Temperature Measurement by Statistical Parameters of Ultrasound Signal Backscattered from Tissue Samples

A novel estimation of temperature changes inside soft tissues has been proposed in sub-ablation range, i.e. ≈ 20°C-48°C. This estimation has been obtained by studying statistical properties of backscattered ultrasonic signals. Two different procedures of heating/cooling have been performed in which the RF echoes have been registered from soft tissue phantom in the first procedure, and from soft tissue in vitro in the second one. Calculated envelopes of signals registered in time points during heating/cooling experiments have been treated as a statistical sample drawn from a random variable with three different distributions, namely the Rayleigh distribution, the Nakagami distribution, and the K-distribution. The histograms obtained in subsequent time moments have been fitted to the three distributions. Dependencies of their shape and scale parameters on temperature have been calculated. It is concluded that the shape parameter of the K-distribution can be chosen as the best marker of temperature changes in both experiments. The choice of the marker has been made by analysis of temperature dependencies of all calculated parameters and by comparing the quality of fitting all histograms to the considered distributions. Besides, the chosen marker as a function of temperature exhibits the closest shape to temperature/time function experimentally measured