Construction and comparative evaluation of different activity detection methods in brain FDG-PET

• DOI: 10.1186/s12938-015-0073-x
• Publication Date: 2015-08-17T10:51:15Z
• Authors: Hans-Georg Buchholz; Fabian Wenzel; Martin Gartenschl...; Frank Thiele; Stewart Young; Stefan Reuss; Mathias Schrecken...
• Keywords: Statistical parametric mapping; Normalization; Concordance; Spatial normalization; Software package

We constructed and evaluated reference brain FDG-PET databases for usage by three software programs (Computer-aided diagnosis dementia (CAD4D), Statistical Parametric Mapping (SPM) NEUROSTAT), which allow a user-independent detection of dementia-related hypometabolism in patients' FDG-PET.Thirty-seven healthy volunteers were scanned order to construct FDG databases, reflect the normal, age-dependent glucose consumption human brain, using either software. Databases compared each other assess impact different stereotactic normalization algorithms used package. In addition, performance new altered brains patients was calculating statistical maps regional 20 with confirmed Alzheimer's (AD) 10 non-AD patients. Extent (hypometabolic volume referred as cluster size) magnitude (peak z-score) detected statistically analyzed.Differences between built CAD4D, SPM or NEUROSTAT observed. Due methods, spatial patterns found. When analyzing patient data created NEUROSTAT, similar characteristic clusters same regions found AD group However, larger z-scores observed CAD4D than those reported SPM. Better concordance achieved spatially normalized images an independent z-score calculation. The packages identified peak region 11 cases, there 16 subjects.The clinical evaluation CAD4D-, SPM- NEUROSTAT-generated from identical dataset showed known be involved AD. extent appeared influenced calculation method package rather parameters.