Evaluation of choroid vascular layer thickness in wet age-related macular degeneration using artificial intelligence

Abstract

Purpose: To facilitate the assessment of choroid vascular layer thickness in patients with wet age-related macular degeneration (AMD) using artificial intelligence (AI).

Methods: We included 194 patients with wet AMD and 225 healthy participants. Choroid images were obtained using swept-source optical coherence tomography. The average Sattler layer-choriocapillaris complex thickness (SLCCT), Haller layer thickness (HLT), and choroidal thickness (CT) were auto-measured at 7 regions centered around the foveola using AI and subsequently compared between the 2 groups.

Results: The SLCCT was lower in the AMD group than in the control group (P < 0.05). The HLT was significantly higher in the AMD group than in the control group at the Tparafovea and T-perifovea in the total population (P < 0.05) and in the ≤70-year subgroup (P < 0.05). The CT was higher in the AMD group than in the control group, particularly at the N-perifovea, T-perifovea, and T-parafovea in the ≤70-year subgroup; Interestingly, it was lower in the AMD group than in the control group at the Nparafovea, N-fovea, foveola, and T-fovea in the >70-year subgroup (P < 0.05).

Conclusion: This novel AI-based auto-measurement was more accurate, efficient, and detailed than manual measurements. SLCCT thinning was observed in wet AMD; however, CT changes depended on the interaction between HLT compensatory thickening and SLCCT thinning.

Keywords: Artificial intelligence; Choroidal thickness; Haller layer thickness; OCT; Sattler layer-choriocapillaris complex thickness; Wet age-related macular degeneration.