Bart Van Berckel, Netherlands
Amsterdam UMC, Vrije Universiteit Amsterdam Department of Radiology and Nuclear MedicineAuthor Of 1 Presentation
VISUAL ASSESSMENT OF [18F]FLUTEMETAMOL PET IMAGES CAN DETECT EARLY AMYLOID PATHOLOGY AND GRADE ITS EXTENT
Abstract
Aims
Visual read (VR) guidelines of amyloid-β PET were developed to capture established pathology in clinical populations, whereas quantification is often used to detect early pathology in research settings. This study investigated the sensitivity of VR to detect early amyloid-β pathology as determined quantitatively, and studied the utility of regional uptake patterns to track progression.
Methods
[18F]flutemetamol PET images of 497 subjects (ALFA+ N=352;ADC N=145) were included, covering both subthreshold and established amyloid pathology. Scans were visually assessed according to product guidelines, recording the number of positive regions (0-5) and a final negative/positive classification. All scans were quantified using the standard Centiloid (CL) method. The agreement between VR-based classification and previously published CL-based cut-offs for early (CL=12) and established (CL=30) pathology was determined. An optimal CL cut-off maximizing sensitivity and specificity (Youden’s index) against VR was derived. The quantitative burden of observed regional patterns was determined and operationalized into 3 stages of VR positivity.
Results
VR showed excellent agreement against CL=12 (κ=.88, 95.0% [472/497]) and CL=30 (κ=.87, 94.8% [471/497]) cut-offs, with 86.5%/99.1% and 100%/93.1% sensitivity/specificity, respectively(Fig.1). ROC analysis resulted in an optimal CL=17 cut-off against VR (sensitivity=93.9%, specificity=98.2%). The posterior cingulate/precuneus and (medial orbito) frontal cortex were reported most frequently positive, either isolated (Stage 1: CL=22.1±6.3), co-occurring (Stage 2: CL=36.3±18.0) or in combination with other regions (Stage 3: CL=83.7±30.5,Fig2).
Conclusions
Our results show that VR is an appropriate method for assessing early amyloid pathology and grading the extent of visual amyloid positivity, which could be valuable to stratify patients.
