Original Research Accuracy of Blood Flow Values Determined by Arterial Spin Labeling: A Validation Study in Isolated Porcine Kidneys Carsten Warmuth, PhD, 1 Stefan Nagel, PhD, 2,4 Oliver Hegemann, MS, 2,4 Waldemar Wlodarczyk, PhD, 3 and Lutz Lu ¨ demann, PhD 3 * Purpose: To validate the accuracy of quantitative blood flow values determined using pulsed arterial spin labeling (ASL) in the preserved and reperfused porcine kidney. Materials and Methods: Ex vivo porcine kidneys were per- fused with whole blood under physiological conditions, in particular including pulsatile flow. Total flow through the kidney was determined using an ultrasound flowmeter. ASL measurements at two different inversion times and four different flow rates in the range of 70 –210 mL/100 mL*minute were performed. Absolute values of blood flow and arterial transit times were determined in the kidney cortex. Results: The quantitative values were in good agreement with the reference values obtained after calibration of the total flow. The greatest difference observed was 13%. Conclusion: Isolated organ hemoperfusion allows validat- ing perfusion imaging techniques. The experimental setup enables long-term radiotherapeutic or toxicological studies using noninvasive ASL to monitor blood flow quantitatively. Key Words: perfusion quantification; arterial spin labeling; isolated porcine kidney perfusion; blood flow; perfusion model J. Magn. Reson. Imaging 2007;26:353–358. © 2007 Wiley-Liss, Inc. MAGNETIC RESONANCE IMAGING (MRI) provides not only morphologic but functional information as well. Perfusion is one of the most important functional pa- rameters. Quite a number of perfusion weighted MRI techniques have been developed, which differ in their invasiveness and the ability to provide accurate quan- titative values, e.g., for blood flow. However, validation of these techniques is cumbersome in practice. Labeled microspheres as a tracer are often used as a gold stan- dard for perfusion quantification—a destructive method that, in addition, requires arterial blood sam- pling. Physiologically accurate perfusion phantoms are difficult to implement. One way to overcome these lim- itations is to investigate explanted organs under he- moperfusion. A previously presented isolated kidney hemoperfusion system (1–2) capable of maintaining physiologic organ function over several hours, was made MR compatible by using pneumatically driven artificial heart pump (Berlin heart AG, Berlin, Ger- many) blood pumps. The setup allows for direct manip- ulation and calibration of total blood flow through the organ. As a truly noninvasive perfusion MRI method, arterial spin labeling (ASL) has become increasingly popular. The basic approach common to all implementations is to manipulate the longitudinal magnetization of arterial blood supplying the tissue of interest (3,4). An excellent review on the topic of ASL is given by Golay et al (5). In theory, the method allows for absolute quantification of regional blood flow. The labeling may be continuous, pulsed, or even periodic; blood magnetization may be saturated or inverted. A certain transit time is needed for blood to travel from the labeling region to the image plane. If blood water is freely diffusible, prepared blood protons exchange their magnetization with tissue pro- tons and can be detected as signal intensity variations. In the simplest case, subtraction of images acquired with and without labeling removes the background sig- nal of static tissue, the remaining difference signal is that of tagged blood that reached the slice. Two condi- tions must be fulfilled for accurate perfusion measure- ment using ASL: 1) all blood that is going to perfuse a voxel must be tagged, in particular spin labeling cannot detect in-plane flow; and 2) the tagged blood must reach the voxel before T 1 relaxation destroys the label. Most organs, like the brain or the kidneys, have a strong directionality in blood flow and sufficiently short transit 1 Department of Radiology, Charite ´–Universitary Medicine Berlin, Cam- pus Charite ´-Mitte (CCM), Berlin, Germany. 2 Department of Comparative Medicine and Experimental Animal Sci- ence, Charite ´–University Medicine Berlin, Campus Virchow Klinikum (CVK), Berlin, Germany. 3 Department of Radiology, Nuclear Medicine, and Radiooncology, Charite ´ –Universitary Medicine Berlin, Campus Virchow Klinikum (CVK), Berlin, Germany. 4 Department of Comparative Medicine and Experimental Animal Sci- ence, Charite ´–University Medicine Berlin, Campus Virchow Klinikum (CVK), Berlin, Germany. Contract grant sponsor: German Research Foundation (Deutsche For- schungsgemeinschaft, DFG); Contract grant number: GRK 238/2. *Address reprint requests to: L.L., Institut fu ¨ r Strahlenheilkunde, Charite ´ Universitary Medicine Berlin, Augustenburger Platz 1, 13353 Berlin, Germany. E-mail: lutz.luedemann@charite.de Received August 3, 2006; Accepted March 8, 2007. DOI 10.1002/jmri.21011 Published online in Wiley InterScience (www.interscience.wiley.com). JOURNAL OF MAGNETIC RESONANCE IMAGING 26:353–358 (2007) © 2007 Wiley-Liss, Inc. 353