Simulation study of noise property of CMRO2 quantitation methods with inhalation of 15O2

Measurements of the oxygen consumption in brain have been studied by PET. Autoradiographic method(ARG) was suggested (Mintum et al.) to yield CMRO₂. This method required separately obtained information about CBF, CBV, thus time of 30-60 min. is required for three separate scans. To decrease the scan time, a new protocol was suggested as a rapid dual table method(ARG-D), in which [¹⁵O]water injection scan and [¹⁵O]O₂ inhalation scan are continuously carried out. Another method of weighted integration(WI) method with single 3 min. ¹⁵O₂ inhalation scan was suggested (Ohta et W.). We modified this method by taking into account the water re-circulation(WI-WR). In this study, the statistical noise properties and effects of error propagated from dispersion, delay and volume of distribution on CMRO₂ image, derived by these methods were evaluated. Tissue time activity curves was generated from typical blood time activity curves. A 80% of noise at a peak in [¹⁵O]water tissue time activity curve was added to study the noise propagation and accuracy in CMRO₂ image. Also dispersion, delay, and volume of distribution was varied and evaluated the error propagation. Methods of ARG, WI-WR and ARG-D, reproduce the given CMRO₂ within 2% accuracy, while method WI gives CMRO₂ 5-15 The effect of noise in unit of %SD was 12 % for ARG, 25 % for WI and WI-WR, and 17 % for ARG-D method. On the basis of simulation study suggests that the ARG-D method developed could be used to estimate the CMRO₂ values in clinic.