Abstract
Purpose :
Flavoprotein fluorescence (FPF) imaging provides a readout of retinal oxidative stress by quantifying natural autofluorescence of oxidized retinal mitochondrial flavoproteins (535 nm) after blue light excitation (467 nm). Natural fluorophores in a patient’s lens get brighter with oxidative stress and so may confound the desired retinal FPF signal, resulting in a combined lens and retinal fluorescence signal that elevates the FPF score. At the same time, natural lens aging can attenuate the signal. The purpose of this study was to generate and evaluate a lens correction algorithm (LCA) for use with FPF imaging.
Methods :
Lens fluorescence and transmission were accounted for in a two-step process. First, interpolation was used to generate an estimate of lens fluorescence and isolate the retinal component. Second, after an adjustment for instrument calibration, natural lens attenuation was compensated using a third-order model. The efficacy of the LCA was assessed in 544 images of healthy control patients with no ocular pathology from different databases obtained automatically generated by Ocumet ® Image Analysis software at the time of patient presentation. A case study is also presented of a patient with images before and after cataract surgery, highlighting clinical relevance of this application.
Results :
After the LCA was applied the raw FPF scores of 554 images, the slope of best fit was attenuated (m= 1.4556 vs m= 0.0584) suggesting the correction algorithm was successful in accounting for native lens autofluorescence. Signals clustered tightly around line of best fit before and after compensation. A positive age-dependent relationship was maintained, although largely adjusted for (R2= 0.7012 vs 0.0185). After adjustment, FPF intensity scores displayed similar strength positive relationships with age when compared to patients with intraocular lens implants (IOL), which display minimal fluorescence (R2= 0.0165 vs 0.0193).
Conclusions :
The LCA reduced the effects of the aging lens and addressed an important confounder in measuring retinal FPF. Future work may test this algorithm on a large cohort of cataract patients after phacoemulsification and IOL implantation to confirm the FPF signal changes from aging lens fluorescence contribution
This abstract was presented at the 2022 ARVO Annual Meeting, held in Denver, CO, May 1-4, 2022, and virtually.