Objectives: ¹⁸F-FDG PET with voxel-based statistical image analysis plays an important role in the diagnosis of Alzheimer's disease (AD). However, the effect of an age-matched and sex-matched or mismatched normal database (NDB) on the diagnostic performance of ¹⁸F-FDG PET has not yet been investigated systematically. The aim of this study was to determine whether an age-matched and sex-matched NDB is necessary for the detection of AD using ¹⁸F-FDG PET.
Methods: We generated 11 NDB sets for ¹⁸F-FDG PET, including six age-specific NDB sets consisting of participants ranging in age from 20 to 70 years, one age-non-specific NDB set, one age-matched NDB set, two sex-specific NDB sets, each consisting of 20 men or 20 women, and one sex-matched NDB set. The average z-scores in predefined AD-specific regions of interest of the PET images were calculated using those NDB sets and a receiver-operating characteristic analysis was carried out to assess the diagnostic performance of ¹⁸F-FDG PET to discriminate 46 patients with AD from 50 normal controls.
Results: There was no significant difference in each area under the receiver-operating characteristic curve using either age-matched/mismatched NDB sets or sex-matched/mismatched NDB sets.
Conclusion: The diagnostic performance of ¹⁸F-FDG PET was rather insensitive to differences in age or sex in the NDB, indicating that exact age-matched or sex-matched NDB may not be essential for discriminating patients with AD from normal participants using ¹⁸F-FDG PET.