Original Research Effect of Lung Inflation on Arterial Spin Labeling Signal in MR Perfusion Imaging of Human Lung Vu M. Mai, PhD, 1 * Qun Chen, PhD, 1 Alexander A. Bankier, MD, 1 Michael Blake, MD, 1 Klaus D. Hagspiel, MD, 2 Jack Knight-Scott, PhD, 3 Stuart S. Berr, PhD, 2 and Robert R. Edelman, MD 1 The effect of lung inflation on arterial spin-labeling signal in lung perfusion is investigated. Arterial spin-labeling schemes, called alternation of selective inversion pulse (ASI) and its hybrid (HASI), which uses blood water as an endogenous, freely diffusible tracer, were applied to mag- netic resonance (MR) perfusion imaging of the lung. Perfu- sion-weighted images of the lung from nine healthy volun- teers were obtained at different time delays. There was a significant signal difference in ASI images acquired at dif- ferent respiratory phases. Greater signal enhancement has been observed when the volunteers performed breath hold- ing on end expiration than on end inspiration. This is in agreement with the normal physiologic effect of lung in- flation on the pressure-flow relationship of pulmonary vas- culature. ASI and HASI perfusion-weighted images show similar lung features and image quality. Preliminary re- sults from pulmonary embolism patients indicate that ar- terial spin labeling is sensitive for the detection of areas of perfusion deficit. J. Magn. Reson. Imaging 2001;13: 954 –959. © 2001 Wiley-Liss, Inc. Index terms: arterial spin labeling; ASI; HASI; lung inflation; pulmonary perfusion; pulmonary embolism; lung parenchyma THE ASSESSMENT OF PULMONARY PERFUSION is of great clinical importance. Detecting regional perfusion change may help evaluate the functionality and viability of lung transplants or diagnose pathologies such as pul- monary embolism. Magnetic resonance imaging (MRI) of the lung has been traditionally difficult because of its unique structure. Its low proton density (approximately 20%–30%) inherently produces a weak signal. The large variation of magnetic susceptibility caused by interfaces between air and tissue induces inhomogeneous intra- voxel fields, which rapidly dephase the already weak sig- nal, making the signal detection difficult with conven- tional MR sequences (1,2). Furthermore, cardiac and respiratory motions potentially introduce significant arti- facts. Arterial spin-labeling (ASL) techniques have been introduced to noninvasively measure cerebral blood flow using blood water as an endogenous, freely dif- fusible tracer (3– 6). ASL techniques generally involve the acquisition of a control and a tag image. The tag image results when the inflowing blood has been magnetically labeled, usually through an inversion pulse. The control image, on the other hand, is ob- tained with the inflowing blood remaining fully re- laxed. Subtracting the tag from the control image produces a perfusion-weighted image, in which signal enhancement reflects the degree of tissue perfusion. The detected ASL signal change is normally on the order of a few percent, and as a result, significant signal averaging is required. With the inherently weak signal of the lung and the small ASL signal change, ASL techniques face significant challenges with respect to pulmonary perfusion imaging. Never- theless, detection of pulmonary perfusion is possible, due to the hypervascularity of the lung, and has been successfully demonstrated using different ASL tech- niques (7,8). In addition, detection of perfusion defi- cits in an animal model of pulmonary embolism (PE) and airway obstruction has been demonstrated (9,10). It is known that lung inflation has a significant effect on the pressure-flow relationship of the pulmonary vas- culature, and as a result, there exists a difference in perfusion between different states of lung inflation (11). Consequently, a change in the ASL signal is expected. In this study, we report the application of the alterna- tion of selective inversion pulse (ASI) technique and its hybrid (HASI) to pulmonary perfusion imaging and the effect of respiratory phases on ASL signals in the lung. The two respiratory phases that were investigated are breath holding on end inspiration and end expiration. 1 Department of Radiology, Beth Israel Deaconess Medical Center and Harvard School of Medicine, Boston, Massachusetts. 2 Department of Radiology, Health Sciences Center, University of Vir- ginia, Charlottesville, Virginia. 3 Biomedical Engineering, Health Sciences Center, University of Vir- ginia, Charlottesville, Virginia. Contract grant sponsor: American Heart Association; Contract grant sponsor: NIH; Contract grant number: R01HL57437. Presented in part at the 8th Annual Meeting of the International Society for Magnetic Resonance in Medicine, Denver, 2000. *Address reprint requests to: V.M.M., Department of Radiology, Evan- ston Hospital, Evanston Northwestern Healthcare, 2650 Ridge Avenue, Evanston, IL 60201. E-mail: vmai@enh.org Received May 12, 2000; Accepted November 30, 2000. JOURNAL OF MAGNETIC RESONANCE IMAGING 13:954 –959 (2001) © 2001 Wiley-Liss, Inc. 954