IC‐P‐136: MODELING AMYLOID, TAU, AND NEURODEGENERATION ACROSS THE ALZHEIMER'S DISEASE SPECTRUM: CROSS‐SECTIONAL STUDY

• DOI: 10.1016/j.jalz.2018.06.2202
• Publication Date: 2018-10-22T10:02:40Z
• Authors: Jungho Cha; Yeong-Hun Park; Jong-Min Lee; Suzanne L. Baker; Adrienne Visani; Viktoriya Bourakova; Julie Pham; Mustafa Janabi; James P. O'Neil; Bruce L. Miller; William J. Jagust; Gil D. Rabinovici
• Keywords: Sigmoid function; Akaike information criterion

Hypothetical models of the evolution Alzheimer's disease (AD) biomarkers predict a non-linear, sigmoid model. However, it is still unclear whether better fit model than linear for progression amyloid, tau, and neurodegeneration across AD spectrum. [11C]PiB-, [18F]AV1451-PET, 3T T1-weighted MRI were obtained 110 subjects, including 49 cognitively normal controls (33 PiB-, 16 PiB+), 61 PiB+ patients (12 MCI, AD) (Table1). PiB 35-90min DVR AV1451 80-100min SUVR images created using cerebellar gray matter (PiB+=cortical DVR>1.065), inferior AV1451. MRIs processed with CIVET-pipeline to extract surfaces measure cortical thickness. Each was mapped onto surfaces, then surface-registered, smoothed. Global measures calculated by averaging imaging values cortex. We modeled global each modality both logistic models. Disease estimated applying tiered that ranks subject first clinical stage, subjects within stage by. Akaike Information Criterion used determine relative likelihood biomarker best or Finally, normalized facilitate comparisons between modalities. Modeled data shown in Figure 1 lines simulating curves. AIC determined high probability (probability=100%) (97%), but probabilities vs. similar thickness (52% 48%) (Table2). 2 compares progression. appeared reach saturation at early MCI phase, while plateaued mid-AD more dynamic through later stages AD.