Correlation of Morphology and Function of Flecks Using Short-Wave Fundus Autofluorescence and Microperimetry in Patients With Stargardt Disease

Abstract

Purpose: The purpose of this study was to evaluate the functional relevance of longitudinal changes in hyperautofluorescent areas and flecks in Stargardt disease (STGD1) using short-wavelength autofluorescence (SW-AF) imaging.

Methods: In this prospective, longitudinal study, 31 patients with STGD1 (56 eyes) underwent microperimetry (MP) and SW-AF imaging twice in 3 to 5 years. A total of 760 MP test points were included in the statistical analysis based on stable fixation and accurate alignment of SW-AF and MP. Autofluorescence intensity was qualitatively assessed in all MP test points. Small circumscriptive hyperautofluorescent lesions were defined as flecks. Longitudinal imaging characteristics observed on SW-AF were classified into the following categories: appearing, disappearing, and stable flecks, stable hyperautofluorescent, and stable background autofluorescence. The relationship between SW-AF intensity changes and MP changes was analyzed using a linear mixed model corrected for baseline sensitivity.

Results: Retinal sensitivity declined most in locations without change in SW-AF intensity. Functional decline per year was significantly larger in flecks that disappeared (-0.72 ± 1.30 dB) compared to flecks that appeared (-0.34 ± 0.65 dB), if baseline sensitivity was high (≥10 dB; P < 0.01). The correlation between the change observed on SW-AF and the sensitivity change significantly depended on the sensitivity at baseline (P = 0.000).

Conclusions: Qualitative longitudinal assessment of SW-AF poorly reflected the retinal sensitivity loss observed over the course of 3 to 5 years.

Translational relevance: When aiming to assess treatment effect on lesion level, a multimodal end point including MP focused on hyperautofluorescent lesions appears essential but needs further studies on optimizing MP grids, eye-tracking systems, and alignment software.

Figure 1.

Examples of quality assessment of the data alignment between both visits, after coregistration of MP with SW-AF images. The data of the first visit (images A1, B1, C1, and D1) was compared to data of the second visit (images A2, B2, C2, and D2).

Figure 2.

At the MP test points SW-AF imaging intensity was graded as background autofluorescence, hyperautofluorescent area, hyperautofluorescent circumscriptive lesion (fleck), or hypo-autofluorescent.

Figure 3.

The linear mixed model shows that the retinal sensitivity change over time not only depends on the longitudinal SW-AF imaging intensity over time but also on the baseline sensitivity as is depicted in (A). Although no SW-AF imaging change over time was observed in lesions of category “stable hyperautofluorescent,” the retinal sensitivity in these loci decreased. (B) Displays the model-estimated marginal means and 95% confidence interval per defined category. Significant differences in functional decline were found between several categories in regions with high initial sensitivity. *P < 0.05.