• View in gallery
    Figure 1

    Schematic illustration of the T2* relaxation time (T2*) signal decay during blood oxygen–dependent (BOLD) MRI. The T2* signal begins to decay immediately after termination of the induction pulse because of the inherent inhomogeneity of the magnetic field, and that decay can be assessed by obtaining multiple echoes at different times. As the deoxyhemoglobin concentration in RBCs increases, the T2* signal decay curve becomes steeper because the increase in deoxyhemoglobin contributes to magnetic resonance inhomogeneity.

  • View in gallery
    Figure 2

    Representative color-coded T2* (A) and RT* relaxation rate (R2*; B) maps of the left kidney of a healthy adult Beagle derived from BOLD MRI. Oxygenation in the cortex is greater than the oxygenation in the medulla. Thus, the cortex is brighter in color on the T2* map and darker in color on the R2* map, compared with the color of the medulla on the T2* and R2* maps.

  • View in gallery
    Figure 3

    Representative color-coded R2* map (A) of the left kidney of a healthy adult Beagle derived from BOLD MRI on which the ROIs for the cortex (solid line) and medulla (dashed line) have been manually drawn and the time-signal curves have been generated for each of those ROIs (B). The T2* and R2* were 53.29 milliseconds and 18.79 seconds–1 in the cortex and 44.15 milliseconds and 22.64 seconds–1, respectively, for the medulla. Notice that the time-signal curve for the medulla (dashed line) has a steeper slope (decay) than that of the cortex owing to a shorter T2*.

  • 1.

    Fine LG, Norman JT. Chronic hypoxia as a mechanism of progression of chronic kidney diseases: from hypothesis to novel therapeutics. Kidney Int. 2008;74(7):867872.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 2.

    Prasad PV, Thacker J, Li L-P, et al. Multi-parametric evaluation of chronic kidney disease by MRI: a preliminary cross-sectional study. PLoS One. 2015;10(6):e0139661.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 3.

    Schley G, Jordan J, Ellmann S, et al. Multiparametric magnetic resonance imaging of experimental chronic kidney disease: a quantitative correlation study with histology. PLoS One. 2018;13(7):e0200259.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 4.

    Hall ME, Jordan JH, Juncos LA, Hundley WG, Hall JE. BOLD magnetic resonance imaging in nephrology. Int J Nephrol Renovasc Dis. 2018;11:103112.

  • 5.

    Luo F, Liao Y, Cui K, Tao Y. Noninvasive evaluation of renal oxygenation in children with chronic kidney disease using blood-oxygen-level–dependent magnetic resonance imaging. Pediatr Radiol. 2020;50(6):848854.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 6.

    Juillard L, Lerman LO, Kruger DG, et al. Blood oxygen level–dependent measurement of acute intra-renal ischemia. Kidney Int. 2004;65(3):944950.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 7.

    Pruijm M, Hofmann L, Piskunowicz M, et al. Determinants of renal tissue oxygenation as measured with BOLD-MRI in chronic kidney disease and hypertension in humans. PLoS One. 2014;9(4):e95895.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 8.

    Li C, Liu H, Li X, Zhou L, Wang R, Zhang Y. Application of BOLD-MRI in the classification of renal function in chronic kidney disease. Abdom Radiol (NY). 2019;44(2):604611.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 9.

    Pohlmann A, Arakelyan K, Hentschel J, et al. Detailing the relation between renal T2* and renal tissue pO2 using an integrated approach of parametric magnetic resonance imaging and invasive physiological measurements. Invest Radiol. 2014;49(8):547560.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 10.

    Textor SC, Glockner JF, Lerman LO, et al. The use of magnetic resonance to evaluate tissue oxygenation in renal artery stenosis. J Am Soc Nephrol. 2008;19(4):780788.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 11.

    Sadowski EA, Fain SB, Alford SK, et al. Assessment of acute renal transplant rejection with blood oxygen level–dependent MR imaging: initial experience. Radiology. 2005;236(3):911919.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 12.

    Djamali A, Sadowski EA, Samaniego-Picota M, et al. Noninvasive assessment of early kidney allograft dysfunction by blood oxygen level–dependent magnetic resonance imaging. Transplantation. 2006;82(5):621628.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 13.

    Gullichsen E, Nelimarkka O, Halkola L, Niinikoski J. Renal oxygenation in endotoxin shock in dogs. Crit Care Med. 1989;17(6):547550.

  • 14.

    Nelimarkka O, Halkola L, Niinikoski J. Renal hypoxia and lactate metabolism in hemorrhagic shock in dogs. Crit Care Med. 1984;12(8):656660.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 15.

    Rosenbaum B, DiScala VA. Renal tubular sodium reabsorption and oxygen consumption in the hypothyroid dog. Metabolism. 1982;31(3):247251.

  • 16.

    Pedersen M, Dissing TH, Mørkenborg J, et al. Validation of quantitative BOLD MRI measurements in kidney: application to unilateral ureteral obstruction. Kidney Int. 2005;67(6):23052312.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 17.

    Li L-P, Lin J, Santos EA, Dunkle E, Pierchala L, Prasad P. Effect of nitric oxide synthase inhibition on intrarenal oxygenation as evaluated by blood oxygenation level–dependent magnetic resonance imaging. Invest Radiol. 2009;44(2):6773.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 18.

    dos Santos EA, Li L-P, Ji L, Prasad PV. Early changes with diabetes in renal medullary hemodynamics as evaluated by fiberoptic probes and BOLD magnetic resonance imaging. Invest Radiol. 2007;42(3):157162.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 19.

    Hirakawa Y, Tanaka T, Nangaku M. Renal hypoxia in CKD; pathophysiology and detecting methods. Front Physiol. 2017;8:99.

  • 20.

    Cox EF, Buchanan CE, Bradley CR, et al. Multiparametric renal magnetic resonance imaging: validation, interventions, and alterations in chronic kidney disease. Front Physiol. 2017;8:696.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 21.

    Morrell GR, Zhang JL, Lee VS. Magnetic resonance imaging of the fibrotic kidney. J Am Soc Nephrol. 2017;28(9):25642570.

  • 22.

    Pruijm M, Mendichovszky IA, Liss P, et al. Renal blood oxygenation level–dependent magnetic resonance imaging to measure renal tissue oxygenation: a statement paper and systematic review. Nephrol Dial Transplant. 2018;33(suppl 2):ii22ii28.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 23.

    Li L-P, Halter S, Prasad PV. Blood oxygen level–dependent MR imaging of the kidneys. Magn Reson Imaging Clin N Am. 2008;16(4):613625.

  • 24.

    Pruijm M, Milani B, Burnier M. Blood oxygenation level–dependent MRI to assess renal oxygenation in renal diseases: progresses and challenges. Front Physiol. 2017;7:667.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 25.

    Hwang SI, Lee HJ, Chin HJ, Chae D-W, Na KY. Evaluation of renal oxygenation in normal Korean volunteers using 3.0 T blood oxygen level–dependent MRI. J Korean Soc Magn Reson Med. 2013;17(1):1925.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 26.

    Niendorf T, Pohlmann A, Arakelyan K, et al. How bold is blood oxygenation level–dependent (BOLD) magnetic resonance imaging of the kidney? Opportunities, challenges and future directions. Acta Physiol (Oxf). 2015;213(1):1938.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 27.

    Bane O, Mendichovszky IA, Milani B, et al. Consensus-based technical recommendations for clinical translation of renal BOLD MRI. MAGMA. 2020;33(1):199215.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 28.

    Warner L, Glockner JF, Woollard J, Textor SX, Romero JC, Lerman LO. Determinations of renal cortical and medullary oxygenation using blood oxygen level–dependent magnetic resonance imaging and selective diuretics. Invest Radiol. 2011;46(1):4147.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 29.

    Woo S, Cho JY, Kim SY, Kim SH. Intravoxel incoherent motion MRI-derived parameters and T2* relaxation time for noninvasive assessment of renal fibrosis: an experimental study in a rabbit model of unilateral ureter obstruction. Magn Reson Imaging. 2018;51:104112.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 30.

    Vivier P-H, Storey P, Yamamoto A, et al. Furosemide and water load challenges in renal BOLD imaging. In: Proceedings of the European Congress of Radiology. European Society of Radiology; 2012.

    • Search Google Scholar
    • Export Citation
  • 31.

    Li LP, Storey P, Pierchala L, Li W, Polzin J, Prasad P. Evaluation of the reproducibility of intrarenal R2* and ΔR2* measurements following administration of furosemide and during waterload. J Magn Reson Imaging. 2004;19(5):610616.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 32.

    Thacker JM, Li L-P, Li W, Zhou Y, Sprague SM, Prasad PV. Renal blood oxygen level-dependent magnetic resonance imaging: a sensitive and objective analysis. Invest Radiol. 2015;50(12):821827.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 33.

