Noninvasive assessment of renal dynamics and pH in a unilateral ureter obstruction model using DCE MR-CEST urography

Abstract

Purpose:

To assess the potential of DCE MR CEST urography for assessing renal function in mice with unilateral ureter obstruction (UUO) by simultaneous pH and renal uptake/clearance measurements following injection of iopamidol.

Methods:

The right ureter of nine mice was obstructed via suture ligation. The animals were imaged at day 1, 2, and 3 post-obstruction on an 11.7T MRI scanner. Ninety-six sets of saturated CEST images at 4.3 and 5.5 ppm were collected. Renal pH values were obtained by calculating the signal ratio for these two frequencies and using a pH calibration curve. Renal time activity curves were measured as a percentage change in the post-injection CEST signal at 4.3 ppm relative to the average pre-injection signal.

Results:

For the healthy mice, the time activity curves of both kidneys were nearly identical and displayed rapid excretion of contrast. For the UUO mice, the dynamic CEST curves for the obstructed kidneys displayed prolonged time to peak (TTP) values and delayed contrast excretion compared with the contralateral (CL) kidneys. Renal pH maps of the healthy animals showed similar acidic values for both kidneys (pH 6.65 ± 0.04 vs 6.67 ± 0.02), whereas in the obstructed kidneys there was a significant increase in pH values compared with the CL kidneys (pH 6.67 ± 0.08 vs 6.79 ± 0.11 in CL and UUO kidneys, respectively).

Conclusion:

Our findings indicate that DCE-MR-CEST urography can detect changes in renal uptake/excretion and pH homeostasis and distinguish between obstructed and unobstructed kidney as early as 1 day after UUO.

1 |. INTRODUCTION

Urinary tract obstructions (UTOs) at any point in the urinary tract that disrupt normal urine flow cause urinary retention and increased retrograde hydrostatic pressure. The most common cause of UTO are kidney stones with a lifetime prevalence of 8.5%. Other common causes include urothelial malignancy, benign prostatic hyperplasia (BPH), strictures, and congenital ureteropelvic junction (UPJ) obstruction.¹ There are as many 1 44 320 surgeries/year performed for kidney stones and 1 01 270 surgeries/year for BPH in the United States alone.² Significant obstructions profoundly impair the excretory function of the kidney as well as its ability to maintain fluid and electrolyte homeostasis. If undiagnosed or left untreated UTOs lead to a progressive tubular atrophy and chronic interstitial inflammation and fibrosis.^1,3,4 Because the extent of recovery of kidney function in obstructive nephropathy (ON) resulting from UTO depends on the severity and duration of the obstruction, early diagnosis and prompt intervention are crucial to prevent chronic kidney disease (CKD) and avoid irreversible kidney damage.

Imaging plays an important role in the initial diagnosis of UTOs in the clinic and several radiologic techniques are available to detect the site, level and cause of the obstruction and to determine whether dilation of the renal pelvis and calices (hydronephrosis) represents a functional obstruction that will lead to kidney injury or a non-obstructive process with preserved kidney function.⁵ While ultrasound and conventional CT have emerged as preferred anatomic imaging modalities when obstruction is suspected, they cannot reliably quantify renal function. Conversely, dynamic nuclear scintigraphy provides a functional assessment of both kidneys after intravenous administration of a radiotracer such as ^99mTc-mercaptoacetyltriglycine (^99mTc-MAG3) and can diagnose upper UTO, but its anatomic definition is poor. Moreover, this technique has other pitfalls related to drawing ROIs (particularly in the setting of hydronephrosis), inadequate SNR, and variation in scanning intervals.¹ Because of their comparatively high temporal and spatial resolution, fast CT variants such as multi-detector CT (MDCT) and electron-beam CT (EBCT) have the potential to provide adequate evaluation of differential renal function via glomerular filtration rate (GFR) measurements. Unfortunately, these novel approaches involve exposure to relatively high levels of ionizing radiation and doses of contrast agent.^6–10

