{"database":"biostudies-literature","file_versions":[],"scores":null,"additional":{"submitter":["Son J"],"funding":["National IT Industry Promotion Agency","National Research Foundation of Korea"],"pagination":["5934"],"full_dataset_link":["https://www.ebi.ac.uk/biostudies/studies/S-EPMC10097752"],"repository":["biostudies-literature"],"omics_type":["Unknown"],"volume":["13(1)"],"pubmed_abstract":["The identification of abnormal findings manifested in retinal fundus images and diagnosis of ophthalmic diseases are essential to the management of potentially vision-threatening eye conditions. Recently, deep learning-based computer-aided diagnosis systems (CADs) have demonstrated their potential to reduce reading time and discrepancy amongst readers. However, the obscure reasoning of deep neural networks (DNNs) has been the leading cause to reluctance in its clinical use as CAD systems. Here, we present a novel architectural and algorithmic design of DNNs to comprehensively identify 15 abnormal retinal findings and diagnose 8 major ophthalmic diseases from macula-centered fundus images with the accuracy comparable to experts. We then define a notion of counterfactual attribution ratio (C"],"journal":["Scientific reports"],"pubmed_title":["An interpretable and interactive deep learning algorithm for a clinically applicable retinal fundus diagnosis system by modelling finding-disease relationship."],"pmcid":["PMC10097752"],"funding_grant_id":["NIPA-C0202-17-1045","NRF-2021R1C1C1014138"],"pubmed_authors":["Shin JY","Park KH","Park J","Kim HD","Jung KH","Son J","Park SJ","Kong ST","Kwon G"],"additional_accession":[]},"is_claimable":false,"name":"An interpretable and interactive deep learning algorithm for a clinically applicable retinal fundus diagnosis system by modelling finding-disease relationship.","description":"The identification of abnormal findings manifested in retinal fundus images and diagnosis of ophthalmic diseases are essential to the management of potentially vision-threatening eye conditions. Recently, deep learning-based computer-aided diagnosis systems (CADs) have demonstrated their potential to reduce reading time and discrepancy amongst readers. However, the obscure reasoning of deep neural networks (DNNs) has been the leading cause to reluctance in its clinical use as CAD systems. Here, we present a novel architectural and algorithmic design of DNNs to comprehensively identify 15 abnormal retinal findings and diagnose 8 major ophthalmic diseases from macula-centered fundus images with the accuracy comparable to experts. We then define a notion of counterfactual attribution ratio (C","dates":{"release":"2023-01-01T00:00:00Z","publication":"2023 Apr","modification":"2025-04-04T14:23:47.23Z","creation":"2025-02-19T03:59:52.243Z"},"accession":"S-EPMC10097752","cross_references":{"pubmed":["37045856"],"doi":["10.1038/s41598-023-32518-3"]}}