Purchase this article with an account.
Rebekka Heitmar; Comparison of different imaging protocols used in retinal vessel oximetry. Invest. Ophthalmol. Vis. Sci. 2016;57(12):3747.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
There is a wide variety of equipment available to obtain retinal vessel oxygen saturation measurements, the same is true for the protocols employed to collect and analyse the data. While some "standards" in regards to measurement location have emerged, there is little known on the influence of (a) number of vessels included and (b) image size (i.e. 30 degree vs 50 degree images).
All participants were free from ocular and systemic disease, normotensive (systolic BP <140mmHg and diastolic BP <90mmHg), with normal IOPs (<21mmHg) and had no previous ocular trauma or surgery.Following a full eye examination to ensure clear media and good visual acuity participants (n=47) were dilated with 1% tropicamide. Once full pupil dilation was reached we obtained a total of 10 retinal images using a ZeissFF450+ fundus camera with a dual-wavelength filter (specification:610 (+/- 10) nm and 548 (+/- 10) nm) in place. All images were obtained with the ONH centred, 5 images with the camera angle set to 30 degrees and the remaining 5 with the camera angle set at 50 degrees.For analysis the three best images of each setting were used to obtain average values of arterial and venous vessel saturation parameters at 30 and 50 degrees respectively. All vessels larger than 70micron in diameter which crossed through the measurement area (Fig. 1) were included for analyses.
There was no differnce between 30 and 50 degree images for arterial (30 degree image: 94% (SD+/-8) vs 50 degree image: 94% (SD+/-8); paired t-test: p=0.752) and venous saturation values (30 degree image: 51% (SD+/-14) vs 50 degree image: 51% (SD+/-14); paired t-test: p=0.505). Agreement between the two image sizes was higher for arterial saturation values (mean bias 0.21% (limits of agreement +/-8.82%) than in veins (mean bias 0.71% (limits of agreement +/-14.11%). Furthermore, average arterial and venous saturation values were independent of number of vessels included.
Overall, there is good agreement between values obtained with different images sizes and between protocols.While there is good agreement study comparisons should still acknowledge the differences in protocols as there might well be differences in some of the protocols, i.e. where only one or two vessels were analysed or only a specific retinal location was included.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.
This PDF is available to Subscribers Only