Abstract
Purpose :
Two-photon (2P) retinal imaging can non-invasively reveal defects in the visual cycle prior to structural damage. A 2P scanning laser ophthalmoscope (SLO) system for imaging the living human eye has been previously developed to capture fundus reflectance and 2P retinal fluorescence images. Here, we aimed to improve the reproducibility of high-quality images and ease of operation of this 2P human imaging system.
Methods :
We developed a standing desk-type optical table that accommodates the entire optical apparatus and is adjustable for different subject heights. Fellow eye fixation, previously achieved by a singular diode, was replaced by an 8x8 LED matrix and mounted on a motor for alignment. A motorized version of the chinrest was developed to allow for more precise control of subject position and stability during measurements. Real-time controls of fixation position and chinrest alignment were incorporated into software (Figure 1).
Results :
Adjustable optical table allowed for accommodation of subjects of different heights. Improvements with the 2P human system’s fixation and chinrest mechanism provided greater control over subject alignment, which is critical for achieving sufficient 2P fluorescence signal. Real-time manipulation allowed the operator to readily realign the subject’s eye without interrupting the imaging session and with minimal reliance on subject feedback which may be unreliable. With our system, an additional 10 healthy individuals were successfully imaged with discernible 2P fundus images (Figure 2).
Conclusions :
Flexible adjustment of optical system and human subject positioning is critical for image acquisition. Automation of the chinrest and digital fixation target adjustments can address the issues of subject alignment and stability for 2P imaging of the human retina. Achieving proper subject alignment in a time efficient manner remains a challenge and needs to be overcome, especially for elderly patients who may have poor fixation. These challenges may be overcome using pupil cameras to ensure proper positioning, additional chinrest modifications and enhanced data processing methods to accelerate image generation.
This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.