MRI can overcome some of these limitations and, unlike CT and nuclear scintigraphy, does not involve exposure to ionizing radiation, which makes it particularly useful for examining children, pregnant women and patients who need frequent scans.¹¹ Furthermore, because of the significant advances made over the past decades, MRI is now capable of providing information on renal physiological function characterized by perfusion, renal excretory function, and intra-renal oxygenation in addition to having high spatial resolution, excellent soft tissue delineation.^12,13 Excretory gadolinium-enhanced MR urography (MRU) can provide functional information on renal perfusion, excretion, and drainage.^14–16 However, the use of gadolinium-based contrast agent requires caution in patients with severe renal insufficiency as it may lead to nephrogenic systemic fibrosis, which is a progressive and potentially fatal fibrotic disorder of the skin and internal organs.^17,18 Moreover, there are increasing concerns related to cumulative gadolinium administration, as the effects of CNS gadolinium deposition remain incompletely understood.¹⁹ As a result of these limitations for all current radiological methods, new technologies are warranted.

Chemical exchange saturation transfer (CEST) is an emerging MRI technique that allows indirect detection of low concentration molecules after applying a saturation pulse on resonance with exchangeable protons and measuring the saturation frequency dependence of the water signal. Chemical exchange between saturated labile protons on the contrast agents and water results in amplification of this saturation signal loss onto the water signal.^20–26 Several recent studies have demonstrated the potential of CEST imaging for detecting molecular and cellular changes associated with renal and urinary diseases such as diabetic nephropathy,²⁷ acute renal allograft rejection,²⁸ sepsis-induced acute kidney injury,²⁹ unilateral ureter obstruction (UUO),³⁰ and kidney fibrosis.³¹ Furthermore, because of its high sensitivity to pH changes, contrast-enhanced CEST imaging has been proposed to measure renal pH^32–34 values and induced pH alterations following an acute renal injury,³⁵ reperfusion ischemia,³⁶ and CKD.³⁷ In the present study, we have applied dynamic contrast-enhanced CEST imaging to a murine model of irreversible UUO, and assessed its utility for evaluating differential renal function in obstructive nephropathy. We have demonstrated that using a dynamic CEST-MRI approach combined with a single injection of iopamidol, which is a clinically approved x-ray contrast agent, it is possible to obtain both the spatially localized time-activity curves that are similar to standard renograms and renal pH maps. This allows for simultaneous assessment of renal perfusion and excretory function, as well as renal pH changes in response to kidney damage in UUO.

2 |. METHODS

2.1 |. Animal preparation

All animal procedures were conducted in accordance with institutional guidelines and approved by the Johns Hopkins University Institutional Animal Care and Use Committee. Thirteen male and female C57BL/6N-mice were fed a standard diet and allowed free access to water. To perform the surgery, nine mice were anesthetized with 2%–2.5% isoflurane and the right kidney was exposed through a midline incision. The ureter was obstructed completely near the renal pelvis using a silk tie suture. The mice were imaged at day 1 (three mice), 2 (three mice), and 3 (three mice) after the procedure. For histologic evaluation, three additional mice underwent UUO surgery, and their kidneys were harvested at the same time points as in the MRI experiments.

2.2 |. In silico CEST protocol optimization

Numerical simulations were performed using an open-source Matlab-based implementation of the Bloch-McConnell equations (https://github.com/cest-sources/BM_sim_fit),³⁸ investigating the effect of saturation power (B₁), saturation time duration (t_sat), recovery time (T_rec, the time between the end of the acquisition and begin of the subsequent saturation pulse), and longitudinal relaxation rate of water (R_1A) on the iopamidol CEST signal. The Bloch-McConnell system consisted of six pools including amide (4.3 ppm), 2-hydrooxypropanamido protons (5.5 ppm), two hydroxyl groups (1.8 and 0.8 ppm), bulk water (0 ppm),³⁹ and an in vivo-like semisolid MT pool (0 ppm). The chemical exchange rates of iopamidol were taken from the literature.^39,40 The relaxation rates of the water CEST pool were set to R_1w = 0.66 1/s and R_2w = 35 1/s mimicking the values reported in the healthy kidney tissue.²⁹ Detailed simulation parameters are given in Table 1. We did not include an NOE pool or any other endogenous CEST pool in our simulations because we assume that the endogenous NOE effects are negligible for the optimization outcome. After the optimal B₁ at t_sec = 3 s was found, t_sat and T_rec were optimized iteratively. Furthermore, the effect of longitudinal water relaxation on the CEST signal simulated with the final set of saturation parameters and T_1A (=1/R_1A) values in the physiological range between 1 and 3 s was explored.

