RT Journal Article
SR Electronic
T1 Interrater variability of image interpretation for 18F-FDG and 11C-PIB PET in the diagnosis of Alzheimer’s disease
JF Journal of Nuclear Medicine
JO J Nucl Med
FD Society of Nuclear Medicine
SP 1259
OP 1259
VO 52
IS supplement 1
A1 Tomohiko Yamane
A1 Hidehito Nagai
A1 Tomoyuki Nishio
A1 Yasuhiko Ikari
A1 Yoko Makishi
A1 Kazunari Ishii
A1 Kenji Ishii
A1 Takashi Kato
A1 Kengo Ito
A1 Michio Senda
YR 2011
UL http://jnm.snmjournals.org/content/52/supplement_1/1259.abstract
AB 1259 Objectives The aim of this study was to evaluate the interrater variability and usefulness of consensus reading for FDG and PIB PET regarding the diagnosis of Alzheimer’s disease (AD). Methods We analyzed 255 FDG and 122 PIB PET obtained in 23 centers as initial baseline scans of Japanese Alzheimer’s Disease Neuroimaging Initiative (J-ADNI) project. Initially, these PET scans were independently interpreted visually by 3 expert raters blinded to the clinical diagnosis. Subsequently, final decisions were made by a consensus of the raters. In FDG PET interpretation, Silverman’s criterion was used for classifying the uptake pattern into 7 categories (Silverman DH, et al. JAMA. 2001; FDG-7), and further into 2 categories: progressive and non-progressive pattern (FDG-2). In PIB PET interpretation, the scans were classified into 3 categories: negative, equivocal and positive amyloid deposition (PIB-3). Concordance among the 3 raters was evaluated by Cohen Κ statistics. Results Complete concordance rates among 3 raters were 61.1% in FDG-7, 75.3% in FDG-2 and 92.6% in PIB-3. Concordance of 2 or 3 out of the 3 raters was 94.1% in FDG-7 and 97.5% in PIB-3. With respect to FDG, concordance rates between clinical diagnosis and PET diagnosis in consensus reading were higher than those in individual reading (49.4% vs 44.9-47.8% in FDG-7, and 52.6% vs 49.4-51.0% in FDG-2). In PIB-3, the concordance rate in consensus reading was not the highest (50.8% vs 49.2-52.5%). Average Κ values were 0.56 in FDG-7, 0.66 in FDG-2 and 0.89 in PIB-3. Conclusions Interrater variability was minimal for PIB PET but was significant for FDG PET, for which a consensus reading is valuable in the diagnosis of AD. Research Support Research Association for Biotechnology (J-ADNI project