Abstract #1550

Kosior J, Frayne R

Our purpose is to develop a digital anthropomorphic perfusion phantom that can be used to evaluate the accuracy and robustness of bolus-tracking perfusion-imaging algorithms in the presence of simulated acquisition artifacts (i.e., noise). This phantom represents a significant advancement compared to previous perfusion-imaging approaches that have relied on simulated 1D-signals for evaluation, since the phantom can also be used to evaluate spatiotemporal algorithms (i.e., noise filters). In this work, we use the phantom to demonstrate the effects of patient motion on cerebral blood flow estimates compared with the results obtained using a registration algorithm to correct for the motion.