Abstract
Purpose :
Off-Axis Full-Field Time-Domain Optical Coherence Tomography (OA-FF-TD-OCT) is a compact OCT technique that can be produced at a very low-cost. It is a potential candidate for home-care OCT solutions for AMD patients, if the patient can perform the OCT scan reliably by himself. In this study, we introduced AMD patients with typical age and vision to our device. Aim of this study was to measure the success rate of OA-FF-TD-OCT for patient self-assessment and to compare image quality to standard devices.
Methods :
In a clinical study, patients with various retinal diseases were placed in front of our OA-FF-TD-OCT device. The patients received a 5-10-minute training in device usage and were then asked to perform a retinal self-scan. This requires the correct alignment of the eye by the patient himself, who can start the measurement by pushing a trigger button. As a reference standard, SD-OCT pictures with a Heidelberg Spectralis OCT were taken.
After acquisition, the images were graded in terms of signal-to-noise ratio, motion artefacts and foveal centering. Only when all of these criteria reached a sufficient score, the picture was marked as “gradable”, i.e. sufficient for determination of disease activity.
Results :
51 patients were included in the study (39 AMD patients, 6 patients with diabetic macular edema, 3 patients with retinal vein occlusion, 3 patients with other diseases). Patient age was between 45 and 86 years, visual acuity was between 20/200 and 20/20. After the brief training, 73% of the patients were able to acquire images that were clinically gradable. Failure to acquire sufficient image quality was mainly correlated with higher age, low visual acuity and geographic atrophy. In successful patients, it could be shown that the most relevant biomarkers for AMD (intraretinal fluid, subretinal fluid and pigment epithelium detachment) could be identified with high sensitivity.
Conclusions :
OA-FF-TD-OCT allows high-resolution retinal self-scanning by patients with retinal diseases. However, not all patients were able to produce images with sufficient image quality. Further improvement of the patient interface and longer patient training could improve usability, but still, presumably not all patients are suitable candidates.
This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.