Intravoxel incoherent motion imaging of the kidney: alterations in diffusion and perfusion in patients with renal dysfunction

Magnetic Resonance Imaging 31 (2013) 414–417

Contents lists available at SciVerse ScienceDirect

Magnetic Resonance Imaging journal homepage: www.mrijournal.com

Intravoxel incoherent motion imaging of the kidney: alterations in diffusion and perfusion in patients with renal dysfunction Shintaro Ichikawa, Utaroh Motosugi, Tomoaki Ichikawa ⁎, Katsuhiro Sano, Hiroyuki Morisaka, Tsutomu Araki Department of Radiology, University of Yamanashi, Yamanashi 409–3898, Japan

a r t i c l e

i n f o

Article history: Received 17 March 2012 Revised 11 July 2012 Accepted 30 August 2012 Keywords: Kidney Diffusion-weighted imaging (DWI) Intravoxel incoherent motion (IVIM) Magnetic resonance imaging (MRI)

a b s t r a c t Purpose: To investigate the relationship between estimated glomerular filtration rate (eGFR) and parameters calculated using intravoxel incoherent motion (IVIM) imaging of the kidneys. Materials and Methods: We studied 365 patients, divided into 4 groups based on eGFR levels (mL/min/ 1.73 m2): group 1, eGFR≥80(n=80); group 2, eGFR 60–80 (n=156); group 3, eGFR 30–60 (n=114); and group 4 ,eGFRb30 (n=15). IVIM imaging was used to acquire diffusion-weighted images at 12 b values. The diffusion coefficient of pure molecular diffusion (D), the diffusion coefficient of microcirculation or perfusion (D*), and perfusion fraction (f) were compared among the groups using group 1 as control. Results: In the renal cortex, D* values were significantly lower in groups 2, 3, and 4 than in group 1. The D value of renal cortex was significantly low in only group 3. In the renal medulla, the D* and D values were significantly lower only in groups 2 and 3, respectively. Conclusion: As renal dysfunction progresses, renal perfusion might be reduced earlier and affected more than molecular diffusion in the renal cortex. These changes are effectively detected by IVIM MR imaging. © 2013 Elsevier Inc. All rights reserved.

1. Introduction Tissue perfusion is typically imaged by scintigraphy, angiography, dynamic contrast-enhanced computed tomography (CT), or magnetic resonance (MR) imaging [1–3]. Arterial spin labeling is also used to evaluate tissue perfusion and does not require use of a contrast agent [4–6]. Intravoxel incoherent motion (IVIM), which was first described by Le Bihan et al. in 1986 [7], is an interesting imaging technique for the estimation of tissue perfusion by calculation of diffusivity parameters using multi-b-value diffusion-weighted MR imaging. Recent advances in MR imaging have facilitated the application of IVIM imaging in the separate estimation of tissue perfusion and diffusivity of protons in abdominal organs [8]. Several recent studies have demonstrated the utility of IVIM imaging for distinguishing between cirrhotic and non-cirrhotic liver, pancreatitis and pancreatic carcinoma, and enhancing and non-enhancing renal lesions [9–12]. Glomerular filtration rate (GFR) is considered the best indicator of renal function. GFR can be estimated by creatinine clearance or renal scintigraphy using 99mTc. One advantage of scintigraphy is the ability to estimate GFR of both kidneys separately, while the relatively high radiation exposure is a drawback. However, several studies have shown that the apparent diffusion coefficient (ADC) of the kidney assessed by diffusion-weighted imaging (DWI) is well correlated with ⁎ Corresponding author. Tel.: +81 55 273 1111; fax: +81 55 273 6744. E-mail address: [email protected] (T. Ichikawa). 0730-725X/$ – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.mri.2012.08.004

