neovascular AMD at baseline are excluded.
4.3. Age range
4.3.1. Age 50 years and over.
Results of findings to be updated as we receive the information