    Gomez SI, Warner L, Haas JA, et al. Increased hypoxia and reduced renal tubular response to furosemide detected by BOLD magnetic resonance imaging in swine renovascular hypertension. Am J Physiol Renal Physiol. 2009;297(4):F981F986.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 34.

    Perinetti G. StaTips part IV: selection, interpretation and reporting of the intraclass correlation coefficient. South Eur J Orthod Dentofac Res. 2018;5(1):35.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 35.

    Pan J, Zhang H, Man F, et al. Measurement and scan reproducibility of parameters of intravoxel incoherent motion in renal tumor and normal renal parenchyma: a preliminary research at 3.0 T MR. Abdom Radiol (NY). 2018;43(7):17391748.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 36.

    Simon-Zoula SC, Hofmann L, Giger A, et al. Non-invasive monitoring of renal oxygenation using BOLD-MRI: a reproducibility study. NMR Biomed. 2006;19(1):8489.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 37.

    Behnia R, Koushanpour E, Goldstick T, Linde HW, Osborn R. Renal tissue oxygenation following induced hypotension in dogs. Br J Anaesth. 1984;56(9):10371043.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 38.

    Zhang JL, Rusinek H, Chandarana H, Lee VS. Functional MRI of the kidneys. J Magn Reson Imaging. 2013;37(2):282293.

  • 39.

    Gloviczki ML, Glockner JF, Crane JA, et al. Blood oxygen level–dependent magnetic resonance imaging identifies cortical hypoxia in severe renovascular disease. Hypertension. 2011;58(6):10661072.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 40.

    Ludens JH, Williamson HE. Effect of furosemide on renal blood flow in the conscious dog. Proc Soc Exp Biol Med. 1970;133(2):513515.

  • 41.

    Nuutinen LS, Tuononen S. The effect of furosemide on renal blood flow and renal tissue oxygen tension in dogs. Ann Chir Gynaecol. 1976;65:272276.

    • Search Google Scholar
    • Export Citation
  • 42.

    Vink EE, de Boer A, Hoogduin HJM, et al. Renal BOLD-MRI relates to kidney function and activity of the renin-angiotensin-aldosterone system in hypertensive patients. J Hypertens. 2015;33(3):597604.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 43.

    Zhou H, Yang M, Jiang Z, Ding J, Di J, Cui LI. Renal hypoxia: an important prognostic marker in patients with chronic kidney disease. Am J Nephrol. 2018;48(1):4655.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 44.

    Chrysochou C, Mendichovszky IA, Buckley DL, Cheung CM, Jackson A, Kalra PA. BOLD imaging: a potential predictive biomarker of renal functional outcome following revascularization in atheromatous renovascular disease. Nephrol Dial Transplant. 2012;27(3):10131019.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 45.

    Hofmann L, Simon-Zoula S, Nowak A, et al. BOLD-MRI for the assessment of renal oxygenation in humans: acute effect of nephrotoxic xenobiotics. Kidney Int. 2006;70(1): 144150.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 46.

    Hall ME, Rocco MV, Morgan TM, et al. Beta-blocker use is associated with higher renal tissue oxygenation in hypertensive patients suspected of renal artery stenosis. Cardiorenal Med. 2016;6(4):261268.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 47.

    Rogers-Smith E, Hammerton R, Mathis A, Allison A, Clark L. Twelve previously healthy non-geriatric dogs present for acute kidney injury after general anaesthesia for non-emergency surgical procedures in the UK. J Small Anim Pract. 2020;61(6):363367.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 48.

    Niles DJ, Gordon JW, Fain SB. Effect of anesthesia on renal R2* measured by blood oxygen level–dependent MRI. NMR Biomed. 2015;28(7):811817.

  • 49.

    Pouchelon JL, Atkins CE, Bussadori C, et al. Cardiovascular–renal axis disorders in the domestic dog and cat: a veterinary consensus statement. J Small Anim Pract. 2015;56(9):537552.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 50.

    Chang D, Wang Y-C, Xu T-T, et al. Noninvasive identification of renal hypoxia in experimental myocardial infarctions of different sizes by using BOLD MR imaging in a mouse model. Radiology. 2018;286(1):129139.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 51.

    Siddiqi L, Hoogduin H, Visser F, Leiner T, Mali WP, Blankestijn PJ. Inhibition of the renin‐angiotensin system affects kidney tissue oxygenation evaluated by magnetic resonance imaging in patients with chronic kidney disease. J Clin Hypertens (Greenwich). 2014;16(3):214218.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 52.

    Aronson LR. Update on the current status of kidney transplantation for chronic kidney disease in animals. Vet Clin North Am Small Anim Pract. 2016;46(6):11931218.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 53.

    Halling KB, Graham JP, Newell SP, et al. Sonographic and scintigraphic evaluation of acute renal allograft rejection in cats. Vet Radiol Ultrasound. 2003;44(6):707713.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 54.

    Halling KB, Ellison GW, Armstrong D, et al. Evaluation of oxidative stress markers for the early diagnosis of allograft rejection in feline renal allotransplant recipients with normal renal function. Can Vet J. 2004;45(10):831837.

    • Search Google Scholar
    • Export Citation
  • 55.

    Neugarten J, Golestaneh L. Blood oxygenation level–dependent MRI for assessment of renal oxygenation. Int J Nephrol Renovasc Dis. 2014;7:421435.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 56.

    Chen F, Li S, Sun D. Methods of blood oxygen level–dependent magnetic resonance imaging analysis for evaluating renal oxygenation. Kidney Blood Press Res. 2018;43(2): 378388.

    • Crossref
    • Search Google Scholar
    • Export Citation

Advertisement

Quantification of renal T2 relaxation rate by use of blood oxygen level–dependent magnetic resonance imaging before and after furosemide administration in healthy Beagles

Sang-Kwon LeeFrom the College of Veterinary Medicine, Kyungpook National University, Daegu 41566, South Korea;

Search for other papers by Sang-Kwon Lee in
Current site
Google Scholar
PubMed
Close
 DVM, PhD
,
Juryeong Leethe College of Veterinary Medicine and BK21 Plus Project Team, Chonnam National University, Gwangju 61186, Republic of Korea;

Search for other papers by Juryeong Lee in
Current site
Google Scholar
PubMed
Close
 DVM
,
Seolyn Jangthe College of Veterinary Medicine and BK21 Plus Project Team, Chonnam National University, Gwangju 61186, Republic of Korea;

Search for other papers by Seolyn Jang in
Current site
Google Scholar
PubMed
Close
 DVM
,
Eunji Leethe College of Veterinary Medicine and BK21 Plus Project Team, Chonnam National University, Gwangju 61186, Republic of Korea;

Search for other papers by Eunji Lee in
Current site
Google Scholar
PubMed
Close
 DVM
,
Chang-Yeop JeonNational Primate Research Center and Futuristic Animal Resource and Research Center, Korea Research Institute of Bioscience and Biotechnology, Cheongju 28116, Republic of Korea;

Search for other papers by Chang-Yeop Jeon in
Current site
Google Scholar
PubMed
Close
,
Kyung-Seob LimNational Primate Research Center and Futuristic Animal Resource and Research Center, Korea Research Institute of Bioscience and Biotechnology, Cheongju 28116, Republic of Korea;

Search for other papers by Kyung-Seob Lim in
Current site
Google Scholar
PubMed
Close
 DVM, PhD
,
Yeung Bae Jinthe College of Veterinary Medicine, Gyeongsang National University, Jinju 52828, Republic of Korea;

Search for other papers by Yeung Bae Jin in
Current site
Google Scholar
PubMed
Close
 DVM, PhD
, and
Jihye Choithe Department of Veterinary Medical Imaging, College of Veterinary Medicine, Seoul National University, Seoul 08826, South Korea.

Search for other papers by Jihye Choi in
Current site
Google Scholar
PubMed
Close
 DVM, PhD

Abstract

OBJECTIVE

To assess the feasibility of blood oxygen level–dependent (BOLD) MRI for measurement of the renal T2* relaxation rate (R2*; proxy for renal oxygenation) before and after furosemide administration and to evaluate the reliability and repeatability of those measurements in healthy dogs.

ANIMALS

8 healthy adult Beagles (4 males and 4 females).

PROCEDURES

Each dog was anesthetized and underwent BOLD MRI before (baseline) and 3 minutes after administration of furosemide (1 mg/kg, IV) twice, with a 1-week interval between scanning sessions. Mapping software was used to process MRI images and measure R2* and the difference in R2* (∆R2*) before and after furosemide administration. The intraclass correlation coefficient was calculated to assess measurement reliability, and the coefficient of variation and Bland-Altman method were used to assess measurement repeatability.

