Dr. Andrei Szigiato was honored with the Retina Society Fellowship Research award for his work at the Cole Eye Institute on comparing half-dose versus full-dose fluorescein for ultra-wide field fundus fluorescein angiography (UWFFA) under the mentorship of Dr. Sunil Srivastava. This project was spurred by the fluorescein angiography shortage at the time.
Can you tell our readers about your study?
This study aimed to compare image quality, clinical utility, and the side effect profile of half-dose (250mg) vs full-dose (500mg) fluorescein in UWFFA for a cohort of stable patients. UWFFA is a diagnostic procedure that allows clinicians to assess the anatomy, physiology, and pathology of retinal and choroidal circulation, where a series of fundus photographs of both eyes are taken after intravenous injection of sodium fluorescein. These photographs aid the retinal specialist in the assessment of ischemia, neovascularization, and disease activity in diabetic retinopathy, retinal vascular occlusion, uveitis, and numerous other conditions. In early 2022, a fluorescein shortage occurred in the United States. To meet the standard of care for patients who required UWFFA, a regimen of half-dose (250mg) sodium fluorescein (10%) was adopted instead of the full-dose (500mg) to ration fluorescein at the Cole Eye Institute (CEI). This study was a formal assessment of the image quality, clinical utility, and the side effect profile between the two doses for a cohort of stable patients.
What patients were included in the study?
We included patients with retinal vascular disease that received both half-dose and full-dose UWFFA (Optos California, Dunfermline, UK) within 6 months at the CEI. Eyes were excluded if they received intraocular injections, laser procedures, new immunosuppression, and worsened or improved inflammation on clinical examination. There were 52 eyes included in the study, 42 (81%) uveitic, 5 (9%) diabetic, and 4 (8%) normal controls.
Tell us more about how UWFFA images differed with half dose vs full dose of fluorescein?
We compared the images qualitatively, with 2 masked retinal specialists assessing the de-identified late phase images (5-7mins) side by side, and qualitatively, with a machine learning enhanced automated segmentation platform. Qualitatively, there was no observable difference in image quality or observed rates of leakage between doses, with a high interrater reliability (78-87% agreement, kappa 0.642). UWFFA quality was deemed equal by reviewers in 54% of images and an equal proportion of full dose and half dose images were deemed to be higher in quality (21% vs 23%), so there was no clear preference for either dose. Automated leakage analysis showed mildly increased leakage in half-dose images overall (3.8% vs 2.8%, p=0.01), and in the macula (1.5% vs 0.6%, p=0.01). However, this difference was small and unlikely to be clinically significant. Further analysis of the images showed that these patients had increased media opacity (ie from a cataract) that required the technician acquiring the images to increase the brightness on the photos, which was interpreted by the automated algorithm as leakage. To the trained retina specialist, the 2 images showed stable retinal disease.
What are your next steps?
There are a number of questions we would still like to answer. Firstly, what would the quality of a quarter dose (125mg) of fluorescein be compared to full or half dose? Do these results also apply for clinically active patients? How do these images compare with patients that had oral fluorescein? Would the same findings apply for ICG? Many of these questions are currently being investigated in follow-up
studies at the CEI.
