Abstract
Purpose :
Self-examination low-cost full-field optical coherence tomography (SELFF-OCT) can place the control of ophthalmic therapies into the hand of the patient. Regular self-examination by volumetric OCT imaging of the retina can detect retinal changes in age-related macular degeneration (AMD) and other diseases in such early stages that optimal treatment is possible. Frist clinical studies showed that 77% [1] to 86% [2] of all self-examined eyes resulted in interpretable retinal OCT volume scans. Necessity for anti-VEGF treatment of AMD could be determined with 94% sensitivity and 95% specificity [2]. During the recording time of 1.3 seconds, motion of the eye can considerably degrade image quality. Depending on severeness, planes in the volume will be shifted laterally or axially, SNR will be reduced or en-face planes are completely missing.
[1] von der Burchard C, et al. Graefes Arch Clin Exp Ophthalmol, 2020, DOI: 10.1007/s00417-020-05035-6.
[2] von der Burchard C, et al. Ophthalmology, 2021, submitted
Methods :
SELFF-OCT records interferograms of en-face plane, which contain not only information on the retinal structure but also on the iris position. The imaging process and the possibility to detect ocular motion from a recorded volume were analyzed theoretically and experimentally in a phantom and in self-measurements of patients.
Results :
The SELFF-OCT data contained accurate position information. Accuracies better than 4 µm laterally and 30 µm axially were demonstrated. An algorithm quantified both lateral and axial changes of the retina position in the recorded images. The accuracy is on par with modern eye-trackers, but was achieved without any additional optical components. A complete loss of the OCT signal was also successfully identified.
Conclusions :
Due its special technology and the intended hand-held use, SELFF-OCT is more sensitive to motion artifacts than clinical OCT devices. Currently, 20% of the recorded volumes cannot be analyzed, and part of them contain motion artifacts. Automated detection can identify invalid volumes. Since acquisition of a volume takes only 1.3 seconds, remeasuring is possible in these cases. Integrated into the self-measurement process (i.e. repeat measurement until a valid scan was obtained) will increase success rate. Furthermore, automated artifact correction in postprocessing in cases of light and especially lateral movement will improve image quality and diagnostic accuracy.
This is a 2021 ARVO Annual Meeting abstract.