Online decoupled estimation of temperature field in biological tissue during invasive laser thermal therapy

Invasive laser-induced thermal therapy (LITT) is an emerging minimally invasive technique for treating deep-seated tumors. Online estimation of the transient temperature field in target biological tissue is crucial for the research and development of LITT technology. Existing state estimation theories, which employ the state-space model to describe the system, essentially belong to coupled estimation methods. They cannot achieve decoupled estimation of system states, leading to numerous challenges in applying current state estimation methods to temperature field estimation during LITT procedures. This paper focuses on the online decoupled estimation of the temperature field in biological tissue during invasive LITT. By constructing an optimal estimation model and a spatiotemporal correlation model, it achieves online decoupled estimation of temperatures at critical discrete points throughout the invasive LITT process. This scheme effectively overcomes the difficulties inherent in existing temperature field estimation techniques based on state-space models. The reliability of the constructed biological model is validated using existing experimental data. Through numerical simulations, the impacts of the influence time domain, measurement noise, and model mismatch are analyzed. Comparison with the standard Kalman filter demonstrates that when only real-time temperature estimation of specific regions (such as tumor regions) is required, the decoupled estimation scheme proposed in this paper exhibits significant advantages in computational efficiency. This research provides reliable technical support for the online estimation of tissue temperature fields.