Artificial Intelligence-Assisted Perfusion Density as Biomarker for Screening Diabetic Nephropathy

Area of Science:

• Ophthalmology
• Nephrology
• Artificial Intelligence
• Medical Imaging

Background:

• Diabetic nephropathy (DN) is a major complication of type 2 diabetes mellitus (T2DM).
• Early detection of DN is crucial for effective management and prevention of kidney damage.
• Current screening methods may be invasive or lack sensitivity for early-stage disease.

Purpose of the Study:

• To identify a reliable biomarker for screening diabetic nephropathy (DN) using artificial intelligence (AI)-assisted ultra-widefield swept-source optical coherence tomography angiography (UWF SS-OCTA).
• To evaluate the correlation between retinal microvascular changes and DN in patients with T2DM.

Main Methods:

• Analysis of data from 169 patients (287 eyes) with T2DM, including demographic, clinical, and UWF SS-OCTA imaging data.
• Application of statistical analysis, 10-fold cross-validation, and a random forest approach for data processing.
• Quantitative assessment of retinal microvasculature, specifically perfusion density (PD).

Main Results:

• Significant retinal microvascular damage was observed in diabetic retinopathy (DR) patients with DN compared to those without DN.
• Strong associations were found between reduced perfusion density (PD) and DN diagnosis in both T2DM and DR populations (P < 0.001).
• The random forest model achieved high classification accuracy (85.84% for T2DM, 82.57% for DR) in identifying DN patients based on PD.

Conclusions:

• Quantitative microvascular analysis using UWF SS-OCTA correlates with DN presence.
• Ultra-widefield perfusion density (UWF PD) shows potential as a significant, noninvasive biomarker for DN evaluation via deep learning.
• AI-assisted UWF SS-OCTA offers a rapid and reliable tool for DN screening.