TH-CD-207-02: Quantification of Dynamic Susceptibility Contrast MRI: Comparison of Calibration Methods Using Venous Output Function and Arterial Spin Labeling

Abstract

Purpose: To evaluate the validity of the calibration of arterial input function (AIF) by using a venous output function (VOF) for the quantification of cerebral blood flow (CBF) in dynamic susceptibility contrast MRI (DSC-MRI), by comparing to the arterial spin labeling (ASL) measurement. Method: Fourteen patients with unilateral ICA stenosis enrolled in the study on a 3T clinical MRI scanner. The ASL-CBF map was calculated by averaging 45 dynamic points acquired by using a Q2TIPS sequence. For DSC perfusion analysis, AIF was selected to derive the relative CBF (rCBF) map and the delay (Tmax) map; VOF with signal saturation was corrected by a small vein which was amplified and time-shifted to match the base and flanks of the original VOF. The DSC-CBF map was corrected by two different patient-specific correction factors (CF): (1) CF1: the ratio of the mean ASL-and DSC-CBF obtained from the mask based on Tmax < 3s, the gray matter mask and the exclusion of large vessels; (2) CF2: the ratio of the area under the first-pass phase in AIF and VOF concentration time curves. Result: Average whole-brain (WB) DSC-CBF calculated by CF1 and CF2 showed the same precision (coefficient of variation [COV] = 22%), while CBFCF2 (100.26 ± 23.36 ml/100g/min [mean ± SD]) was higher than CBFCF1 (37.4 ± 8.1 ml/100g/min). Significant correlation (R= 0.632; P<0.05) was found between the results obtained from these two calibration approaches. Conclusion: The degree of linear correlation between the WB DSC-CBF estimates obtained by AIF PV correction and ASL-based correction was significant. Comparing with CBF values obtained using ASL, the DSC-CBF estimates were high even after applying AIF PV correction. The over-estimation may come from the partial volume effect in VOF and other sources of errors in the DSC-MRI quantification.