Automation improves repeatability of retinal oximetry measurements.
Robert Arnar Karlsson1,2, Olof Birna Olafsdottir1,3, Vedis Helgadottir1
1Faculty of Medicine, University of Iceland, Reykjavik, Iceland.
Plos One
|December 16, 2021
Summary
Automated retinal oximetry measurements significantly improve repeatability for both vessel oxygen saturation and diameter compared to manual methods. This advancement enhances the reliability of this spectrophotometry technique for clinical applications.
Area of Science:
- Ophthalmology
- Medical Imaging
- Spectrophotometry
Background:
- Retinal oximetry uses spectrophotometry to measure oxygen saturation and vessel diameter.
- Manual selection of retinal vessels for analysis has been a limitation.
Purpose of the Study:
- To evaluate the impact of automating retinal oximetry measurements.
- To compare the repeatability of automated versus semiautomated analysis for retinal vessel oxygen saturation and diameter.
Main Methods:
- Five operators analyzed retinal images using semiautomated (v2.5.1) and automated (v3.0) Oxymap Analyzer software.
- Inter- and intra-operator variability were assessed using intraclass correlation coefficients (ICC).
Main Results:
- Automated saturation measurements showed high inter-rater ICCs (0.97 arterioles, 0.96 venules).
- Automated diameter measurements also demonstrated high inter-rater ICCs (0.97 arterioles, 0.95 venules).
- Automated ICCs were significantly higher (p < 0.01) than semiautomated ICCs for both parameters.
Conclusions:
- Automated retinal oximetry offers superior repeatability compared to manual vessel selection.
- This automation enhances the reliability of retinal oximetry for clinical use.
