DiaNet v2 deep learning based method for diabetes diagnosis using retinal images

Abstract

Diabetes mellitus (DM) is a prevalent chronic metabolic disorder linked to increased morbidity and mortality. With a significant portion of cases remaining undiagnosed, particularly in the Middle East North Africa (MENA) region, more accurate and accessible diagnostic methods are essential. Current diagnostic tests like fasting plasma glucose (FPG), oral glucose tolerance tests (OGTT), random plasma glucose (RPG), and hemoglobin A1c (HbA1c) have limitations, leading to misclassifications and discomfort for patients. The aim of this study is to enhance diabetes diagnosis accuracy by developing an improved predictive model using retinal images from the Qatari population, addressing the limitations of current diagnostic methods. This study explores an alternative approach involving retinal images, building upon the DiaNet model, the first deep learning model for diabetes detection based solely on retinal images. The newly proposed DiaNet v2 model is developed using a large dataset from Qatar Biobank (QBB) and Hamad Medical Corporation (HMC) covering wide range of pathologies in the the retinal images. Utilizing the most extensive collection of retinal images from the 5545 participants (2540 diabetic patients and 3005 control), DiaNet v2 is developed for diabetes diagnosis. DiaNet v2 achieves an impressive accuracy of over 92%, 93% sensitivity, and 91% specificity in distinguishing diabetic patients from the control group. Given the high prevalence of diabetes and the limitations of existing diagnostic methods in clinical setup, this study proposes an innovative solution. By leveraging a comprehensive retinal image dataset and applying advanced deep learning techniques, DiaNet v2 demonstrates a remarkable accuracy in diabetes diagnosis. This approach has the potential to revolutionize diabetes detection, providing a more accessible, non-invasive and accurate method for early intervention and treatment planning, particularly in regions with high diabetes rates like MENA.

Figure 1

ROC plot for the modified models (a) and backbone models (b).

Figure 2

Performance of different deep learning models based on gender-stratified groups.

Figure 3

Performance of different deep learning models based on age-stratified groups.

Figure 4

Retinal images with overlaid heatmap. Images a the top (a) to (d) show examples of diabetic images. Images at the bottom (e) to (h) show examples of control (non-diabetic) image. Images on the left are the original input images while those on the right are the corresponding class activation map (CAM).

Figure 5

Summary statistics of images used in the study from HMC and QBB.

Figure 6

Sample of retinal images used in this study. Presented pathologies include: hemorrhage (a), microaneurysm (b), mild NPDR (c), glaucoma (d), laser caused trauma (e) and retinal detachment (f).

Figure 7

A diagram showing the overall flow of the deep learning model to diagnose diabetes. Up: all images are cropped and then processed by subtracting the local mean from a 4 $\times$ 4 neighboring pixels. Down: the architecture of the deep learning model with VGG-11 as the backbone of the model.