Clinical Radiology 69 (2014) 1117e1122

Contents lists available at ScienceDirect

Clinical Radiology journal homepage: www.clinicalradiologyonline.net

Assessment of renal fibrosis in chronic kidney disease using diffusion-weighted MRI J. Zhao a, Z.J. Wang b, M. Liu c, J. Zhu a, X. Zhang a, T. Zhang c, S. Li a, Y. Li c, * a

Department of Radiology, The Third Hospital of Hebei Medical University, Shijiazhuang, China Department of Radiology and Biomedical Imaging, University of California, San Francisco (UCSF), San Francisco, CA, USA c Department of Nephrology, The Third Hospital of Hebei Medical University, Shijiazhuang, China b

art icl e i nformat ion Article history: Received 12 January 2014 Received in revised form 19 May 2014 Accepted 10 June 2014

AIM: To assess the performance of diffusion-weighted magnetic resonance imaging (MRI) for the assessment of renal fibrosis in chronic kidney disease (CKD), with histopathology as a reference standard. MATERIALS AND METHODS: Forty patients with CKD and 30 healthy volunteers were recruited for the study. All participants underwent diffusion-weighted MRI. Renal biopsy was performed in 25 patients with CKD. Mean renal medullary and cortical apparent diffusion coefficient (ADC) values were compared between CKD patients and the healthy volunteers. Pearson’s correlation coefficient was calculated to investigate the relationship between ADC values, serum creatinine (SCr), estimated glomerular filtration rate (eGFR), 24 h urinary protein (24h-UPRO), and renal histopathological scores. RESULTS: Cortical and medullary ADC values in the CKD group were significantly lower compared to those in the healthy controls. In the CKD group, a significant negative correlation was found between cortical ADC values and SCr/24h-UPRO, and significant positive correlation was found between cortical ADC and eGFR. There was also a significant negative correlation between medullary ADC values and SCr. Both cortical and medullary ADC values were significantly correlated with histopathological fibrosis score. CONCLUSION: Renal ADC values strongly correlate with histological measures of fibrosis, and have the potential to enhance the non-invasive monitoring of chronic kidney disease. Ó 2014 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.

Introduction Chronic kidney disease (CKD) is characterized by progressive loss of kidney function due to tubular atrophy, interstitial fibrosis, and reduced blood flow. There are about 26 million Americans diagnosed with CKD, which brings an

* Guarantor and correspondent: Y. Li, Department of Nephrology, The Third Hospital of Hebei Medical University, 139 Ziqiang Road, Shijiazhuang, Hebei PC 050051, China. Tel.: þ86 311 88603000; fax: þ86 311 88603610. E-mail address: [email protected] (Y. Li).

enormous burden to both the individual patient and society.1 CKD may be caused by a variety of renal diseases. The final common pathway leading to CKD and end-stage renal disease is renal fibrosis.2 The ability to accurately and noninvasively evaluate the presence and extent of renal fibrosis would enhance the monitoring of CKD progression, and may better predict the long-term outcome of patients with CKD.3 However, currently available clinical indicators of kidney disease lack the sensitivity or specificity to measure renal fibrosis. Although renal biopsy can diagnose fibrosis, it is invasive and prone to sampling errors, and is not used routinely to monitor the progression of fibrosis.

http://dx.doi.org/10.1016/j.crad.2014.06.011 0009-9260/Ó 2014 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.

1118

J. Zhao et al. / Clinical Radiology 69 (2014) 1117e1122

Functional MRI techniques have been applied to assess early renal damage non-invasively and quantitatively.4e6 One of these techniques is diffusion-weighted (DW) magnetic resonance imaging (MRI). DW imaging (DWI) is sensitive for the detection of the combined effects of blood microcirculation and Brownian motion of water molecules, which is expressed as an apparent diffusion coefficient (ADC), within the tissues. High ADC values suggest a large amount of water movement, whereas low ADC values suggest restricted water movement. DWI technique is particularly promising for non-invasive assessment of renal disease given the large blood flow in the kidneys. Several studies have reported using DWI methods to evaluate renal function in both native and transplanted kidneys.7,8 Other studies have demonstrated the relationship between DWI and renal fibrosis or renal function.6,9e16 However, those previous studies either were focused on experimental models or did not have histopathology reference for clinical studies. To the present authors’ knowledge, the assessment of renal fibrosis in patients with CKD using 3 T MRI with histopathology as a reference standard has not been previously reported. The aim of the present study was to evaluate whether ADC values from DW MRI are correlated with renal fibrosis in CKD.

Materials and methods The study was approved by the institutional ethics committee. Written informed consent was obtained from all participants.

