Diffusion-attenuated MRI signal of renal allografts: comparison of two different statistical models.

Abstract

Contrast-enhanced MRI is considered problematic in renal allograft recipients because of the development of nephrogenic systemic fibrosis. Therefore, we assessed the clinical value of a monoexponential model and a distribution function model of diffusion-weighted imaging (DWI) in renal allografts. A total of 23 patients were divided into three groups, as follows: group A, stable renal allograft function for at least 6 months; group B, transplantation within the past 30 days, with good renal allograft function; and group C, an acute deterioration or decrease in renal allograft function. T2-weighted axial, T1-weighted coronal, and a paracoronal DWI sequences with 16 b values (b = 0-750 s/mm(2)) were performed on a 1.5-T scanner. Region of interest-based analysis of the apparent diffusion coefficient (ADC) of the renal cortex was used. Monoexponential analysis showed mean (± SD) ADC values of 1932 ± 98, 2095 ± 246, and 1636 ± 200 10-(6) mm(2)/s for patient groups A, B, and C, respectively. The distribution function revealed a mean ADC of 2487 ± 185, 2850 ± 325, and 2142 ± 31410-(6) mm(2)/s for groups A, B, and C, respectively. The difference between groups A and B combined and group C (p < 0.005) was statistically significant for both models. R(2) yielded the best regression of mathematic fitting for the distribution function model (p < 0.0001). Unenhanced evaluation of renal allografts with DWI correlated well with renal function for both the monoexponential analysis and the distribution function model. There was no statistically significant difference in ADC values and renal allograft function between both types of analysis, but the distribution function showed the best regression.