TL;DR
This paper introduces a novel method for monitoring Selective Retina Therapy (SRT) using speckle variance optical coherence tomography (svOCT). By analyzing changes in speckle patterns, the system can reliably correlate svOCT signal intensity with the degree of retinal lesion formation, enabling precise control over laser energy for safer and more effective treatments.
What Problem Does This Paper Address?
Selective Retina Therapy (SRT) is a crucial treatment for various retinal diseases, targeting the retinal pigment epithelium (RPE) to minimize damage to surrounding tissues. However, a significant challenge in SRT is precisely controlling the laser energy. The RPE lesions are often invisible ophthalmoscopically, and variations in melanin concentration among patients and even within the same eye make it difficult to establish a universal energy threshold for effective treatment. Existing monitoring methods, like fundus fluorescence angiography, are not real-time, and other approaches lack visual feedback. This paper aims to address this gap by providing a real-time, non-invasive monitoring system that can directly assess the impact of laser pulses on retinal tissue.
Methods
The researchers developed an integrated system combining a swept-source OCT (ssOCT) imaging system with a microsecond pulsed laser used for SRT. The core of their monitoring technique is speckle variance OCT (svOCT). svOCT works by calculating the interframe intensity variance of a sequence of OCT images. This variance is sensitive to changes in the scattering properties of the tissue, which are induced by the thermal effects of the laser pulse.
Specifically, they used M-scans (time-resolved sequences of A-scans) to capture rapid changes. The svOCT images were then calculated from these sequences. They identified regions of interest (ROIs) in the photoreceptor and RPE layers, averaging the svOCT values along the axial direction. The peak values of these averaged svOCT signals were analyzed in relation to the laser pulse energy. To validate their findings, they also performed numerical simulations using COMSOL Multiphysics to model the spatial and temporal temperature distributions in the retina and correlated these with the svOCT measurements and observed tissue damage.
Datasets
The study utilized ex vivo bovine eyes obtained from a local butcher. These eyes were used to acquire OCT images of treated retinal spots. The researchers created 39 treated spots across two eyes. Microscopic images of these treated spots were also captured before and after removing the neural retinal layer to assess the extent of damage.
Key Insights and Ideas
- Speckle Variance as a Biomarker: The study demonstrates that speckle variance in OCT images is a sensitive indicator of thermal-induced changes in retinal tissue caused by laser pulses.
- Direct Correlation with Lesion Formation: A key finding is the reliable correlation between the peak svOCT values and the degree of retinal lesion formation, suggesting svOCT can directly reflect the therapeutic effect.
- Energy Proportionality: The peak svOCT values were found to be proportional to the laser pulse energy, providing a quantitative measure of the laser's impact.
- Temperature Estimation: By correlating svOCT peak values with simulated retinal temperatures, the researchers could estimate the temperature reached in the RPE and neural retina, which aligned with known thresholds for tissue damage.
- Potential for Real-time Dosimetry: The method shows promise for real-time dosimetry control, allowing for automatic adjustment of laser energy during treatment to achieve desired outcomes.
Key Results
- svOCT images showed abrupt speckle variance changes immediately following laser pulse irradiation.
- The peak svOCT values in the RPE and photoreceptor layers increased with increasing laser pulse energy, becoming distinct even at low energies (as low as 54 μJ).
- A reliable correlation was established between the average svOCT peak values and the extent of retinal lesion formation. For successful RPE-only lesions, average svOCT peak values between 1.88 and 15.3 showed a lower type I error (20%) and type II error (0.083%) compared to using energy level thresholds alone.
- Simulated temperatures in the RPE reached over 150°C for higher energy levels, consistent with microvaporization, and were well-correlated with svOCT peak values.
- The study suggests that using a power-ramped pulsed mode with svOCT feedback could enable automatic dosimetry control by stopping irradiation when a predetermined svOCT threshold is reached.
Limitations and Future Work
The authors acknowledge that their study using ex vivo bovine eyes might have missed minor tissue damages when assessing photocoagulation and RPE cell damage from microscopic images alone. They suggest that further studies incorporating histology or fluorescence microscopy would be beneficial for a more accurate correlation between svOCT signals and retinal damage.
The most significant future direction proposed is to conduct in vivo studies using animal models. This would be a crucial step in validating the utility of this svOCT-based method for automatic dosimetry control in clinical SRT systems and assessing its performance in a living organism.
Summary updated