Project Title:

FRB! 2.0: Linking imaging to clinical registry outcomes to improve understanding of progression of age-related macular degeneration

Chief Investigator:

Dr Hemal Mehta

Co-Investigators:

Prof Adam Dunn, Vuong Nguyen

Aim

To compare how two different approaches to measure macula atrophy (MA) progression, total MA lesion size and distance of the foveal centre point (FCP) to the nearest MA lesion edge, predict future visual acuity (VA) decline.

Methods

Patients provided opt-in consent. A computer scientist assisted in the process of tagging retinal images with an FRB! identifier to allow linkage with the clinical data. Heidelberg Engineering optical-coherence tomography (OCT) images were assessed by two certified graders in a reading centre setting. MA lesion area was measured semi-automatically from fundus autofluorescence images using Region Finder software. The distance from the edge of the nearest MA lesion to the FCP was manually measured using in-built calipers. Each eye was assessed at two visits, at baseline and three years later, with VA data from the Fight Retinal Blindness! registry.

Key results

There were 19 eyes of the initial 70 that met inclusion criteria for this pilot study. Linear regression analysis identified a statistically significant relationship between distance change and VA loss (coefficient (95% CI) = -4.31 (-6.28, -2.35), R2 = 0.56, p < 0.001), however there was no statistically significant relationship found between area change and VA loss (coefficient (95% CI) = -3.50 (-12.8, 5.84), R2 = 0.04, p > 0.1). Similar findings were made with Pearson correlation calculations, with log transformed change in distance having a negative correlation with VA (Pearson r = -0.746, p < 0.001), and no correlation with area change using square-root transformation (r = -0.189, p > 0.1).

 

Implications for Clinical Practice/Science and Future Research

Change in area of geographic atrophy was used as the primary endpoint for phase 3 clinical trials of 2 intravitreal complement factor inhibitors that have been approved by the FDA in the USA. The European Medicines Agency has declined an application for pegcetacoplan in Europe because not enough evidence of functional benefit has been provided. It could be that the primary endpoint of clinical trials for therapies that slow GA needs to be modified.

We are currently increasing the sample size. The plan is to submit this data for publication once the sample size reaches 50 eyes. Once data is published, we will be looking towards applying for a NHMRC partnership grant to scale the project.

Conclusion

While there was statistically significant moderate negative predictive value in changes in distance in relation to VA loss, there was no correlation between any measure of area and VA over three years. Distance to the FCP may be a more useful clinical parameter than MA area in predicting VA loss. Larger scale studies are required to corroborate these findings.

 

Lay Summary of Outcomes

There are treatments in development to slow down progression of the most common form of late-stage age-related macular degeneration, related to wear- and-tear of the central retina. This study suggests that the way we measure disease progression could better relate to patient function if we measured the distance of the nearest lesion to the centre of vision rather than the total area of lesions. This has implications for how we design clinical trials as well as select patients most likely to benefit from emerging treatments.

 

This research project was funded by NSW RANZCO in 2023