Geographic atrophy topography and quality of life: What the central 2.5-mm region reveals

The topographic distribution of geographic atrophy (GA) within the central 2.5-mm diameter region of the macula is more strongly associated with vision-related quality of life (VR-QoL) than foveal involvement alone, according to a new analysis published in Ophthalmology.¹ The findings, drawn from the phase 3 Chroma (NCT02247479) and Spectri (NCT02247531) trials of lampalizumab, have implications for how GA extent is evaluated in both clinical trials and clinical practice.

Study design and participants

Anegondi and colleagues analyzed baseline data from 856 participants with bilateral GA secondary to age-related macular degeneration (AMD) who completed the National Eye Institute Visual Function Questionnaire 25 (NEI VFQ-25) across 23 countries. Participants had a median age of 78 years (interquartile range, 73 to 84 years); 62% were female. GA lesions were automatically segmented using a multimodal deep learning model applied to fundus autofluorescence (FAF) and near-infrared reflectance (NIR) images. The extent of GA within central regions of varying diameters (from 0.25- to 6.00-mm, in 0.25-mm intervals) relative to the fovea was then derived for each eye.

Rather than using raw NEI VFQ-25 scores, the investigators derived VR-QoL estimates using calibrated item measures and rating category thresholds from Rasch analysis. This approach addresses a limitation of raw scores, which treat differences between ordinal response categories as equivalent and do not account for item difficulty. The resulting person measures were reported on an invariant logit scale, with positive values indicating higher levels of function. Separate person measures were derived for the visual function (NEI VFQ-VF) and socioemotional (NEI VFQ-SE) domains, as well as for all 30 items combined (NEI VFQ-25C).

The minimum eye-level GA extent within an individual for each central region diameter—defined as the lower value of the 2 eyes for the region being evaluated—was the primary structural parameter of interest, referred to throughout as the “minimum GA extent.”

Results

Minimum GA extent within central regions across varying diameters from 0.25- to 6.00-mm were all significantly associated with NEI VFQ-25C (P < .001 for all), NEI VFQ-VF (P < .001 for all) and NEI VFQ-SE (P ≤ .003 for all) person measures. The highest proportion of variance explained (R²) for all 3 person measures was observed when evaluating the central 2.50-mm diameter region, with R² values of 0.09, 0.11 and 0.04 for the NEI VFQ-25C, NEI VFQ-VF and NEI VFQ-SE, respectively.

In multivariable analysis, only minimum GA extent within the central 2.50-mm diameter region was independently associated with worsening NEI VFQ-VF and NEI VFQ-SE person measures (P < .001 for both). GA extent within the 2.50- to 6.00-mm annulus was not independently associated with either measure (P ≥ .205).

The investigators also found that models considering the number of eyes with subfoveal GA explained roughly half the variance in VR-QoL compared with models based on minimum GA extent within the central 2.50-mm region (R² = 0.06 for NEI VFQ-VF and R² = 0.03 for NEI VFQ-SE vs R² = 0.11 and 0.04, respectively). Maximum and minimum person-level best-corrected visual acuity (BCVA) explained a larger proportion of variance in VR-QoL than structural parameters alone; however, mediation analyses showed that only 36% to 39% of the association between minimum GA extent in the central 2.50-mm region and VR-QoL was mediated by BCVA, supporting the value of disease-specific structural evaluation.

Clinical and trial implications

The authors identified several implications of these findings. First, VR-QoL impairment associated with GA extends beyond subfoveal involvement or the amount of GA within the central 1-mm diameter region, underscoring the importance of preventing disease progression to and within the central 2.50-mm region. Second, while VR-QoL was largely explained by GA extent in the eye with the lower GA extent (the “better” eye), the authors caution that this does not mean slowing GA growth in the “worse” eye lacks clinical value, given that GA progression rates between eyes can be asymmetric and new areas of GA can emerge over time.

Third, the authors note that for eyes where GA already occupies the entire central 2.50-mm region, slowing growth only outside that region may not significantly impact VR-QoL decline, though it may still affect functional vision or aspects of vision-related functioning not captured by the NEI VFQ-25.

A key limitation is that the cohort was restricted to individuals meeting the eligibility criteria for the Chroma and Spectri trials, including having bilateral GA with a specific perilesional autofluorescence pattern, a total lesion size of 1 to 7 disc areas and BCVA of 20/100 or better. The findings may not be fully generalizable to patients with unilateral GA, smaller central GA or more extensive non-subfoveal GA.

Reference

1. Anegondi N, Lam D, Guymer RH, et al. Vision-related quality of life in geographic atrophy: association with topographic lesion distribution. Ophthalmology. Accessed May 28, 2026. doi:10.1016/j.ophtha.2026.04.019