{"doi":"10.1016/j.mcpdig.2024.08.005","title":"Color Fundus Photography and Deep Learning Applications in Alzheimer Disease","abstract":"Objective: To report the development and performance of 2 distinct deep learning models trained exclusively on retinal color fundus photographs to classify Alzheimer disease (AD). Patients and Methods: Two independent datasets (UK Biobank and our tertiary academic institution) of good-quality retinal photographs derived from patients with AD and controls were used to build 2 deep learning models, between April 1, 2021, and January 30, 2024. ADVAS is a U-Net-based architecture that uses retinal vessel segmentation. ADRET is a bidirectional encoder representations from transformers style self-supervised learning convolutional neural network pretrained on a large data set of retinal color photographs from UK Biobank. The models' performance to distinguish AD from non-AD was determined using mean accuracy, sensitivity, specificity, and receiving operating curves. The generated attention heatmaps were analyzed for distinctive features. Results: =.04). No major differences were noted between the original and binary vessel segmentation and between both eyes vs single-eye models. Attention heatmaps obtained from patients with AD highlighted regions surrounding small vascular branches as areas of highest relevance to the model decision making. Conclusion: A bidirectional encoder representations from transformers style self-supervised convolutional neural network pretrained on a large data set of retinal color photographs alone can screen symptomatic AD with high accuracy, better than U-Net-pretrained models. To be translated in clinical practice, this methodology requires further validation in larger and diverse populations and integrated techniques to harmonize fundus photographs and attenuate the imaging-associated noise.","journal":"Mayo Clinic Proceedings Digital Health","year":2024,"id":441039,"datarank":0.38638976803753927,"base_score":2.302585092994046,"endowment":2.302585092994046,"self_citation_contribution":0.3453877639491069,"citation_network_contribution":0.04100200408843236,"self_endowment_contribution":0.3453877639491069,"citer_contribution":0.04100200408843236,"corpus_percentile":52.99760191846523,"corpus_rank":6077,"citation_count":9,"citer_count":6,"citers_with_citation_signal":2,"citers_with_endowment":2,"datacite_reuse_total":0,"is_dataset":true,"is_dataset_confidence":0.7273,"is_data_producer":false,"deposit_databanks":null,"is_oa":true,"file_count":0,"downloads":0,"has_version_chain":false,"published_date":"2024-01-01","fair_score":null,"fair_percentile":null,"algorithm_id":"datarank_citation_only_1hop_v6","ranking_scope":"data_only","authors":[{"id":896086,"name":"Xin Li","orcid":"0000-0001-6449-4044","position":1,"is_corresponding":false},{"id":952006,"name":"Wenhui Zhu","orcid":"0009-0000-5207-6283","position":2,"is_corresponding":false},{"id":589377,"name":"Bryan K. Woodruff","orcid":"0000-0002-4728-7294","position":3,"is_corresponding":false},{"id":737060,"name":"Simona Nikolova","orcid":"0000-0002-9203-1159","position":4,"is_corresponding":false},{"id":1161836,"name":"Jacob Sobczak","orcid":null,"position":5,"is_corresponding":false},{"id":1253462,"name":"Amal Youssef","orcid":"0000-0003-2870-0191","position":6,"is_corresponding":false},{"id":1253463,"name":"Siddhant Saxena","orcid":"0000-0001-7402-1785","position":7,"is_corresponding":false},{"id":1254017,"name":"Janine Andreev","orcid":null,"position":8,"is_corresponding":false},{"id":407159,"name":"Richard J. Caselli","orcid":"0000-0001-5784-2987","position":9,"is_corresponding":false},{"id":782615,"name":"John J. Chen","orcid":"0000-0003-0170-0407","position":10,"is_corresponding":false},{"id":317852,"name":"Yalin Wang","orcid":"0000-0002-6241-735X","position":11,"is_corresponding":false},{"id":241438,"name":"Oana M. Dumitrascu","orcid":"0000-0003-2033-449X","position":0,"is_corresponding":true}],"reference_count":78,"raw_metadata":null,"created_at":"2026-07-19T02:01:06.281460Z","pmid":"39748801","pmcid":"PMC11695061","fwci":null,"citation_percentile":null,"influential_citations":0,"oa_status":null,"license":null,"views":0,"total_file_size_bytes":0,"version_count":0,"fair_f":null,"fair_a":null,"fair_i":null,"fair_r":null,"fair_zscore":null,"fair_rationale":null,"fair_model":null,"fair_agent_version":null,"fair_fulltext_source":null,"fair_has_llm":null,"fair_computed_at":null,"clinical_trials":[],"software_tools":[],"db_accessions":[],"linked_datasets":[],"topics":[]}