Our current investigation focused on pathophysiological changes of renal tissue in patients with LN via DWI-MRI. Distinct from the previous related studies, our study focused on the features and distributions of water molecular diffusion rather than merely assessing average levels. After we analyzed ADC data from renal DWI images, two principle results were found. First, conspicuous changes of ADC values existed in the renal cortical zone. ADC values of the renal parenchyma in LN patients were significantly lower than those in healthy volunteers. Moreover, ADC data from LN patients fluctuated more drastically; the corresponding ADC data from healthy volunteers exhibited more homogeneous values with smaller fluctuations in magnitude. By reconstructing ADC data of ROIs into tridimensional maps that were akin to geographical maps, distinctive ADC features were revealed. In the LN group, the majority of the cortical-zone portions of the tridimensional ADC maps displayed a geographic terrain that was canyon-like and/or valley-like. In contrast, the cortical-zone geographical ADC maps from the healthy-volunteers group were flatland-like and/or plateau-like. Secondly, another discovery of our study was the heterogeneity of impaired water molecular motion in LN. By comparing bilateral renal ADC values and configurations of a total of 21 patients, we found that twelve patients showed significantly different ADC values across their two bilateral kidneys. The tridimensional geographic maps also displayed distinctive non-analogous terrain on both sides of the kidneys. To our surprise, the pattern of different ADC values across the two kidneys within the same individual could also be found in healthy volunteers. Specifically, the water molecular motion in left kidneys was slightly greater than that in right kidneys.
Although the precise mechanism underlying the reduced ADC values and their abnormal distributions in LN patients was not elucidated in this study, we speculate that pathophysiological intracellular and/or intercellular changes in renal tissue might impact the motion of water molecules. Theoretically, several categories of proliferative lesions can be found in renal tissues. These lesions include proliferation of endothelial and mesangial cells, proliferation of parietal epithelial cells, cellular or fibrotic crescent formation, and interstitial fibrosis [23]. Since increased cellular densities and extracellular matrix present similar barriers that obstruct the free movement of water, lower ADC values may be detected by DWI-MRI. For example, Li and colleagues detected lower ADC values in LN patients, especially in proliferative LN such as type-III and type-IV subclasses [15]. Another previous study also corroborated this finding. Li et al. investigated the relationship of renal ADC values and renal pathological scores in 71 patients with CKD who had been inspected by renal biopsy. They found that higher renal pathological scores correlated with lower renal ADC values [7]. Whether type-III or type-IV LN subclasses or higher pathological scores were examined, proliferative lesions were the prominent characteristics in the renal tissue.
A functional asymmetry of bilateral kidneys has been documented by nephrologists for many years. Oh et al. detected the glomerular filtration rate (GFR) in each side of the kidney using technetium-99m diethylenetriaminepentaacetic acid (99mTcDPTA). The left kidneys showed greater renal function, and the functional ratio of the left kidney and the right kidney was 52.5% and 47.5%, respectively [24]. Another related finding came from data of living-donor nephrectomies. Hsu et al. reviewed a large sample of donor nephrectomies comprising 27,942 cases of laparoscopic donor nephrectomies (LDNs) and 8,048 cases of open-donor nephrectomies (ODNs). They found that left-sided donor nephrectomies were associated with a lower risk of allograft failure, regardless of open or laparoscopic approaches [25]. Our current study found obvious heterogeneity of ADC configurations in both healthy volunteers and LN patients. To our surprise, the discrepancy of bilateral-renal ADC values in healthy volunteers was opposite to the pattern found in those in LN patients. At present, there are no published studies to explain our findings because relevant asymmetrical studies of bilateral kidneys are lacking. However, we speculate that there was slight asymmetry of renal water molecular diffusion or microcirculation perfusion in the bilateral kidneys. In general, right kidneys may have a superior ability to deliver or distribute liquids compared to left kidneys under healthy physiological conditions. Since autoantibodies and pathogenic cytokines are principle components of the pathogenesis of LN, these autoantibodies or cytokines may be preferentially transported into right kidneys on the basis of left-right asymmetry. Subsequently, proliferative lesions may be more serious in right kidneys and subsequently give rise to decreased ADC values.
Our study also investigated discrepancies in ADC values among the three subgroups of LN patients. ADC values in the type-V subclass were the highest among the three subgroups. Theoretically, lower renal-cortical ADC values should be detected in type-IV-subclass patients with LN. To our surprise, ADC values of the cortical zone in the type-IV subclass were significantly higher than those in the type-III subclass, although pathological injuries in the type-IV subclass were more serious than those in the type-III subclass. According to previous studies, renal-parenchyma ADC values are closely correlated with the magnitude of proliferative lesions. For example. Xu et al. explored the relationship between renal ADC values and histopathologic changes in 52 patients with CKD. They found negative correlations between renal ADC values and scores of tubulointerstitial lesions and severity of interstitial fibrosis [13]. Inoue et al. detected the ADC values in 142 patients with either diabetic nephropathy, CKD without diabetes, or acute-kidney injury. They found that increased fibrosis was significantly correlated with ADC values [26]. As such, it is unclear why lower ADC values were measured in type-III-subclass LN patients in our present study. This surprising result could be due to different constituents of intercellular and extracellular matrices between these two subclasses of LN patients. These renal tissue matrices have different capacities for impacting the molecular motion of water. The tendency of ADC values in the medullary zone was quite different among the three subclasses of LN patients. ADC values in type-V patients gradually declined in the renal medulla. The shape of ADC values in the renal parenchyma was similar with that in healthy volunteers. We speculate that the main reason for this phenomenon may to maintain an increasing osmolality in the renal-medullary zone. In physiological conditions, kidneys should maintain a continuously ascending osmolality for urine concentration. In renal outer medulla, the thick ascending limb can prevent water reabsorption due to low-water permeability in the tubular wall. Although the descending limb is highly permeable to water, the direction of water movement is primarily from the inner tubular to outer tubular region due to aquaporin-1 influences. Moreover, water permeability of the inner-medullary collecting duct is also low. When vasopressin exerts influence on the water-channel aquaporin-2, water permeability of the collecting-duct wall increases. Subsequently, water can then be reabsorbed along the direction from the inner duct to the outer duct. This may explain why lower molecular motion of water is found in the medulla. However, the normal microstructure and physiological functions will be destroyed or absent in the face of tubular-interstitial injuries. Neither the loop of Heles nor the collecting duct would be able to maintain the low-water permeability under the conditions of damaged tissue. This may account for why elevated ADC values were detected in the medullary zone. Furthermore, higher renal-medullary ADC values in type-IV-subclass patients may be attributed to more serious tubular-interstitial pathological injuries.
Our study still had some inherent limitations. First, there are no corresponding studies of renal pathological changes and local regional ADC values distributions available in the present literature. As such, the renal-parenchyma locations of ROIs for our ADC-value measurements did not rigorously match the biopsy sites. This potential intrinsic bias may influence the reliability of our study. Second, the heterogeneity of ADC values in the renal parenchyma was easily found in our LN patients. Even in healthy volunteers, markedly lower ADC values were sporadically distributed in renal tissues. Since our study did not include functional MRI (fMRI) data such as blood-oxygen level differences (BOLDs) or diffusion-tensor imaging (DTI), we did not investigate this matter further. Third, the sample size in our current study was relatively small, which might influence the applicability of our conclusions.