RESULTS

Mean ± SD baseline R2* in the renal medulla (24.5 ± 3.8 seconds−1) was significantly greater than that in the renal cortex (20.6 ± 2.7 seconds−1). Mean R2* in the renal cortex (18.6 ± 2.6 seconds−1) and medulla (17.8 ± 1.5 seconds−1) decreased significantly after furosemide administration. Mean ∆R2* in the medulla (6.7 ± 2.4 seconds−1) was significantly greater than that in the renal cortex (2.1 ± 0.7 seconds−1). All R2* and ∆R2* values had good or excellent reliability and repeatability, except the cortical ∆R2*, which had poor repeatability.

CONCLUSIONS AND CLINICAL RELEVANCE

Results suggested that BOLD MRI, when performed before and after furosemide administration, was noninvasive and highly reliable and repeatable for dynamic evaluation of renal oxygenation in healthy dogs.

Abstract

OBJECTIVE

To assess the feasibility of blood oxygen level–dependent (BOLD) MRI for measurement of the renal T2* relaxation rate (R2*; proxy for renal oxygenation) before and after furosemide administration and to evaluate the reliability and repeatability of those measurements in healthy dogs.

ANIMALS

8 healthy adult Beagles (4 males and 4 females).

PROCEDURES

Each dog was anesthetized and underwent BOLD MRI before (baseline) and 3 minutes after administration of furosemide (1 mg/kg, IV) twice, with a 1-week interval between scanning sessions. Mapping software was used to process MRI images and measure R2* and the difference in R2* (∆R2*) before and after furosemide administration. The intraclass correlation coefficient was calculated to assess measurement reliability, and the coefficient of variation and Bland-Altman method were used to assess measurement repeatability.

RESULTS

Mean ± SD baseline R2* in the renal medulla (24.5 ± 3.8 seconds−1) was significantly greater than that in the renal cortex (20.6 ± 2.7 seconds−1). Mean R2* in the renal cortex (18.6 ± 2.6 seconds−1) and medulla (17.8 ± 1.5 seconds−1) decreased significantly after furosemide administration. Mean ∆R2* in the medulla (6.7 ± 2.4 seconds−1) was significantly greater than that in the renal cortex (2.1 ± 0.7 seconds−1). All R2* and ∆R2* values had good or excellent reliability and repeatability, except the cortical ∆R2*, which had poor repeatability.

CONCLUSIONS AND CLINICAL RELEVANCE

Results suggested that BOLD MRI, when performed before and after furosemide administration, was noninvasive and highly reliable and repeatable for dynamic evaluation of renal oxygenation in healthy dogs.

Introduction

Renal hypoxia is a key factor in the development and progression of chronic kidney disease (CKD).1 It promotes tubular injury, interstitial fibrosis, inflammation, and deterioration of renal function, which leads to further hypoxia and progression of CKD. The effects of changes in renal oxygenation on the pathophysiologic progression of CKD have been investigated in kidney transplant recipients and animals with experimentally induced and human patients with diabetes, hypertension, and renal artery stenosis.2,3,4,5,6,7,8,9,10,11,12 A few studies13,14,15 have evaluated renal oxygenation and hypoxia by use of invasive catheterization methods in dogs with sepsis, hemorrhagic shock, and hypothyroidism. To our knowledge, the effects of anesthesia, heart failure, chronic dehydration, and NSAID administration, all of which are associated with a risk for kidney injury in dogs, on renal oxygenation have not been investigated. Thus, the role of renal hypoxia on the progression of CKD in dogs with underlying risk factors remains unclear.

Although renal blood flow regulates the supply of oxygen to the kidneys, measurement of renal blood flow alone does not accurately represent renal oxygenation because it does not account for oxygen consumption by renal tissues.10 A decrease in renal blood flow reduces the oxygen supply to the kidney and decreases the glomerular filtration rate (GFR), which in turn decreases oxygen consumption. Therefore, evaluation of renal oxygenation independent of renal blood flow is necessary for assessment of kidney disease.

Various methods can be used to evaluate renal oxygenation, including oxygen-sensitive microelectrodes, detection of activated hypoxia-inducible factor, phosphorescence lifetime measurement, and blood oxygen level–dependent (BOLD) MRI.16,17,18,19 Among those techniques, BOLD MRI is considered the only method appropriate for clinical use because it is noninvasive and can be used for serial measurements. Multiple studies2,4,5,7,8,20,21,22,23,2425 have used BOLD MRI to noninvasively estimate renal oxygenation in humans and animals. Studies2,3,5,8,10,11,16 have also evaluated the usefulness of BOLD MRI as a diagnostic and prognostic tool for various kidney disorders such as CKD, renal arterial stenosis, unilateral ureteral obstruction, and kidney transplant rejection.

Blood oxygen level–dependent MRI assesses changes in regional blood concentrations of oxyhemoglobin and deoxyhemoglobin.4,16,22,26,27 Oxyhemoglobin has diamagnetic properties and deoxyhemoglobin has paramagnetic properties, which result in an inhomogeneous magnetic field. This shortens the T2* relaxation time (T2*) of protons, leading to signal loss in the surrounding tissues because of increased dephasing of magnetic spins. The spin dephasing T2* relaxation rate (R2*; 1/T2*) is closely related to the concentration of deoxyhemoglobin in tissue and can be measured in inverse seconds (seconds−1). The ratio of oxyhemoglobin to deoxyhemoglobin is the major determinant of the Po2 in the blood. If it is assumed that Po2 in blood is in equilibrium with the Po2 in tissue, then the R2* is an indirect reflection of tissue oxygenation (ie, a low R2* is indicative of high tissue oxygenation).

Blood oxygen level–dependent MRI can be used to assess the change in Po2 in the renal cortex and medulla.9,16,17,18,28 Results of a study16 involving pigs in which the R2* was compared with the Po2 measured by oxygen-sensitive microelectrodes indicate that there is a strong linear correlation between R2* and the Po2 in renal tissues. In a study19 involving rats, the T2* was significantly correlated with changes in renal tissue Po2 measured by an invasive laser probe during various experimentally induced conditions such as aortic occlusion, hypoxia, and hyperoxia. Results of other studies involving rats with experimentally induced diabetes18 or hypertension17 indicate that changes in R2* as determined by BOLD MRI are consistent with changes in renal tissue Po2 as determined by the fluorescent lifetime method. Although early research involving BOLD MRI yielded inconsistent results owing to the absence of standard protocols and analysis methods, most recent studies2,5,20,24 have consistently demonstrated an increase in renal cortical R2* in human patients with CKD. Additionally, the R2* is significantly associated with renal function and histologic severity of renal fibrosis.3,8,29

The R2* derived by BOLD MRI can be affected by other factors such as shimming, renal blood flow rate, tissue water content, and magnetic susceptibility effects unrelated to deoxyhemoglobin.22,30 Moreover, the R2* may not accurately reflect renal dysfunction, depending on the patient’s physiologic state. For example, in patients with end-stage kidney disease, the baseline R2* may not increase or decrease relative to oxygen consumption because oxygen consumption decreases as GFR decreases.10 Consequently, evaluation of the dynamic change in R2* (∆R2*) in response to various physiologic and pharmacological challenges is preferred to compensate for this limitation of R2* measurement.27,31

A furosemide challenge protocol is commonly used in conjunction with BOLD MRI.2,10,12,32 Furosemide administration induces active tubular sodium reabsorption in the thick ascending loop of Henle in the outer portion of the renal medulla. Furosemide administration causes an increase in oxygenation within the renal medulla despite a decrease in medullary blood flow and oxygen delivery owing to a decrease in oxygen consumption, The magnitude of the response to furosemide administration in patients with CKD is less than that in healthy subjects.7 Although a subject’s response to furosemide administration as determined by BOLD MRI may not represent absolute renal oxygenation, it is considered a sensitive functional marker for diagnosis of kidney disease.7,33

Renal oxygenation is rarely measured in veterinary patients, despite the fact that kidney disease is common in dogs and cats. Magnetic resonance imaging is becoming increasingly available in veterinary practice, and it is believed that BOLD MRI can be adapted to evaluate renal oxygenation in dogs and cats. The purpose of the study reported here was to assess the feasibility of BOLD MRI for measurement of the R2* before and after furosemide administration in healthy dogs. Interobserver and intraobserver reliability and test-retest repeatability of those measurements were also assessed. Our goal was to establish preliminary guidelines for T2*, R2*, and ∆R2* in the kidneys of healthy dogs against which those measurements in dogs with kidney disease can be compared.

