Related Experiment Video
Updated: Jul 8, 2026
Using Retinal Imaging to Study Dementia
Published on: November 6, 2017
Retinal imaging and AI for non-invasive detection of diabetic macrovascular complications
1Department of Information Technology, Thiagarajar College of Engineering, Madurai, Tamil Nadu, India.
Insights
An AI framework using retinal images can detect Peripheral Arterial and Aortic Diseases (PAAD) in diabetic patients. This non-invasive tool offers high accuracy for early diagnosis, potentially preventing serious cardiovascular complications.
Area of Science:
- Ophthalmology
- Cardiology
- Artificial Intelligence
Background:
- Diabetes mellitus accelerates vascular degeneration, increasing risks of Peripheral Arterial and Aortic Diseases (PAAD).
- PAAD is linked to adverse cardiovascular outcomes but often goes undiagnosed in diabetic populations due to asymptomatic progression.
- Limited access to early screening hinders timely intervention for PAAD in diabetic patients.
Purpose of the Study:
- To develop and validate a non-invasive AI screening framework for early PAAD detection.
- Utilize retinal fundus imaging and microvascular features as systemic biomarkers for PAAD.
- Exploit artificial intelligence to identify PAAD risk through retinal analysis.
Main Methods:
- A hybrid AI pipeline integrated simulated OCT-like structural features and handcrafted vascular biomarkers.
- Employed an attention-enhanced VGG16 backbone with Convolutional Block Attention Modules (CBAM) and Multiple Instance Learning (MIL).
- Utilized multi-level feature fusion and interpretable AI techniques (SHAP, Grad-CAM) on public and clinical fundus datasets.
Main Results:
- The AI framework achieved 94.6% accuracy, 90.5% sensitivity, 96.2% specificity, and an AUC-ROC of 0.973.
- Retinal thickness and texture entropy were identified as dominant predictors by SHAP.
- Grad-CAM highlighted vessel bifurcations and arteriolar narrowing, consistent with PAAD pathophysiology, with a 150ms inference time.
Conclusions:
- An interpretable AI system demonstrates high diagnostic performance for PAAD detection via retinal imaging.
- This non-invasive, cost-effective approach offers a scalable alternative to conventional vascular assessments.
- The AI framework may facilitate earlier diagnosis and improved prevention of cardiovascular and cerebrovascular complications.
Background:
Diabetes mellitus accelerates vascular degeneration and increases the risk of major macrovascular complications, including Peripheral Arterial Disease (PAD) and aortic pathologies, collectively termed Peripheral Arterial and Aortic Diseases (PAAD). These conditions are strongly associated with adverse cardiovascular outcomes but often remain underdiagnosed in diabetic populations due to asymptomatic progression and limited access to early screening.
Objective:
This study aims to develop and validate a non-invasive, artificial intelligence (AI)-based screening framework using retinal fundus imaging for early detection of PAAD by exploiting retinal microvascular features as systemic biomarkers.
Methods:
A hybrid diagnostic pipeline integrated simulated Optical Coherence Tomography (OCT)-like structural features (retinal thickness, texture entropy, vessel density factor, and layer separation index), handcrafted vascular biomarkers, and an attention-enhanced VGG16 backbone with Convolutional Block Attention Modules (VGG16 + CBAM). Multiple Instance Learning (MIL) improved lesion-level discrimination in weakly labeled datasets. Multi-level feature fusion aggregated spatial, physiological, and morphological descriptors. High-resolution fundus datasets from public and clinical cohorts were used for training and validation. Model interpretability was ensured using SHapley Additive exPlanations (SHAP) and Gradient-weighted Class Activation Mapping (Grad-CAM).
Results:
The framework achieved an accuracy of 94.6%, sensitivity of 90.5%, specificity of 96.2%, and an AUC-ROC of 0.973 on an independent test set. SHAP identified retinal thickness and texture entropy as dominant predictors, while Grad-CAM highlighted vessel bifurcations and arteriolar narrowing, consistent with PAAD pathophysiology. The average inference time was 150 ms per image on GPU, enabling real-time use.
Conclusion:
This interpretable AI-based system demonstrates high diagnostic performance for PAAD detection using retinal imaging. It offers a non-invasive, cost-effective, and scalable alternative to conventional vascular assessments and may support earlier diagnosis and improved prevention of cardiovascular and cerebrovascular complications.
