Abstract #2726

Towards Robust and Automated Identification of Vascular Input Function in DCE-MRI

Kim Mouridsen¹, Dominique Jennings¹, Elisa Gelasca¹, Elizabeth Gerstner¹, Tracy Batchelor², Gregory Sorensen¹

¹Athinoula A. Martinos Center for Biomedical Imaging, Department of Radiology, Massachusetts General Hospital, Charlestown, MA, United States; ²Stephen E. & Catherine Pappas Center for Neuro-Oncology, Massachusetts General Hospital, Boston, MA, United States

Dynamic contrast enhanced MRI (DCE-MRI) holds potential for characterizing key physiological markers of tumor vascularity such as blood brain-barrier-permeability. Reproducibility of pharmacokinetic parameters in multicenter settings is contingent on reliable characterization of the vascular input function (VIF). This is compromised by signal attenuating T2* effects at high concentrations and insensitivity of typical T1-weighted sequences at peak bolus passage, as well as non-reproducible manual identification of VIFs. We demonstrate that a completely automatic VIF identification procedure combined with T2* based estimation of peak concentration yields VIF reproducibility comparable to expert manual selection in two pre-treatment baseline scans of 10 glioblastoma patients.