Materials and Methods

Animals

The study protocol was reviewed and approved by the Chonnam National University Institutional Animal Care and Use Committee, and the care of the dogs used for the study adhered to the Guidelines for Animal Experiments of Chonnam National University (CNU IACUC-YB-R-2020-32). Eight purpose-bred Beagles (4 sexually intact males and 4 sexually intact females) with a median age of 2 years (range, 1 to 3 years) and body weight of 10.2 kg (range, 8.7 to 12.6 kg) were used for the study. All dogs were considered healthy on the basis of results of a physical examination, blood pressure values, CBC, serum biochemical analysis, urinalysis (including urine dipstick evaluation and measurement of urine specific gravity), thoracic and abdominal radiography, abdominal ultrasonography, and echocardiography.

MRI

For each dog, food but not water was withheld for 24 hours prior to anesthesia induction for MRI. A 20-gauge catheter was aseptically placed into a cephalic vein, and anesthesia was induced with alfaxalone (Alfaxa; Careside Co Ltd; 3 mg/kg, IV). Following endotracheal intubation, anesthesia was maintained with isoflurane (Terrell; Piramal Critical Care; 2% to 4%) delivered in oxygen (1 to 2 L/min). The dog was positioned in dorsal recumbency, and all MRI sequences were obtained by use of a 3.0-T whole-body scanner (Achieva; Philips Healthcare) with a 32-channel sensitivity encoding torso-cardiac coil. Three orthogonal plane images were obtained with 3-D T1-weighted MRI as a localizer. Transverse BOLD MRI images were acquired during breath holding induced by hyperventilation. During BOLD image acquisition, the respiratory rate, heart rate, and blood oxygen saturation were continuously monitored with a respiratory sensor, ECG tracing, and pulse oximetry, respectively.

Baseline BOLD images were obtained by use of a multi-echo gradient echo sequence with 9 different echoes in the transverse plane. For each BOLD image, the signal was collected at 9 different echo times (TEs) after excitation pulses were introduced, which allowed the degree of T2* signal decay to be analyzed over time (Figure 1). The presence of deoxyhemoglobin makes the surrounding magnetic field inhomogeneous and further accelerates the T2* signal decay over time, thereby resulting in a steeper T2* signal decay curve. Each set of 9 T2* images was acquired during a single breath hold of 18 seconds. The scan parameters were as follows: repetition time, 180 milliseconds; number of echoes, 12; initial TE, 2.3 milliseconds; echo spacing, 4.8 milliseconds; flip angle, 10°; slice thickness, 3 mm; field of view, 200 × 200 × 99 mm; and matrix, 68 × 66 mm.

Figure 1
Figure 1

Schematic illustration of the T2* relaxation time (T2*) signal decay during blood oxygen–dependent (BOLD) MRI. The T2* signal begins to decay immediately after termination of the induction pulse because of the inherent inhomogeneity of the magnetic field, and that decay can be assessed by obtaining multiple echoes at different times. As the deoxyhemoglobin concentration in RBCs increases, the T2* signal decay curve becomes steeper because the increase in deoxyhemoglobin contributes to magnetic resonance inhomogeneity.

Citation: American Journal of Veterinary Research 82, 11; 10.2460/ajvr.82.11.880

After the baseline BOLD MRI images were acquired, furosemide (Lasix; Handok Inc; 1 mg/kg, IV) was administered followed by saline (0.9% NaCl) solution (5 mL, IV) to ensure that the entire dose of furosemide was flushed through the catheter. Three minutes after furosemide administration, BOLD MRI images (postdiuretic images) were acquired as described for the baseline BOLD images. The protocol was repeated 1 week later to acquire a second set of baseline and postdiuretic BOLD MRI images.

Mapping T2* and R2* from BOLD data

Image processing and data analysis were conducted with relaxation mapping software (Express; Philips Healthcare). All analyses were performed by use of baseline BOLD images, followed by analysis of the paired postdiuretic BOLD images. Color-coded T2* and R2* maps of baseline BOLD images were generated in which bright yellow represented a high value and dark gray represented a low value. The renal cortex and medulla were presented in different colors on the T2* and R2* color-coded maps (Figure 2). The regions of interest (ROIs) in the renal cortex and medulla were manually traced on the color-coded R2* map to reduce overlap between the cortex and medulla (Figure 3). To prevent partial volume effect, the outermost part of the renal cortex and the inner most part of the renal medulla were excluded from the ROIs. Voxels representing susceptibility artifacts or other artifacts were not included in the ROIs. Placement of the ROIs was performed separately at the hilar region of each kidney. The signal intensity at multiple TEs was fit to the logarithm of the exponential signal decay, and the T2* and R2* were calculated by use of the following 2 equations: S(TE) = S0e–TE/T2* and R2* = 1/T2*, where S is signal intensity and S0 is the proton density. The ROIs used during analysis of the baseline images were saved and used for analysis of the postdiuretic images, and the change in R2* (∆R2* = baseline R2* − postdiuretic R2*) due to furosemide administration was calculated.

Figure 2
Figure 2

Representative color-coded T2* (A) and RT* relaxation rate (R2*; B) maps of the left kidney of a healthy adult Beagle derived from BOLD MRI. Oxygenation in the cortex is greater than the oxygenation in the medulla. Thus, the cortex is brighter in color on the T2* map and darker in color on the R2* map, compared with the color of the medulla on the T2* and R2* maps.

Citation: American Journal of Veterinary Research 82, 11; 10.2460/ajvr.82.11.880

Figure 3
Figure 3

Representative color-coded R2* map (A) of the left kidney of a healthy adult Beagle derived from BOLD MRI on which the ROIs for the cortex (solid line) and medulla (dashed line) have been manually drawn and the time-signal curves have been generated for each of those ROIs (B). The T2* and R2* were 53.29 milliseconds and 18.79 seconds–1 in the cortex and 44.15 milliseconds and 22.64 seconds–1, respectively, for the medulla. Notice that the time-signal curve for the medulla (dashed line) has a steeper slope (decay) than that of the cortex owing to a shorter T2*.

Citation: American Journal of Veterinary Research 82, 11; 10.2460/ajvr.82.11.880

Reproducibility and repeatability of BOLD MRI

The first set of baseline and postdiuretic BOLD images were evaluated by a fourth-year PhD student (S-KL) and veterinarian with 1 year of radiology experience (JRL) under the supervision of a veterinary radiologist (JHC). Placement of each ROI over the renal cortex and medulla was performed on the basis of color derived from the MRI processing software and was repeated 3 times. The mean for the 3 measurements was calculated and used for subsequent analyses. Two observers independently measured the R2* and ∆R2*, and interobserver reliability was assessed. After at least 7 days, the first set of baseline and postdiuretic images was reevaluated by the PhD student (S-KL), who was blinded to the previous results, and interobserver reliability was assessed.

The second set of baseline and postdiuretic images was analyzed by the PhD student (S-KL), who was blinded to the measurements for the first set of baseline and postdiuretic images, and the test-retest repeatability of BOLD MRI was assessed. Measurements obtained by that observer were used to compare the R2* and ∆R2 between the left and right kidneys and between the renal cortex and medulla.

Statistical analysis

The Kolmogorov-Smirnov test was used to assess data for normality. All variables were normally distributed except for postdiuretic R2* of the right renal cortex and mean ∆R2* for the renal cortex, and results were summarized as the mean ± SD. The intraobserver and interobserver reliability of R2* and ∆R2* were evaluated by calculating the respective intraclass correlation coefficients (ICCs). An ICC < 0.50, between 0.50 and < 0.75, between 0.75 and < 0.90, and ≥ 0.90 was considered poor, fair, good, and excellent reliability, respectively.34 The repeatability of R2* and ∆R2* was assessed by calculation of the coefficient of variation (CV) and with the Bland-Altman method. A CV > 30%, between 20% to 30%, between 10% and 20%, and < 10% was considered poor, fair, good, and excellent repeatability, respectively.35 The R2* and ∆R2* between the first and second scans, before and after furosemide administration, between the renal cortex and medulla in each kidney, and between the left and right kidneys of each dog were assessed by use of the paired t test for normally distributed variables or the Wilcoxon signed rank test for variables that were not normally distributed. The 95% CIs for the R2* and ∆R2* were calculated to serve as preliminary guidelines for future comparisons. All analyses were performed by use of statistical software (SPSS Statistics version 25; IBM Corp), and values of P < 0.05 were considered significant.

Results

Blood oxygen level–dependent MRI scans were successfully obtained before and after furosemide administration twice for all 8 dogs. Thus, 32 sets of BOLD MRI scans were available for evaluation. The mean BOLD MRI scan duration under each condition (baseline or postdiuretic scan) was 166 seconds, and although not recorded, it typically took approximately 5 minutes (including the breath-hold induction time) for BOLD image acquisition under each condition. On the color-coded R2* maps created for the baseline images, there was a clear distinction between the renal cortex and medulla, which allowed manual tracing the ROIs in the renal cortex and medulla. The distinction between the renal cortex and medulla was slightly blurred on the color-coded R2* maps created for the postdiuretic images, but that did not cause any problems during analysis because the same ROIs were used for both the baseline and postdiuretic images.

