In Vivo Determination of Absolute Cerebral Blood Volume Using Hemoglobin as a Natural Contrast Agent: An MRI Study Using Altered Arterial Carbon Dioxide Tension

Abstract

Summary:The ability of the magnetic resonance imaging transverse relaxation time, R2 = IIT2, to quantify cerebral blood volume (CBV) without the need for an exogenous con trast agent was studied in cats (n = 7) under pentobarbital anesthesia. This approach is possible because R2 is directly affected by changes in CBF, CBV, CMR02, and hematocrit (Hct), a phenomena better known as the blood-oxygenation level-dependent (BOLD) effect. Changes in CBF and CBV were accomplished by altering the carbon dioxide pressure, Paco2, over a range from 20 to 140 mm Hg. For each Paco2 value, R2 in gray and white matter were determined using MRI, and the whole-brain oxygen extraction ratio was obtained from arteriovenous differences (sagittal sinus catheter). Assuming a Recent advances in functional magnetic resonance im aging (MRI) have allowed the design of MRI techniques for quantitative determination of hemodynamic param- oxygen extraction ratio; R2, transverse relaxation time constant; R�. effective transverse relaxation time constant, equals R2 plus coherent dephasing effects owing to local field inhomogeneities and suscepti bility differences; S, signal attenuation; SE, spin echo; TE, echo time; TR, repetition time; Xlicoxy' fraction of deoxygenated hemoglobin; Yo' arterial oxygenation fraction; T, lifetime between exchanges; �w, sus ceptibility shift difference; T C PMG' delay for a single echo refocusing in a Carr-Purcell-Meiboom-Gill experiment. 809constant CMR02, the microvascular CBV was obtained from an exact fit to the BOLD theory for the spin-echo effect. The resulting CBV values at normal Paco2 and normalized to a common total hemoglobin concentration of 6.88 mmollL were 42 ± 18 fLLlg (n = 7) and 29 ± 19 fLLlg (n = 5) for gray and white matter, respectively, in good agreement with the range of literature values published using independent methodologies. The present study confirms the validity of the spin-echo BOLD theory and, in addition, shows that blood volume can be quan tified from the magnetic resonance imaging spin relaxation rate R2 using a regulated carbon dioxide experiment. Key Words: Spin-echo-BOLD-MRI-Cerebral blood volume-Carbon dioxide-Transverse relaxation rate R2.eters. Approaches to quantify cerebral blood volume (CBV) generally involve injection of a paramagnetic contrast agent (Belliveau et a!., 1990), whereas CBF quantification has used either contrast agents (Hamberg et a!., 1993; Kucharczyk et aI., 1993; Guckel et ai, 1994) or frequency labeling of arterial water nuclei Kim, 1995). Analogous to radioactive tracer methods, these new MRI techniques require either the assumption or measurement of an ar terial input function. Eliminating the need for such a function using a natural contrast agent studied under equilibrium conditions would be beneficial. Magnetic resonance imaging has the capability of doing this through the so-called blood-oxygenation-Ievel dependent (BOLD) effect (Ogawa et aI., 1990(Ogawa et aI., , 1992(Ogawa et aI., , 1993a(Ogawa et aI., , 1993bMoonen et a!., 1990;Prielmey...

The interaction between magnetization transfer and blood‐oxygen‐level‐dependent effects

“…These voxel components each have individual tissue relaxation and MT parameters, and their relative contributions depend on the size and location of the voxel and the MRI acquisition parameters used. For homogeneously perfused tissue (parenchyma), the normalized imaging signal intensity as a function of echo time (TE) is a multiexponential decaying process (12,16,17),…”

“…We assume the tissue model used in Refs. (12,16,17), in which parenchyma consists of tissue, arterioles, venules, and capillaries. However, contrary to these papers, we now assume a constant total parenchymal water content (18), as shown to be the case in recent experiments (28).…”

“…Because arterioles, capillaries, and venules occupy approximately 21, 33, and 46% of the total CBV (11), respectively, one hasx blood,a ϭ 0.21CBV/V par , x blood,c ϭ 0.33CBV/V par , and x blood,v ϭ 0.46CBV/V par . To correlate CBV with the oxygen extraction ratio (OER), we used the OER-CBV relationship based on a cylindrical vessel model (12,16): CBV ϭ CBV norm (OER norm /OER) 0.5 , in which OER norm ϭ 0.55 (16). The blood water density of 0.873 g water/mL blood (or 0.873 mL water/mL blood) at the hematocrit fraction (Hct) of 0.35 (20) was used for CBV unit conversion from milliliters of blood to grams of water (or milliliters of water).…”

“…[A3] measured for the cat (16). Finally, we assumed that a previously reported (16) empiric OER-P a CO 2 relationship in the cat brain, OER ϭ 6.848 ⅐ P a CO 2 Ϫ 0.737 (see Appendix I), is applicable to the rat brain.…”

“…We used the OER-CBV relationship for cylindrical vessels (12,16): CBV ϭ CBV norm ͑OER norm /OER͒ 0.5 , in which CBV norm ϭ 5.2 mL blood/100 g brain for GM and 2.8 mL blood/100 g brain for WM (17), OER norm ϭ 0.55 (16). The total water contents of GM and WM in the brain were previously measured to be 84.69 and 70.37 g water/100 g brain (40), respectively.…”