Monitoring intracellular cavitation during selective targeting of the retinal pigment epithelium

Abstract

PURPOSE Selective targeting of the Retinal Pigment Epithelium (RPE), by either applying trains of microsecond laser pulses or, in our approach, by repetitively scanning a tightly focused spot across the retina, achieves destruction of RPE cells while avoiding damage to the overlying photoreceptors. Both techniques have been demonstrated as attractive methods for the treatment of retinal diseases that are caused by a dysfunction of the RPE. Because the lesions are ophthalmoscopically invisible, an online control system that monitors cell death during irradiation is essential to ensure efficient and selective treatment in a clinical application. MATERIALS AND METHODS Bubble formation inside the RPE cells has been shown to be the cell damage mechanism for nano- and picosecond pulses. We built an optical system to investigate whether the same mechanism extends into the microsecond regime. The system detects changes in backscattered light of the irradiating beam during exposure. Samples of young bovine eyes were exposed in vitro using single pulses ranging from 3 μs to 50 μs. Using the viability assay calcein-AM the ED50 threshold for cell death was determined and compared to the threshold for bubble formation. We also set up a detection system on our slit lamp adapted scanning system in order to determine the feasibility of monitoring threshold RPE damage during selective laser treatment in vivo. RESULTS AND DISCUSSION Intracellular cavitation was detected as a transient increase in backscattering signal, either of an external probe beam or of the irradiation beam itself. Monitoring with the irradiation beam is both simpler and more sensitive. We found the threshold for bubble formation to coincide with the threshold for cell damage for pulse durations up to 20 μs, suggesting that cavitation is the main mechanism of cell damage. For pulse widths longer than 20 μs, the cell damage mechanism appears to be increasingly thermal as the two thresholds diverge. We conclude that bubble detection can be used to monitor therapeutic endpoint for pulse durations up to 20 μs (or equivalent dwell time in a scanning approach). We have integrated a detection module into our slit lamp adapted laser scanner in order to determine threshold RPE damage during selective laser treatment in vivo