BOLD MRI measurements

The mean T2* and R2* values for the renal cortex and medulla of the left and right kidneys before and after furosemide administration were summarized (Table 1). The baseline R2* did not differ significantly between the left and right renal cortex (P = 0.563) or between the left and right renal medulla (P = 0.587). The mean ± SD baseline T2* was 49.2 ± 5.6 milliseconds (95% CI, 45.4 to 53.1 milliseconds) for the renal cortex and 41.7 ± 5.8 milliseconds (95% CI, 37.7 to 45.7 milliseconds) for the renal medulla. The mean ± SD baseline R2* was 20.6 ± 2.7 seconds–1 (95% CI, 18.8 to 22.5 seconds–1) for the renal cortex and 24.5 ± 3.8 seconds–1 (95% CI, 21.9 to 27.1 seconds–1) for the renal medulla. The baseline R2* was significantly greater in the medulla than in the cortex of both the left (P < 0.001) and right (P = 0.001) kidneys. The baseline R2* did not differ significantly between the left (P = 0.886) or right (P = 0.343) renal cortices of males and females or between the left (P = 0.343) or right (P = 0.114) renal medullas of males and females.

Table 1

Blood oxygen level–dependent (BOLD) MRI–derived values for T2* relaxation time (T2*) and T2* relaxation rate (R2*) before (baseline) and after (postdiuretic) furosemide (1 mg/kg, IV) administration and the change in R2* (∆R2*) for 8 healthy adult Beagles.

Measurement site Kidney T2* (ms) R2* (s–1) ∆R2* (s–1)
Baseline Postdiuretic Baseline Postdiuretic
Cortex Left 49.7 ± 6.1 54.3 ± 6.2 20.5 ± 2.8 18.7 ± 2.6 1.8 ± 0.9
(45.4–53.9) (50.0–58.6) (18.5–22.5) (16.9–20.5) (1.2–2.4)
Right 48.8 ± 5.6 55.1 ± 6.7 20.8 ± 2.7 18.5 ± 2.7 2.3 ± 0.7
(44.9–52.6) (50.4–59.8) (19.0–22.7) (16.6–20.4) (1.9–2.8)
Medulla Left 41.7 ± 5.0 56.7 ± 3.3 24.4 ± 3.3 17.7 ± 1.1 6.7 ± 2.5
(38.2–45.1) (54.4–59.0) (22.1–26.7) (17.0–18.5) (5.0–8.4)
Right 41.7 ± 6.7 56.3 ± 6.4 24.7 ± 4.3 17.9 ± 2.1 6.8 ± 2.3
(37.0–46.3) (51.9–60.7) (21.7–27.6) (16.5–19.4) (5.2–8.4)

Values represent the mean ± SD (95% CI). Each dog underwent BOLD MRI twice, with a 1-week interval between scanning sessions. Baseline and postdiuretic images were acquired during both scanning sessions.

Value differs significantly (P ≤ 0.05) from the corresponding baseline value.

Value differs significantly (P ≤ 0.05) from the corresponding value for the renal cortex.

= Baseline R2* – postdiuretic R2*. s−1 = Inverse seconds.

The mean values did not differ significantly between the left and right kidneys or between the first and second scans for any of the 3 parameters.

The postdiuretic R2* was significantly lower than the baseline R2* in both the renal cortex and medulla (Table 1). However, the postdiuretic R2* did not differ significantly between the left and right renal cortex (P = 0.657) or between the left and right renal medulla (P = 0.665). The mean postdiuretic R2* was 18.6 ± 2.6 seconds−1 (95% CI, 16.8 to 20.4 seconds−1) in the renal cortex and 17.8 ± 1.5 seconds−1 (95% CI, 16.8 to 18.8 seconds−1) in the medulla. Unlike the baseline R2*, the postdiuretic R2* did not differ significantly between the cortex and medulla in either the left (P = 0.213) or right (P = 0.216) kidneys.

The mean ∆R2* did not differ significantly between the left and right renal cortex (P = 0.149) or between the left and right renal medulla (P = 0.847). The mean ∆R2* was 2.1 ± 0.7 seconds−1 (95% CI, 1.6 to 2.5 seconds−1) in the renal cortex and 6.7 ± 2.4 seconds−1 (95% CI, 5.1 to 8.4 seconds−1) in the medulla. The mean ∆R2* in the medulla was significantly greater than that in the cortex in both the left (P = 0.001) and right (P = 0.001) kidneys.

Reliability and test-retest repeatability of R2* and ∆R2*

The intraobserver and interobserver ICCs for baseline and postdiuretic R2* and ∆R2* in the cortex and medulla of both the left and right kidneys were summarized (Table 2). There was good or excellent intraobserver and interobserver reliability (ICC > 0.70) for all parameters.

Table 2

Intraobserver and interobserver ICCs (95% CI) for the baseline R2*, postdiuretic R2*, and ∆R2* measurements in Table 1.

Reliability type Measurement site Kidney Baseline R2* Post diuretic R2* ∆R2*
Intraobserver Cortex Left 0.976 (0.879–0.995) 0.961 (0.806–0.992) 0.887 (0.436–0.977)
Right 0.972 (0.860–0.994) 0.997 (0.984–0.999) 0.776 (−0.119–0.955)
Medulla Left 0.976 (0.882–0.995) 0.785 (−0.076–0.957) 0.956 (0.780–0.991)
Right 0.991 (0.954–0.998) 0.955 (0.773–0.991) 0.969 (0.846–0.994)
Interobserver Cortex Left 0.986 (0.931–0.997) 0.969 (0.844–0.994) 0.843 (0.218–0.969)
Right 0.958 (0.790–0.992) 0.963 (0.813–0.993) 0.809 (0.044–0.962)
Medulla Left 0.959 (0.793–0.992) 0.864 (0.321–0.973) 0.862 (0.312–0.972)
Right 0.983 (0.913–0.996) 0.948 (0.741–0.990) 0.968 (0.838–0.994)

An ICC < 0.50, between 0.50 and < 0.75, between 0.75 and < 0.90, and ≥ 0.90 was considered poor, fair, good, and excellent measurement reliability, respectively.

The mean baseline and postdiuretic R2* and ∆R2* did not differ significantly between the first and second BOLD MRI scans for the cortex and medulla of either the left or right kidneys. The CVs and results of the Bland-Altman analyses for each parameter were summarized (Table 3). Results of the Bland-Altman analyses suggested that there were no significant test-retest differences for any of the parameters. The CVs indicated that all baseline and postdiuretic R2* measurements and ∆R2* measurements in the renal medulla had excellent (CV < 10%) repeatability. However, the repeatability of the ∆R2* measurements in the cortex was only fair for the left kidney (CV = 25.1%) and poor for the right kidney (CV = 31.85%).

Table 3

Coefficients of variation (CV) and Bland-Altman analyses results for comparison of the baseline R2*, postdiuretic R2*, and ∆R2* between the first and second BOLD MRI scans for the dogs of Table 1.

Parameter Measurement site Kidney CV (%) Bland-Altman analysis
Bias 95% limits of agreement
Baseline R2* Cortex Left 2.87 0.11 −2.41 to 2.64
Right 2.73 −0.19 −2.68 to 2.29
Medulla Left 3.40 −0.23 −4.00 to 3.53
Right 3.89 0.13 −4.68 to 4.94
Postdiuretic R2* Cortex Left 4.38 0.65 −2.48 to 3.79
Right 3.89 0.53 −2.87 to 3.94
Medulla Left 3.80 0.29 −3.27 to 3.84
Right 5.11 0.28 −4.47 to 5.03
∆R2* Cortex Left 25.10 −0.54 −2.49 to 1.41
Right 31.85 −0.73 −3.04 to 1.59
Medulla Left 9.20 −0.52 −3.09 to 2.04
Right 8.91 −0.15 −2.79 to 2.50

A CV > 30%, between 20% to 30%, between 10% and 20%, and < 10% was considered poor, fair, good, and excellent repeatability, respectively.

Discussion

To our knowledge, the present study was the first to evaluate the BOLD MRI–derived T2*, R2* and ∆R2* in the renal cortex and medulla of healthy dogs before (baseline) and after (postdiuretic) furosemide administration. The mean baseline R2* and ∆R2*, but not the postdiuretic R2*, were significantly greater in the renal medulla than in the renal cortex. All measurements had good or excellent reliability, and all measurements except the ∆R2* in the renal cortex had excellent repeatability. The repeatability of the ∆R2* in the renal cortex was only fair to poor.

