3. MR Method  To minimize exchange effects, we used a fast gradient echo method with 1/TR >> R e , i.e., exchange-minimized regime [2]. RBV is then calculated using equations for slow exchange.  In the slow exchange limit, RBV is calculated: where: SI lung = signal from ROI of lung before the contrast agent SI lung + Fe = signal from ROI of lung after contrast agent SI blood = signal from control blood, without contrast agent SI blood + Fe = signal from control blood with contrast agent V b is the mean volume occupied by the blood within a voxel Accurate Quantification of Fractional Blood Volume in Lung Tissue G. P. Topulos 1 , S. Patz 2 , J. P. Butler 3 , L. L. Tsai 4 , R. W. Mair 4 , M. S. Rosen 4 , R. L. Walsworth 4 1 Department of Anaesthesia, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA, USA 2 Department of Radiology, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA, USA 3 Harvard School of Public Health, Boston, MA, USA 4 Harvard-Smithsonian Center for Astrophysics, Cambridge, MA, USA 1. Introduction  Measures of regional blood volume fraction (RBV) are important for studying physiological changes in many organs.  RBV is difficult to measure generally, however recent attempts using MRI have successfully measured RBV in organs such as the kidneys and the heart [1,2].  RBV in the lung has not been attempted by MRI, due to low signal and susceptibility artifacts.  Our goal was to demonstrate for the first time that MRI can be used to quantitatively determine RBV in lungs.  Determinations of RBV allow calculation of the lung compliance, an important physiological parameter. 2. Background  Measuring RBV with MRI requires measurement of MRI signal before and after administering an intra-vascular contrast agent.  The fast exchange limit method assumes sufficiently fast water exchange between extra- and intra-vascular compartments such that during TR, water protons fully sample the environment of both compartments.  The “slow exchange limit” method assumes negligible proton exchange during TR.  Biological exchange rates R e (1-10Hz) are same order as typical 1/TR (2.5-10Hz). Therefore neither the fast nor slow exchange models are appropriate [2]. 5. Phantom system RBV method tested using a dialysis canister as a phantom.  Canister contains a large number of parallel capillary fibers.  Fibers and tube can be filled independently with different solutions.  Solution inside fibers serves as the intra-vascular volume.  Dialysate region (outside the fibers) is the model for the extra-vascular space (tissue).  Compartments physically isolated but can exchange water.  Feridex does not dialyze, remaining “intra-vascular”.  Supplier: Fresenius Medical Inc., Model F8. 6. Phantom Characterization  Inner solution volume fraction of dialysis canister was determined independently from confocal microscopy images of a cross section of the canister (right).  From the confocal images we found: - 12,819 fibers in the canister - fiber average inner diameter = 189 µm - Measured ID of the entire canister = 4.80 cm.  Inner solution volume fraction = = 0.199 ± 0.001. 10. Animal Preparation RBV was measured in isolated perfused lungs from 3 rabbits.  Rabbits (3-4 kg.) were anesthetized, exsanguinated, heart and lungs excised en bloc. Pulmonary artery (PA) and vein (PV) were cannulated via right ventricle and left atrial appendages.  The PA and PV cannulae were each connected to separate reservoir bags filled with blood. The height of the reservoir bags could be adjusted to control the PA pressure (Ppa) and PV pressure (Ppv).  Entire preparation placed in 20cm solenoidal extremity coil.  Tracheal cannula connected to constant pressure oxygen source. Before imaging, lung was set at target transpulmonary pressure (Ptm) to control volume.  During imaging, PV cannula was clamped, and intravascular pressure (Pao) was controlled by adjusting PA reservoir height.  After imaging pre-Feridex, all blood in the vasculature was collected, and doped at a dose of 0.001 ml Feridex/ml blood. Lungs were re-perfused with doped blood, and a sample collected for calibration. 9. Animal Preparation 4. Chosen MR Contrast Agent Contrast agent chosen for these experiments was Feridex  Solution of colloidal iron oxide particles  contains Fe 2+ and Fe 3+ with a core size of 4.3 - 5.6 nm, covered in dextran to give large particle size ~ 50 nm  obtained from Advanced Magnetics, Inc., Cambridge, MA.  Very long (~2 hour) intra-vascular lifetime  In slow exchange limit, only the signal from blood changes after adding Feridex since it remains intra-vascular  Primarily cleared by liver macrophage activity 7. Phantom Characterization 8. MRI measurement of Canister RBV  Instrument: GE Profile 0.2 T open-access scanner at Brigham & Women’s Hospital.  