Prediction of Chronic Kidney Disease Progression by Magnetic Resonance Imaging: Where Are We?

Chronic kidney disease (CKD) is characterized by a glomerular filtration rate (GFR) under 60 ml/min/1.73m² or kidney damage as defined by structural or functional abnormalities other than decreased GFR [1]. CKD is a worldwide health issue, afflicting around 14% of the population [2]. This progressive and often irreversible process of renal function decline may reach an endpoint of end-stage renal failure, requiring renal replacement therapy. The capability to diagnose progressive CKD at an early stage, and predict which patients are prone to progress, is critical for timely therapeutic interventions. As such, researchers have expended enormous efforts in development of novel biomarkers that may identify subjects with early CKD at risk of progression.

Chronic hypoxia has been recognized to plays a pivotal role in CKD [3]. Tissue injury factors, like capillary dropout, mitochondrial dysfunction, and inefficient oxygen utilization, may decrease renal perfusion and induce tissue hypoxia, which subsequently trigger tubular injury and interstitial fibrosis. These in turn damage adjacent capillaries, glomeruli, and tubules, which exacerbates kidney hypoxia and fibrosis and sets in motion a vicious cycle leading to end-stage renal failure. Therefore, accurate assessment of renal parameters such as perfusion, oxygenation, and fibrosis, may allow gauging of CKD stage and potentially prediction of CKD progression.

Magnetic resonance imaging (MRI) provides powerful tools for noninvasive assessment of renal function. Arterial spin labeling (ASL) uses endogenous blood as a tracer to measure renal perfusion. Blood oxygenation-level-dependent (BOLD) MRI probes renal oxygenation level by taking advantage of the paramagnetic property of deoxyhemoglobin, which increases the effective transverse relaxation rate R₂*. Diffusion MRI measures tissue apparent diffusion coefficient (ADC), which is considered to be lowered by renal fibrosis. In the current issue of American Journal of Nephrology, Prasad and colleagues evaluated the utility of these MRI techniques in assessment of renal function assessment and prediction of CKD progression in patients with diabetes and stage 3 CKD, an early stage CKD with important clinical implications [4]. The investigators enrolled 41 patients with moderate CKD and 13 healthy controls. In the renal cortex, tissue perfusion by ASL was found dramatically decreased in CKD patients compared to the controls, but no statistical significant differences were observed in ADC and R₂*. The authors also investigated the predictive value of MRI parameters on annual loss in estimated GFR (eGFR_slope), and reported moderate to good predictability of basal cortical perfusion and furosemide-induced medullary R₂* changes. The strength of the paper is in recruiting 41 patients with mild-moderate diabetic kidney disease. While their study raise some questions regarding the specificity of non-contrast enhanced MRI techniques for renal assessment in CKD, it also sheds lights on promising application of MR renography for CKD management.

The ASL technique successfully detected a large drop in renal cortical perfusion in CKD, which was also significantly associated with eGFR_slope. These observations are in line with the notion that renal function loss is triggered by capillary dropout-induced perfusion insufficiency in CKD. Although ASL does not require exogenous contrast injection, the use of endogenous blood as perfusion tracer actually makes it a contrast-enhanced MR technique. Therefore, ASL offers high specificity to renal perfusion and was able to capture its early changes in CKD.

By contrast, diffusion MRI and BOLD-MRI are non-contrast enhanced, and therefore have limited specificity to their biological targets, i.e., renal fibrosis and hypoxia. Specifically, the ADC is an overall measurement of the water mobility in tissue, which is literally affected by a composite of factors, including tissue perfusion, vascularity, tubular flow, and true water diffusion. Notably, the perfusion-induced water mobility has been reported to be an order magnitude larger than the true water diffusivity in human kidneys [5]. Although the authors were careful to use relatively high b-values in this study, this does not preclude the potential influence of perfusion on diffusion measurement, especially considering the reported significant association between the ADC and renal perfusion [4]. Similarly, R₂* by BOLD-MRI depends on many factors other than deoxyhemoglobin, such as renal perfusion, vascular volume, presence of inflammatory cells and edema, hydration status, and anemia. Changes in these factors may offset the effect of deoxyhemoglobin in decreasing T₂* [6]. This may account for the observation that BOLD-MRI is effective in measuring tissue hypoxia in acute kidney diseases with less confounding factors, but is less consistent when applied in CKD [6]. Therefore, the mere fact that diffusion MRI and BOLD-MRI failed to detect renal fibrosis and tissue hypoxia in CKD does not rule out their contribution to the pathogenesis of early CKD.

One method to increase the specificity of a non-contrast enhanced MRI techniques in measuring renal function is to perform a stress test, such as water loading or furosemide challenge. This diuretic inhibits active solute transport in the thick ascending limb of Henle’s loops, which results in decrease in medullary oxygen consumption and thus increases medullary oxygenation [7]. Hence, a robust response to furosemide mandates adequate delivery of furosemide to the tubules, but mainly indicates available mass of active medullary tubules engaged in solute transport. Prasad et al. applied the furosemide stress test (FST) in CKD patients and reported a good correlation between the measured medullary ΔR₂* and eGFR_slope. Subsequent receiver operating characteristic analysis showed an area under curve of 0.88, indicating a good predictive power of FST on CKD progression [4]. The reason that FST performed better than absolute R₂* values may be related to the modest reliability and reproducibility of basal R₂* alone measured by MRI. Interestingly, tubular function was found somehow disassociated from renal filtration function, as indicated by the poor correlation between medullary ΔR₂* and eGFR [4], which warrants further investigation.

Nevertheless, the inability of BOLD MRI to distinguish patients from controls might be related to the modest degree of CKD in these patients, with eGFR of merely 51.2±12.6 mL/min/1.73m². In contrast, a recent multicenter study showed decreased both cortical and medullary absolute R₂* and ADC in 127 patients with advanced CKD (eGFR 33.4±7.2 mL/min/1.73m²), despite a more mixed underlying etiology of CKD [8]. Whether the ability to detect these parameters in advanced CKD was related to the larger number of patients or the greater severity of CKD needs to be determined.

Taken together, MR renography offers exciting opportunities for monitoring progression of CKD. However, different MR techniques may show variations in their performance, depending on the specificity to their respective physiological targets and the severity of the disease. Plausibly, a fall in renal perfusion precedes loss of oxygenation and development of fibrosis in CKD. Furthermore, the combination of FST and BOLD-MRI is useful in assessing tubular function and predicting CKD progression. Finally, as CKD progression is multi-factorial, a combination of multi-parametric MRI and other clinical tests may provide more reliable assessment of CKD stage and prediction of future progression.