Institute of Diagnostic and Interventional Radiology, University Hospital Zurich, University of Zurich, Raemistr. 100, 8091, Zurich, Switzerland.
Department of Urology, University Hospital Zurich, University of Zurich, Raemistr. 100, 8091, Zurich, Switzerland.
Abdom Radiol (NY). 2018 Jun;43(6):1432-1438. doi: 10.1007/s00261-017-1309-y.
To apply texture analysis (TA) in computed tomography (CT) of urinary stones and to correlate TA findings with the number of required shockwaves for successful shock wave lithotripsy (SWL).
CT was performed on thirty-four urinary stones in an in vitro setting. Urinary stones underwent SWL and the number of required shockwaves for disintegration was recorded. TA was performed after post-processing for pixel spacing and image normalization. Feature selection and dimension reduction were performed according to inter- and intrareader reproducibility and by evaluating the predictive ability of the number of shock waves with the degree of redundancy between TA features. Three regression models were tested: (1) linear regression with elimination of colinear attributes (2), sequential minimal optimization regression (SMOreg) employing machine learning, and (3) simple linear regression model of a single TA feature with lowest squared error.
Highest correlations with the absolute number of required SWL shockwaves were found for the linear regression model (r = 0.55, p = 0.005) using two weighted TA features: Histogram 10th Percentile, and Gray-Level Co-Occurrence Matrix (GLCM) S(3, 3) SumAverg. Using the median number of required shockwaves (n = 72) as a threshold, receiver-operating characteristic analysis showed largest area-under-the-curve values for the SMOreg model (AUC = 0.84, r = 0.51, p < 0.001) using four weighted TA features: Histogram 10 Percentile, and GLCM S(1, 1) InvDfMom, S(3, 3) SumAverg, and S(4, -4) SumVarnc.
Our in vitro study illustrates the proof-of-principle of TA of urinary stone CT images for predicting the success of stone disintegration with SWL.
将纹理分析（TA）应用于尿路结石的计算机断层扫描（CT）中，并将 TA 结果与碎石术（SWL）所需的冲击波数量相关联。
在体外环境下对 34 颗尿路结石进行 CT 检查。对尿路结石进行 SWL，记录结石碎裂所需的冲击波数量。TA 在像素间距和图像归一化的后处理后进行。根据读者间和读者内的可重复性以及通过评估 TA 特征之间的冗余度对冲击波数量的预测能力，进行特征选择和降维。测试了三种回归模型：（1）消除共线性属性的线性回归（2）采用机器学习的序贯最小优化回归（SMOreg），以及（3）具有最低平方误差的单个 TA 特征的简单线性回归模型。
与所需 SWL 冲击波的绝对数量相关性最高的是使用两个加权 TA 特征的线性回归模型（r=0.55，p=0.005）：直方图第 10 百分位和灰度共生矩阵（GLCM）S（3，3）SumAverg。使用所需冲击波的中位数（n=72）作为阈值，接收器操作特征分析显示 SMOreg 模型的曲线下面积最大（AUC=0.84，r=0.51，p<0.001），使用四个加权 TA 特征：直方图第 10 百分位和 GLCM S（1，1）InvDfMom、S（3，3）SumAverg 和 S（4，-4）SumVarnc。
我们的体外研究说明了 TA 分析尿路结石 CT 图像以预测 SWL 结石碎裂成功率的原理证明。