renal function or renal fibrosis [13–15]. Tissue perfusion, however, is expected to be strongly affected by changes in renal function such that the renal changes in ADCs may result from perfusion effects rather than changes in true diffusivity of the protons in the tissue [16–19]. In fact, ADCs measured in previous studies using DWI were highly variable based on the b values used, perhaps due to the perfusion effect largely obscuring the signal changes in DWI [20]. Because IVIM MR imaging can estimate tissue perfusion and diffusivity of the protons separately, we postulated that the procedure would more accurately assess renal function than conventional DWI [21,22]. Calculation of the estimated glomerular filtration rate (eGFR) was proposed as a simplified method to estimate renal function using only serum creatinine level, sex, and age of the patient [23–25]. We adopted the eGFR as a reference standard in this preliminary study of the application of IVIM MR imaging in estimating renal function. The purpose of the study was to investigate the relationship between eGFRs and parameters calculated using IVIM MR imaging of the kidneys. 2. Materials and methods 2.1. Patients The applicable institutional review board approved the study. Written informed consent was waived by the board. We retrospectively evaluated 365 patients (210 men and 155 women; age range 13– 92 years, mean age 67.4 years) who underwent abdominal magnetic resonance imaging (MRI) with IVIM-DWI and the blood test of serum

S. Ichikawa et al. / Magnetic Resonance Imaging 31 (2013) 414–417 Table 1 Baseline characteristics of patient groups

eGFR (mL/min/1.73 m2) No. of patients Age (years) Sex (men:women) Weight (kg)

Group 1

Group 2

Group 3

Group 4

N80 80 60.5±13.5 42:38 56.7±13.7

60–80 156 67.3±10.3 90:66 54.9±16.7

30–60 114 72.8±9.28 66:48 54.7±14.1

b30 15 67.1±10.8 11:4 59.1±12.0

creatinine level at our institute between May 2010 and January 2011. The patients’ clinical charts were reviewed to obtain serum creatinine levels. The time interval between MR examination and blood test was 3.02±3.95 days (mean±standard deviation). Estimated GFR (eGFR) was calculated based on serum creatinine level (Cr) using the following equations proposed by the Japanese Society of Nephrology:   2 −0:287 −1:094 Men : eGFR mL = min = 1:73 m = 194 × Age × Cr   2 −0:287 −1:094 Women : eGFR mL = min = 1:73 m = 194 × Age × Cr × 0:739 The patients were divided into 4 groups according to the eGFR levels: group 1, estimated eGFR≥80 mL/min/1.73 m 2 (n=

415

80); group 2, 60≤eGFRb80 mL/min/1.73 m 2 (n=156); group 3, 30≤eGFRb60 mL/min/1.73 m 2 (n=114); group 4, eGFRb30 mL/ min/1.73 m 2 (n=15). See Table 1 for demographic data. 2.2. MRI MRI was performed on all patients using a superconducting magnet operating at 1.5 T (Signa EXCITE HD; GE Medical Systems, Milwaukee, WI, USA) and an 8-channel phased-array coil. The IVIM-DWI was acquired in the transverse plane by respiratory triggered spin echo-echo planar imaging (EPI). A section thickness of 7 mm was applied, while the intersection gap was varied (3–7 mm) to cover the whole liver and right kidney. The repetition time (TR; ms)/echo time (TE; ms) wasN5000/95; flip angle (FA), 90°; number of signals acquired (NSA), 3; field of view (FOV), 40– 50×30–40 cm; matrix size, 96×96; scan time, 240–300 s; parallel imaging (ASSET) factor, 2; MPG pulse, x-axis. Twelve b values were used: 0, 10, 20, 30, 40, 50, 80, 100, 200, 400, 800, and 1000 s/mm 2. 2.3. Image analysis Three IVIM parameters were set as follows: the diffusion coefficient of slow or non-perfusion-based molecular diffusion

Fig. 1. Parameter maps generated from IVIM MR images of a 65-year-old man in the group A (eGFR=86.49 mL/min/1.73 m2). (A) ADC map (B) D map reflecting pure molecular diffusion. (C) D* map reflecting perfusion or microcirculation. (D) f map reflecting perfusion fraction. (E) DW image obtained with b value of 0 s/mm2.

