Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) allows in vivo measurement of many physiological parameters of clinical interest (1-3). Following intravenous administration, the exogenous paramagnetic contrast agent gadopentetate-dimeglumine (Gd) is distributed throughout the body, except in the central nervous system (CNS) tissue because of the blood-brain barrier (BBB) and the blood-spinal cord barrier (BSCB). In the presence of CNS pathology, the barrier is regionally disrupted; Gd no longer remains intravascular and leaks into the interstitial tissue spaces. Measuring the rate of this leakage facilitates quantitative evaluation of barrier permeability. For example, the BBB in brain tumors or injuries was successfully characterized in vivo using Gd concentration data obtained from DCE-MRI (4). This technique has not yet been applied to evaluate the status of BSCB following traumatic spinal cord injury (SCI) or disease. Quantitative measurements of changes in BSCB permeability in vivo can be used to assess the severity of injury, monitor the evolution of cord pathology, identify the effective time interval for delivering drugs to the cord tissue across the compromised barriers, and observe the endogeneous repair processes (5).In this paper, we report results from in vivo DCE-MRI experiments performed on rats in the acute phase of SCI with Gd administered intravenously as a bolus. The time course and distribution of Gd concentration in injured spinal cord (SC) tissue and cerebrospinal fluid (CSF) were determined from intensity enhancements on the postcontrast T 1 -weighted images that were acquired sequentially. To represent the concentration data, a pharmacokinetic model was developed under limited permeability conditions for BSCB (6). This model consists of three compartments for plasma, injured cord, and CSF. Inclusion of a separate compartment for CSF was necessary because even though in humans CSF intensity remains nearly unchanged after Gd delivery in normal subjects, intensity enhances almost immediately in the CSF of normal rats (see the Discussion section). In this model, Gd transport between compartments was represented by different transfer rate constants. A window-based simulation analysis and modeling software was used to estimate the transfer parameters from the concentration data in plasma, CSF, and injured cord. The estimates associated with injured cord were ultimately used to evaluate the status of BSCB. Unlike the two-compartment pharmacokinetic models suggested specifically for studying BBB (4), the proposed model with three compartments takes into account the exchange of Gd between plasma and CSF. The utility of the model is discussed in the context of primary and secondary pathobiological processes that take place in the acute phase of SCI and the time window of opportunity for any therapeutic intervention.
MATERIALS AND METHODS
Animals and SCIA total of 11 (eight injured, two laminectomy controls, and one naïve control) male Sprague-Dawley rats weighing 350 -400 g were used...