Purchase this article with an account.
Mohamed Kamel Soliman, Salman Sarwar, Muhammad Hassan, Muhammad Fayez Jawed, Mohammad Ali Sadiq, Aniruddha Agarwal, Rubbia Afridi, Kanika Aggarwal, Quan Dong Nguyen, Yasir Jamal Sepah; Repeatability and Reproducibility of Retinal Vessel Caliber Measurements Using Adaptive Optics Fundus Camera. Invest. Ophthalmol. Vis. Sci. 2016;57(12):5916. doi: https://doi.org/.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Assessment of the retinal vessels caliber is considered a valuable biomarker in several retinal vascular diseases. Several methods have been developed to quantify vessel caliber. However, none are able to identify or measure the vascular walls. Adaptive optics imaging can capture high-resolution images of the retinal vasculature, allowing an objective quantification of wall-to-lumen ratio. In the index study, the reproducibility and repeatability of retinal vessel caliber measurements using adaptive optics (AO) fundus camera were assessed.
An AO retinal camera (rtx1; Imagine Eyes, Orsay, France) was used to acquire images of retinal vessels from eleven subjects (21 eyes) with no known ocular disease. Retinal vessel images were captured at 2 different locations within the central 8°x10° of the posterior pole of each eye by 2 independent operators. Measurement of the retinal vessels lumen and wall were quantified using AOdetectArtery 2.0 beta 0.8 (Figure 1). Manual adjustments of the software measurements were made by 2 independent graders (different from operators). Axial length measurements were obtained using non-contact biometry. Assessment of intra-operator repeatability was done by evaluating 5 consecutive images obtained by the same operator in the same session. Inter-operator reproducibility was assessed by comparing results obtained from 2 trained operators. The coefficient of repeatability (CoR) and reproducibility and intraclass correlation coefficients (ICC) were calculated.
The CoR was 2.6μm and 1.9μm for the vessel lumen and wall, respectively. The ICC values were 0.99 (CI 0.98 – 0.99) for vessel lumen and 0.94 (CI 0.91 – 0.97) for vessel wall. Inter-operator reproducibility showed high ICC of 0.97 (CI 0.96-0.99) and 0.97 (CI 0.56-0.95) for the retinal vessel lumen and wall, respectively. The limit of agreement between the 2 graders using Bland-Altman plot were −0.30 and +0.35 and -0.94 and +0.8 for the vessel lumen and wall, respectively (Figure 2).
Retinal vessel wall and lumen measurements using AO camera are repeatable and reproducible. The technology may be employed to monitor changes in retinal vascular disease.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.
Adaptive optics image of retinal vessel depicting the method of vessel wall and lumen measurement by AOdetectArtery 2.0 software.
Bland-Altman plot of retinal vessel wall and lumen as measured by 2 graders
This PDF is available to Subscribers Only