Towards temperature controlled retinal laser treatment with a single laser at 10 kHz repetition rate

Abstract

Laser photocoagulation is one of the most frequently used treatment approaches in ophthalmology for a variety of retinal diseases. Depending on indication, treatment intensity varies from application of specific micro injuries down to gentle temperature increases without inducing cell damage. Especially for the latter, proper energy dosing is still a challenging issue, which mostly relies on the physician’s experience. Pulsed laser photoacoustic temperature measurement has already proven its ability for automated irradiation control during laser treatment but suffers from a comparatively high instrumental effort due to combination with a conventional continuous wave treatment laser. In this paper, a simplified setup with a single pulsed laser at 10 kHz repetition rate is presented. The setup combines the instrumentation for treatment as well as temperature measurement and control in a single device. In order to compare the solely pulsed heating with continuous wave (cw) tissue heating, pulse energies of 4 µJ were applied with a repetition rate of 1 kHz to probe the temperature rise, respectively. With the same average laser power of 60 mW an almost identical temporal temperature course was retrieved in both irradiation modes as expected. The ability to reach and maintain a chosen aim temperature of 41 °C is demonstrated by means of model predictive control (MPC) and extended Kalman filtering at a the measurement rate of 250 Hz with an accuracy of less than ±0.1 °C. A major advantage of optimization-based control techniques like MPC is their capability of rigorously ensuring constraints, e.g., temperature limits, and thus, realizing a more reliable and secure temperature control during retinal laser irradiation.

State and parameter estimation for retinal laser treatment

“…The results of this work hence pave the way for successful application of model-based control approaches, in particular model predictive control. A first proof-of-concept in this direction was recently shown by us in [14], [15], where we use MPC in combination with EKF.…”

Parameter estimation and model reduction for retinal laser treatment

“…The RPE is the major absorber in the eye. For more details on the experimental setup and measurement routines, we refer to [18]. The volume temperature that can be obtained as a measurement from the setup depicted in Figure 1 can only be used as a control variable to a limited extent, since the peak temperature within the tissue affects the treatment outcome significantly.…”

“…Further, we presented first results regarding observers for the reduced models states and the absorption coefficient, such as an extended Kalman filter and a moving horizon estimator in [16]. In [18], we proposed a Model Predictive Control (MPC) scheme to ensure a safe and effective treatment. As a core feature, MPC enables us to directly formulate bounds on the peak temperature in the underlying optimization problem that is used to evaluate the feedback controller.…”

Model predictive control for retinal laser treatment at 1 kHz

“…Again, the resulting pressure transients are measured and the process is repeated. For further details considering the experimental setup, we refer to our previous work [17].…”

“…In [17], we recently presented preliminary MPC closed-loop results at 250 Hz in experiments, using a reduced order model based on Taylor-approximation of the parametric dependency as suggested earlier in [12]. Regarding the MPC stage-cost design, we considered quadratic penalization of the control size of the form |u| 2 .…”