Visible-light optical coherence tomography oximetry based on circumpapillary scan and graph-search segmentation

Area of Science:

• Ophthalmology
• Biomedical Optics
• Medical Imaging

Background:

• Visible-light optical coherence tomography (vis-OCT) allows for retinal oximetry, measuring hemoglobin oxygen saturation (sO2) in individual retinal blood vessels.
• Accurate sO2 calculation depends on precise estimation of backscattered light spectra from the posterior vessel wall.
• Subject motion and image noise complicate spectral estimation, leading to inaccurate sO2 measurements.

Purpose of the Study:

• To develop and validate a novel algorithm for reliable spectral extraction of backscattered light in vis-OCT.
• To improve the accuracy and reliability of retinal oximetry (sO2) measurements.
• To assess the impact of repeated measurements on sO2 accuracy.

Main Methods:

• Utilized circumpapillary scanning for repeated sampling of retinal vessels at identical locations.
• Employed a combined cross-correlation and graph-search based segmentation approach to identify posterior vessel wall locations.
• Validated the algorithm against 100 B-scans, considered a gold standard.

Main Results:

• The developed algorithm demonstrated highly accurate sO2 measurements with minimal bias.
• The study quantified the effect of the number of repeated measurements on the precision of sO2 estimation.
• Successful extraction of backscattered light spectra was achieved despite motion and noise.

Conclusions:

• The novel algorithm significantly enhances the reliability of retinal oximetry using vis-OCT.
• This method provides a foundation for future large-scale investigations into retinal oxygenation in both animal models and human subjects.
• Improved spectral estimation paves the way for more robust clinical and research applications of vis-OCT.