Participants Forty patients (25 male, 15 female, mean age 41.6  17.1 years, range 16e71 years) with a clinical diagnosis of CKD at The Third Hospital of Hebei Medical University were recruited for participation in the study. Thirty healthy volunteers (20 male and 10 women; average age 38.3  14.1 years, range 24e60 years) were also recruited and served as controls. All participants underwent both standard MRI and DW MRI. No participants were taking medications, such as angiotensin converting enzyme inhibitors (ACEi) or angiotensin receptor blockers (ARBs), prior to their MRI examination. SCr and 24 h urinary protein (24h-UPRO) were obtained within 2 weeks of MRI. Estimated glomerular filtration rate (eGFR) were calculated from serum creatinine (SCr) measurements for all participants using the Chronic

Kidney Disease Epidemiology Collaboration (CKD-EPI) equation.17 The units of SCr were milligrams per decilitre. Renal biopsy was performed in 25 patients with CKD within 2 weeks following the MRI examination. The grades of CKD for each patient are: grade I, 25 cases; II, eight cases; III, five cases; IV, two cases; V, none. Detailed demographics of CKD patients who underwent biopsy are summarized in Table 1.

Imaging protocol MRI was performed using a 3 T Imager (Verio; Siemens, Germany) and a six-channel body coil. The following images were acquired: coronal T2-weighted half-Fourier single-shot fast spin-echo sequence (HASTE), coronal T1weighted fast low angle shot gradient-echo sequence (T1WI), and coronal multisection echo-planar DWI. The parameters for HASTE were: 1500 ms TR, 89 ms TE; 4 mm section thickness, 0.8 mm intersection gap; 400  400 mm field of view, 256  256 matrix, one signal average, two concatenations, and partial Fourier factor ¼ 5/8. The parameters for T1WI were: 240 ms TR, 2.31 ms TE; 4 mm section thickness, 0.8 mm intersection gap, 400  400 mm field of view, 256  256 matrix, one signal average, two concatenations, and partial Fourier factor ¼ 5/8. DW-MRI was performed with the following parameters: 5600 ms TR, 76 ms TE; 4 mm section thickness, 0.8 mm intersection gap, 400  400 mm field of view, 160  160 matrix, five signal average; 1644 Hz/pixel bandwidth, and partial Fourier factor ¼ 6/8. Diffusion gradient b-values were 0 and 800 s/mm2. The gradients were applied in three directions and subsequently averaged to minimize the effects of diffusion anisotropy. A parallel imaging technique with a reduction factor of 2 was applied. Respiratory gating was used to minimize the motion artefacts. The total scan time for each DWI acquisition was 4e5 min.

Image analysis All renal DW images were reviewed by two abdominal radiologists, and intraclass correlation coefficients were calculated to test the reproducibility of renal ADC values. Both readers were blinded to the participants’ clinical information. ADC maps of the kidney were generated using software on the MRI workstation. For each participant, the coronal sections closest to the renal hilum of the left and right kidneys were selected for region of interest (ROI) analysis. A

Table 1 Clinical characteristics of patients with chronic kidney disease (CKD) who underwent biopsy.

Membranous nephropathy Minimal change nephropathy Atypical membranous nephropathy Focal proliferative IgA nephropathy Mesangial proliferative glomerulonephritis Hypertensive nephropathy

Case (#)

Male (%)

Age (year)

6 5 5 5 3 1

83.3 60 33.3 80 100 100

55.8 26.6 43.3 33.6 44.7 37

    

14.4 9.1 17.6 16.8 25.4

SCr (mmol/l) 57.7 64.4 79.3 81.9 68.0 101

    

13.9 19.3 31.1 25.6 10.6

24h-UPRO (g)

eGFR (ml/min/1.73 m2)

7.5  5.0  8.1  1.6  7.1  0.31

110.9 124.0 91.5 108.2 121.1 81.7

5.5 4.2 6.0 1.5 10.7

Results are expressed as mean  standard deviation. F, female; M, male; #, case number; SCr, serum creatinine; 24h-UPRO, 24 h-urinary protein; eGFR, estimated glomerular filtration rate.

    

10.9 24.6 45.5 35.3 38.8

J. Zhao et al. / Clinical Radiology 69 (2014) 1117e1122

minimum of 10 pixels was included in each ROI to reduce noise. For each image section, ROIs (range 10e35 pixels) were placed at the upper, middle, and lower pole of each kidney in the medulla and cortex on the ADC map (Fig 1). Three ROIs were placed in medulla and three ROIs were placed in cortex of each kidney, with care to avoid vessels, renal sinus, renal cysts, and susceptibility artefacts. For atrophic kidneys with indistinct corticomedullary boundary, ROIs were placed as close to the outer-edge of the cortex as possible. Medullary ROIs were placed between the expected region of the cortex and renal pelvis (Fig 2). For each participant, mean cortical ADC values were obtained by averaging six cortical ROIs obtained from both kidneys. Similarly, mean medullary ADC values were obtained by averaging six medullary ROIs obtained from both kidneys. The average ADC values of the two observers were used for data analysis.

