Segmentation errors in OCT‐angiography associated with epiretinal membranes
Purpose To explore segmentation errors, image quality and motion associated artifacts in eyes with idiopathic epiretinal membrane (ERM). Methods Retrospective comparative study. We included 32 eyes from 32 patients affected by ERM, as well as 32 eyes from 32 healthy, age‐matched subjects. Optical co...
Gespeichert in:
Veröffentlicht in: | Acta ophthalmologica (Oxford, England) England), 2019-12, Vol.97 (S263), p.n/a |
---|---|
Hauptverfasser: | , , , , , |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | Purpose
To explore segmentation errors, image quality and motion associated artifacts in eyes with idiopathic epiretinal membrane (ERM).
Methods
Retrospective comparative study. We included 32 eyes from 32 patients affected by ERM, as well as 32 eyes from 32 healthy, age‐matched subjects. Optical coherence tomography‐angiography (OCT‐A) was performed in the right eye of healthy controls (65.82 ± 17.43 years) and on the eye with the most advanced clinically diagnosed ERM (69.53 ± 13.12 years). The device used in this study provides a set of reference planes automatically segmented by the integrated software including the internal limiting membrane (ILM), the inner plexiform layer (IPL), and the retinal pigment epithelium (RPE) reference. Segmentation was regarded as inaccurate if either segmentation deviated from the correct plane by more than 50 μm. Presence of motion artifacts (blink lines, displacement, stretch artifacts, quilting, vessel doubling) as well as image quality were reported.
Results
Mean image quality was 7.2 ± 0.9 for the ERM patients. Phakic eyes had a mean quality of 7.4 ± 0.8, and pseudo‐phakic eyes had a mean quality of 7.2 ± 0.6 (p = 0.107). Motion artifacts were 1.22 ± 0.7 in the study cohort. Segmentation was accurate in all healthy subjects (n = 32). Segmentation errors occurred in 65.6% of ERM patients. The IPL was the segmentation boundary most prone to inaccurate segmentation, followed by the ILM. Segmentation of RPE layer was accurate in 96.7% of all cases.
Conclusions
OCT‐A image quality is an essential aspect in clinical practice, however, in several cases it cannot be reproduced in pathological conditions and is prone to motion artifacts and segmentation errors. Incorrect segmentation results in anatomically incorrect en‐face OCT‐A images and consequently in false quantification measures. Even though manual segmentation can be performed, it is rather laborious and suffers from interobserver errors. |
---|---|
ISSN: | 1755-375X 1755-3768 |
DOI: | 10.1111/j.1755-3768.2019.5134 |