Quantitative MR cerebral blood flow using ARMA-based deconvolution: preliminary results

A system for determining quantitative cerebral blood flow (CBF) from dynamic susceptibility contrast (DSC) images was developed. The system uses a novel deconvolution method, based on autoregressive and moving-average (ARMA) processing, to extract CBF from the DSC images. The ARMA-based method is expected to he less sensitive to many factors compared to the more commonly used singular-value decomposition (SVD) method. In selected acute stroke patients, the SVD- and ARMA-derived CBF results were equivalent.