Local Time : 10:58 CET

EIBIR 1 - VPH-DARE@IT: Novel biomarkers and platforms for earlier dementia diagnosis

Thursday, March 2, 08:30 - 10:00 Room: X Session Type: EIBIR Session Topics: Nuclear Medicine, Computer Applications, Neuro Moderator: M. van Gils (Tampere/FI) Add session to my schedule In your schedule (remove)

A-173 08:30

Introduction

M. van Gils1, Z. A. Taylor2; 1 Tampere/FI 2 Sheffield/UK

Learning Objectives

1. To learn about the VPH-DARE@IT project and the use of its results.

Abstract

Dementia, in all its forms, is recognised as one of the key healthcare problems facing Western societies, especially as populations age, and concomitant economic and societal costs of the condition expand. Enabling more objective, earlier, predictive and individualised diagnosis and prognosis of dementias will support health systems worldwide to cope with this burden. The VPH-DARE@IT integrated project aims at precisely this. To this end, it pursues, first, a unique programme of biomarker discovery founded on a combination of patient-specific mechanistic and phenomenological models of disease. Second, it supports development of two new IT infrastructures aimed at clinical research and clinical decision support, respectively, which facilitate the complete pipeline from biomarker discovery and validation, to clinical deployment and evaluation. In this session, we will present latest developments in these areas.

A-174

Patient care platform

M. van Gils; Tampere/FI

Learning Objectives

1. To learn about dementia patient's needs.
2. To understand the VPH-DARE@IT patient care platform.

Abstract

Diagnosing complex diseases, such as dementia, is a challenging task. Clinicians need to combine in their minds a multitude of data from different sources, to consider a high number of different possible reasons behind clinical symptoms, to take into account the background information based on the precision medicine concepts, and finally to apply existing economic constraints. This complexity easily leads to suboptimal decision making, even with profound clinical expertise available. Clinical decision support systems (CDSS) based on the principles of data-driven medicine hold potential for making the process more quantitative and objective. In VPH-DARE@IT, the Patient Care Platform (PCP) implements a CDSS integrating heterogeneous biomarkers from medical images, neuropsychological tests and other measurements, and helping to form a holistic view of the patient’s status based on the principles of data-driven medicine. The PCP uses a data-driven approach where large databases of previously diagnosed patients are used for building models of several dementing diseases, including Alzheimer’s disease (AD), frontotemporal dementia (FTD), vascular dementia (VaD) and dementia with Lewy bodies (LBD). When a new patient arrives at a clinic, studies and tests are done, after which all available patient data are contrasted to the disease models, revealing the most likely reasons for dementia. The software architecture enables the CDSS to access heterogeneous patient data from several data sources. Furthermore, it communicates with the VPH-DARE@IT Clinical Research Platform to allow incorporation of novel biomarkers into the decision making.

A-176

Mechanistic model-based biomarkers

Y. Ventikos; London/UK

Learning Objectives

1. To learn about mechanistic model-based biomarkers for dementia.
2. To understand the biomarkers used in the VPH-DARE@IT project.

Abstract

Chronic cerebral hypoperfusion has been identified as a possible initiator of neurodegenerative processes leading to dementia. Lifestyle and environmental factors (LEFs), such as smoking, have an effect on the vascular system physiology. Pathophysiological changes in this system are, therefore, going to have an effect on the functioning of the brain. We propose to relate LEFs that act via the cardiovascular system and that have been shown to associate with AD. These LEFs are garnered from the preliminary data of a cross-sectional case-control study (Lido study). This multi-modal dataset is used to drive a patient-specific circulation model that provides predictions of CBF. The additional mechanistic components include a poroelasticity model used to acquire 4D maps of perfusion, clearance and ICP, in addition to a high-throughput imaging pipeline capable of providing anisotropic tissue permeability to capture the patterns of interstitial flow in addition to the parenchymal tissue and cerebroventricular representations required by the poroelasticity model. Preliminary simulation results indicate that predisposition to cognitive loss may be linked to modifiable risk factors, such as those witnessed relating to ICP, perfusion and clearance. Based on the preliminary data obtained from the Lido study, the CBF parameters that were found to be associated with MCI status were: reduced total CBF, reduced total perfusion (in women), increased APP (in very old subjects) and increased PI. The multiscale model of brain fluid transport developed in VPH-DARE@IT will be used to account for the long-term response to hypothetical LEF patterns and timings.

A-177

Phenomenological model-based biomarkers

S. Klein; Rotterdam/NL

Learning Objectives

1. To learn about phenomenological model-based biomarkers for dementia.
2. To understand the biomarkers used in the VPH-DARE@IT project.

Abstract

Both normal ageing and neurodegenerative disorders such as Alzheimer’s disease cause morphological changes (i.e. atrophy) in the brain, which can be visualised using magnetic resonance imaging (MRI). However, it is often difficult to distinguish normal ageing from neurodegenerative disorders by visual inspection of these images. In the VPH-DARE@IT project, we have developed a novel method for distinguishing between normal and abnormal brain morphology, based on a quantitative phenomenological modelling approach. A comprehensive spatiotemporal model of normal brain morphology is derived, in a data-driven and hypothesis-free fashion, from MRI brain data collected in a large-scale population study (the Rotterdam scan study). Based on this model, percentile curves are constructed that characterise the distribution of brain morphology in the general population as a function of age. These percentile curves serve as reference charts to which patient data can be compared. Experiments using data of the Alzheimer’s disease neuroimaging initiative (ADNI) show that the model generalises across study populations, and that it could be used to quantify a patient’s brain health.

This website uses cookies. Learn more