Abstract： objective　To explore the effect on semiquantitive analysis for brain glucose metabolism with 2 or 3 dimensional PET scan respectively.Methods　36 normal persons were divided into two group,15 for 2D,21 for 3D.PET scans were done after 40 minutes of injection of FDG .The transaxial,the coronal,the saggital were gotten through computer reconstruction. The transaxial slice density was 0.5cm. The borders ofbilateral frontal lobes , bilateral parietal lobes , bilateral tempt lobes ,bilateral occipital lobes and bilateral cerebellum were drawn. According to ROI, weight and injection dose of FDG,mean SUVs in all above lobes were gotten. R(L/R)s from Mean SUVs in the left lobes via those in the right were caculated. R(C/CB)s from mean SUVs in different brain lobes via those different cerebellum lobes were also caculated. Results　There was no significance between 2D and 3D about SUV ratios of R(L/R)s and R(C/CB)s . Conculsions　The semiquantitive analysis for brain glucose metabolism could not be affected by PET scan methods. According to patients condition , 2D or 3D could be chosen to be used without affecting semiquantitive analysis.