LAUSR

Objectives To evaluate the value of renal diffusion kurtosis imaging (DKI) in the diagnosis of early diabetic nephropathy (DN) in a rat model. Materials and methods Forty male Zucker diabetic fatty rats that spontaneously developed type 2 diabetes mellitus (DM) and 20 age-matched nondiabetic lean Zucker rats were included. Renal DKI scans and histological examinations were performed on the rats in batches at the end of the 4th, 8th, 12th, 16th, and 20th week after DM model was built. Based on renal histopathological appearance, included animals were divided into three groups: a nondiabetic control group, a DM group without DN, and an early DN group. Mean kurtosis (MK) and mean diffusivity (MD) values of renal cortex and medulla were analyzed statistically. Results MK values of renal cortex and medulla tended to increase from the control group to the early DN group, respectively, while MD values tended to decrease. The cutoff MD and MK values of renal cortex and medulla showed different values in discriminating early DN from controls. Among them, cutoff MK value of medulla of 0.62 was the best parameter (sensitivity, 93.9%; specificity, 96.4%; and area under the curve, 0.95). For discriminate early DN from DM without DN and DM without DN from controls, cutoff MK value of renal cortex or medulla achieved an area under the curve of 0.76–0.85. Conclusions MR DKI may be valuable for the noninvasive detection of early DN, and MK value might serve as a more sensitive biomarker of early DN than MD value. Key Points • In this article, diffusion kurtosis imaging (DKI) was used to detect the changes in the kidneys due to early diabetic nephropathy (DN). • MR DKI may be valuable for the noninvasive detection of early DN. • The mean kurtosis values of renal cortex and medulla might serve as a more sensitive biomarker of early DN than the mean diffusivity values.