Radiomics identifies distinct cortical bone texture alterations in patients with CKD using HR-pQCT.

Abstract

Standard clinical imaging metrics perform poorly in predicting skeletal fragility in chronic kidney disease (CKD), particularly due to the complex and heterogeneous cortical deterioration that characterizes advanced disease. Here, this study aimed to identify radiomic features derived from high-resolution peripheral quantitative computed tomography (HR-pQCT) in tibial cortical bone that distinguish CKD-related differences and may serve as markers of subtle cortical alterations undetected by conventional imaging. HR-pQCT image stacks were obtained from 72 participants (38 non-CKD and 34 with CKD stage 5D) at 7.3% (distal) and 30% (diaphyseal) proximally from the tibial endplate, resulting in a total of 24 192 slices. In non-CKD cases, features were largely derived from first-order statistics, while complex features from higher-order statistics were more prominent in CKD cases. Although conventional HR-pQCT outcomes, such as volumetric bone mineral density, showed limited ability to differentiate CKD from non-CKD cortical bone in our population of stage 5D patients, the top features, such as Minimum and Strength, provided a significant distinction between the two groups (P < 0.001, Effect size r = from 0.813 to 0.856). Our findings demonstrate that radiomic analysis identifies cortical bone differences associated with CKD that were not distinguished by conventional HR-pQCT metrics, highlighting its potential to improve bone quality assessment in this high-risk population.