Premium
Assessment of the spectralis spectral domain oct segmentation software in a retrospective cohort study of exudative amd patients
Author(s) -
TILLEUL J,
QUERQUES G,
CANOUIPOITRINE F,
LEVEZIEL N,
SOUIED EH
Publication year - 2012
Publication title -
acta ophthalmologica
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.534
H-Index - 87
eISSN - 1755-3768
pISSN - 1755-375X
DOI - 10.1111/j.1755-3768.2012.2613.x
Subject(s) - bruch's membrane , inner limiting membrane , optical coherence tomography , medicine , limiting , ophthalmology , segmentation , retrospective cohort study , external limiting membrane , retinal , retinal pigment epithelium , computer science , pathology , artificial intelligence , mechanical engineering , engineering
Purpose To assess the ability of the Spectralis optical coherence tomography (OCT) segmentation software to identify the inner limiting membrane and the Bruch’s membrane in exudative AMD patients. Methods Thirty‐eight eyes of 38 naive exudative AMD patients were retrospectively included. They all had a complete ophthalmologic examination including Spectralis OCT at baseline, at month‐1 and at month‐2. Reliability of the segmentation software was was defined as good if both inner limiting membrane and Bruch’s membrane were correctly drawn. Results A total of 38 patients charts were reviewed (114 scans). The inner limiting membrane was correctly drawn in 114/114 SD‐OCT scans (100 %). Conversely, the Bruch’s membrane was correctly drawn in 59/114 scans (51.8 %). The software was less reliable in locating the Bruch’s membrane in case of pigment epithelium detachment (PED) than without PED (42.5% vs 73.5 % respectively, p =0.049). Conclusion Segmentation of inner limiting membrane was constantly trustworthy but Bruch’s membrane segmentation was poorly reliable using the automatic SD‐OCT Spectralis segmentation software. Based on this software, evaluation of retinal thickness may be incorrect, particularly in case of PED.