LAUSR

PURPOSE To quantify the extent of topographic correspondence between baseline (BSL) choroidal neovascularization (CNV) and macular atrophy (MA) at follow-up in eyes with neovascular age-related macular degeneration (NVAMD). DESIGN Post hoc analysis of randomized controlled clinical trial data. METHODS Sixty treatment-naïve NVAMD patients from the TREX-AMD trial were followed for 18 months. Regions of month 18 macular atrophy (MA) were graded on fundus autofluorescence (FAF) with guidance of spectral-domain optical coherence tomography (SDOCT). CNV lesions were graded manually on fluorescein angiography (FA) with lesion components including classic and occult CNV delineated. FAF and FA images were registered to quantitate area and location of overlap between CNV and MA. Outcome measures included overlap of month 18 MA to BSL CNV subtype and progression of MA from BSL to month 18. RESULTS Twenty-six eyes had both MA at month 18 and CNV at BSL. A total of 84.6% of eyes showed evidence of MA and CNV overlap. MA appeared by month 18 in regions corresponding to BSL classic CNV in 36.4% of eyes and occult CNV in 40.9%, and in both regions in 22.7%, with more area of MA (AMA) in regions of occult than classic CNV. MA position at BSL corresponded to BSL classic CNV in 76.9% of eyes and occult CNV in 61.5%, and to both regions in 15.4%, with more AMA in regions of occult than classic CNV. Among eyes with MA and CNV at BSL but with no overlap, 50% progressed to involve regions with BS -CNV. Six eyes had no BSL MA but developed MA at month 18 within regions of BSL CNV. CONCLUSIONS In ranibizumab-treated eyes with NVAMD, more MA lesions develop within the region of baseline CNV (type 1, CNV-based MA) than outside (type 2, CNV-independent MA). Baseline-MA also tends to be located within regions of CNV in the pretreatment phase.