In the present study, the mean baseline R2* for the renal medulla was significantly greater than that for the renal cortex, which suggested that oxygenation of the medulla was less than the oxygenation of the cortex (ie, the medulla was more hypoxic than the cortex). That finding was consistent with results of studies20,25,31,36 involving human subjects in whom renal oxygenation was assessed by BOLD MRI and a study37 involving dogs in which renal oxygenation was assessed by oxygen-sensitive microelectrodes. In the dog study,37 the Po2 in the renal cortex ranged from 38.6 to 63.8 kPa, whereas the Po2 in the renal medulla ranged from 10.4 to 13.3 kPa.37 Renal medullary hypoxia might be associated with low oxygen delivery, shunt diffusion of oxygen from the arteries to the venous vasa recta, or high oxygen consumption owing to active transport of sodium.23,25,26,38 Because the oxygen dissociation curve has a sigmoidal shape, the change in hemoglobin oxygen saturation is considered more pronounced in tissues with a low Po2 than in tissues with a high Po2. Therefore, BOLD MRI might be more useful for monitoring oxygenation changes in the renal medulla than in the renal cortex, although many researchers agree that clinical use of BOLD MRI has value for assessing oxygenation in both the renal cortex and medulla.4,27

For the dogs of the present study, the R2* decreased significantly from baseline following administration of furosemide in both the renal cortex and medulla, although the magnitude of that decrease (ie, ∆R2*) was significantly greater in the medulla than in the cortex. Investigators of other studies7,16,28,30,31 likewise report that the ∆R2* is greater in the medulla than in the cortex following furosemide administration and that the ∆R2* for the cortex is generally very small. The renal medullary response to furosemide administration is reduced in patients with kidney disease relative to healthy subjects.2,10,39 Although several mechanisms for the reduced medullary response to furosemide administration in patients with kidney disease have been proposed,2,32,39 the exact mechanism remains unclear. In diseased kidneys, the GFR is decreased, which decreases delivery of fluid to be absorbed in the thick ascending loop of Henle; thus, oxygen consumption required for active solute transport may also decrease.9,12,33 A decrease in GFR might be associated with a decrease in furosemide concentration within renal tissues; therefore, the response to furosemide in diseased kidneys with an impaired GFR may be less robust than in healthy kidneys.2,7 Diseased kidneys may also have altered sodium handling, mitochondrial metabolism, oxygen consumption, or tubular concentration capacity, all of which can reduce the response to furosemide administration.7,39 Results of 1 study7 indicate that the ∆R2* in the renal medulla is more sensitive than baseline R2* for distinguishing between human patients with CKD or hypertension and healthy volunteers. In patients with end-stage kidney disease, the baseline R2* may not increase or may decrease because oxygen consumption decreases as GFR decreases.10 Thus, the response to furosemide administration might be useful for identification of patients with renal dysfunction.

The clinical usefulness of renal cortical ∆R2* has not been established. Although the exact reason for the decrease in renal cortical R2* following furosemide administration is unknown, several potential mechanisms have been proposed. Because of partial volume or indistinct boundaries between the cortex and medulla on BOLD MRI images, the outer portion of the medulla might be inadvertently included in the ROI for the renal cortex. Moreover, although the effect of furosemide on renal perfusion remains controversial, it is clear that changes in renal perfusion affect cortical R2*. In dogs, renal blood flow as measured by an electromagnetic flowmeter40,41 and renal cortical oxygen tension as measured by oxygen electrodes41 increase after furosemide administration. It is also possible that, after furosemide administration, more oxygenated blood returns to the renal cortex owing to a decrease in medullary oxygen consumption.28

Blood oxygen level–dependent MRI has been used to evaluate pathophysiologic changes in patients with various kidneys disorders, such as CKD, renal arterial stenosis, diabetes mellitus, and hypertension, as well as kidney transplant recipients.2,3,4,5,7,8 For example, in rats with experimentally induced diabetes mellitus, the renal R2* increased while renal perfusion remained unchanged, which indicated that an increase in renal oxygen consumption was an important factor in the progression of CKD. In human patients with hypertension, the respective relationships between renal R2* and plasma aldosterone concentration and renin activity suggest that activation of the renin-angiotensin-aldosterone system is associated with renal hypoxia.42 Blood oxygen level–dependent MRI has also been proposed for the diagnosis of early-stage kidney disease and patient prognosis. For example, in human patients with CKD, the renal cortical R2* increases over time, and renal R2* is positively associated with the stage of CKD and negatively associated with the estimated GFR.2,4,5,7,8,24,43 In pigs6 and rats9 with experimentally induced renal artery stenosis, the R2* of the renal cortex and medulla increased as the severity of the stenosis increased and normalized after interventions to alleviate the stenosis. In human medicine, BOLD MRI can also be used to identify patients with renal disease who will most likely benefit from intervention.10,44 It has also been used to differentiate kidney transplant recipients who are experiencing transplant rejection from those with healthy kidneys. Among human kidney transplant recipients, the renal R2* in patients with acute transplant rejection or acute tubular necrosis is generally lower than the renal R2* in patients with normally functioning transplants owing to decreased oxygen consumption in delayed graft function, and the renal R2* in patients with acute transplant rejection is typically lower than that in patients with acute tubular necrosis.11,12,22,23 Blood oxygen level–dependent MRI has also be used to understand potential toxic mechanisms of drugs such as NSAIDs, renin-angiotensin-aldosterone inhibitors, and iodinated radiocontrast agents.45,46 In healthy human volunteers, the renal medullary R2* increased after injection of an iodinated radiocontrast agent but did not change after oral administration of the NSAID indomethacin.45

In veterinary medicine, BOLD MRI could be used to understand renal oxygenation changes in dogs with various kidney disease and other conditions (eg, dehydration) that can induce renal hypoxia. Renal insult is a major concern for clinical patients undergoing prolonged or multiple anesthetic sessions.47 In an experimental study48 in which healthy mice underwent BOLD MRI while anesthetized with each of 4 different anesthetic protocols, the renal R2* varied significantly among the anesthetic protocols, likely owing to variations in renal oxygen delivery secondary to alterations in systemic circulation induced by the anesthetic protocols.48 Therefore, BOLD MRI can be used to evaluate changes in renal oxygenation on the basis of duration or type of anesthesia.

Advances in the understanding of cardiorenal syndrome in veterinary patients has led to concerns about the link between heart disease and acute and chronic kidney injury in dogs and cats.49 The pathophysiologic mechanisms by which heart disease leads to kidney insult remain unclear in both human and veterinary medicine. Renal R2* increases in rats with experimentally induced myocardial infarction, which suggests that myocardial infarction induces renal hypoxia, which in turn can lead to tubulointerstitial injury.50 In dogs and cats with heart diseases such as myxomatous mitral valve disease and hypertrophic cardiomyopathy, BOLD MRI can be used to assess changes in renal oxygenation in response to disease progression or treatment. In human patients with hypertensive CKD, the renal R2* decreases following administration of captopril (an angiotensin converting enzyme inhibitor) or aliskiren (a renin inhibitor), which suggest that inhibitors of the renin-angiotensin system alleviate renal hypoxia and improve renal oxygenation.51 Although renin-angiotensin system inhibitors are commonly used in dogs with heart or kidney disease, the effect of those drugs on the kidney is unclear.

Measurement of renal oxygenation can help predict the life of transplanted kidneys. Kidney transplantation is uncommon in veterinary medicine but has been performed in cats for which the acute allograft rejection rate ranges from 13% to 26%.52 Although the diagnostic efficacy of venous oxidative markers and the resistive index of renal tissue as determined by ultrasonography have been investigated, renal biopsy remains the definitive method for diagnosing acute allograph rejection in cats following kidney transplantation.53,54 Blood oxygen level–dependent MRI is an emerging diagnostic modality, and the clinical indications for its use have not been established in human or veterinary medicine. Nevertheless, BOLD MRI has the potential to clarify many aspects of CKD and be of clinical benefit for predicting renal hypoxia.

Several factors require further investigation before BOLD MRI is recommended for assessment of renal oxygenation in patients with kidney disease. Multiple factors can affect or confound the BOLD MRI signal, and oxygenation is not homogenous within the kidneys.23,26,55 Because the R2* is dependent on tissue deoxyhemoglobin concentration, the BOLD MRI signal is affected by alterations in oxygen supply and consumption and the oxygen dissociation curve. Factors that affect renal oxygen supply and consumption include renal blood flow, Hct, water diuresis, diuretic administration, hydration status, dietary salt intake, carbon dioxide and oxygen breathing, pH, and body temperature. Those confounding physiologic factors can be calibrated with blood sampling and multiparametric MRI, such as diffusion-weighted imaging for probing tubular water, arterial spin rebelling for assessing renal blood flow, and contrast-enhanced imaging for monitoring renal blood volume.26 However, it is difficult to consider all confounding factors before BOLD MRI, and the extent to which these factors affect the renal R2* has yet to be established. Consequently, it is difficult to directly correlate renal R2* with renal Po2, and the relative change in R2* is the BOLD MRI parameter that has been most widely investigated. Further research is necessary to establish a mathematical model that estimates renal Po2 by use of R2* while controlling for potential confounders.

