Premium
IC‐P‐034: PiB PET Classification: A model to determine those who are at risk of developing Alzheimer's disease
Author(s) -
Fripp Jurgen,
Villemagne Victor L.,
Bourgeat Pierrick,
Raniga Parnesh,
Acosta Oscar,
Szoeke Cassandra,
Ourselin Sebastien,
Ames David,
Ellis Kathryn A.,
Masters Colin L.,
Rowe Christopher C.,
Salvado Olivier
Publication year - 2009
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1016/j.jalz.2009.05.054
Subject(s) - voxel , receiver operating characteristic , principal component analysis , nuclear medicine , pittsburgh compound b , spatial normalization , cognitive impairment , pattern recognition (psychology) , normalization (sociology) , medicine , psychology , pathology , artificial intelligence , disease , radiology , computer science , sociology , anthropology
Background: Dementia is often of mixed etiology, and additional comorbidity is frequently present. Some features / patterns can be missed by statistical methods due to the difficulty of obtaining adequate normative databases given the large variance in some areas of the brain in normals. The usefulness of a combined visual-semiquantitative and quantitative Brain SPECT protocol was evaluated. Methods: Two Institutions using same triple head gamma camera, and 99mTc-HMPAO. Each place has its own visual evaluation: one is based on the camera’s own software package. The other is based on a discrete color code supplemented by 4 levels of thresholded volumes and, 8 color Talairach normalized surface views (Neurostat). The quantitative protocol was based on an automated analysis software (BRASS) applied to asses relative blood flow (cerebellar ratios). Population : 30 cases referred for memory problems and / or possible dementia differential, and for whom, clinical, followup of > 1 year and /or other information was available; ages: 48-81 (55 % males). Results: The semiquantitative evaluation included: 1) hemispheric cortex (including orbitofrontal and lateral temporal areas; 2) limbic and paralimbic structures; 3) subcortical structures ; 4) cerebellum & pons. The quantitative algorithm provided the slice display and tabulated results of 47 regions . In 22/30 cases abnormalities suggested a mixed etiology and in some, suggestions of patterns of superimposed comorbidity (depression, anxiety, psychosis etc.). In 4/30 a single etiology was suggested and in 2/30 no definite cortical nor posterior cingulate abnormalities but only subcortical abnormal features. In 2/30 no definite abnormality found. Type of abnormalities: AD predominant,variants of Fronto Temporal Lobar Degeneration, Lewy Body disease, and Vascular. As far as cortical abnormalities a good correlation between the 2 methods was found in 24/30 cases. Most subcortical abnormalities were detected only by the visual method. Conclusions: If properly performed and displayed, visual-semiquantitative Brain SPECT is a clear, practical, useful, and cost effective adjuvant to the quantitative methods. Indeed the optimized displays provide the detailed information needed while waiting for more refined normative databases and, in addition, serve as quality control for quantitative evaluations.