Volumetric analysis of MRI data monitoring the treatment of polycystic kidney disease in a mouse model.

Abstract

OBJECT: The human condition autosomal dominant polycystic kidney disease (ADPKD) is characterized by the growth of cysts in the kidneys that increase renal volume and lead to kidney failure. Mice studies are performed for treatment development monitored with imaging. The analysis of the imaging data is typically manual, which is costly and potentially biased. This paper presents a reliable and reproducible method for the automated segmentation of polycystic mouse kidneys. MATERIALS AND METHODS: Treated and untreated mice have been imaged longitudinally with high field anatomic MRI. The region of interest (ROI) of the kidneys in the images is identified and restored for artifacts. It is then analyzed statistically and geometric models are estimated for each kidney. The statistical and geometric information are provided to the graph cuts algorithm that delineates the kidneys. RESULTS: The accuracy of the analysis has been demonstrated by showing consistency with results obtained with previous methods as well as by comparing with manual segmentations. CONCLUSION: The method developed can accelerate and improve the accuracy of kidney volumetry in preclinical treatment trials for ADPKD.