416

S. Ichikawa et al. / Magnetic Resonance Imaging 31 (2013) 414–417

3. Results

Table 2 IVIM parameters by group eGFR (mL/min/ 1.73 m2)

Group 1

Group 2

Group 3

Group 4

N80

60–80

30–60

b30

Cortex ADC (×10-3 mm2/s) p D⁎ (×10-3 mm2/s) p D (×10-3 mm2/s) p f (%) p

2.28±0.16 Control 110.8±12.4 Control 2.14±0.32 control 23.4±10.6 Control

2.28±0.14 1.0000 106.6±12.1⁎ 0.0237 2.10±0.26 1.0000 24.6±9.07 0.7980

2.17±0.15⁎ b .0001 97.8±10.9⁎ b.0001 1.99±0.26⁎ 0.0007 25.1±10.2 0.7565

2.08±0.17⁎ 0.0001 97.3±12.7⁎ 0.0008 2.08±0.33 1.0000 22.0±8.76 1.0000

Medulla ADC (×10-3 mm2/s) p D⁎ (×10-3 mm2/s) p D (×10-3 mm2/s) p f (%) p

1.89±0.19 Control 40.7±9.82 Control 2.10±0.34 Control 29.7±9.34 Control

1.88±0.15 1.0000 38.6±7.88⁎ 0.0477 2.02±0.30 0.1630 30.3±8.57 1.0000

1.85±0.19 0.1659 39.5±7.52 0.7043 1.91±0.27⁎ b.0001 29.9±9.49 1.0000

1.83±0.21 0.5205 39.0±4.29 1.0000 1.98±0.37 0.5092 35.5±11.2 0.2265

3.1. Renal cortex The D* values of renal cortex were significantly lower in group 2 (106.6±12.1, p=0.0237), group 3 (97.8±10.9, pb.0001), and group 4 (97.3±12.7, p=0.0008) than in group 1 (110.8±12.4). (Table 2). The D value of renal cortex was significantly lower in group 3 (1.99±0.26, p =0.0007) than in group 1 (2.14±0.32), whereas no significant differences were observed between group 2 (2.10±0.26, p=1.0000) and group 4 (2.08±0.33, p=1.0000). No significant differences in the f value of renal cortex were observed between group 1(23.4±10.6) and group 2 (24.6±9.07, p=0.7980), group 3 (25.1±10.2, p=0.7565), or group 4 (22.0± 8.76, p=1.0000). The ADC values of renal cortex were significantly lower in group 3 (2.17±0.15, pb.0001) and group 4 (2.08±0.17, p=0.0001) compared to group 1 (2.28±0.16), whereas group 2 (2.28±0.14, p= 1.0000) was not significantly different from group 1.

Statistical analysis was employed using the Dunn test. Data presented as mean±SD. ADC, apparent diffusion coefficient; D, true diffusion coefficient; D*, diffusion coefficient for perfusion; f, perfusion fraction. ⁎ Significant difference from the control (group A), pb0.05. -3

2

(D;×10 mm /s), which represents pure molecular diffusion; the diffusion coefficient of fast or perfusion-based molecular diffusion (D*;×10 -3 mm 2/s), which represents intravoxel microcirculation or perfusion; and perfusion fraction (f; %). These IVIM parameters were calculated using the IVIM model equations described by Le Bihan et al.: Sb/S0=(1 − f)×exp (−bD)+f×exp (−b (D+D*)). D was initially determined by monoexponential fitting using DWI with b values more than 200 s/mm 2. D* and f were estimated using the least squares approach. The calculation was performed by SYNAPSE VINCENT software (FUJIFILM Medical, Tokyo, Japan), which provided the D, D*, and f parameters mapped on a pixel-by-pixel basis. For comparison, the conventional apparent diffusion coefficient (ADC) was also calculated using 2 b values: 0 and 1000 s/mm 2. To obtain IVIM parameters and ADC values, 3 regions of interest (ROIs) were placed in each cortex and medulla of the right kidney by an experienced abdominal radiologist. A total of 6 ROIs were analyzed for each patient; each ROI was set to contain 5 pixels (Fig 1).

3.2. Renal medulla A significantly low D* value of renal medulla was observed only in group 2 (38.6±7.88, p =0.0477) compared with group 1 (40.7± 9.82), whereas no significant differences were observed for group 3 (39.5 ±7.52, p=0.7043) or group 4 (39.0±4.29, p=1.0000) (Table 2). The medullary D value of was significantly lower in group 3 (1.91± 0.27, p b.0001) compared with group 1 (2.10±0.34), whereas no significant differences were observed for group 2 (2.02±0.30, p=0.1630) or group 4 (1.98±0.37, p=0.5092). No significant differences were observed between the medullary f value for group 1(29.7±9.34) with those for group 2 (30.3±8.57, p=1.0000), group 3(29.9±9.49, p=1.0000), and group 4 (35.5± 11.2, p=0.2265). No significant differences were observed among medullary ADC values of the 3 groups with that of group 1 (1.89±0.19). ADC values for groups 2, 3, and 4 were 1.88±0.15, p=1.0000;1.85±0.19, p= 0.1659; and 1.83±0.21, p=0.5205; respectively.

