Ultrasound thermal mapping based on a hybrid method combining physical and statistical models

Non-invasive temperature measurement of tissues deep inside the body has great potential for clinical applications, such as temperature monitoring during thermal therapy and early diagnosis of diseases. We developed a novel method for both temperature estimation and thermal mapping that uses ultrasound B-mode radiofrequency data. The proposed method is a hybrid that combines elements of physical and statistical models to achieve higher precision and resolution of temperature variations and distribution. We propose a dimensionless combined index (CI) that combines the echo shift differential and signal intensity difference with a weighting factor relative to the distance from the heat source. In vitro experiments verified that the combined index has a strong linear relationship with temperature variation and can be used to effectively estimate temperature with an average relative error <5%. This algorithm provides an alternative for imaging guidance-based techniques during thermal therapy and could easily be integrated into existing ultrasound systems.