SIMULATION STUDY OF CEREBRAL BLOOD-FLOW MEASUREMENTS IN XENON-CT - EVALUATION OF WASHIN WASHOUT PROCEDURES

Simulation programs have been created that allow one to vary image pixel noise, the number and the distribution of scans with time, cerebral tissue parameters, and the type of xenon CT inhalation procedure in order to investigate CBF measurements with respect to accuracy and signal-to-noise ratio (SNR). In particular, standard washin studies were compared to washin/washout studies. Based on the results of these simulations, a new protocol is suggested; it consists of only 3 min of xenon inhalation (washin phase) and 3 min of washout, monitored by one reference and six enhancement scans taken at 1-min intervals. Compared with a standard 8-min washin study of equal total dose, flow standard deviation (s.d.) for an unconstrained least-squares algorithm is reduced by factors of 2.2 and 1.2 for gray and white matter, respectively; for flow distributed uniformly from 20 to 80 ml/min/100 g, an average s.d. reduction factor of 1.7 is achieved. This was confirmed experimentally in a volunteer study using noise power spectrum analysis. In addition, effects of tissue heterogeneity have been investigated; both the bias and s.d. of flow estimates due to varying proportions of white and gray matter in a given volume element are reduced in washin/washout protocols. When compared to a short washin-only study of 4.5 min, the 3-min washin/3-min washout study provides an improvement of flow s.d. by a factor of 1.6 and 1.9 for gray and white matter, respectively. The very short xenon inhalation time and the short total examination time of the suggested new protocol are expected to significantly reduce the patient movement which is considered to be a main source of errors of xenon-CT in clinical CBF measurements.