Selective retina therapy (SRT) targets the retinal pigment epithelium (RPE) with pulsed laser irradiation by inducing microbubble formation (MBF) at the intracellular melanin granula, which leads to selective cell disruption. The following wound healing process rejuvenates the chorio-retinal junction. Pulse energy thresholds for selective RPE effects vary intra- and interindividually. We present the evaluation of an algorithm that processes backscattered treatment light to detect MBF as an indicator of RPE cell damage since these RPE lesions are invisible during treatment. Eleven patients with central serous chorioretinopathy and four with diabetic macula edema were treated with a SRT system, which uses a wavelength of 527 nm, a repetition rate of 100 Hz, and a pulse duration of 1.7 μs. Fifteen laser pulses with stepwise increasing pulse energy were applied per treatment spot. Overall, 4626 pulses were used for algorithm parameter optimization and testing. Sensitivity and specificity were the metrics maximized through an automatic optimization process. Data were verified by fluorescein angiography. A sensitivity of 1 and a specificity of 0.93 were achieved. The method introduced in this paper can be used for guidance or automatization of microbubble-related treatments like SRT or selective laser trabeculoplasty.
Selective retina therapy (SRT) is an ophthalmological laser technique, targeting the retinal pigment epithelium (RPE) with repetitive microsecond laser pulses, while causing no thermal damage to the neural retina, the photoreceptors as well as the choroid. The RPE cells get damaged mechanically by microbubbles originating, at the intracellular melanosomes. Beneficial effects of SRT on Central Serous Retinopathy (CSR) and Diabetic Macula Edema (DME) have already been shown. Variations in the transmission of the anterior eye media and pigmentation variation of RPE yield in intra- and inter- individual thresholds of the pulse energy required for selective RPE damage. Those selective RPE lesions are not visible. Thus, dosimetry-systems, designed to detect microbubbles as an indicator for RPE cell damage, are demanded elements to facilitate SRT application. Therefore, a technique based on the evaluation of backscattered treatment light has been developed. Data of 127 spots, acquired during 10 clinical treatments of CSR patients, were assigned to a RPE cell damage class, validated by fluorescence angiography (FLA). An algorithm has been designed to match the FLA based information. A sensitivity of 0.9 with a specificity close to 1 is achieved. The data can be processed within microseconds. Thus, the process can be implemented in existing SRT lasers with an automatic pulse wise increasing energy and an automatic irradiation ceasing ability to enable automated treatment close above threshold to prevent adverse effects caused by too high pulse energy. Alternatively, a guidance procedure, informing the treating clinician about the adequacy of the actual settings, is possible.
Photocoagulation is a treatment modality for several retinal diseases. Intra- and inter-individual variations
of the retinal absorption as well as ocular transmission and light scattering makes it impossible to achieve
a uniform effective exposure with one set of laser parameters. To guarantee a uniform damage throughout
the therapy a real-time control is highly requested. Here, an approach to realize a real-time optical feedback
using dynamic speckle analysis in-vivo is presented. A 532 nm continuous wave Nd:YAG laser is used for
coagulation. During coagulation, speckle dynamics are monitored by a coherent object illumination using a
633 nm diode laser and analyzed by a CMOS camera with a frame rate up to 1 kHz. An algorithm is presented
that can discriminate between different categories of retinal pigment epithelial damage ex-vivo in enucleated
porcine eyes and that seems to be robust to noise in-vivo. Tissue changes in rabbits during retinal coagulation
could be observed for different lesion strengths. This algorithm can run on a FPGA and is able to calculate a
feedback value which is correlated to the thermal and coagulation induced tissue motion and thus the achieved
Photocoagulation is a laser treatment widely used for the therapy of several retinal diseases. Intra- and inter-individual
variations of the ocular transmission, light scattering and the retinal absorption makes it impossible
to achieve a uniform effective exposure and hence a uniform damage throughout the therapy. A real-time
monitoring and control of the induced damage is highly requested. Here, an approach to realize a real time
optical feedback using dynamic speckle analysis is presented. A 532 nm continuous wave Nd:YAG laser is
used for coagulation. During coagulation, speckle dynamics are monitored by a coherent object illumination
using a 633nm HeNe laser and analyzed by a CMOS camera with a frame rate up to 1 kHz. It is obvious that
a control system needs to determine whether the desired damage is achieved to shut down the system in a
fraction of the exposure time. Here we use a fast and simple adaption of the generalized difference algorithm
to analyze the speckle movements. This algorithm runs on a FPGA and is able to calculate a feedback value
which is correlated to the thermal and coagulation induced tissue motion and thus the achieved damage. For
different spot sizes (50-200 μm) and different exposure times (50-500 ms) the algorithm shows the ability to
discriminate between different categories of retinal pigment epithelial damage ex-vivo in enucleated porcine
eyes. Furthermore in-vivo experiments in rabbits show the ability of the system to determine tissue changes in
living tissue during coagulation.
Laser coagulation of the retina is an established treatment for several retinal diseases. The absorbed laser energy and thus the induced thermal damage varies with the transmittance and scattering properties of the anterior eye media and with the pigmentation of the fundus. The temperature plays the most important role in the coagulation process. An established approach to measure a mean retinal temperature rise is optoacoustics, however it provides limited information on the coagulation. Phase sensitive OCT potentially offers a three dimensional temporally resolved temperature distribution but is very sensitive to slightest movements which are clinically hard to avoid. We develop an optical technique able to monitor and quantify thermally and coagulation induced tissue movements (expansions and contractions) and changes in the tissue structure by dynamic laser speckle analysis (LSA) offering a 2D map of the affected area. A frequency doubled Nd:YAG laser (532nm) is used for photocoagulation. Enucleated porcine eyes are used as targets. The spot is 100μm. A Helium Neon laser (HeNe) is used for illumination. The backscattered light of a HeNe is captured with a camera and the speckle pattern is analyzed. A Q-switched Nd:YLF laser is used for simultaneous temperature measurements with the optoacoustic approach. Radial tissue movements in the micrometer regime have been observed. The signals evaluation by optical flow algorithms and generalized differences tuned out to be able to distinguish between regions with and without immediate cell damage. Both approaches have shown a sensitivity of 93% and a specificity above 99% at their optimal threshold.
Selective Retina Therapy (SRT) targets the Retinal Pigment Epithelium (RPE) without effecting neighboring layers as the photoreceptors or the choroid. SRT related RPE defects are ophthalmoscopically invisible. Owing to this invisibility and the variation of the threshold radiant exposure for RPE damage the treating physician does not know whether the treatment was successful or not. Thus measurement techniques enabling a correct dosing are a demanded element in SRT devices. The acquired signal can be used for monitoring or automatic irradiation control. Existing monitoring techniques are based on the detection of micro-bubbles. These bubbles are the origin of RPE cell damage for pulse durations in the ns and μs time regime 5μs. The detection can be performed by optical or acoustical approaches. Monitoring based on an acoustical approach has already been used to study the beneficial effects of SRT on diabetic macula edema and central serous retinopathy. We have developed a first real time feedback technique able to detect micro-bubble induced characteristics in the backscattered laser light fast enough to cease the laser irradiation within a burst. Therefore the laser energy within a burst of at most 30 pulses is increased linearly with every pulse. The laser irradiation is ceased as soon as micro-bubbles are detected. With this automatic approach it was possible to observe invisible lesions, an intact photoreceptor layer and a reconstruction of the RPE within one week.