Sequence: 2D spoiled GRASS (Gradient Echo Steady State) matrix: 256 μ256, FOV = 40 cm, TR/TE =18/6 ms, NS = 8, slice thickness = 1 cm. Total scan time = 37s.  Pre-contrast images taken with both compartments filled with saline, post contrast images acquired after perfusing the intra-vascular compartment with Feridex-doped saline.  Syringes filled with saline and Feridex-doped saline were simultaneously imaged. Doped saline contained 0.2mM Fe, such that water T 1 ~ 200 ms and T 2 ~50 ms.  The MRI determined RBV (see panel 3) was 0.200 ± 0.003. 11. MR Method  Instrument: GE Profile 0.2 T scanner.  Sequence: 3D SPGR (Spoiled Gradient Echo) matrix: 128 μ128μ26, FOV = 17 cm, NS = 1, TR/TE = 25/8 ms, flip angle=30 o , resn: 1.5μ1.5μ5 mm. Total scan time = 1 min 46 s . 12. Image Analysis  In each 2D plane that contained lung tissue, ROI's were chosen over, avoiding large vessels (light) or airways (dark)  Generally, ROIs were chosen at four locations on each plane, upper and lower right lobe, upper and lower left lobe.  Total voxels in all ROI’s from each 3D data set ~ 1000.  ROI graphical boundary was pasted from the pre-Feridex to corresponding post-Feridex plane to insure sampling of identical regions before and after contrast.  Signal intensity SI for each ROI was recorded. Mean SI ± standard error of the means was calculated for each dataset.  Control and Feridex-doped blood SI obtained from an ROI containing control and doped blood syringes in each image.  The RBV results are given in Table 1 13. RBV Results Table 1. Measured RBV at different Pao and Ptm values. Pao = airway opening pressure = lung recoil pressure (cm of H 2 0) Ptm = capillary transmural pressure = vascular pressure (cm of H 2 0)  Our results are close to previously measured values. - RBV in dog lungs near resting volume (microscopy) = 0.08 [3] - RBV in humans (PET scans) ≈ 0.15 [4,5].  These data validate our proposed RBV measurement method. 14. 3D Images 15. Lung Compliance  Lung compliance is measured from the change in RBV with varying vascular pressure at constant lung recoil pressure. Thus ∆RBV was calculated for Ptm = 3 and 20 cm H 2 O.  The results show that at higher lung recoil pressures, the lung gets stiffer and is less compliant.  Table 2: ∆RBV for a change in vascular pressure of 15 cm H 2 O at two different lung recoil pressures. 16. Conclusion We have demonstrated that MRI can be used to accurately measure RBV of the lung by using an intra-vascular contrast agent together with a simple 3D gradient echo sequence. In addition, we have demonstrated the utility of this technique in measuring a functional lung parameter, the lung compliance. References  1. C. Schwarzbauer, J. Syha, and A. Haase. Magn Reson Med, 29, 709-12 (1993).  2. K. Donahue, R. Weisskoff, D. Chesler, K. Kwong, A. Bogdanov, Jr., J. Mandeville, and B. Rosen. Magn Reson Med, 36, 858-67 (1996).  3. R. Crapo, J. Crapo, and A. Morris. Respir Physiol, 49, 175-86 (1982).  4. L. Brudin, C. Rhodes, S. Valind, T. Jones, and J. Hughes. J Appl Physiol, 76, 1205-10 (1994).  5. L. Brudin, C. Rhodes, S. Valind, T. Jones, B. Jonson, and J. Hughes. J Appl Physiol, 76 1195-204 (1994). RBV V V SI SI SI SI b voxel lung Fe lung blood Fe blood = = − − + + Figure 1. Cross section confocal microscope image of dialysis canister (id = 4.8 cm). Figure 2. Cross section confocal microscope image showing diameter in µm of individual capillary fibers. RBV N r r m cm fibers fiber canister ≈ × = × ( ) ( ) π π π µ π 2 2 2 2 12819 94 5 24 . . Figure 3. GE 0.2T Profile scanner with excised perfused rabbit lung. Figure 4. Rabbit lungs in MRI coil with pulmonary artery and vein cannulas. Tracheal cannula is out of view. Saline doped with Feridex Undoped blood Blood doped with Feridex Figure 5. 2D plane from 3D image of rabbit lung with vascular system doped with Feridex. ROI analysis conducted on these images plane by plane (see panel 12). Pulmonary artery and vein cannulae Figure 6. Volume rendered image of lung surface from 3D data set. Pulmonary artery and vein cannulae Figure 7. Threshold-rendered image of 3D data showing vascular tree and heart. 0.105 ± 0.001 0.101 ± 0.001 0.218 ± 0.003 0.160 ± 0.002 3 0.157 ± 0.002 0.126 ± 0.002 0.192 ± 0.003 0.152 ± 0.003 2 0.197 ± 0.004 0.183 ± 0.004 0.276 ± 0.005 0.205 ± 0.004 1 Pao=20 Ptm=20 Pao=20 Ptm=5 Pao=3 Ptm=20 Pao=3 Ptm=5 Animal p < 0.025 p < 0.005 p value - 1 tail T test 0.016 ± 0.002 0.056 ± .003 Mean ∆RBV 0.004 ± 0.002 0.058 ± .004 3 0.031 ± 0.002 0.040 ± .004 2 0.014 ± 0.006 0.071 ± .007 1 Pao=20 ∆Ptm=15 Pao=3 ∆Ptm=15 Animal