A model of the kidneys and the systemic blood pool is constructed to derive the renal transit time spectrum from data obtained from three-probe renography. The transit time spectrum is derived by a deconvolution technique based on the maximum likelihood method. The computations were done on a microcomputer, and the c.p.u. time for solving two transit time spectra was mostly less than 5 min when 50 averageo sample points per measuring detector were used. The method is shown to be very insensitive to statistical noise, which is generally a problem using deconvolution techniques. The averaging of samples was done to save memory and reduce c.p.u. time, but no further filtering procedure or blood background subtraction was needed prior to execution of the algorithm.
Download Full PDF Version (Non-Commercial Use)