Hassan Bagher-Ebadian1,2, Tavarekere N. Nagaraja3, Robert Knight1,2, Ramesh Paudyal1, Siamak P. Nejad-Davarani1, Stephen Brown4, Sawyam Panda1, Polly Whitton1, Joseph D. Fenstermacher3, James R. Ewing1,2
1Neurology, Henry Ford Hospital, Detroit, MI, United States; 2Physics, Oakland University, Rochester, MI, United States; 3Anesthesiology, Henry Ford Hospital, Detroit, MI, United States; 4Radiation Oncology, Henry Ford Hospital, Detroit, MI, United States
Estimating the arterial input function of a contrast agent, the time-concentration curve in plasma, has long presented a problem in MR dynamic contrast Enhanced and dynamic susceptibility studies. A set of time courses of radiolabeled-Gd-DTPA CA-concentrations in arterial plasma (timeactivity curve) of 13 animals has been recently measured and named as Standard-Radiological-AIF (SRAIF). Herein, an adaptive model was used for predicting the time trace of AIF from Dual-Gradient-Echo (DGE) signal measured in the normal area of rat brains. A set of physically-meaningful extracted features from DGE sequences, an artificial neural network was trained to estimate the time-trace of SRAIF.