3.3. Comparison of renal cortex and medulla 2.4. Statistical analysis (1) IVIM parameters and ADC values of the renal cortex and medulla were compared using the Dunn test with group 1 as control. (2) These values of renal cortex and renal medulla were compared within each group using the t-test. A pb0.05 was considered statistically significant. JMP software (Ver. 9, SAS institute, Cary, NC, USA) was used for the analyses.

The ADC value of renal cortex was higher than that of renal medulla as well as D* and f values in each group (p b 0.0001 in all except for the f value of group 4 [p=0.0043]). The D value of renal cortex was higher than that of renal medulla in group 2 (p= 0.0031) and group 3 (p= 0.0021), whereas there were no significant differences in group 1 (p=0.3184) and group 4 (p= 0.2393) (Table 3).

Table 3 Comparison of renal cortex and medulla in each group ADC Cortex Group 1 (eGFRN80) Group 2 (eGFR60–80) Group 3 (eGFR30–60) Group 4 (eGFR) b30

D* Medulla

2.28±0.16 1.89±0.19 (p=b.0001) 2.28±0.14 1.88±0.15 (p=b.0001) 2.17±0.15 1.85±0.19 (p=b.0001) 2.08±0.17 1.83±0.21 (p=b.0001)

Cortex

D Medulla

110.8±12.4 40.7±9.82 (p=b.0001) 106.6±12.1 38.6±7.88 (p=b.0001) 97.8±10.9 39.5±7.52 (p=b.0001) 97.3±12.7 39.0±4.29 (p=b.0001)

Cortex

f Medulla

2.14±0.32 2.10±0.34 (p=0.3184) 2.10±0.26 2.02±0.30 (p=0.0031) 1.99±0.26 1.91±0.27 (p=0.0021) 2.08±0.33 1.98±0.37 (p=0.2393)

Statistical analysis was employed using the t test. Data presented as mean±SD. A pb0.05 was considered statistically significant.

Cortex

Medulla

23.4±10.6 29.7±9.34 (p=b.0001) 24.6±9.07 30.3±8.57 (p=b.0001) 25.1±10.2 29.9±9.49 (p=b.0001) 22.0±8.76 35.5±11.2 (p=0.0043)

