Spectral diffusion analysis of kidney intravoxel incoherent motion MRI in healthy volunteers and patients with renal pathologies

Purpose: To assess the feasibility of measuring tubular and vascular signal fractions in the human kidney using nonnegative least-square (NNLS) analysis of intravoxel incoherent motion data collected in healthy volunteers and patients with renal pathologies. Methods: MR imaging was performed at 3 Tesla in 12 healthy subjects and 3 patients with various kidney pathologies (fibrotic kidney disease, failed renal graft, and renal masses). Relative signal fractions f and mean diffusivities of the diffusion components in the cortex, medulla, and renal lesions were obtained using the regularized NNLS fitting of the intravoxel incoherent motion data. Test–retest repeatability of the NNLS approach was tested in 5 volunteers scanned twice. Results: In the healthy kidneys, the NNLS method yielded diffusion spectra with 3 distinguishable components that may be linked to the slow tissue water diffusion, intermediate tubular and vascular flow, and fast blood flow in larger vessels with the relative signal fractions, f_slow, f_interm and f_fast, respectively. In the pathological kidneys, the diffusion spectra varied substantially from those acquired in the healthy kidneys. Overall, the renal cyst showed substantially higher f_interm and lower f_slow, whereas the fibrotic kidney, failed renal graft, and renal cell carcinoma demonstrated the opposite trend. Conclusion: NNLS-based intravoxel incoherent motion could potentially become a valuable tool in assessing changes in tubular and vascular volume fractions under pathophysiological conditions.