TABLE 1.

Detailed parameters used for numerical Bloch-McConnell simulations

	Water	Semi-solid MT	Iopamidol I	Iopamidol II	Iopamidol III	Iopamidol IV
Chemical shift [ppm]	0		4.3	5.5	1.8	0.8
Fraction f	1	0.05	2/3700	1/3700	1/3700	4/3700
R1 [1/s]	0.66	0.5	1.6	1.6	1.6	1.6
R2 [1/s]	34	109 890	66	66	66	66
Exchange rate [1/s]	-	40	380	1200	1104	1715

2.3 |. In vivo MR imaging

Each mouse was placed in a mouse restrainer and a needle catheter was inserted into a lateral tail vein for intravenous injection. The in vivo MRI experiments were conducted on an 11.7 T Bruker horizontal bore animal scanner (Bruker Biospin GmbH, Ettlingen, Germany) using an eight-channel mouse body phased-array receive coil. The animals were positioned on an imaging cradle and anesthetized with 0.5%–2% isoflurane. The respiratory rate was continuously monitored and the mice body temperature was maintained at 37°C using a water-heated animal bed. Twenty-one axial and 12 coronal contiguous slices were collected using a multi-slice T2w RARE sequence with the following parameters: TE/TR: 20/4000 msec, matrix size: 128 × 128, slice thickness: 1.5 mm, and RARE factor: 8. A single coronal slice of 1.5 mm thickness containing the center of two kidneys was selected for CEST imaging. To obtain B₀ inhomogeneity maps, the water saturation shift referencing (WASSR) approach⁴¹ was used. The WASSR images were collected at 42 frequency offsets between −1.5 and 1.5 ppm using 10 saturation pulses with a duration of 300 ms, an inter-pulse delay of 10 μs and B₁ = 1.2 μT. For the dynamic CEST experiments, the CEST saturation module consisted of 10 rectangular-shaped RF pulses, each 300 ms long with an inter-pulse delay of 10 μs and an amplitude of 3.6 μT. This amplitude was chosen based on the results of our numerical simulations, which showed strong CEST contrast at 4.3 ppm and better separation between two amide peaks of iopamidol than the 4.0 μT we used previously. The readout parameters were: TE/TR = 3.49/5125 ms, matrix size: 64 × 64, FOV: 28 × 20 mm², slice thickness: 1.5 mm, RARE factor: 33. A total of two-hundred and four CEST images were collected, including 12 images at 40 ppm and 96 sets of images at 4.3 and 5.5 ppm. Iopamidol (Bracco Imaging S.p.A, Milan, Italy) at a weight-controlled dose of 1.5 g of iodine per kilogram was injected approximately 3 min and 30 s after the start of CEST data acquisition. Total CEST scan time was 17 min and 25 s.

2.4 |. In vivo dynamic CEST-MRI data analysis and pH mapping

Image post-processing and dynamic CEST data analysis were performed using custom-written MATLAB (Mathworks, Natick, MA, USA) procedures. Parenchymal ROIs, excluding the renal hilum, were drawn manually around each kidney on high-resolution T2w images. To remove motion-related signal fluctuations, a moving average filter (MATLAB function filter) with the span of four was applied at each frequency offset. A region of interest (ROI)-based time-course curves of signal enhancement were obtained by plotting percentage change in the CEST-prepared signal intensity at 4.3 ppm after iopamidol injection (M_z,post) relative to the averaged signal from seven images measured before the injection (M_z,pre): %ST_enh(t) = 100 × (M_z,pre—M_z,post(t))/M_z,pre. The time-to-peak (TTP) in units of seconds and percent peak CEST signal enhancement (Peak %ST_enh) were computed from the time-activity curves of each kidney.

