Background: Head and neck Magnetic Resonance (MR) Images are vulnerable to the arterial blood in-flow effect. To compensate for this effect and enhance accuracy and reproducibility, dynamic tracer concentration in veins was proposed and investigated for quantitative dynamic contrast-enhanced (DCE) MRI analysis in head and neck.
Methodology: 21 patients with head and neck tumors underwent DCE-MRI at 3T. An automated method was developed for blood vessel selection and separation. Dynamic concentration-time-curves (CTCs) in arteries and veins were used for the Tofts model parameter estimations. The estimation differences by using CTCs in arteries and veins were compared. Artery and vein voxels were accurately separated by the automated method. Remarkable inter-slice tracer concentration differences were found in arteries while the inter-slice concentration differences in veins were moderate. Tofts model fitting by using the CTCs in arteries and veins produced significantly different parameter estimations. The individual artery CTCs resulted in large (>50% generally) inter-slice parameter estimation variations. Better inter-slice consistency was achieved by using the vein CTCs.
Conclusions: The use of vein CTCs helps to compensate for arterial in-flow effect and reduce kinetic parameter estimation error and inconsistency for head and neck DCE-MRI.