Renal histopathology Ultrasound-guided kidney biopsy was performed in 25 patients with CKD. The specimens were fixed in buffered 10% formalin, embedded in paraffin, and sectioned into

1119

3 mm-thick sections by using a standard procedure. The periodic acidesilver methenamine (PASM) stain, periodic acideSchiff (PAS), haematoxylin and eosin (HE) stain, and the Masson stain were used for histopathological scoring of renal fibrosis. All slides were read by a pathologist with more than 20 years of experience in renal disease. In order to better quantify the histological lesions, a scoring system based on the standardized scoring system for chronicity developed for lupus nephritis was employed. This scoring system was extended to other glomerulopathies because biological processes that take place in vasculitis or lupus nephritis are also present in other types of primary or secondary glomerular diseases.18 The glomeruli, tubules, interstitium, and peri-tubular vessels were assessed separately for fibrotic lesions. The results were used to derive the fibrosis score. Glomeruli were divided into eight segments to assess the presence of glomerular fibrotic lesions. A segment was recorded positive if there was increased mesangial matrix or sclerosis. The numbers of positive segments in the glomeruli were used to calculate the percentage of the glomeruli affected. A score of 1 corresponds to80% affected glomeruli. The presence of cellular crescents was also recorded. The percentage of glomeruli either partially or completely surrounded by crescents was calculated. Fibrous crescents were counted only when they partially surround the glomeruli, circumferential fibrous crescents were recorded as a sclerosed glomerulus. Tubulo-interstitial lesions, such as interstitial fibrosis, tubular atrophy and casts, were assessed in a semiquantitatively manner.18 Arterial vascular lesions and additional abnormalities were semi-quantitatively evaluated (Table 2).

Statistical analysis All values were expressed as means  standard deviations. Data on intra-observer agreement were determined by intraclass correlation coefficients (ICC). ICC is one of the indicators to evaluate interobserver reliability and testeretest reliability. It is a better reliability test for both

Data from 35 out of 40 patients were successfully analysed. Data from five patients were not included due to severe respiratory artefacts in four patients, and increased susceptibility due to air-filled colon loops leading to poor image quality in one patient. Data from all of the healthy controls could be analysed. The ICC of renal cortical ADC values and medullary ADC values were 0.79 and 0.82, respectively. CKD patients show a wider range of ADC values than controls. The ranges of ADC values for CKD patients were 1.72e2.98 mm2/s for the cortex, and 1.46e2.97 mm2/s for the medulla. For controls, the values were 2.09e2.70 mm2/s for the cortex, and 1.9e2.45 mm2/s for the medulla. Cortical and medullary ADC values in patients with CKD were significantly lower when compared to those from healthy controls (Table 3). In the CKD group, a significant negative correlation was found between cortical ADC values and SCr/24h-UPRO, and significant positive correlation was found between cortical ADC values and the eGFR. There was also a significant negative correlation between the medullary ADC values and the SCr values. No significantly

J. Zhao et al. / Clinical Radiology 69 (2014) 1117e1122

1121

Table 3 Comparison of ADC values of renal cortex and medulla between controls and CKD patients.

Cortical ADC (103 mm2/s) Medullary ADC (103 mm2/s)

Controls

CKD patients

t

p-Value

2.57  0.22

2.41  0.33

2.25

0.028

2.37  0.27

2.17  0.34

2.57

0.012

Two-tailed Student’s t-test was used for the analysis. ADC, apparent diffusion coefficient; CKD, chronic kidney disease.

correlation was found between medullary ADC values and eGFR/24h-UPRO. There was a significant inverse correlation between cortical ADC and medullary ADC values and the histological fibrosis score (Table 4; Figs 3 and 4). Figure 3 Scatter plot with interpolation line showing linear correlation between renal cortical ADC values and histological fibrosis score. R ¼ e0.78, p < 0.05.