For pH mapping, post-injection magnetization transfer ratio (MTR) at 4.3 and 5.5 ppm was quantified from three averaged images collected at the peak enhancement time. Averaged pre-injection MTR maps were then subtracted from the post-injection MTR images to remove endogenous CEST signals. Subsequently, renal pH values were obtained by calculating the concentration-independent saturation transfer ratio R_ST = (1—MTR_4.3ppm) × MTR_5.5ppm/((1—MTR_4.3ppm) × MTR_5.5ppm)⁴² and using the pH calibration curve as described previously.³⁷

2.5 |. Histologic examination

Formalin-fixed paraffin-embedded kidney tissues were cut into 7 μm-thick coronal sections and stained with hematoxylin and eosin (H&E) and Mason trichrome staining (MTS). Percentage of fibrosis was estimated from the Trichrome stain at 40x magnification using ImageJ Software for pixel analysis.

2.6 |. Statistical analysis

A Shapiro-Wilk test was used to determine whether the null hypothesis of composite normality of the CEST-derived parameters (pH, TTP, and %ST_enh) estimated in obstructed and contralateral kidneys is true at a significance level α = 0.05. Paired t-test, Wilcoxon signed rank test, or Tukey-Kramer test for post-hoc analysis was applied to compare the parameters between groups where appropriate.

3 |. RESULTS

3.1 |. CEST protocol optimization

In order to optimize our CEST imaging sequence for detecting the iopamidol signals at 4.3 and 5.5 ppm in kidney tissue, we performed numerical simulations based on a six-pool Bloch-McConnell model with simulated MTR_asym curves shown in Figure 1 and MTR curves shown in Supporting Information Figure S1, which is available online. With t_sat set to 3 s, the CEST contrast at 4.3 ppm increases with saturation field strength (B₁) while in contrast the 5.5 ppm signal decreases above 2.6 μT (Figure 1A). In addition, as expected higher saturation B₁ also leads to a broadening of both peaks. Trying to maximize the size of both signals, a B₁ = 3.6 μT was selected. We then simulated at this B₁ how t_sat will impact contrast (Figure 1B). Based on our simulations, t_sat >1.8 s will result in strong CEST signals at both frequencies. To account for the potential loss in labeling efficiency due to B₀ and B₁ inhomogeneities and differences in relaxation rates, a slightly longer saturation pulse t_sat = 3 s was chosen. Our simulations also indicate that for the optimized set of saturation parameters (B₁ = 3.6 μT, t_sat = 3 s), the maximal iopamidol contrast can be achieved with a short recovery time (T_rec) of around 1 s (Figure 1C). Nevertheless, the optimal T_rec depends on the longitudinal relaxation rate of solute and solvent pools (R_1B and R_1A) and is therefore affected by deviations in R₁ from the one used in our simulations. A T_rec of 2 s was eventually selected for our in vivo study to reduce the specific absorption rate (SAR) to 60% since we will acquire for prolonged periods to monitor iopamidol excretion. As shown in Figure 1D, the CEST signals at 4.3 and 5.5 ppm and their ratio were also only slightly affected by the longitudinal water relaxation in the range of R_1A values observed in vivo. These minor differences in R₁ can be effectively removed by ratiometric analysis.³⁴ With the optimized pre-saturation parameters of t_sat = 3 s and T_rec = 2 s combined with an acquisition matrix of 64 × 64 and a single-shot RARE readout, the TR was 5125 ms yielding a temporal resolution of ~10 s for the two-offset protocol, which has been demonstrated to be sufficient to monitor intra-renal signal changes in previous studies.⁴³ As we aimed to use a single CEST sequence to obtain both renal pH maps and time-activity curves, we decided to acquire CEST images at only two frequency offsets that are needed for the ratiometric pH analysis. Based on our results of the BM simulations, the larger CEST signal (4.3 ppm) was used to assess renal perfusion and excretory function.

FIGURE 1.

MTR_asym curves obtained using numerical Bloch-McConnell simulations at different (A) B₁, (B) t_sat, (C) T_rec, and (D) R_1A

