Abstract
Purpose :
Clinical trials involving novel therapies in choroideremia are complicated by the choice of outcome measure. Best corrected visual acuity (BCVA) is mostly insensitive to changes in disease state, until late stages, and hence also to potential therapeutic gains after gene therapies at earlier stages. Ranked importance of other potential functional outcome measures remains unknown.
Methods :
Retrospective analysis of twelve patients with choroideremia (mean age (SD) at baseline=37 (7.9) years) over five years was conducted. Collected data included: BCVA using Early Treatment Diabetic Screening chart (ETDRS) (mean BCVA (SD) at baseline=83 (2.7) letters), low luminance visual acuity (LLVA), Pelli-Robson contrast sensitivity, Full-threshold stimulus testing (FST), electroretinography (ERG), Cambridge contrast sensitivity function (CSF), Cambridge color test, blue autofluorescence (BAF). BAF was used to set the dependent variable and a surrogate measure of disease severity. All other variables were set as predictor variables. Analysis of variable importance was performed using dominance analysis (Shapley regression) with a combination of r squared fit and Akaike information criteria (AIC) for determining variable importance within a linear mixed effect model framework to correctly model the longitudinal repeated measures.
Results :
The strongest association between disease severity and all possible covariates was from microperimetry explaining 63% of the variance and was strongly ranked first as the predictor of most importance. All other predictors were significantly lower in importance with time from baseline (months), explaining only approximately 5% of the variance. CSF was ranked closely behind with spatial frequency gratings of 4 and 10 cycles/degree, in particular, warranting further investigation. BCVA and LLVA were low on the rankings (7th and 10th, respectively).
Conclusions :
This study supports the use of microperimetry in studies of choroideremia and explicitly demonstrates the insensitivity of BCVA with a low importance ranking in this cohort of mid-stage patients. Time since baseline visit was a surprisingly poor predictor likely due to a greater importance of time since disease onset. Contrast sensitivity may be a promising future outcome measure demonstrating a far higher sensitivity to early disease changes than BCVA or LLVA.
This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.