肾脏射频消融中超声图像特征参数与温度关系的研究

为探索肾脏射频消融中超声图像特征参数与温度的相关性,设计并进行了新鲜离体猪肾脏射频消融实验。记录热消融过程中的温度（侵入式热电偶测量）和B超影像数据,采用将温度探针视为圆形的方法和阈值分割法去除感兴趣区域中的探针,提取灰度均值、灰度标准差和灰度梯度值等特征参数,使用MATLAB软件对得到的灰度均值、灰度标准差和温度作回归分析,采用连续分段函数模型进行数据拟合。证明了阈值分割法可用于探针去除,肾脏射频消融中的超声图像特征参数与温度具有相关性,得到拟合良好的灰度均值-温度、灰度标准差-温度模型。但所得模型中的参数会随图像增益、消融组织差异变化,在后续研究中应排除此类干扰,得到更具有临床价值的温度测量模型。

Correlation between renal ultrasound image features and temperature during radiofrequencyablation

To explore the correlation between ultrasonic image feature parameters and temperature during renal radiofrequencyablation, a radiofrequency ablation experiment of fresh pig kidney in vitro is designed and carried out. The temperature measuredby invasive thermocouple and B-Mode images during ablation are collected. The feature parameters such as average gray-level(AVGL), grayscale standard deviation and grayscale gradient value are calculated after removing the probe in the region of interestby threshold segmentation method or taking the temperature prove as a circle. ThenMATLAB software is used to performregressionanalysis on the obtained AVGL, grayscale standard deviation and temperature, and continuous piecewise function model is usedfor data fitting. It is proved that threshold segmentation method can be used for probe removal, and the feature parameters ofultrasound image in renal radiofrequency ablation are found to be correlated with temperature, and a well-fittedAVGL-temperaturemodel and a grayscale standard deviation-temperature model are obtained. However, the parameters in the obtained models willchange with the difference of image gain and ablation tissue. Such interference should be excluded in the follow-up research toobtain a more clinically valuable temperature measurement model.