What is Image Slices

Shock wave lithotripsy (SWL) is a noninvasive and safe treatment for small renal stones. In unsuccessful cases, retreatment procedures are needed after SWL. According to the previous studies, patient and stone descriptors are good predictors of SWL success. Some stone and kidney descriptors are measured from renal Non-Contrast Computed Tomography (NCCT) images. It is a tedious, time-consuming and error-prone process with large inter-user variability when performed manually. In this study, novel features are proposed automatically extracted from NCCT images to describe morphology and location of renal stones and kidneys to predict retreatments after SWL. The proposed features can distinguish between different kidney and stone morphologies and locations while being less sensitive to image segmentation errors. These features are added to other stone and patient features to predict retreatment within 3 months after SWL. The experimental results show that using the proposed stone features extracted from NCCT images can improve the accuracy of predicting retreatment.