Effect of blood curve smearing on the accuracy of parameter estimates obtained for 82Rb/PET studies of blood-brain barrier permeability.

82Rb in conjunction with positron emission tomography (PET) has been used to estimate the blood to brain transport rate constant (K1) for Rb and the regional brain/tumour blood volume (Vb). Errors in K1 and Vb depend upon the accuracy of the measured arterial blood radioactivity and PET-monitored brain radioactivity. Arterial blood is usually sampled by placing a catheter in the radial artery and measuring the radioactivity in blood passing continuously in front of a detector or by counting discrete blood samples in a well scintillation detector. In either case, the passage of blood through catheter/pump tubing produces a smearing of the waveform as well as a delay in the arrival of radioactivity at the blood sampling site. The change in shape of the blood curve is significant for bolus-type injections and results in large errors in those model parameters which contribute substantially to the initial phase of the brain activity curve. We report here the results of computer simulations and an analysis of patient data which suggest that parameter estimation errors due to smearing and time shift may be large (greater than 50%) but that these errors can be minimised by the use of deconvolution techniques.