3.2 |. Dynamic iopamidol-enhanced MR-CEST urography

Taking inspiration from how MAG3 nuclear scintigraphy data is used to characterize differential kidney function, we proceeded with an in vivo evaluation of injecting iopamidol and performing CEST imaging in such a way that we could examine the relative CEST contrast in each kidney and how the contrast evolves after injection, i.e. time-activity curves. Specifically, we decided to characterize the uptake and excretion of iopamidol on nine mice with UUO (three for each time point after surgery) and three control mice. Representative T2w images of a healthy mouse and UUO mice are shown in Figure 2 A–D. These display a gradual dilatation of the renal pelvis and ureter in the obstructed kidney (hydronephrosis) with time after ligation, confirming progression in renal injury caused by UUO. By 3 days, a disappearance of the papilla and thinning of the medulla could be observed. We manually segmented both obstructed and contralateral kidneys excluding the collecting system to depict how contrast varies with time after injection. As presented in Figure 2 E–H, prolonged contrast was measured in UUO compared to native kidneys. This can also be observed in MTR maps at 4.3 ppm as is shown in Figure 3 for a representative UUO mouse at day 1 after obstruction. In the contralateral kidney, peak medullar enhancement occurred between 1 and 1.5 min after injection, whereas in the UUO kidney maximum enhancement in the medulla was reached between 1.5 and 2 min. Based on this data, we could conclude that our surgery was successful in obstructing the ureter, and this obstruction resulted in an injury that evolved with time.

FIGURE 2.

T2w images and dynamic CEST signal enhancement curves measured in (A, E) a control mouse, (B, F) UUO mice at day 1, (C, G) day 2, and (D, H) day 3 after obstruction. Blue and red contour indicate ROIs in the contralateral and UUO kidney, respectively

FIGURE 3.

MTR images at 4.3 at different time points following iopamidol injection at day 1 after UUO. Blue and red contour indicate ROIs in the contralateral and UUO kidney, respectively

We next performed a group analysis of our time activity data on these mice. As can be seen in Figure 4, CEST MRI time-activity curves showed clear differences between the healthy and UUO kidneys. For the healthy mice, the dynamic CEST curves of both kidneys were nearly identical and displayed rapid excretion of the contrast agent. In the UUO mice, the time-activity curve for the obstructed kidneys showed prolonged contrast excretion with decreased ST_enh% values with time after ligation compared with the contralateral kidneys. As shown in Figure 4, TTP was significantly longer in the UUO kidney than in the contralateral kidney (p < 0.001). Differences in TTP increased further as the disease progressed. At the same time, the peak %ST_enh values were slightly decreased in the obstructed kidney, although this was not statistically significant (p > 0.05).

FIGURE 4.

A comparison of maximal CEST signal enhancement (%ST_enh) (A) and mean TTP values (C) and %ST_enh (B) and TTP distribution (D) in obstructed and contralateral kidneys at different time points post-surgery

3.3 |. CEST-MRI pH mapping

In addition to the time-activity curves, iopamidol-enhanced dynamic CEST MRI also provides truly quantitative and spatially resolved information about tissue pH. To assess whether pH might be a valuable biomarker for early diagnosis of obstructive nephropathy, we performed the ratiometric CEST-MRI pH imaging on our mice and compared the mean pH values and pH distribution between the control and UUO kidneys. Representative renal pH maps overlaid on M₀ images, and pH histograms obtained in a healthy mouse, and UUO mice at day 1, day 2, and day 3 after obstruction are displayed in Figure 5. Whereas for the healthy kidneys a pH gradient from the cortex and outer medulla (more neutral pH) to the inner medulla and papilla (more acidic pH) could be observed, the obstructed kidneys showed a slight alkalinization of the regions corresponding to the inner and outer medulla. The pH histograms demonstrated that both kidneys of the control mouse had very similar pH distributions with a peak at pH 6.7, while in the UUO mice the distribution was shifted towards higher pH values in the obstructed kidney compared with the unobstructed one. We also performed group analysis on these mice with the results presented in Figure 6. In the control mice, both kidneys showed similar pH values. As early as 1 day after UUO, there was an increase in pH values in the obstructed kidney compared with the contralateral one. The same trend was observed at day 2, and day 3 after obstruction. The box-plot analysis for all nine UUO mice revealed significantly increased pH values in the UUO kidneys compared with the native kidneys (p < 0.01). All the CEST-derived semi-quantitative parameters are summarized in Table 2. Based on this mouse data, we conclude that DCE-MR CEST urography based pH imaging can detect differential function between kidneys after UTO.

FIGURE 5.

CEST pH maps and pH histograms measured in a control mouse (A,E), UUO mice at day 1 (B,F), day 2 (C,G), and day 3 (D,H) after obstruction

FIGURE 6.

A comparison of mean pH values (A) and pH differences between UUO and contralateral kidneys (B)

TABLE 2.

