Safety Analysis of Repeated Low-Level Red Light Therapy on Cone Mosaic and Microvasculature by Adaptive Optics Scanning Laser Ophthalmoscopy

• RLRL therapy for 8 months does not exacerbate cone density loss or alter retinal microvasculature in myopic children. • RLRL-treated eyes uniquely showed reduced cone dispersion, a potential marker of photobiomodulation. • Cone density and spacing strongly correlate with axial length and refraction. • AOSLO imaging resolution (∼3 μm) approaches the minimal foveal cone spacing (∼2.5 μm), which may limit detection of subtle cone regularity changes at the foveal center and should be considered when interpreting certain conclusions. Despite the growing use of repeated low-level red light (RLRL) therapy for myopic children, its safety remains controversial. This study aimed to analyze macular cone morphology and retinal microvascular hemodynamics using adaptive optics scanning laser ophthalmoscopy (AOSLO) to evaluate retinal microstructural impacts. Cross-sectional study. Three groups were recruited: Myopia-RLRL (20 eyes; ≥8 months treatment), Myopia-Untreated (32 eyes), and Non-Myopic (14 eyes). RLRL therapy was administered using a 650 nm laser at 0.29 mW, twice daily for 3 minutes per session. AOSLO was used to assess cone morphology (density, spacing, regularity, dispersion) and retinal microvasculature parameters (vessel diameter, lumen diameter, blood flow velocity). Cone density was higher in the Non-Myopic group (45,649 ± 5,092/mm²) compared to the Myopia-Untreated group (40,202 ± 7,365/mm², P =0.013), with no significant difference versus the Myopia-RLRL group (41,274 ± 6,228/mm², P =0.627). Cone spacing was enlarged in myopic groups relative to non-myopic ( P <0.05). Cone dispersion was lower in the Myopia-RLRL group (16.47 ± 2.43%) than in the Myopia-Untreated (17.36 ± 2.14%, P =0.030) and Non-Myopic (17.80 ± 2.25%, P =0.024) groups. No significant intergroup differences were found in cone regularity or microvasculature. Cone density positively correlated with spherical equivalent refraction (r=0.541) and negatively with axial length (r=−0.594; both P <0.001). After 8 months of RLRL therapy, there was no significant loss of cone density or issues with retinal microvasculature. Assessment of retinal arteriole and venule diameters and flow velocity also showed no significant differences. The reduced cone dispersion in the RLRL group suggests potential photobiomodulation effects on photoreceptor organization, indicating the therapy's safety and potential benefits.