The nephron structure and epithelial organization are essential for kidney function. Diffusion tensor imaging (DTI) and susceptibility tensor imaging (STI) can assess the integrity of the nephron where STI provides additional molecular information (1,2). STI has demonstrated sensitivity to changes in kidney disease models (3). The source of susceptibility anisotropy is hypothesized to be the organized tubules, basement membrane, and the organized lipids (2). Here, susceptibility was found to be most diamagnetic when tubules were parallel with the magnetic field, suggesting that a diamagnetic material points along the tubular long axis. Ischemia reperfusion is one particular disease model with well known cellular and lipid disorganization in specific nephron segments (4). In these cells, the microvilli, basolateral infoldings, and mitochondria become disorganized and swollen (Fig. 1). In the present study, we applied STI in a model of ischemia perfusion to demonstrate changes in susceptibility anisotropy and to provide additional evidence that the cellular organization is a major contributor.

Male C57Bl/6 mice (14 weeks) were used for ischemia reperfusion (n=6 control, n=6 ischemic reperfused). The right kidney was excised while the left renal pedicle was clamped (45 min). No surgery was performed on the control animals. Animals were perfusion fixed 14 days after surgical completion. Kidneys were immersion enhanced with 2.5 mM ProHance to decrease T1 and improve SNR.

Imaging was performed in a vertical bore 9.4T Oxford magnet (Agilent Direct Drive console). The specimen was placed in a sphere to facilitate multiple orientations for STI. The holder was placed in a high Q loop gap resonator. STI data were acquired using 3D multi-echo gradient echo sequence (12 directions total). DTI data were acquired using a 3D diffusion-weighted spin echo sequence. One b0 image and 12 diffusion weighted (1500 s/mm²) images were acquired at 55×55×55 μm³.

Registered phase images were used to solve the 6 independent elements of the susceptibility tensor (χ₁₁, χ₁₂, χ₁₃, χ₂₂, χ₂₃, and χ₃₃) following (5). Eigenvalue decomposition was performed on the tensor to define the principal susceptibilities and eigenvectors. Anisotropy was measured using a susceptibility index, SI=(x₁-x₃+y)/x, where x₁-x₃ is a direct measure of susceptibility anisotropy, y is adjustable parameter, and x is the mean susceptibility (6). Tractography was performed on major diamagnetic eigenvector, seeded by mask images, and filtered by renal regions and track lengths (TrackVis). Angle threshold was 30°. DTI fractional anisotropy (FA) and STI susceptibility index (SI) were set a threshold of 0.2 to 0.9.

DTI and STI in normal and ischemic reperfused kidneys are shown in Fig. 2. Size and shape changes are noticeable in the ischemic reperfused kidneys. DTI revealed some reduced anisotropy and tractography in the inner medulla. STI demonstrated a greater reduction in anisotropy and tractography in the same region (indicated by green arrows in Fig. 2). Anisotropy was measured in specific regions of the kidney (cortex, outer medulla, and inner medulla) and compared between DTI and STI (Table 1). The difference in anisotropy between normal and ischemic reperfused kidneys was greater in STI compared to DTI (Table 1). The anisotropy reduction was as great as 0.13 (38% reduction) with STI and 0.06 (21% reduction) with DTI. In the cortex, STI revealed an anisotropy change of 0.06 (32% reduction) while DTI showed a change of 0.01 (9% reduction).In this study, we selected a kidney disease model with well-characterized cellular injuries. These include mitochondrial and cell swelling, most evident in the proximal tubules of the cortex (4). Here, microvilli are very long at the brush border and can have the greatest amount of lipid organization. If disrupted, the cellular components can greatly affect the anisotropy detected with STI. We found that one of the greatest reductions of susceptibility anisotropy was in the cortex of the kidney. Considering that the cortex consists of mostly tortuous nephron segments, the reduction in anisotropy is significant. Susceptibility anisotropy change was greatest in the inner medulla where nephron segments are most straight and coherent. Consequently, the presence of diffusion anisotropy and performance of DTI tractography demonstrate that the tubular segments are mostly intact. This provides additional evidence that the source of susceptibility anisotropy originates from the cellular structures in the renal epithelia. STI can be very sensitive in detecting nephron cells and walls, while DTI is mostly limited to the anisotropic water diffusion inside the tubules. In conclusion, we demonstrated that the susceptibility anisotropy in coherent tubules of diseased kidneys is considerably reduced, while diffusion anisotropy remains similar. STI offers a method in detecting renal diseases with subtle microstructural damages.