A summary of CEST-derived parameters (TTP, %ST_enh, and pH) in the obstructed and native kidneys

	Kidney	TTP (s)	%STenh	pH
Control (n = 3)	Left	72 ± 27	17 ± 13	6.65 ± 0.04
	Right	72 ± 27	17 ± 13	6.67 ± 0.02
UUO Day 1 (n = 3)	CL	75 ± 6	13 ± 10	6.63 ± 0.03
	UUO	85 ± 12	12 ± 8	6.77 ± 0.12
UUO Day 2 (n = 3)	CL	69 ± 22	9 ± 6	6.61 ± 0.05
	UUO	89 ± 12	8 ± 5	6.69 ± 0.02
UUO Day 3 (n = 3)	CL	55 ± 5	12 ± 12	6.77 ± 0.06
	UUO	79 ± 12	7 ± 5	6.89 ± 0.02

3.4 |. Histologic evaluation

H&E staining demonstrates progressive glomerular atrophy, tubular dilation, and tubular atrophy in the obstructed kidney (Supporting Information Figure S2). Masson’s trichrome staining displayed in Figure 7 show that the extent of interstitial fibrosis (blue stain) gradually increases as the disease progresses. The percentage of fibrosis was significantly higher in the obstructed kidney compared to the contralateral kidney on day 3 (p = 0.0007). The percentage of fibrosis of the obstructed kidney was significantly higher on day 3 compared to both kidneys in the control mouse and ipsilateral kidneys at day 1 and day 2 post UUO (p = 0.0004, p = 0.0004, p = 0.0074, respectively).

FIGURE 7.

Masson trichrome staining of the kidneys of a control mouse, and UUO mice at day 1, day 2, and day 3 post obstruction

4 |. DISCUSSION

In the present study, we have applied DCE-MR CEST urography to a mouse model of obstructive nephropathy and non-invasively assessed kidney function over 3 days after UUO. Our results demonstrate that DCE MR CEST acquisitions combined with a single injection of iopamidol allow simultaneous measurements of both renal uptake/clearance and pH as opposed to the conventional diagnostic techniques which provide only one type of metric to characterize kidney function. Similar CEST MRI measurements have recently been used for monitoring renal perfusion and pH homeostasis in an glycerol-induced bilateral kidney injury³⁵ murine model.⁴⁴ In contrast to that study, we were able to obtain both renal pH maps and time-activity curves using a single CEST acquisition and two-offset protocol, which has shown promising performance in our previous study characterizing perfusion and pH changes associated with a progression to CKD in an methylmalonic academia mouse model.³⁷ The two-offset CEST approach achieves a much higher temporal resolution (10 s in our case) than the conventional full z-spectra acquisition which is required to measure signal changes during the wash-in/wash-out process in the kidney. Moreover, it also allows the use of a moving average filter and signal averaging to reduce the impact of motion on our pH maps, improve SNR and hence the sensitivity of pH measurements. Compared to our previous study, here we increased the spatial resolution while maintaining the same high temporal resolution and acquired a coronal slice instead of axial slice to better depict the renal pelvis and ureter. These improvements were felt to be necessary for detecting the subtle changes observed from suture obstructions.

Urographic contrast media such as iopamidol and ^99mTc-MAG3 are primarily filtered through the glomerulus and primarily not reabsorbed or secreted in the renal tubule,^6,45 therefore, it is possible to monitor their transit through vascular and tubular compartments through DCE-CT, MRI and renal scintigraphy. The resulting contrast kinetics provides a measure of renal function. In patients with tubular dysfunction, ^99mTc-MAG3 renograms show radiotracer retention in the renal parenchyma and delayed probe detection in the pelvis and bladder.^46,47 Similar observations have been reported in a renal scintigraphy study on a mouse model of UUO by Tantawy et al.⁴³ Our results appear to confirm these findings, revealing altered TTP values and prolonged iopamidol clearance in the obstructed kidneys compared with the contralateral kidneys. The histological evaluation showed typical patterns of UUO, including progressive glomerular atrophy, tubular dilation, and tubular atrophy. These findings are consistent with previous histological findings^43,48,49 and indicate that structural changes in UUO kidneys could be assessed by measuring iopamidol excretion via CEST MRI. However, the application of DCE CEST MR urography for estimating perfusion-related parameters may be hampered by the nonlinear relationship between CEST signal enhancement and iopamidol concentration and the influence of relaxation on our measurements which may result in more variability than seen in measurements based on detecting x-ray scattering or detecting emitted gamma radiation.

