GFR measurement by simulating constant infusion with data acquired using a CdTe detector. A feasibility study.
Conventional nuclear medicine methods for measuring glomerular filtration rate, GFR, are invalidated if the initial injection of radiopharmaceutical is tissued. An analytical method proposed by Veall and Gibbs in 1982 overcomes this difficulty but does require multiple blood samples. This paper investigates whether it is feasible to use the analytical method on data acquired with a CdTe detector system and thereby reducing to one the number of blood samples required. Synthetic data were used to investigate the effect of (a) duration of sampling (b) statistical noise and (c) non-stationarity on the GFR values derived by the Veall and Gibbs method. A sampling duration of 100 min is shown to be adequate and statistical noise is not a limiting factor. However, data fluctuations, mainly due to detector motion, do reduce the accuracy of the method. When the analytical technique is applied to data obtained using a CdTe detector on 40 patients the GFR values are similar to those derived by the conventional method (r = 0.957, Syx = 9.1 ml min-1). The technique is sufficiently attractive to warrant further studies to establish its absolute accuracy.