Influence of temperature variations on the average grayscale of B-mode images of HIFU-induced lesions

The aim of this work is to understand how the variations in grayscale values of B-mode ultrasound (US) images can be used as an approach for non-invasive temperature estimation. In order to obtain real-time monitoring of HIFU treatment, an US imaging system and HIFU were synchronized. Images were acquired using an electronic convex imaging probe. An 8% BSA tissue-mimicking polyacrylamide gel was used for the experiments. First, the HIFU power was set to 10 W. The application of HIFU resulted in the appearance of hyperechoic regions that were used to place a thermocouple tip at the focal spot by means of US imaging. Afterwards, the power was gradually increased up to 40 W for 4 min. The temperature sampling rate was set to 5 Hz. For each temperature sample the synchronization system captured one frame. The region of interest (ROI) was manually selected and a set of morphological operations were implemented in order to obtain the shape and size of the thermal lesion. From here, the average grayscale (AGS) and area of the thermal lesion were calculated to assess temperature quantification. The AGS parameter showed a maximum correlation coefficient of 0.6626 as a function of temperature whereas the thermal lesion appeared. In contrast, area values showed a greater correlation coefficient of 0.9122. In conclusion, temperature shows a non-linear behavior with respect to the parameters estimated due to the nature of the thermal lesion formation by HIFU exposure.