Besides providing information on renal uptake and clearance of iopamidol, dynamic CEST MRI enables spatially resolved pH measurements. In the present study, no clear difference in peak %ST_enh between the UUO and CL kidney was observed at day 1 and 2 post-obstruction, suggesting that minimally dilated obstructive nephropathy might not cause a significant decrease in perfusion.⁹ Furthermore, the degree of dilation partially depends on the patient’s hydration status and thus in dehydrated patients hydronephrosis might be less evident despite progressing ON. In contrast, as early as day one after UUO we observed increased pH values in the obstructed kidney compared with the contralateral one, suggesting that pH might be a more useful MRI biomarker for the early diagnosis of ON. Similar findings were reported in previous studies in different murine models murine models, which demonstrated elevated renal pH levels in the affected kidneys as a result of glycerol-induced bilateral injury,³⁵ and an unilateral ischemia-reperfusion damage.³⁶ Impaired urinary acidification has also been documented in patients with various types of ON.⁵⁰ Furthermore, animal studies with experimentally induced obstructive nephropathy have reported a transport defect in the distal nephron accompanied by impaired sodium reabsorption and reduced excretion of hydrogen and potassium.^51,52 In both animals and patients following release of obstruction, there was no decrease in urinary pH in response to acid loading, indicating that obstruction substantially impact the ability of the distal nephron to acidify the urine.^53–55

The structural and molecular changes associated with obstructive renal damage have previously been assessed in a murine model of UUO using endogenous CEST MRI instead of applying a contrast agent such as iopamidol.³⁰ While no regional changes in CEST signals between 1 and 3 ppm were detected, signals from 3 to 5 ppm and −5 to −1 ppm were higher in medulla and papilla of UUO kidney compared with CL kidney. Nevertheless, the interpretation of pH- and concentration-dependent CEST effects originating from various kidney metabolites is challenging.⁵⁶ In contrast, iopamidol contrast enhanced CEST pH imaging enables quantitative assessment of renal function, and has been proven feasible on clinical scanners.^5,57 Conversely, while iodinated contrast agents are considered safe and widely used, there are concerns regarding their potential nephrotoxicity in patients with preexisting renal impairment.⁵⁸ Although multiple retrospective studies have shown that iopamidol demonstrates a reasonable safety profile,^59,60 there is an increasing interest in developing novel non-iodinated pH-responsive MRI probes with reduced risk of nephrotoxicity.^61,62

Our study has several limitations worth noting. First, temporal resolution of the dynamic CEST acquisition was too low to ensure an adequate sampling of the rapid signal variations occurring during the first past of iopamidol through the vascular and each of tubular compartments. Second, DCE-CEST MR urography provides only semi-quantitative measures of renal perfusion and excretory function, which are hardly comparable with those obtained using conventional gadolinium-enhanced DCE-MRI or multi-detector CT. Third, differences in peak signal enhancement observed between the kidneys and different animals might be partially due to B₀/B₁ field inhomogeneities. Last but not least, unlike the human kidney, the mouse kidney is unilobar with only a single papilla that extends deep into the renal pelvis. Consequently, the translational capability of this animal work to human UTOs remains unknown.

5 |. CONCLUSIONS

Our findings indicate that DCE-MR-CEST urography can detect changes in renal excretion and pH homeostasis and distinguish between obstructed and unobstructed kidney as early as 1 day after UUO. Overall, our results suggest that CEST imaging might be particularly useful for detecting and monitoring the progression of renal injury caused by UUO. On the basis of these findings, we believe that our dynamic CEST MRI protocol is promising for early assessment of upper UTOs and could be translated to patients with obstructive nephropathy.

Supplementary Material

Supp Info

Figure S1: MTR curves obtained using numerical Bloch-McConnell simulations at different (A) B₁, (B) t_sat, (C) T_rec, and (D) R_1A.

Figure S2: H&E stained sections of the kidneys of a control mouse, and UUO mice at day 1, day 2 and day 3 post obstruction.