Translation of Color Fundus Photography into Fluorescein Angiography using Deep Learning for Enhanced Diabetic Retinopathy Screening
نویسندگان
چکیده
PurposeTo develop and validate a deep-learning model that can transform color fundus photography (CF) into corresponding venous late-phase fluorescein angiography (FFA) imagesDesignCross-sectional studySubjectsWe included 51,370 CF-venous FFA pairs 14,644 CF-late from 4,438 patients for development. External testing involved 50 eyes with CF-FFA two public datasets diabetic retinopathy (DR) classification, 86952 CF EyePACs, 1744 MESSIDOR2.MethodsWe trained to images. The translated images’ quality was evaluated quantitatively on the internal test set subjectively 100 paired images (50 external), based realisticity of global image, anatomical landmarks (macula, optic disc, vessels), lesions. Moreover, we validated clinical utility classifying five-class DR macular edema (DME) in EyePACs MESSIDOR2 datasets.Main Outcome MeasuresImage generation assessed by Structural Similarity Measures (SSIM), experts five-point scale (1 refers real FFA), intra-grader agreement kappa. classification accuracy Area under Receiver Operating Characteristic curve (AUC).ResultsThe SSIM were >0.6, subjective scores ranged 1.37-2.60. Both reported similar substantial (all kappas>0.8). Adding generated top improved datasets, AUC increased 0.912 0.939 dataset 0.952 0.972 dataset. DME also 0.927 0.974 dataset.ConclusionOur CF-to-FFA framework produced realistic adding screening. These results suggest translation could be used as surrogate method when examination is not feasible simple add-on improve
منابع مشابه
Fundus photography and fluorescein angiography Fluorescein angiography
Fluorescein angiography is a clinical diagnostic test procedure in which rapid-sequence photography is performed after intravenous injection of sodium fluorescein. The procedure allows the documentation of the retinal circulation. Of equal importance is the ability of the angiogram to outline the entire retinal vascular bed. Previous clinical methods of examination of the fundus enabled the phy...
متن کاملDeep Learning for Automated Quality Assessment of Color Fundus Images in Diabetic Retinopathy Screening
192 words] Purpose: To develop a computer based method for the automated assessment of image quality in the context of diabetic retinopathy (DR) to guide the photographer. Methods: A deep learning framework was trained to grade the images automatically. A large representative set of 7000 color fundus images were used for the experiment which were obtained from the EyePACS (http://www.eyepacs.co...
متن کاملFundus fluorescein angiography in retinopathy of prematurity.
In this issue of the Journal of Pediatric Ophthalmology & Strabismus, Ng, Lanigan, and O’Keefe present their findings regarding fundus fluorescein angiography in the management and screening of retinopathy of prematurity (ROP). Recently, there has been interest in photographic documentation of ROP in hospitals lacking sufficient ophthalmologic staff. Images created in the intensive care nursery...
متن کاملDiabetic retinopathy screening using digital non-mydriatic fundus photography and automated image analysis.
PURPOSE To investigate the use of automated image analysis for the detection of diabetic retinopathy (DR) in fundus photographs captured with and without pharmacological pupil dilation using a digital non-mydriatic camera. METHODS A total of 83 patients (165 eyes) with type 1 or type 2 diabetes, representing the full spectrum of DR, were photographed with and without pharmacological pupil dil...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: Ophthalmology science
سال: 2023
ISSN: ['2666-9145']
DOI: https://doi.org/10.1016/j.xops.2023.100401