Brain Clearance of Contrast Agent in Intravenous <scp>DCE</scp> ‐ <scp>MRI</scp> Is Measurable and Cannot Be Explained by Clearance Across the <scp>BBB</scp> Alone

ABSTRACT Purpose To determine the feasibility of measuring brain clearance of gadolinium contrast agent using standard intravenous DCE‐MRI acquisitions and evaluate the contribution of blood–brain barrier (BBB) and non‐BBB clearance routes. Methods Uptake and extended Tofts models were fit to DCE‐MRI data from people with Parkinson's disease, post‐stroke and controls. Models were compared using the Akaike information criterion. Key parameters ( K trans , v p , v e , and k ) were extracted from gray and white matter ROIs. In mice, two‐photon microscopy of intravenously injected Sulforhodamine 101 was used to confirm small tracer clearance kinetics without the confounds of partial volume effects or water exchange. Results The extended Tofts model provided a superior fit compared to all uptake‐only models. Extended Tofts estimates of the extravascular extracellular volume fraction v e were underestimated by a factor of 10–20 compared to known literature values (1.5% vs. 20%–30%). We hypothesized this bias may be due to an additional competing non‐BBB clearance mechanism not accounted for by the extended Tofts model. In both MRI and two‐photon data, we found that v e estimates trended toward literature values as BBB permeability (and thus BBB clearance) increased, supporting our hypothesis. Conclusion Modeling clearance of extravasated contrast agent from the brain improves fit quality compared to models that neglect clearance. Estimates of interstitial volume fraction were underestimated at low K trans but tended toward literature values as BBB permeability (and clearance) increased. This study indicates that estimating brain clearance of contrast agent using DCE‐MRI is feasible, and that estimates reflect a combination of BBB and non‐BBB clearance pathways.