Magnetic resonance imaging with arterial spin labeling (ASL) is a noninvasive approach to measure organ perfusion. The purpose of this study was to evaluate the reproducibility of ASL kidney perfusion measurements with semiautomatic segmentation, which allows separate quantification of cortical and medullary perfusion. The right kidneys of 14 healthy volunteers were examined 6 times on 2 occasions (3 times at each occasion). There was a 10-minute pause between each examination and a 14-day interval between the 2 occasions. Cortical, medullary, and whole kidney parenchymal perfusion was determined with customized semiautomatic segmentation software. Coefficient of variances (CVs) and intraclass correlations (ICCs) were calculated. Mean whole, cortical, and medullary kidney perfusion was 307.26 ± 25.65, 337.10 ± 34.83, and 279.61 ± 26.73 mL/min/100 g, respectively. On session 1, mean perfusion for the whole kidney, cortex, and medulla was 307.08 ± 26.91, 336.79 ± 36.54, and 279.60 ± 27.81 mL/min/100 g, respectively, and on session 2, 307.45 ± 24.65, 337.41 ± 33.48, and 279.61 ± 25.94 mL/min/100 g, respectively (P > 0.05; R² = 0.60/0.59/0.54). For whole, cortical, and medullary kidney perfusion, the total ICC/CV were 0.97/3.43 ± 0.86%, 0.97/4.19 ± 1.33%, and 0.96/4.12 ± 1.36%, respectively. Measurements did not differ significantly and showed a very good correlation (P > 0.05; R² = 0.75/0.76/0.65). ASL kidney measurements combined with operator-independent semiautomatic segmentation revealed high correlation and low variance of cortical, medullary, and whole kidney perfusion.

Total kidney volume (TKV) is an important marker for the presence or progression of chronic kidney disease, however, routine ultrasonography underestimates renal volume to a high and varying degree. The aim of this work was to adapt and evaluate a semi-automatic unimodal thresholding method for volumetric analysis of the kidney in native T2-weighted magnetic resonance (MR) images. In a group of healthy volunteers (n = 24; 48 kidneys), we defined a region of interest (ROI) by manually tracing the outline of the kidney in every MR image. An automatic unimodal thresholding algorithm with visual feedback was applied to the probability distribution function of voxel intensities in the ROI to remove intrarenal non-parenchyma volume. For comparison, reference volumes were created by manual segmentation. Intra- and inter-observer reliability was evaluated. There was a small, significant mean difference of 1.5 ml between semi-automatically and manually segmented TKV (p = 0.009, 95% CI [0.4, 2.7]). While intra-observer reliability was good (mean difference 2.9 ml, p < 0.01, 95% CI [1.5, 4.2]) there was a small but significant mean difference of 4.8 ml (p < 0.01, 95% CI [3.6, 5.9]) between the TKV results of different observers. Reference volume correlations were excellent (r = 0.97–0.98). Semi-automated segmentation was significantly faster than manual segmentation; mean difference = 234 s [91–483 s]; p < 0.05. Automatic unimodal thresholding removed a considerable mean volume of 18.7 ml (13.1%) from the coarse manual pre-segmentations. Unimodal thresholding of native MR images is a robust and sufficiently reliable method for kidney segmentation and volumetric analysis. The manual pre-segmentation can be done by non-experts with little introduction.