Hybrid Reference Tissue Calibrated Dual-Bolus 3D Quantitative Dynamic Contrast-Enhanced MRI in a Rabbit VX2 Tumor Model

Introduction: Dynamic contrast-enhanced (DCE) MRI can be used to quantitatively measure pharmacokinetic parameters that indicate tumor angiogenesis and perfusion. The ability of quantitative DCE-MRI to generate physiological parameter maps has important implications in staging tumors and in evaluating the cancer therapeutic response. The purpose of this study was to develop an innovative hybrid reference tissue calibrated dual-bolus 3D quantitative DCE-MRI method. Methods: In this study, we implanted six VX2 tumors total in the uteri of six rabbits [1]. 3 weeks after implantation, rabbits were imaged using a 1.5T clinical MRI scanner (Siemens Magnetom Sonata) with a dual-bolus DCE-MRI approach [2]. Before dynamic scan, an in vivo 3D B1 map was generated using the double-angle method, and a baseline 3D R10 map was acquired using the variable flip-angle (FA) GRE method. We performed 3D dynamic GRE MRI to record the tissue enhancement over time after the first IV bolus injection of 0.04125 mmol/kg Gd-DTPA contrast agent. 10 minutes later, we used a 2D dynamic saturation recovery (SR) GRE sequence to characterize the shape of arterial input function (AIF) in the aorta after the second IV bolus injection of 0.005 mmol/kg Gd-DTPA, assuming a linear relationship between signal intensity and contrast agent concentration (Fig 1a). Imaging parameters included: 200 mm FOV; 3D GRE: TR/TE = 6/1.66 ms, 128×72×8 matrix, baseline FA = 2°, 9°, 19°, 4 averages; dynamic FA = 9°, 1.6 sec sampling rate; 2D SR GRE: TR/TE/TI = 3.2/1.2/47 ms, 128×64 matrix, FA = 15°, 0.4 sec sampling rate. With B1 calibration and a baseline R10 map, 3D R1 map time series and further contrast concentration map series were derived from the 3D DCE-MR image series [3]. The extended Tofts model was used for perfusion analysis, generating K (volume transfer coefficient, min), ve (fractional volume occupied by extracellular extravascular space), and vp (fractional volume of plasma space) parameter maps [4]. A lookup method was used to estimate the delay time during curve fitting (Fig 1b) [5]. Region of interest (ROI) analysis was used to compare the tumor hypervascular region, tumor core, and back muscle regions on the parameter maps for all six animals (Fig 2). Literature reference and measured ve ratios from the back muscle were used to construct the final calibration for overall measurements [6].