Purchase this article with an account.
Marco Mazzola, Ekaterina Semenova, Wei Wei, R Theodore Smith, Paul T. Finger; Quantitative autofluorescence characteristics of choroidal nevi. Invest. Ophthalmol. Vis. Sci. 2019;60(9):727.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Differentiating choroidal nevi from choroidal melanomas remains challenging. Quantitative autofluorescence (qAF) is performed by calibrating the AF image to an embedded reference of known fluorescence efficiency (RT Smith et al, IOVS 2011). We performed an observational cross-sectional pilot study to first assess the utility of qAF in the evaluation of choroidal nevi.
Choroidal nevi were diagnosed through standard ophthalmic evaluation including indirect ophthalmoscopy, color fundus photography (CFP), qAF, ultrasonography, fluorescein angiography (FA), and spectral domain OCT (SD-OCT). Twelve patients identified as having a choroidal nevus were included. Presence of commonly accepted risk factors (RF) for malignancy (thickness ≥ 2 mm, orange pigmentation (OP), focal fluorescence on FA, as well as the presence of subretinal fluid (SRF) and dysmorphic ellipsoid zone (EZ) on the SD-OCT) was assessed.Choroidal nevi with 0-1 RF were labelled as low risk (LR), whereas those with 2 or more RF were labelled as high risk (HR). CFP and qAF images were co-registered using a Matlab algorithm. Lesion area (LA) was manually identified on CFP using Photoshop. Vessel within the LA were eliminated on the qAF image. Mean qAF and standard deviation were then computed for every LA. Mean qAF levels and qualitative AF features were correlated with the clinical characteristics.
Six nevi were classified as LR, and 6 as HR. Mean qAF did not differ between the 2 groups (118.51 ± 14.16 for the LR group, 109.78 ± 19.84 for the HR group; unpaired t test, p=0.4). Five out of 6 of the LR nevi were not visible on the AF images, whereas all of the HR nevi were completely or partially visible. AF visibility distinguished HR nevi from LR nevi with a sensitivity of 100%, a specificity of 83.3%, a positive predictive value of 85.7%, and a negative predictive value of 100%. The odds ratio (OR) between AF visibility and thickness ≥ 2 mm was 5; with SRF was 5; with FA staining was 8, and with OP was 8.55. The OR between AF visibility and EZ disruption was 47.67, and was the only one statistically significant (Fishers’s exact test, p=0.015).
AF is a promising method to determine metabolic activity or cell death over the surface of the tumor, and could therefore be used as a non-invasive screening tool to evaluate choroidal nevi risk. However, clinical utility of the quantification of AF levels remains to be established in future studies.
This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.
This PDF is available to Subscribers Only