Classification and growth rate of retinal atrophy after voretigene neparvovec-rzyl for RPE65-mediated retinal dystrophy

Abstract

Purpose : Classify the appearance of retinal atrophy and quantify the growth rate in patients who received voretigene neparvovec-rzyl (VN) for RPE65-mediated retinal dystrophy

Methods : Atrophy was qualitatively classified into different subtypes by expert consensus then manually segmented in ImageJ by two graders. Area of atrophy was calculated and fit to linear models of atrophy and square root of atrophy using Python. Main outcome measures were number of eyes, and slopes and coefficients of determination of linear regression models for each subtype of atrophy. Values are reported as mean ± standard deviation.

Results : 20 eyes from 10 patients across 4 centers developed progressive chorioretinal atrophy after subretinal administration of VN. A mean of 4.9 ± 3.0 images per eye obtained over 1.7 ± 0.7 years were reviewed, and atrophy was categorized into touchdown site (8 eyes), perifoveal (13 eyes), and nummular (11 eyes) subtypes (Fig. 1). 10 eyes demonstrated more than one type of atrophy (mixed). All eyes demonstrated atrophy involving the macula.

The growth rate of all atrophy by area and by square root of area was 18.0 ± 11.0 mm²/year (R² 0.9 ± 0.1) and 2.6 ± 1.2 mm/year (R² 0.8 ± 0.2). Growth rates of the atrophy subtypes were 2.0 ± 1.7 mm²/year (R² 0.9 ± 0.2) and 0.8 ± 0.4 mm/year (R² 0.9 ± 0.2) for touchdown atrophy, 17.2 ± 12.5 mm²/year (R² 0.9 ± 0.1) and 2.7 ± 1.4 mm/year (R² 0.8 ± 0.2) for perifoveal atrophy, and 3.8 ± 2.7 mm²/year (R² 0.9 ± 0.0) and 1.5 ± 0.4 mm/year (R² 0.9 ± 0.1) for nummular atrophy (Fig. 2)

Conclusions : Chorioretinal atrophy following subretinal administration of VN for RPE65-mediated retinal dystrophy developed according to a touchdown site, perifoveal, and/or nummular pattern. Perifoveal atrophy trended toward growing the most rapidly whereas touchdown site atrophy trended toward growing the least rapidly. Atrophy growth was more strongly correlated with atrophy area rather than square root of area, in contrast to other causes of macular atrophy such as dry AMD. These findings suggest distinct mechanisms of atrophy development and demonstrate the need for close observation in studies involving other gene therapies using similar vectors.

This abstract was presented at the 2022 ARVO Annual Meeting, held in Denver, CO, May 1-4, 2022, and virtually.

 

Touchdown site (Panel A), perifoveal (Panel B), nummular (Panel C), and mixed (Panel D) atrophy.

 

All atrophy (Panel A), touchdown site (Panel B), perifoveal (Panel C), and nummular (Panel D) atrophy over time.