The normal heterogeneity in oxygenation within the kidneys may impair the diagnostic sensitivity of BOLD MRI to detect nonuniform alterations in renal oxygenation.55 Several analytic methods (eg, the fractional hypoxia method, compartmental method, and semiautomatic 12-layer concentric objects method) have been suggested to overcome the limitations associated with the normal heterogeneity in renal oxygenation, but each of those methods have disadvantages, and it is unknown which method is best for use in conjunction with BOLD MRI.22,55,56 Most researchers prefer ROI methods for BOLD MRI analysis owing to their widespread availability. Further research is warranted to standardize analytic methods for BOLD MRI.

In the present study, the intraobserver and interobserver reliability and test-retest repeatability of R2* and ∆R2* were generally good to excellent. Those findings might have been attributable to the selection of a consistent slice at the level of the renal hilus for BOLD image analysis and the use of large ROIs. However, the ∆R2* in the renal cortex had poor test-retest repeatability. The cortical ∆R2* had an extremely small ∆R2* (≤ 2.3 seconds–1); therefore, it may have been sensitive to small changes in R2* owing to physiologic changes or differences in the ROI settings between the first and second scans. Given the mechanism of furosemide action (induction of active tubular sodium reabsorption in the thick ascending loop of Henle in the outer portion of the renal medulla) and the low repeatability of cortical ∆R2*, the medullary ∆R2* is believed to be more useful than the cortical ∆R2* for clinical analysis of renal oxygenation by BOLD MRI.

Unlike human patients, dogs must be anesthetized to undergo BOLD MRI, and oxygen is typically administered to dogs during maintenance anesthesia. Oxygen administration can increase blood and tissue oxygen saturation, which may affect BOLD MRI signal intensity. However, studies involving human subjects20 and mice48 indicate that oxygen supplementation has little to no effect on BOLD MRI signal intensity. Because oxygen is considered necessary for maintaining stable anesthesia in dogs undergoing MRI, supplemental oxygen is not considered a critical limitation for performing BOLD MRI. It is generally recommended that administration of anesthetics that substantially affect the cardiovascular system be avoided in dogs undergoing BOLD MRI to minimize the effect of anesthesia on renal blood flow and oxygenation.48 For the dogs of the present study, anesthesia was induced with alfaxalone and maintained with isoflurane, 2 anesthetics that have minimal effects on the cardiovascular system.

The present study had several limitations. Actual (true) renal oxygenation was not measured owing to the invasiveness of the methods required to obtain that measurement. That was not considered a major limitation because BOLD MRI measurements consider relative oxygen availability rather than absolute oxygenation. Further research is necessary to elucidate the relationship between BOLD MRI and true renal oxygenation. The number of dogs evaluated in this study was small, which may have limited the power and ability to detect significant differences between measurements. Additionally, all 8 study dogs were young (mean age, 2 years; range, 1 to 3 years) healthy Beagles. In human subjects, the medullary ∆R2* decreases with age. Considering that dogs with kidney disease tend to be older, further investigation is warranted to evaluate the R2* and ∆R2* measurements derived for the dogs of this study and the usefulness of BOLD MRI for diagnosis of kidney disease in older dogs.

Results of the present study indicated that BOLD MRI was a feasible method that can be used to analyze renal cortical and medullary oxygenation and performed in dogs with a 3.0-T scanner to yield highly reliable and repeatable measurements. Administration of furosemide in conjunction with BOLD MRI can be useful for dynamic evaluation of oxygenation, particularly in the renal medulla. The T2*, R2*, and ∆R2* data generated for the heathy dogs of this study before and after furosemide administration can be used as preliminary guidelines against which measurements in dogs with kidney disease can be compared.

Acknowledgments

Supported by the Animal Medical Institute of Chonnam National University and the Basic Science Research Program through the National Research Foundation of Korea, funded by the Ministry of Science, ICT, and Future Planning (NRF-2018R1A2B6006775).

The authors declare that there were no conflicts of interest.

References

  • 1.

    Fine LG, Norman JT. Chronic hypoxia as a mechanism of progression of chronic kidney diseases: from hypothesis to novel therapeutics. Kidney Int. 2008;74(7):867872.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 2.

    Prasad PV, Thacker J, Li L-P, et al. Multi-parametric evaluation of chronic kidney disease by MRI: a preliminary cross-sectional study. PLoS One. 2015;10(6):e0139661.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 3.

    Schley G, Jordan J, Ellmann S, et al. Multiparametric magnetic resonance imaging of experimental chronic kidney disease: a quantitative correlation study with histology. PLoS One. 2018;13(7):e0200259.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 4.

    Hall ME, Jordan JH, Juncos LA, Hundley WG, Hall JE. BOLD magnetic resonance imaging in nephrology. Int J Nephrol Renovasc Dis. 2018;11:103112.

  • 5.

    Luo F, Liao Y, Cui K, Tao Y. Noninvasive evaluation of renal oxygenation in children with chronic kidney disease using blood-oxygen-level–dependent magnetic resonance imaging. Pediatr Radiol. 2020;50(6):848854.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 6.

    Juillard L, Lerman LO, Kruger DG, et al. Blood oxygen level–dependent measurement of acute intra-renal ischemia. Kidney Int. 2004;65(3):944950.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 7.

    Pruijm M, Hofmann L, Piskunowicz M, et al. Determinants of renal tissue oxygenation as measured with BOLD-MRI in chronic kidney disease and hypertension in humans. PLoS One. 2014;9(4):e95895.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 8.

    Li C, Liu H, Li X, Zhou L, Wang R, Zhang Y. Application of BOLD-MRI in the classification of renal function in chronic kidney disease. Abdom Radiol (NY). 2019;44(2):604611.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 9.

    Pohlmann A, Arakelyan K, Hentschel J, et al. Detailing the relation between renal T2* and renal tissue pO2 using an integrated approach of parametric magnetic resonance imaging and invasive physiological measurements. Invest Radiol. 2014;49(8):547560.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 10.

    Textor SC, Glockner JF, Lerman LO, et al. The use of magnetic resonance to evaluate tissue oxygenation in renal artery stenosis. J Am Soc Nephrol. 2008;19(4):780788.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 11.

    Sadowski EA, Fain SB, Alford SK, et al. Assessment of acute renal transplant rejection with blood oxygen level–dependent MR imaging: initial experience. Radiology. 2005;236(3):911919.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 12.

    Djamali A, Sadowski EA, Samaniego-Picota M, et al. Noninvasive assessment of early kidney allograft dysfunction by blood oxygen level–dependent magnetic resonance imaging. Transplantation. 2006;82(5):621628.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 13.

    Gullichsen E, Nelimarkka O, Halkola L, Niinikoski J. Renal oxygenation in endotoxin shock in dogs. Crit Care Med. 1989;17(6):547550.

  • 14.

    Nelimarkka O, Halkola L, Niinikoski J. Renal hypoxia and lactate metabolism in hemorrhagic shock in dogs. Crit Care Med. 1984;12(8):656660.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 15.

    Rosenbaum B, DiScala VA. Renal tubular sodium reabsorption and oxygen consumption in the hypothyroid dog. Metabolism. 1982;31(3):247251.

  • 16.

    Pedersen M, Dissing TH, Mørkenborg J, et al. Validation of quantitative BOLD MRI measurements in kidney: application to unilateral ureteral obstruction. Kidney Int. 2005;67(6):23052312.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 17.

    Li L-P, Lin J, Santos EA, Dunkle E, Pierchala L, Prasad P. Effect of nitric oxide synthase inhibition on intrarenal oxygenation as evaluated by blood oxygenation level–dependent magnetic resonance imaging. Invest Radiol. 2009;44(2):6773.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 18.

    dos Santos EA, Li L-P, Ji L, Prasad PV. Early changes with diabetes in renal medullary hemodynamics as evaluated by fiberoptic probes and BOLD magnetic resonance imaging. Invest Radiol. 2007;42(3):157162.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 19.

    Hirakawa Y, Tanaka T, Nangaku M. Renal hypoxia in CKD; pathophysiology and detecting methods. Front Physiol. 2017;8:99.

  • 20.

