Complementarity of visual and voxel-based FDG-PET analysis to detect MCI-like hypometabolic pattern in elderly patients with hypertension and isolated memory complaints

• Published in: Acta radiologica (1987) 2015-08, Vol.56 (8), p.980-989
• Main Authors: Van Der Gucht, Axel, Verger, Antoine, Yagdigul, Yalcin, Poussier, Sylvain, Joly, Laure, Watfa, Ghassan, Benetos, Athanase, Karcher, Gilles, Marie, Pierre-Yves
• Format: Article
• Language: English
• Subjects: Aged; Brain - diagnostic imaging; Cognitive Dysfunction - diagnostic imaging; Diagnosis; Elderly people; Female; Fluorodeoxyglucose F18; Geriatric Assessment - methods; Human health and pathology; Humans; Hypertension; Hypertension - diagnostic imaging; Image Interpretation, Computer-Assisted - methods; Imaging, Three-Dimensional - methods; Life Sciences; Male; Memory Disorders - diagnostic imaging; Middle Aged; Observer Variation; Patients; Positron-Emission Tomography - methods; Quantitative analysis; Radiopharmaceuticals; Recall; Reproducibility of Results; Sensitivity and Specificity; Visual
• ISSN: 0284-1851; 1600-0455
• DOI: 10.1177/0284185114542366

Background 18F-FDG PET can be used to aid in the diagnosis of Alzheimer’s disease (AD) and clarify the diagnosis and prognosis of patients with mild cognitive impairment (MCI). Purpose To compare the results of a quantitative analysis of FDG-PET brain images to a standard visual analysis (SVA) with regards to the detection of MCI-like hypometabolic pattern in elderly patients with hypertension and subjective, isolated memory complaints. Material and Methods FDG-PET brain was performed in 71 patients (mean age, 76.4 ± 5.1 years; women, 53.5%). Images were analyzed for the presence of an MCI-like hypometabolic pattern using an SVA by 2 physicians and a voxel-based statistical procedure (statistical parametric mapping [SPM]) that compared each patient’s images to normal reference samples from 19 elderly individuals obtained using the same PET camera. The reliability of these analyses was evaluated according to neuropsychological assessment results, including the Grober & Buschke Free and Cued Selective Reminding Test, and a combined analysis by a neuropsychologist. Results An MCI-like hypometabolic pattern was documented in 5 patients (7%) by SVA and 7 patients (10%) by SPM analysis; however, only 2 of these patients were selected by both methods. The group characteristics of the 7 patients identified by the quantitative method were consistent with the MCI pattern, which included a higher rate of abnormal GB-FCSRT in Free Recall (57% vs. 9%, p