Screening and identifying hepatobiliary diseases through deep learning using ocular images: a prospective, multicentre study

Abstract

Background Ocular changes are traditionally associated with only a few hepatobiliary diseases. These changes are non-specific and have a low detection rate, limiting their potential use as clinically independent diagnostic features. Therefore, we aimed to engineer deep learning models to establish associations between ocular features and major hepatobiliary diseases and to advance automated screening and identification of hepatobiliary diseases from ocular images.Methods We did a multicentre, prospective study to develop models using slit-lamp or retinal fundus images from participants in three hepatobiliary departments and two medical examination centres. Included participants were older than 18 years and had complete clinical information; participants diagnosed with acute hepatobiliary diseases were excluded. We trained seven slit-lamp models and seven fundus models (with or without hepatobiliary disease [screening model] or one specific disease type within six categories [identifying model]) using a development dataset, and we tested the models with an external test dataset. Additionally, we did a visual explanation and occlusion test. Model performances were evaluated using the area under the receiver operating characteristic curve (AUROC), sensitivity, specificity, and F1* score.

Identifying diabetes from conjunctival images using a novel hierarchical multi-task network

“…Besides the lack of training, we think, this discrepancy is also related to the ocular surface. In a multicenter study involving thousands of ocular surface images in identifying hepatobiliary diseases 3 , the accuracy of identification was only 74% (71%–76%), which is almost similar to our results. Although the HMT-Net model has low specificity, its sensitivity to diabetes detection is high.…”

“…Because of its high learning ability and computing ability, deep learning can even extract hidden information that doctors cannot perceive from the data. For ophthalmic image analysis, deep learning can detect several systemic diseases through eye images, such as anemia 1 , cardiovascular disease 2 , hepatobiliary disease 3 , traumatic brain injury 4 etc. This is because the visual blood vessels and nerves in the eyes are closely related to the health of the whole body.…”

Beyond the Liver: Liver-Eye Communication in Clinical and Experimental Aspects

“…Therefore, ophthalmological examination is helpful to screen some specific hepatobiliary diseases. In view of this, the AI algorithm for screening liver diseases using ocular models has also been applied and reported for the first time (Xiao et al, 2021). In this AI algorithm, the DL system utilized patient information, including slit-lamp anterior segment photos, fundus photos and diagnostic data of hepatic diseases, to train and adjust to form a reliable DL model.…”

A systematic review on application of deep learning in digestive system image processing

“…Xiao led a prospective multi-center study using a slit lamp to conduct DL on the fundus and iris of patients with several common liver diseases and finally achieved excellent results in identifying liver cancer and chronic cirrhosis. In future, ophthalmology imaging may be used as a tool for the early screening of liver and biliary diseases [ 108 ]. This project is innovative because linking two seemingly distant organs together, allows this kind of interdisciplinary computer-aided research to discover biological phenomena that have not been discovered before.…”