Discussion DW MRI has been applied to evaluate the Brownian motion of water molecules in vivo. The ADC values of renal parenchyma are dependent on Brownian motion of water and other factors, such as capillary perfusion and tubular flow. Several studies have reported that patients with renal failure have significantly lower ADC values in the cortex and medulla than healthy subjects.12,13 Similar to those studies, lower renal ADC values were also observed in CKD patients in the present study. A significant negative correlation was found between cortical ADC values and SCr/24h-UPRO, and a significant positive correlation was found between cortical ADC and eGFR values. Decreases in ADC values have been previously shown to reflect interstitial fibrosis in a mouse unilateral ureteral obstruction model.11 However, the histological correlation to ADC values in patients with CKD has not been specifically investigated before. In the present study, DW MRI was applied to examine its potential for detecting renal fibrosis in patients with varying degrees of renal impairment, and an inverse relationship was found between renal ADC values and the histopathological fibrosis score in patients with CKD. The predominant cause of the lower ADC values in fibrotic kidneys may be due to higher cell density. The accumulation of cells would lead to higher cell density, including fibroblasts in the interstitial space, which typically are observed during renal fibrogenesis.11 High cell density could decrease the extracellular water fraction and increase the intracellular water fraction, which in turn may lead to lower ADC valuess.20 Although controversial, collagen deposition in the kidney may be another factor Table 4 Pearson’s correlation coefficient.

Cortical ADC Medullary ADC

SCr

24h-UPRO

eGFR

Histopathological fibrosis score

0.55a 0.35a

0.34a 0.20

0.43a 0.31

0.78a 0.64a

ADC, apparent diffusion coefficient; SCr, serum creatinine; 24h-UPRO, 24 hurinary protein; eGFR, estimated glomerular filtration rate. a p < 0.05.

that may contribute to reduce water motion and ADC values.21 It is interesting that medullary ADC negatively correlated with SCr, but did not positively correlate with eGFR. This implies that age is not an independent predictor of renal fibrosis. The present study showed that ICC values of the renal cortex and medulla were 0.79 and 0.82, respectively, which indicate good reliability between the two observers. Previous studies have shown that the average ADC values in the kidney are reproducible. The ADC values from the present study were in good agreement with those reported in the literature.22,23 Previous studies indicated that ADC at low b-values (0e300 s/mm2) could be affected by local blood and urinary flow.24 To minimize capillary perfusion and tubular flow, b-values of 0 and 800 s/mm2 were used to calculate the ADC values. Furthermore, three ADC values calculated in each orthogonal direction were averaged to reduce possible anisotropic effects that have been shown to be associated with renal ADC.6 Motion artefacts were minimized by using respiratory gating. The present study was performed using a 3 T system, in order to take advantage of the higher signal-to-noise ratio (SNR) and higher temporal and spatial resolution at higher field. To avoid potential drawbacks of higher field strength, including larger chemical shift and more prominent susceptibility effect,25 parallel imaging techniques were employed, which reduce the echo spacing and echo train length,6 and minimize susceptibility artefacts. The results of the present study showed significant correlation between renal ADC values and pathological fibrosis score in patients with CKD. This suggests that diffusion of water molecules is restricted in the kidney of CKD patients with fibrosis. DWI, therefore, has the potential to quantitatively and noninvasively detect renal fibrosis in patients with CKD, and long-term non-invasive monitoring is possible. The present study has limitations. Each patient was only studied with DW MRI at a single time point, and longitudinal studies are needed to determine whether this technique can monitor the progression of fibrosis over time and

1122

J. Zhao et al. / Clinical Radiology 69 (2014) 1117e1122

Figure 4 Scatter plot with interpolation line showing linear correlation between renal medullary ADC values and histological fibrosis score. R ¼ e0.64, p < 0.05.

following treatment. Another limitation was the number of patients with the same abnormality is relatively small. A larger sample size is needed to analyse different abnormalities in more detail in future studies. In conclusion, ADC values obtained using a 3 T MRI system can be used to quantitatively assess the degree of renal fibrosis in patients with CKD. This technique is of particular relevance to future clinical trials addressing earlier detection and quantification of renal fibrosis, and evaluation of response to novel therapies. Larger clinical studies are needed to further validate the utility of DW MRI for longitudinal monitoring of renal fibrosis in patients with CKD.

Acknowledgement The authors thank Benjamin M. Yeh, MD, and Songling Liu, MD, for the help with reviewing the manuscript. The authors thank Ya Li, MD, for the help with the histological procedures.

References 1. Coresh J, Selvin E, Stevens LA, et al. Prevalence of chronic kidney disease in the United States. JAMA 2007;298:2038e47. 2. Hewitson TD. Renal tubulointerstitial fibrosis: common but never simple. Am J Physiol Ren Physiol 2009;296:F1239e44. 3. Shiiki H, Saito T, Nishitani Y, et al. Prognosis and risk factors for idiopathic membranous nephropathy with nephrotic syndrome in Japan. Kidney Int 2004;65:1400e7. 4. Chandarana H, Lee VS. Renal functional MRI: are we ready for clinical application? AJR Am J Roentgenol 2009;192:1550e7.

