AUTHOR=Huang Xiaoqin , Sun Jian , Gupta Krati , Montesano Giovanni , Crabb David P. , Garway-Heath David F. , Brusini Paolo , Lanzetta Paolo , Oddone Francesco , Turpin Andrew , McKendrick Allison M. , Johnson Chris A. , Yousefi Siamak 

TITLE=Detecting glaucoma from multi-modal data using probabilistic deep learning

JOURNAL=Frontiers in Medicine

VOLUME=Volume 9 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/medicine/articles/10.3389/fmed.2022.923096

DOI=10.3389/fmed.2022.923096

ISSN=2296-858X

ABSTRACT=Objective: To assess the accuracy of probabilistic deep learning models to discriminate normal eyes and eyes with glaucoma from fundus photographs and visual fields.
Design: Algorithm development for discriminating normal and glaucoma eyes using data from multicenter, cross-sectional, case-control study. 
Subjects and Participants: Fundus photograph and visual field data from 1655 eyes of 929 normal and glaucoma subjects to develop and test deep learning models and an independent group of 196 eyes of 98 normal and glaucoma patients to validate deep learning models. 
Main Outcome Measures: Accuracy and area under the receiver-operating characteristic curve (AUC). 
Methods Fundus photographs and OCT images were carefully examined by clinicians to identify glaucomatous optic neuropathy (GON). When GON was detected by the reader, the finding was further evaluated by another clinician. Three probabilistic deep convolutional neural network (CNN) models were developed using 1655 fundus photographs, 1655 visual fields, and 1655 pairs of fundus photographs and visual fields collected from Compass instruments. Deep learning models were trained and tested using 80% of fundus photographs and visual fields for training set and 20% of the data for testing set. Models were further validated using an independent validation subset. 
Results: The AUC of the deep learning model in detecting glaucoma from fundus photographs, visual fields, and combined modalities using development and testing subset were 0.90 (95% confidence interval: 0.89-0.92), 0.89 (0.88-0.91), and 0.94 (0.92-0.96), respectively. The AUC of the deep learning model in detecting glaucoma from fundus photographs, visual fields, and both modalities using the independent validation subset were 0.94 (0.92-0.95), 0.98 (0.98-0.99), and 0.98 (0.98-0.99), respectively. The uncertainty level of the correctly classified eyes is much lower in the combined model compared to the model based on visual fields only. 
Conclusions and Relevance: Probabilistic deep learning models can detect glaucoma development from multi-modal data with higher accuracy and can provide higher confidence level on the generated outcome. Our findings suggest that models based on combined visual field and fundus photograph modalities detects glaucoma with higher accuracy. Models based on fundus photographs only also showed a higher detection rate compared to models based on visual fields only.