A deep learning model for detection of Alzheimer's disease based on retinal photographs: a retrospective, multicentre case-control study

作者：Carol Y. Cheung, An Ran Ran, Shujun Wang, Victor T.T. Chan, Kaiser Sham, Saima Hilal, Narayanaswamy Venketasubramanian, Ching‐Yu Cheng, Charumathi Sabanayagam, Yih Chung Tham, Leopold Schmetterer, Gareth J. McKay, Michael Williams, Adrian Wong, Lisa Au, Zhihui Lu, Jason C. Yam, Clement C. Tham, John J. Chen, Oana Dumitrașcu, Pheng‐Ann Heng, Timothy Kwok, Vincent Mok, Dan Miléa, Christopher Chen, Tien Yin Wong · 发表于：The Lancet Digital Health · 年份：2022 · DOI：10.1016/s2589-7500(22)00169-8 · 被引用次数：261 · 研究领域：Retinal Imaging and Analysis、Dementia and Cognitive Impairment Research、Retinal Diseases and Treatments

BACKGROUND: There is no simple model to screen for Alzheimer's disease, partly because the diagnosis of Alzheimer's disease itself is complex-typically involving expensive and sometimes invasive tests not commonly available outside highly specialised clinical settings. We aimed to develop a deep learning algorithm that could use retinal photographs alone, which is the most common method of non-invasive imaging the retina to detect Alzheimer's disease-dementia. METHODS: In this retrospective, multicentre case-control study, we trained, validated, and tested a deep learning algorithm to detect Alzheimer's disease-dementia from retinal photographs using retrospectively collected data from 11 studies that recruited patients with Alzheimer's disease-dementia and people without disease from different countries. Our main aim was to develop a bilateral model to detect Alzheimer's disease-dementia from retinal photographs alone. We designed and internally validated the bilateral deep learning model using retinal photographs from six studies. We used the EfficientNet-b2 network as the backbone of the model to extract features from the images. Integrated features from four retinal photographs (optic nerve head-centred and macula-centred fields from both eyes) for each individual were used to develop supervised deep learning models and equip the network with unsupervised domain adaptation technique, to address dataset discrepancy between the different studies. We tested the trained model...