Early detection of functional changes in transplanted kidneys can play an important role in investigating renal impairments and initiating accurate treatments. In recent years, there have been several studies in native and transplanted kidneys investigating the impact of the functional MRI methods (DTI, BOLD, & ASL MRI) for the assessment of renal function (1-9). However, there are only a few studies investigating the accuracy and repeatability of functional MRI techniques in transplanted kidneys, especially applying several methods in concert. The main aim of this study was therefore to evaluate the reproducibility of DTI, BOLD, and ASL MRI parameters derived from two scans in transplanted kidneys. A secondary aim was to investigate whether there are significant correlations between renal parameters obtained from these three functional MRI methods in transplanted kidneys.

DTI, BOLD, and ASL MRI were performed in 12 renal allografts with normal kidney function (47.3±14.4 years) using 3T MR scanners (Verio, Trio, Siemens). Renal allografts were scanned two times with the same protocol, back-to-back with 30 minutes break. A DW single shot echo-planar measurement was performed without triggering in transplanted kidneys with ten different b-values (0-700s/mm²) and (acquisition number=2, TR=3300ms, TE=56ms, slice-thickness=5mm, Matrix=192x192, FOV=300x300mm²). BOLD MRI was performed in a single breathold of 17 seconds per slice with 12 echoes (6-52.3ms) and following parameters: TR=65ms, FOV=400x400mm², Matrix=256x256, slice-thickness=5mm. The FAIR true fast imaging with steady state precession (True-FISP) ASL (10) was performed in coronal orientation with following parameters: TR=4.0ms, TE=2.0ms, slice-thickness=7mm, Matrix=128x128, FOV=350x350mm², TI=1200ms, T1=1150ms. All data were analyzed using in-house custom scripts written in IDL and MATLAB. DTI yielded ADC, fractional anisotropy (FA) and perfusion fraction (F_P). BOLD MRI yielded the relaxation rate R2* and ASL yielded perfusion values. To assess reproducibility, coefficients of variations within (CV_w) and between (CV_b) subjects were calculated and presented in percent of the mean. Pearson linear regression method was used to assess correlations between parameters.

DTI and BOLD MRI measurements were successfully completed in 10 out of 12 patients and ASL measurements were completed in 9 subjects. One patient was excluded due to polycystic kidney disease. The CV_w for medullary and cortical ADC were 2.5% and 2.9%, respectively and the CV_b for medullary and cortical ADC were 4.2%, and 6.1%, respectively. The variations were higher for FA (<28%) as well as for F_P, especially between subjects, CV_b (<35%, Table 1). The CV_w for R2* in medulla and cortex were 6.5% and 6.9%, respectively, while the CV_b for R2* were clearly higher (in medulla: 16.6%; in cortex: 14.4%). The CV_w and CV_b of perfusion results obtained from ASL in cortex were 9.3% and 10.0%, respectively and in medulla 16.2% and 35.5%, respectively. Figures 1 & 2 show that the values of the second scans for each subject are close to those obtained from the same subject in the first scans. There were no significant correlations between R2* and diffusion parameters. Similarly, the correlation between R2* and perfusion values calculated from ASL was not found significant and no significant correlations were found between diffusion parameters and perfusion.High reproducibility was obtained for most parameters derived from DTI, BOLD and ASL scans. FA showed lower variance in medulla than in cortex. The high variance for F_P is in concordance with the high variance between subjects found in the other studies. F_P values are lower than reported before (6,11), which may be due to shorter TE used in the present study and improved signal stability. The cortical perfusion derived in ASL showed low variance between and within subjects. However, medullary perfusion showed greater variation than cortical perfusion. Low reproducibility in medullary perfusion has been reported before (7). This increased variability in medulla is most likely due to the lower SNR as compared to the cortex. ASL signal is directly proportional to blood flow, and medullary perfusion is much lower than cortical perfusion (7). There were no significant correlations between renal perfusion and the other parameters obtained from DTI and BOLD MRI. This is in contrast to another study showing a significant correlation between perfusion in ASL and F_P in DTI for transplanted kidney (11). In conclusion, our results suggest that DTI, BOLD MRI, and ASL techniques provide reproducible results for different parameters reflecting renal function. The clearly lower CV_w for some parameters suggest performing longitudinal studies instead of cross-sectional studies if possible. No significant correlations between parameters in different methods indicate that these techniques may yield complementary information in transplanted kidneys. These measurements build the basis for planning future clinical studies.