S. Ichikawa et al. / Magnetic Resonance Imaging 31 (2013) 414–417

4. Discussion DWI is a technique for imaging molecular movement or diffusivity. The values of molecular diffusivity or ADC in conventional DWI are affected by 2 types of molecular movement: molecular diffusion and microcirculation in vessels (perfusion) [26]. When applying high b values, the influence of perfusion is largely cancelled out, and the ADC value can approximately represent diffusion. The effect of perfusion, however, cannot be completely eliminated, even using high b values, when employing a clinical MR scanner. Thus, DWI performed using a range of low (i.e., b50 s/ mm 2) and high (i.e., N200 s/mm 2) b values, or IVIM imaging, was proposed to measure diffusion and perfusion separately [26–30]. IVIM comprises 2 factors: D value (true diffusion that reflects intraand inter-cellular molecular movement) and D* value (microcirculation in the vessels or perfusion); these 2 factors can be separated using a biexponential fitting of the DWI data [7,8,28]. An earlier study focusing on DWI of the kidney showed that the ADC value decreases as renal function decreases [13–15]. Similarly, we measured the IVIM parameters of the renal cortex and medulla and correlated them with the eGFR. In our study, D* value of the renal cortex was significantly low even in kidneys showing mild renal dysfunction. On the other hand, the ADC value of the renal cortex, as reported in previous work, was not reduced in mild renal dysfunction. These results suggest that tissue perfusion is decreased first in mild renal dysfunction, before it affects proton diffusion. Although the D value of the renal cortex was significantly reduced in moderate renal dysfunction, no significant decrease was observed in severe renal dysfunction in our study. This is probably due to the small number of patients with severe renal dysfunction in our study population. Moreover, atrophic renal cortex in these severe cases might affect the accuracy of the ROI measurements. Furthermore, there is conflicting data in the literature on ADC differences from cortex to medulla [13]. Our comparison of renal cortex and medulla in each group seemed to indicate that the ADC difference was due to perfusion effect, not true diffusion. Our data revealed no consistent pattern of variation in any of the parameters in the renal medulla. Both diffusion and perfusion of the renal medulla seem to be less influenced in cases of renal dysfunction than those of the renal cortex. A limitation of this study is the fact that we used single gradient direction(x-axis). As previously reported, fractional anisotropy is large particularly in the medullary parenchyma of the kidney [31]. It would be better for evaluating diffusivity of the tissue with large fractional anisotropy to use orthogonal MPG gradient. Further study would be expected to reveal the effect of anisotropic diffusivity of the tissue on IVIM parameters. 5. Conclusion As renal dysfunction progresses, renal perfusion might be reduced earlier and affected more than molecular diffusion in the renal cortex. These changes are effectively detected by IVIM MR imaging. Acknowledgment This work was supported by KAKENHI (Grant-in-Aid for Young Scientists [B] No.23791404). References [1] Notohamiprodjo M, Pedersen M, Glaser C, Helck AD, Lodemann KP, Jespersen B, et al. Comparison of Gd-DTPA and Gd-BOPTA for studying renal perfusion and filtration. J Magn Reson Imaging 2011;34(3):595-607. [2] Lemoine S, Papillard M, Belloi A, Rognant N, Fouque D, Laville M, et al. Renal perfusion: noninvasive measurement with multidetector CT versus fluorescent microspheres in a pig model. Radiology 2011;260(2):414-20.