5. Kim S, Naik M, Sigmund E, et al. Diffusion-weighted MR imaging of the kidneys and the urinary tract. Magn Reson Imaging Clin N Am 2008;16:585e96. viieviii. 6. Thoeny HC, De Keyzer F, Oyen RH, et al. Diffusion-weighted MR imaging of kidneys in healthy volunteers and patients with parenchymal diseases: initial experience. Radiology 2005;235:911e7. 7. Namimoto T, Yamashita Y, Mitsuzaki K, et al. Measurement of the apparent diffusion coefficient in diffuse renal disease by diffusionweighted echo-planar MR imaging. J Magn Reson Imaging 1999;9:832e7. 8. Thoeny HC, Zumstein D, Simon-Zoula S, et al. Functional evaluation of transplanted kidneys with diffusion-weighted and BOLD MR imaging: initial experience. Radiology 2006;241:812e21. 9. Xu X, Fang W, Ling H, et al. Diffusion-weighted MR imaging of kidneys in patients with chronic kidney disease: initial study. Eur Radiol 2010;20:978e83. 10. Zhang JL, Sigmund EE, Chandarana H, et al. Variability of renal apparent diffusion coefficients: limitations of the monoexponential model for diffusion quantification. Radiology 2010;254:783e92. 11. Togao O, Doi S, Kuro-o M, et al. Assessment of renal fibrosis with diffusion-weighted MR imaging: study with murine model of unilateral ureteral obstruction. Radiology 2010;255:772e80. 12. Carbone SF, Gaggioli E, Ricci V, et al. Diffusion-weighted magnetic resonance imaging in the evaluation of renal function: a preliminary study. Radiol Med 2007;112:1201e10. 13. Goyal A, Sharma R, Bhalla AS, et al. Diffusion-weighted MRI in assessment of renal dysfunction. Indian J Radiol Imaging 2012;22:155e9. 14. Hueper K, Hartung D, Gutberlet M, et al. Magnetic resonance diffusion tensor imaging for evaluation of histopathological changes in a rat model of diabetic nephropathy. Invest Radiol 2012;47:430e7. 15. Haque ME, Franklin T, Bokhary U, et al. Longitudinal changes in MRI markers in a reversible UUO mouse model: preliminary experience. J Magn Reson Imaging 2013. 16. Inoue T, Kozawa E, Okada H, et al. Noninvasive evaluation of kidney hypoxia and fibrosis using magnetic resonance imaging. J Am Soc Nephrol 2011;22:1429e34. 17. Levey AS, Stevens LA, Schmid CH, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med 2009;150:604e12. 18. Neumann I, Kain R, Regele H, et al. Histological and clinical predictors of early and late renal outcome in ANCA-associated vasculitis. Nephrol Dial Transplant 2005;20:96e104. 19. de Vet HC, Terwee CB, Knol DL, et al. When to use agreement versus reliability measures. J Clin Epidemiol 2006;59:1033e9. 20. Chin CL, Wehrli FW, Hwang SN, et al. Biexponential diffusion attenuation in the rat spinal cord: computer simulations based on anatomic images of axonal architecture. Magn Reson Med 2002;47:455e60. 21. Luciani A, Vignaud A, Cavet M, et al. Liver cirrhosis: intravoxel incoherent motion MR imagingdpilot study. Radiology 2008;249:891e9. 22. Cutajar M, Clayden JD, Clark CA, et al. Test-retest reliability and repeatability of renal diffusion tensor MRI in healthy subjects. Eur J Radiol 2010;80:e263e268. 23. Damasio MB, Tagliafico A, Capaccio E, et al. Diffusion-weighted MRI sequences (DW-MRI) of the kidney: normal findings, influence of hydration state and repeatability of results. Radiol Med 2008;113:214e24. 24. Yamada I, Aung W, Himeno Y, et al. Diffusion coefficients in abdominal organs and hepatic lesions: evaluation with intravoxel incoherent motion echo-planar MR imaging. Radiology 1999;210:617e23. 25. Inan N, Akhun N, Akansel G, et al. Conventional and diffusionweighted MRI of extrahepatic hydatid cysts. Diagn Interv Radiol 2010;16:168e74.

Assessment of renal fibrosis in chronic kidney disease using diffusion-weighted MRI.

To assess the performance of diffusion-weighted magnetic resonance imaging (MRI) for the assessment of renal fibrosis in chronic kidney disease (CKD),...
610KB Sizes 0 Downloads 4 Views