Abstract
In this paper we analyse the performance of machine learning methods in predicting patient information such as age or sex solely from retinal imaging modalities in a heterogeneous clinical population. Our dataset consists of N = 135,667 fundus images and N = 85,536 volumetric OCT scans. Deep learning models were trained to predict the patient’s age and sex from fundus images, OCT cross sections and OCT volumes. For sex prediction, a ROC AUC of 0.80 was achieved for fundus images, 0.84 for OCT cross sections and 0.90 for OCT volumes. Age prediction mean absolute errors of 6.328 years for fundus, 5.625 years for OCT cross sections and 4.541 for OCT volumes were observed. We assess the performance of OCT scans containing different biomarkers and note a peak performance of AUC = 0.88 for OCT cross sections and 0.95 for volumes when there is no pathology on scans. Performance drops in case of drusen, fibrovascular pigment epitheliuum detachment and geographic atrophy present. We conclude that deep learning based methods are capable of classifying the patient’s sex and age from color fundus photography and OCT for a broad spectrum of patients irrespective of underlying disease or image quality. Non-random sex prediction using fundus images seems only possible if the eye fovea and optic disc are visible.
Highlights
In this paper we analyse the performance of machine learning methods in predicting patient information such as age or sex solely from retinal imaging modalities in a heterogeneous clinical population
Given the large spectrum of diseases typically found in a clinical setting, we investigate if machine learning based models are capable of extracting the gender and age from a patient’s fundus image or optical coherence tomography (OCT) image data regardless of any ocular disease or severity
We here demonstrate that deep learning classifiers can predict gender and age based on fundus images, and based on OCT volume- and on individual Bscans for a broad spectrum of patients and pathologies
Summary
In this paper we analyse the performance of machine learning methods in predicting patient information such as age or sex solely from retinal imaging modalities in a heterogeneous clinical population. Deep learning models were trained to predict the patient’s age and sex from fundus images, OCT cross sections and OCT volumes. A large focus of previous work has been in performing tasks which may augment or substitute human experts These tasks typically require transferring knowledge from domain experts, in the form of annotations and their respective images, to a machine learning process, whereby implying that a human is capable of doing the task. Given the large spectrum of diseases typically found in a clinical setting, we investigate if machine learning based models are capable of extracting the gender and age from a patient’s fundus image or OCT image data regardless of any ocular disease or severity
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
