Automated quantification of Haller's layer in choroid using swept-source optical coherence tomography

Area of Science:

• Ophthalmology
• Medical Imaging
• Biomedical Engineering

Background:

• Choroidal structural changes are linked to various eye diseases.
• Existing methods for choroidal analysis, like thickness and vascularity, lack specificity for sublayers.
• Haller's layer, a sublayer of the choroid, is of significant interest due to its large blood vessels, but manual quantification is challenging.

Purpose of the Study:

• To develop an automated algorithm for precise quantification of Haller's layer in the choroid.
• To enable detailed analysis of choroidal sublayers, particularly Haller's layer, using swept-source optical coherence tomography (OCT).
• To overcome the limitations of manual quantification, which is time-consuming and prone to errors.

Main Methods:

• Developed an algorithm to detect the boundary between Haller's and Sattler's layers.
• Employed exponentiation-based binarization to estimate vessel cross-sections and identify large vessels.
• Validated the algorithm on 50 OCT B-scans from healthy eyes, comparing results with intra-observer variability using Dice coefficient (DC), correlation coefficient (CC), and absolute difference (AD).

Main Results:

• The algorithm achieved a mean Dice coefficient (DC) of 89.48%, closely matching intra-observer repeatability (89.12%).
• Mean absolute difference (AD) was 17.54 μm and mean correlation coefficient (CC) was 98.10%, also approximating intra-observer values.
• The quantitative results demonstrate high agreement between algorithmic and manual measurements.

Conclusions:

• The developed algorithm provides accurate and reliable quantification of Haller's layer.
• The high correlation with manual delineations suggests its suitability for clinical analysis of choroidal details.
• This automated approach can significantly aid in the finer analysis of choroidal structures, especially in diseased eyes.