OCT Hyperreflective Retinal Foci in Diabetic Retinopathy: A Semi-Automatic Detection Comparative Study

Area of Science:

• Ophthalmology
• Medical Imaging
• Cell Biology

Background:

• Optical coherence tomography (OCT) identifies novel morphological entities in retinal layers as potential clinical biomarkers for retinal diseases.
• Hyperreflective foci (HRF), small (<30 µm) retinal lesions, are recognized as in vivo biomarkers of intraretinal inflammation, potentially representing activated microglial cells.
• Diabetic retinopathy (DR) is a key model for studying intraretinal inflammation using HRF, but manual counting is time-consuming and requires expertise.

Purpose of the Study:

• To develop and validate a semi-automatic method for detecting and quantifying HRF in OCT scans of patients with DR.
• To overcome the limitations of manual HRF counting, enabling reliable and efficient clinical application.

Main Methods:

• High-resolution spectral domain OCT scans from DR patients were analyzed.
• HRF were initially counted manually by trained examiners.
• Semi-automatic HRF counting was performed using ImageJ software with four distinct setting profiles.

Main Results:

• Statistical analysis demonstrated an excellent intraclass correlation coefficient (ICC) between manual and all four semi-automatic counting methods.
• The second setting profile yielded a bias of -0.2 foci and limits of agreement of ±16.3 foci on Bland-Altman analysis.
• The developed method provides a reliable and clinically applicable approach for HRF quantification.

Conclusions:

• The validated semi-automatic method facilitates accurate and efficient quantification of HRF in OCT scans.
• This advancement supports better assessment of intraretinal inflammation in DR and other degenerative retinal diseases.
• Improved HRF measurement can lead to a deeper understanding of inflammation's role, progression, and regression in diabetic retinopathy.