JosE Mª Ruiz-Moreno

Quantification of geographic atrophy associated to AMD: automatic software.


Prof. José Mª Ruiz-Moreno,1,2,3,4 Jorge Ruiz Medrano.5,6

1 President of the Spanish Society of Retina and Vitreous (SERV)

2 Professor of Ophthalmology, Castilla La Mancha University (UCLM)

3 Head of the Ophthalmology Department, Universitary Hospital Puerta de Hierro-Majadahonda. Madrid

4 Medical Director Vissum Corporation

5 Fellow of the Programa de Formación de Expertos en Retina SERV. Bellvitge University Hospital. Barcelona.

6Fellow Jules Gonin Eye Hospital, Fondation Asile des Aveugles, Lausanne





To determine the accuracy and repeatability of the new software to automatically quantify Geographic Atrophy (GA) areas associated to Age-related Macular Degeneration (AMD) using swept-source optical coherence tomography (SS-OCT).

Patients and Methods.

Cross-sectional, non-interventional study. Eyes from AMD patients with GA area, without previous choroidal neovascularization, were scanned using a Triton SS-OCT (Topcon Corporation, Japan). The study included: Best corrected visual acuity, complete ophthalmological examination, colour phundus photography, fundus autofluorescence (FAF), 7x7 mm and 5x12 mm cross-line OCT scan protocols with eye tracking (three consecutive times). Three observers independently determined the area and perimeter using FAF images as “gold standard”, performing the corrections of the automatic segmentation of the atrophic area. All tasks were performed in a masked fashion.


Mean GA area was not statistically different when comparing automatic vs. manual techniques (12.39 mm2 vs. 12.26 mm2; p=0.23). Coefficient of variation was smaller using automatic procedures (11.94% vs. 39.27% respectively; p=0.01).


The new software allows us to automatically quantify GA areas and perimeters associated to AMD. The image analysis is fast, objective, with a smaller coefficient of variation than manual procedures. The accuracy of the new software is better in regular atrophic lesions when compared to irregular ones, which makes it necessary to improve its segmentation for the latter.