1,721,055 research outputs found
Development of novel computational algorithms for quantitative voxel-wise functional brain imaging with positron emission tomography
Full text linkPositron Emission Tomography (PET) allows to study, in vivo, important biological processes in both animals and humans. In particular, it is widely used for receptor studies, where it allows quantitative functional imaging of physiological parameters in terms of receptor binding, volume of distribution, and/or receptor occupancy.
In this thesis an overview on kinetic modeling in PET is presented, as well as some novel methodologies for voxel wise quantification of PET data, applied on various datasets.
The proposed methods are a good alternative for the generation of reliable parametric maps, and applied to clinical data are expected to simplify the detection also of small/specific pathological areas.
Moreover, as additional results, new compartmental models were developed for [11C]SCH442416 and [11C]MDL100907 data and a new clustering approach which allow segmenting the brain volume even for PET data with high level of noise was implemented. This new method was also applied for the selection of the optimal reference region for [11C]MDL100907 data.La Tomografia ad Emissione di Positroni (PET) permette di studiare, in vivo, l'interazione dei traccianti con specifici siti di legame (trasportatori, recettori, etc.). Inoltre permette un imaging funzionale quantitativo di importanti parametri fisiologici quali la densità di recettori, volume di distribuzione e/o occupazione recettoriale.
In questa tesi si espone una panoramica dei principali metodi modellistici in PET e si propongono nuovi approcci Bayesiani sviluppati per la quantificazione a livello di voxel di immagini PET, applicati a vari dataset.
I metodi proposti costituiscono una robusta alternativa per la generazione di mappe parametriche affidabili ed applicati a dati clinici renderanno più semplice il riconoscimento di piccole zone patologiche specifiche.
Come ulteriore risultato, sono stati sviluppati nuovi modelli compartimentali per i dati di [11C]SCH442416 e [11C]MDL100907. Inoltre è stato implementato un nuovo metodo di clustering che permette di segmentare il volume cerebrale anche per dati PET con un alto livello di rumore. Questo nuovo approccio è stato applicato per la selezione della migliore regione di riferimento per dati di [11C]MDL100907
Novel hierarchical approach for voxel-wise PET quantification: application on multi-compartmental models
No full text
An integrated tool for quantification of dynamic PET studies: Spectral Analysis Kinetic Estimation
No full text
Deriving physiological information from PET images: from SUV to compartmental modelling
Full text linkPositron emission tomography (PET) imaging has made it possible to detect the in vivo concentration of positron-emitting compounds accurately and non-invasively. In order to relate the radioactivity concentration measured using PET to the underlying physiological or biochemical processes, the application of mathematical models to describe tracer kinetics within a particular region of interest is necessary. Image analysis can be performed both by visual interpretation and quantitative assessment and, depending on the ultimate purposes of the analysis, several alternatives are available. In clinical practice, PET quantification is routinely performed using the standard uptake value (SUV), a semi-quantitative index in use since the 1980s. Its computation is very simple since it requires only the PET measure at a pre-fixed sample time and the injected dose normalised to some anthropometric characteristic of the subject (generally body weight or body surface area). An alternative to the SUV is the tissue-to-plasma ratio (ratio). As its name indicates, this index is computed as the ratio between the tracer activity measured in the tissue and in the plasma pool within a pre-fixed time window. Moving from static to more informative dynamic PET acquisition, three model classes represent the most frequently used approaches: compartmental models, the spectral analysis modelling approach, and graphical methods. These approaches differ in terms of application assumptions (e.g. reversibility of tracer uptake, model structure, etc.) and computational complexity. They also produce different information about the system under study: from a macro-description of tracer uptake to a full quantitative characterisation of the physiological processes in which the tracer is involved. The application of these approaches to clinical routine is restricted by the need for invasive blood sampling. In order to avoid arterial cannulation and blood sample management, different alternative approaches have been developed for quantification of PET kinetics, including reference tissue methods. Although these approaches are appealing, the results obtained with several tracers are questionable. This review provides a complete overview of the semi-quantitative and quantitative methods used in PET analysis. The pros and cons of each method are evaluated and discussed
Multi-scale hierarchical approach for parametric mapping: assessment on multi-compartmental models
No full text
- …
