A Comparison of the Tofts and Linear Reference Region Models for DCE MRI When Monitoring Vascular Disruption in a...

Dynamic contrast-enhanced magnetic resonance imaging (DCE MRI) can assess changes in tumor vascular perfusion. We compared the Tofts model and the linear reference region model (LRRM) for analyses of DCE MRI for evaluating changes in vascular perfusion to combretastatin A-4 phosphate (CA4P) in a 4T1 tumor model to determine which model can better evaluate changes in vascular perfusion induced by CA4P. Five female BALB/c mice inoculated with 4T1 murine breast cancer cells underwent DCE MRI at 24 h before CA4P treatment and at 3-and 24-h post-treatment. DCE MRI was performed using a 7-T preclinical MRI scanner. Pharmacokinetics parameters were derived using the Tofts model that relies on an arterial input function (AIF) and the LRRM that employs muscle as a reference tissue. We measured K^trans, RK^trans, and k_ep across whole tumors, tumor peripheries, and cores at each time point. Both models identified vascular disruption 3-h post-treatment with partial recovery at 24 h, as expected. The LRRM measured treatment-induced changes in RK^trans with more significance relative to changes in K^trans measured by the Tofts model, whereas the Tofts model measured changes in k_ep with more significance relative to the LRRM, based on the % standard deviation of measurements within the cohort, the p-values, and the effect sizes in treatment-induced changes for these measurements. Measurements of treatment-induced changes in K^trans and RK^trans were more significant than measurements of changes in k_ep. Therefore, our study showed that the LRRM is more reliable than the Tofts model when assessing tumor vascular perfusion. Considering that the LRRM uses a muscle region of interest in place of an AIF, the LRRM has strong advantages for clinical translation.