417

[3] Aburano T, Shuke N, Yokoyama K, Matsuda H, Takayama T, Michigishi T, et al. Renal perfusion with Tc-99 m DTPA – simple noninvasive determination of extraction fraction and plasma flow. Clin Nucl Med 1993;18(7):573-7. [4] Artz NS, Sadowski EA, Wentland AL, Djamali A, Grist TM, Seo S, et al. Reproducibility of renal perfusion MR imaging in native and transplanted kidneys using noncontrast arterial spin labeling. J Magn Reson Imaging 2011;33(6):1414-21. [5] Winter JD, St Lawrence KS, Margaret Cheng HL. Quantification of renal perfusion: comparison of arterial spin labeling and dynamic contrast-enhanced MRI. J Magn Reson Imaging 2011;34(3):608-15. [6] Lanzman RS, Wittsack HJ, Martirosian P, Zgoura P, Bilk P, Kropil P, et al. Quantification of renal allograft perfusion using arterial spin labeling MRI: initial results. Eur Radiol 2010;20(6):1485-91. [7] Le Bihan D, Breton E, Lallemand D, Grenier P, Cabanis E, Laval-Jeantet M. MR imaging of intravoxel incoherent motions: application to diffusion and perfusion in neurologic disorders. Radiology 1986;161(2):401-7. [8] Le Bihan D. Intravoxel incoherent motion perfusion MR imaging: a wake-up call. Radiology 2008;249(3):748-52. [9] Patel J, Sigmund EE, Rusinek H, Oei M, Babb JS, Taouli B. Diagnosis of cirrhosis with intravoxel incoherent motion diffusion MRI and dynamic contrastenhanced MRI alone and in combination: preliminary experience. J Magn Reson Imaging 2010;31(3):589-600. [10] Lemke A, Laun FB, Klauss M, Re TJ, Simon D, Delorme S, et al. Differentiation of pancreas carcinoma from healthy pancreatic tissue using multiple b-values: comparison of apparent diffusion coefficient and intravoxel incoherent motion derived parameters. Invest Radiol 2009;44(12):769-75. [11] Luciani A, Vignaud A, Cavet M, Nhieu JT, Mallat A, Ruel L, et al. Liver cirrhosis: intravoxel incoherent motion MR imaging – pilot study. Radiology 2008;249(3): 891-9. [12] Chandarana H, Lee VS, Hecht E, Taouli B, Sigmund EE. Comparison of biexponential and monoexponential model of diffusion weighted imaging in evaluation of renal lesions: preliminary experience. Invest Radiol 2011;46(5):285-91. [13] Thoeny HC, De Keyzer F, Oyen RH, Peeters RR. Diffusion-weighted MR imaging of kidneys in healthy volunteers and patients with parenchymal diseases: initial experience. Radiology 2005;235(3):911-7. [14] Taouli B, Thakur RK, Mannelli L, Babb JS, Kim S, Hecht EM, et al. Renal lesions: characterization with diffusion-weighted imaging versus contrast-enhanced MR imaging. Radiology 2009;251(2):398-407. [15] Xu Y, Wang X, Jiang X. Relationship between the renal apparent diffusion coefficient and glomerular filtration rate: preliminary experience. J Magn Reson Imaging 2007;26(3):678-81. [16] Artunc F, Rossi C, Boss A. MRI to assess renal structure and function. Curr Opin Nephrol Hypertens 2011;20(6):669-75. [17] Grenier N, Basseau F, Ries M, Tyndal B, Jones R, Moonen C. Functional MRI of the kidney. Abdom Imaging 2003;28(2):164-75. [18] Huang AJ, Lee VS, Rusinek H. MR imaging of renal function. Radiol Clin North Am 2003;41(5):1001-17. [19] Itoh K. 99mTc-MAG3: review of pharmacokinetics, clinical application to renal diseases and quantification of renal function. Ann Nucl Med 2001;15(3):179-90. [20] Thoeny HC, Grenier N. Science to practice: can diffusion-weighted MR imaging findings be used as biomarkers to monitor the progression of renal fibrosis? Radiology 2010;255(3):667-8. [21] Wittsack HJ, Lanzman RS, Mathys C, Janssen H, Modder U, Blondin D. Statistical evaluation of diffusion-weighted imaging of the human kidney. Magn Reson Med 2010;64(2):616-22. [22] Zhang JL, Sigmund EE, Chandarana H, Rusinek H, Chen Q, Vivier PH, et al. Variability of renal apparent diffusion coefficients: limitations of the monoexponential model for diffusion quantification. Radiology 2010;254(3):783-92. [23] Levey AS, Coresh J, Greene T, Stevens LA, Zhang YL, Hendriksen S, et al. Using standardized serum creatinine values in the modification of diet in renal disease study equation for estimating glomerular filtration rate. Ann Intern Med 2006; 145(4):247-54. [24] Matsuo S, Imai E, Horio M, Yasuda Y, Tomita K, Nitta K, et al. On behalf of the collaborators developing the Japanese equation for estimated GFR: Revised equations for estimated GFR from serum creatinine in Japan. Am J Kidney Dis 2009;53(6):982-92. [25] Stevens LA, Coresh J, Schmid CH, Feldman HI, Froissart M, Kusek J, et al. Estimating GFR using serum cystatin C alone and in combination with serum creatinine: a pooled analysis of 3,418 individuals with CKD. Am J Kidney Dis 2008;51(3):395-406. [26] Le Bihan D, Breton E, Lallemand D, Aubin ML, Vignaud J, Laval-Jeantet M. Separation of diffusion and perfusion in intravoxel incoherent motion MR imaging. Radiology 1988;168(2):497-505. [27] Muller MF, Prasad PV, Edelman RR. Can the IVIM model be used for renal perfusion imaging? Eur J Radiol 1998;26(3):297-303. [28] Thoeny HC, De Keyzer F. Diffusion-weighted MR imaging of native and transplanted kidneys. Radiology 2011;259(1):25-38. [29] Thoeny HC, Zumstein D, Simon-Zoula S, Eisenberger U, De Keyzer F, Hofmann L, et al. Functional evaluation of transplanted kidneys with diffusion-weighted and BOLD MR imaging: initial experience. Radiology 2006;241(3):812-21. [30] Wirestam R, Borg M, Brockstedt S, Lindgren A, Holtas S, Stahlberg F. Perfusionrelated parameters in intravoxel incoherent motion MR imaging compared with CBV and CBF measured by dynamic susceptibility-contrast MR technique. Acta Radiol 2001;42(2):123-8. [31] Heuper K, Gutberlet M, Rodt T, Gwinner W, Lehner F, Wacker F, et al. Diffusion tensor imaging and tractography for assesment of renal allograft dysfunctioninitial results. Eur Radiol 2011;21(11):2427-33.