Fundus autofluorescence imaging using red excitation light

**Background:** Fundus autofluorescence (AF) imaging is a standard method for diagnosing and monitoring retinal diseases, typically using short-wavelength (SW) blue or green excitation light (e.g., 486 nm). However, SW-AF has limitations including glare, reduced image quality in cataract, poor foveal visualization due to macular pigment absorption, and potential retinal light toxicity. Longer wavelength excitation, such as near-infrared (785 nm), has lower signal intensity. This study investigates a novel red-excitation AF (R-AF) technique using 642 nm and 705 nm lasers to overcome these limitations.

**Methods:** This prospective, cross-sectional study (NCT03592017) was conducted at the Oxford Eye Hospital. A custom-built confocal scanning laser ophthalmoscope (cSLO) equipped with 642 nm and 705 nm excitation lasers was used. 80 participants were enrolled: 15 healthy subjects and 65 patients with retinal diseases (including age-related macular degeneration, Stargardt disease, and others). Image acquisition included 30° field of view, 1536×1536 pixels, with 12 successive images per series. Perceived glare was assessed using the de Boer scale (9-point scale) in 21 participants. Test-retest repeatability was evaluated by having subjects move away and refocusing. Quantitative AF signal intensity was measured using an in-house image processing pipeline, with correction for vessel effects and dark offset. Statistical analyses included Bland-Altman methods and Kruskal-Wallis tests.

**Key Results:** R-AF images showed a consistent pattern in normal eyes, with AF intensity increasing slightly from the optic disc to the fovea. The foveal signal was similar to eccentric areas with 642 nm excitation and slightly increased with 705 nm. No significant bleaching of the AF signal was observed over 30 seconds. Defocus of ±0.42 D (642 nm) and ±0.30 D (705 nm) resulted in only a 5% decrease in signal. Test-retest repeatability was ±12% for 642 nm and ±9% for 705 nm. Perceived glare was significantly lower with longer wavelengths: median de Boer scores were 5.0 (IQR 4-6) for 486 nm, 6 (IQR 5-7) for 642 nm, 7 (IQR 6.5-9) for 705 nm, and 9 (IQR 7-9) for 788 nm (p < 0.0001). In a patient with cataract, R-AF images showed minimal quality reduction compared to SW-AF, which was severely degraded. In Stargardt disease, R-AF (642 nm) showed increased AF intensity similar to SW-AF, while 705 nm showed a less pronounced increase. Case studies demonstrated that R-AF could reveal additional pathology, such as central AF loss in AMD not seen on SW-AF or OCT, and improved foveal assessment in vitelliform lesions and Stargardt disease.

**Clinical Implications:** R-AF imaging offers several advantages over conventional SW-AF: improved patient comfort due to reduced glare, better image quality in patients with cataract, and enhanced visualization of the fovea by avoiding macular pigment absorption. The technique provides quantitative AF measurements that may serve as a safe outcome measure in clinical trials targeting lipofuscin accumulation. R-AF may also reveal additional diagnostic information, as seen in AMD and Stargardt disease, potentially aiding in disease stratification and prognostication. However, limitations include inability to perform fluorescein angiography and lack of detection of macular pigment abnormalities. Further disease-specific studies are needed to confirm generalizability.