Towards improved PET quantification of amyloid-beta pathology and cerebral blood flow in Alzheimer's disease
30 april 2019
UAntwerp - Campus Drie Eiken - Building Q - Promotiezaal - Universiteitsplein 1 - 2610 WILRIJK (route: UAntwerpen, Campus Drie Eiken
Prof S. Staelens, Prof C. Van Broeckhoven, Prof J. Verhaeghe
PhD defence Julie Ottoy - Faculty of Medicine and Health Sciences (in English)
Abstract (Presentation in English)
In recent years, it has become clear that brain damage from Alzheimer’s disease (AD) can occur decades before clinical symptoms arise. Pathological changes in the brain include accumulation of amyloid-β (Aβ) plaques and neurofibrillary tangles, leading to reduced cerebral blood flow (CBF), neuronal cell death, and cognitive decline. Thanks to the advances in non-invasive medical imaging, the individual hallmarks of AD can now be visualized in-vivo using positron emission tomography (PET). Moreover, following its quantitative nature, PET paved the way to more precise evaluation of brain changes caused by disease progression and drug therapy trials.
However, the role of PET in precision medicine has been challenged by the wide application of ‘simplified’ semi-quantitative metrics such as the standardized uptake value ratio (SUVR). SUVR is obtained as the ratio of PET-tracer uptake in an Aβ-rich region (target region) versus an Aβ-free region (reference region), and is derived from a short static scan. Although useful in clinical practice, SUVR has been widely implemented for various PET tracers without a thorough validation. Therefore, this thesis validated SUVR against gold standard techniques to quantify Aβ plaque deposition and CBF in AD.
To this end, we performed 60min dynamic [18F]AV45-PET with continuous arterial sampling and pharmacokinetic modeling in 10 controls, 18 mild cognitive impairment, and 10 AD dementia subjects. From this, the gold standard measurements of Aβ plaque deposition and CBF were derived as the volume of distribution (VT) and the plasma-to-tissue tracer delivery rate (R1), respectively, for each brain region. First, by validating Aβ SUVR against VT, we found that SUVR was dependent on reference region selection and CBF variations. The subcortical white matter was found to be the preferred reference region for SUVR (instead of the widely used cerebellum), but also made the SUVR most vulnerable to (therapy-induced) changes in global CBF. Second, by validating CBF SUVR against R1, we found that the latter was more sensitive to detect the disease severity.
In conclusion, due to the SUVR shortcomings, we suggest one dynamic [18F]AV45-PET protocol to quantify both Aβ deposition and CBF for studies in which high accuracy is needed (e.g. therapy-monitoring studies). However, the ultimate goal would be the development of a clinically feasible amyloid-PET protocol with short scanning time and without the need for continuous sampling. Therefore, we set up a bolus-plus-constant-infusion [18F]AV45 protocol in which the patient is sent to the waiting room while being constantly infused, after which a short static scan and one blood sample is taken. From this, the gold standard VT is calculated as a simple ratio of Aβ in tissue versus blood at equilibrium - independent from CBF and reference region selection.
Contact e-mail: Nicolaas.VanLeeuwen@uantwerpen.be