Database deconvolution

A new deconvolution technique is described. It regards the Rutland-Patlak plot as the integral of the retention function, and performs deconvolution by creating and then differentiating the Rutland-Patlak plot. The reasoning behind this approach is that in producing the Rutland-Patlak plot, both the content function and the integral of the input function are divided by the input function (i.e. the blood curve), and that this has the effect of producing data equivalent to that which would be produced if the input function did not vary (i.e. if there were a constant blood level of tracer). The concept was tested against data created by modelling techniques. It was able to reproduce retention functions with a variety of different shapes that were used to produce the artificial renogram data. The new deconvolution method is sufficiently simple to be used in either a database or a spreadsheet and does not require any special program units to be written. It does not appear to make any assumptions about the data nor the shape of the retention function, and does not appear to be vulnerable to the oscillating errors that sometimes occur in iterative deconvolution.