Motion-invariant SRT treatment detection from direct M-scan OCT imaging

Area of Science:

• Ophthalmology
• Biomedical Engineering
• Medical Imaging

Background:

• Selective Retina Therapy (SRT) targets the retinal pigment epithelium (RPE) for precise treatment.
• Current methods lack real-time monitoring of SRT efficacy due to RPE invisibility.
• Assessing SRT outcomes non-invasively is crucial for treatment optimization.

Purpose of the Study:

• To develop and evaluate an online, non-invasive method for assessing Selective Retina Therapy (SRT) effectiveness.
• To utilize optical coherence tomography (OCT) imaging coupled with SRT for real-time treatment evaluation.
• To identify key OCT scan features for predicting SRT treatment success.

Main Methods:

• Analysis of time-resolved OCT scans acquired during SRT procedures.
• Application of a convolutional neural network (CNN) to correlate OCT scans with treatment outcomes.
• Focusing on a small segment of the OCT scan at the end of treatment for efficient analysis.

Main Results:

• High accuracy in predicting adequate SRT energy levels using immediate post-treatment OCT signals.
• Demonstrated robustness of the method despite potential eye movements.
• Successful correlation of OCT scan characteristics with microscopic evidence of SRT efficacy.

Conclusions:

• A strategy for online, non-invasive SRT assessment has been proposed.
• The method provides a reliable evaluation of treatment status in real-time.
• This approach can inform treatment continuation planning for improved patient outcomes.