    Cox EF, Buchanan CE, Bradley CR, et al. Multiparametric renal magnetic resonance imaging: validation, interventions, and alterations in chronic kidney disease. Front Physiol. 2017;8:696.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 21.

    Morrell GR, Zhang JL, Lee VS. Magnetic resonance imaging of the fibrotic kidney. J Am Soc Nephrol. 2017;28(9):25642570.

  • 22.

    Pruijm M, Mendichovszky IA, Liss P, et al. Renal blood oxygenation level–dependent magnetic resonance imaging to measure renal tissue oxygenation: a statement paper and systematic review. Nephrol Dial Transplant. 2018;33(suppl 2):ii22ii28.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 23.

    Li L-P, Halter S, Prasad PV. Blood oxygen level–dependent MR imaging of the kidneys. Magn Reson Imaging Clin N Am. 2008;16(4):613625.

  • 24.

    Pruijm M, Milani B, Burnier M. Blood oxygenation level–dependent MRI to assess renal oxygenation in renal diseases: progresses and challenges. Front Physiol. 2017;7:667.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 25.

    Hwang SI, Lee HJ, Chin HJ, Chae D-W, Na KY. Evaluation of renal oxygenation in normal Korean volunteers using 3.0 T blood oxygen level–dependent MRI. J Korean Soc Magn Reson Med. 2013;17(1):1925.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 26.

    Niendorf T, Pohlmann A, Arakelyan K, et al. How bold is blood oxygenation level–dependent (BOLD) magnetic resonance imaging of the kidney? Opportunities, challenges and future directions. Acta Physiol (Oxf). 2015;213(1):1938.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 27.

    Bane O, Mendichovszky IA, Milani B, et al. Consensus-based technical recommendations for clinical translation of renal BOLD MRI. MAGMA. 2020;33(1):199215.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 28.

    Warner L, Glockner JF, Woollard J, Textor SX, Romero JC, Lerman LO. Determinations of renal cortical and medullary oxygenation using blood oxygen level–dependent magnetic resonance imaging and selective diuretics. Invest Radiol. 2011;46(1):4147.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 29.

    Woo S, Cho JY, Kim SY, Kim SH. Intravoxel incoherent motion MRI-derived parameters and T2* relaxation time for noninvasive assessment of renal fibrosis: an experimental study in a rabbit model of unilateral ureter obstruction. Magn Reson Imaging. 2018;51:104112.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 30.

    Vivier P-H, Storey P, Yamamoto A, et al. Furosemide and water load challenges in renal BOLD imaging. In: Proceedings of the European Congress of Radiology. European Society of Radiology; 2012.

    • Search Google Scholar
    • Export Citation
  • 31.

    Li LP, Storey P, Pierchala L, Li W, Polzin J, Prasad P. Evaluation of the reproducibility of intrarenal R2* and ΔR2* measurements following administration of furosemide and during waterload. J Magn Reson Imaging. 2004;19(5):610616.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 32.

    Thacker JM, Li L-P, Li W, Zhou Y, Sprague SM, Prasad PV. Renal blood oxygen level-dependent magnetic resonance imaging: a sensitive and objective analysis. Invest Radiol. 2015;50(12):821827.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 33.

    Gomez SI, Warner L, Haas JA, et al. Increased hypoxia and reduced renal tubular response to furosemide detected by BOLD magnetic resonance imaging in swine renovascular hypertension. Am J Physiol Renal Physiol. 2009;297(4):F981F986.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 34.

    Perinetti G. StaTips part IV: selection, interpretation and reporting of the intraclass correlation coefficient. South Eur J Orthod Dentofac Res. 2018;5(1):35.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 35.

    Pan J, Zhang H, Man F, et al. Measurement and scan reproducibility of parameters of intravoxel incoherent motion in renal tumor and normal renal parenchyma: a preliminary research at 3.0 T MR. Abdom Radiol (NY). 2018;43(7):17391748.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 36.

    Simon-Zoula SC, Hofmann L, Giger A, et al. Non-invasive monitoring of renal oxygenation using BOLD-MRI: a reproducibility study. NMR Biomed. 2006;19(1):8489.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 37.

    Behnia R, Koushanpour E, Goldstick T, Linde HW, Osborn R. Renal tissue oxygenation following induced hypotension in dogs. Br J Anaesth. 1984;56(9):10371043.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 38.

    Zhang JL, Rusinek H, Chandarana H, Lee VS. Functional MRI of the kidneys. J Magn Reson Imaging. 2013;37(2):282293.

  • 39.

    Gloviczki ML, Glockner JF, Crane JA, et al. Blood oxygen level–dependent magnetic resonance imaging identifies cortical hypoxia in severe renovascular disease. Hypertension. 2011;58(6):10661072.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 40.

    Ludens JH, Williamson HE. Effect of furosemide on renal blood flow in the conscious dog. Proc Soc Exp Biol Med. 1970;133(2):513515.

  • 41.

    Nuutinen LS, Tuononen S. The effect of furosemide on renal blood flow and renal tissue oxygen tension in dogs. Ann Chir Gynaecol. 1976;65:272276.

    • Search Google Scholar
    • Export Citation
  • 42.

    Vink EE, de Boer A, Hoogduin HJM, et al. Renal BOLD-MRI relates to kidney function and activity of the renin-angiotensin-aldosterone system in hypertensive patients. J Hypertens. 2015;33(3):597604.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 43.

    Zhou H, Yang M, Jiang Z, Ding J, Di J, Cui LI. Renal hypoxia: an important prognostic marker in patients with chronic kidney disease. Am J Nephrol. 2018;48(1):4655.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 44.

    Chrysochou C, Mendichovszky IA, Buckley DL, Cheung CM, Jackson A, Kalra PA. BOLD imaging: a potential predictive biomarker of renal functional outcome following revascularization in atheromatous renovascular disease. Nephrol Dial Transplant. 2012;27(3):10131019.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 45.

    Hofmann L, Simon-Zoula S, Nowak A, et al. BOLD-MRI for the assessment of renal oxygenation in humans: acute effect of nephrotoxic xenobiotics. Kidney Int. 2006;70(1): 144150.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 46.

    Hall ME, Rocco MV, Morgan TM, et al. Beta-blocker use is associated with higher renal tissue oxygenation in hypertensive patients suspected of renal artery stenosis. Cardiorenal Med. 2016;6(4):261268.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 47.

    Rogers-Smith E, Hammerton R, Mathis A, Allison A, Clark L. Twelve previously healthy non-geriatric dogs present for acute kidney injury after general anaesthesia for non-emergency surgical procedures in the UK. J Small Anim Pract. 2020;61(6):363367.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 48.

    Niles DJ, Gordon JW, Fain SB. Effect of anesthesia on renal R2* measured by blood oxygen level–dependent MRI. NMR Biomed. 2015;28(7):811817.

  • 49.

    Pouchelon JL, Atkins CE, Bussadori C, et al. Cardiovascular–renal axis disorders in the domestic dog and cat: a veterinary consensus statement. J Small Anim Pract. 2015;56(9):537552.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 50.

    Chang D, Wang Y-C, Xu T-T, et al. Noninvasive identification of renal hypoxia in experimental myocardial infarctions of different sizes by using BOLD MR imaging in a mouse model. Radiology. 2018;286(1):129139.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 51.

    Siddiqi L, Hoogduin H, Visser F, Leiner T, Mali WP, Blankestijn PJ. Inhibition of the renin‐angiotensin system affects kidney tissue oxygenation evaluated by magnetic resonance imaging in patients with chronic kidney disease. J Clin Hypertens (Greenwich). 2014;16(3):214218.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 52.

    Aronson LR. Update on the current status of kidney transplantation for chronic kidney disease in animals. Vet Clin North Am Small Anim Pract. 2016;46(6):11931218.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 53.

    Halling KB, Graham JP, Newell SP, et al. Sonographic and scintigraphic evaluation of acute renal allograft rejection in cats. Vet Radiol Ultrasound. 2003;44(6):707713.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 54.

    Halling KB, Ellison GW, Armstrong D, et al. Evaluation of oxidative stress markers for the early diagnosis of allograft rejection in feline renal allotransplant recipients with normal renal function. Can Vet J. 2004;45(10):831837.

    • Search Google Scholar
    • Export Citation
  • 55.

    Neugarten J, Golestaneh L. Blood oxygenation level–dependent MRI for assessment of renal oxygenation. Int J Nephrol Renovasc Dis. 2014;7:421435.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 56.

    Chen F, Li S, Sun D. Methods of blood oxygen level–dependent magnetic resonance imaging analysis for evaluating renal oxygenation. Kidney Blood Press Res. 2018;43(2): 378388.

    • Crossref
    • Search Google Scholar
    • Export Citation

Contributor Notes

Address correspondence to Dr. Choi (imsono@snu.ac.kr).