Automated assessment of renal cortical surface roughness from computerized tomography images and its association with age

Rationale and Objectives: Nephrosclerosis occurs with aging and is characterized by increased kidney subcapsular surface irregularities at autopsy. Assessments of cortical roughness invivo could provide an important measure of nephrosclerosis. The purpose of this study was to develop and validate an image-processing algorithm for quantifying renal cortical surface roughness invivo and determine its association with age. Materials and methods: Renal cortical surface roughness was measured on contrast-enhanced abdominal computed tomography (CT) images of potential living kidney donors. A roughness index was calculated based on geometric curvature of each kidney from three-dimensional images and compared to visual observation scores. Cortical roughness was compared between the oldest and youngest donors, and its interaction with cortical volume and age assessed. Results: The developed quantitative roughness index identified significant differences in kidneys with visual surface roughness scores of 0 (minimal), 1 (mild), and 2 (moderate; P<.001) in a random sample of 200 potential kidney donors. Cortical roughness was significantly higher in the 94 oldest (64-75 years) versus 91 youngest (18-25 years) potential kidney donors (P<.001). Lower cortical volume was associated with older age but not with roughness (r=-0.03, P=75). The association of oldest age group with roughness (odds ratio [OR]=1.8 per standard deviation [SD] of roughness index) remained significant after adjustment for total cortex volume (OR=2.0 per SD of roughness index). Conclusions: A new algorithm to measure renal cortical surface roughness from CT scans detected rougher surface in older compared to younger kidneys, independent of cortical volume loss. This novel index may allow quantitative evaluation of nephrosclerosis invivo using contrast-enhanced CT.