Comment on the Paper by Agnoli et al

The possibility of using a very short 133Xe inhalation time in order to simulate a bolus injection is a tempting one. We tried this modification and abandoned the idea because it introduces a major source of error. Provided the arterial input function is measured and the deconvolution carried out on the head curve the values obtained for the fast component clearance rate are independent of the breathing period if it is varied between 1 and 15 min. For shorter periods the values obtained for the fast component are higher. This is to the fact that in order to achieve adequate count rates it is necessary to use a much higher concentration of 133Xe in the breathing mixture. Under these conditions the contribution to the observed count rate due to scattered radiation from 133Xe in the nasopharynx, which is usually negligible, is considerably accentuated. This contribution can be measured quite simply by inhaling 50 to 100 ml of 133Xe at the end of a normal inspiration so that only the respiratory dead space is filled with the tracer.