Advanced perfusion quantification methods for dynamic PET and MRI data modelling [Tesi di dottorato]

The functionality of tissues is guaranteed by the capillaries, which supply the microvascular network providing a considerable surface area for exchanges between blood and tissues. Microcirculation is affected by any pathological condition and any change in the blood supply can be used as a biomarker for the diagnosis of lesions and the optimization of the treatment. Nowadays, a number of techniques for the study of perfusion in vivo and in vitro are available. Among the several imaging modalities developed for the study of microcirculation, the analysis of the tissue kinetics of intravenously injected contrast agents or tracers is the most widely used technique. Tissue kinetics can be studied using different modalities: the positive enhancement of the signal in the computed tomography and in the ultrasound dynamic contrast enhancement imaging; T1-weighted MRI or the negative enhancement of T2* weighted MRI signal for the dynamic susceptibility contrast imaging or, finally, the uptake of radiolabelled tracers in dynamic PET imaging. Here we will focus on the perfusion quantification of dynamic PET and MRI data. The kinetics of the contrast agent (or the tracer) can be analysed visually, to define qualitative criteria but, traditionally, quantitative physiological parameters are extracted with the implementation of mathematical models. Serial measurements of the concentration of the tracer (or of the contrast agent) in the tissue of interest, together with the knowledge of an arterial input function, are necessary for the calculation of blood flow or perfusion rates from the wash-in and/or wash-out kinetic rate constants. The results depend on the acquisition conditions (type of imaging device, imaging mode, frequency and total duration of the acquisition), the type of contrast agent or tracer used, the data pre-processing (motion correction, attenuation correction, correction of the signal into concentration) and the data analysis method. As for the MRI, dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) is a non-invasive imaging technique that can be used to measure properties of tissue microvasculature. It is sensitive to differences in blood volume and vascular permeability that can be associated with tumour angiogenesis. DCE-MRI has been investigated for a range of clinical oncologic applications (breast, prostate, cervix, liver, lung, and rectum) including cancer detection, diagnosis, staging, and assessment of treatment response. Tumour microvascular measurements by DCE-MRI have been found to correlate with prognostic factors (such as tumour grade, microvessel density, and vascular endothelial growth factor expression) and with recurrence and survival outcomes. Furthermore, DCE-MRI changes measured during treatment have been shown to correlate with outcome, suggesting a role as a predictive marker. The accuracy of DCE-MRI relies on the ability to model the pharmacokinetics of an injected contrast agent using the signal intensity changes on sequential magnetic resonance images. DCE-MRI data are usually quantified with the application of the pharmacokinetic two-compartment Tofts model (also known as the standard model), which represents the system with the plasma and tissue (extravascular extracellular space) compartments and with the contrast reagent exchange rates between them. This model assumes a negligible contribution from the vascular space and considers the system in, what-is-known as, the fast exchange limit, assuming infinitely fast transcytolemmal water exchange kinetics. In general, the number, as well as any assumption about the compartments, depends on the properties of the contrast agent used (mainly gadolinium) together with the tissue physiology or pathology studied. For this reason, the choice of the model is crucial in the analysis of DCE-MRI data. The value of PET in clinical oncology has been demonstrated with studies in a variety of cancers including colorectal carcinomas, lung tumours, head and neck tumours, primary and metastatic brain tumours, breast carcinoma, lymphoma, melanoma, bone cancers, and other soft-tissue cancers. PET studies of tumours can be performed for several reasons including the quantification of tumour perfusion, the evaluation of tumour metabolism, the tracing of radiolabelled cytostatic agents. In particular, the kinetic analysis of PET imaging has showed, in the past few years, an increasing value in tumour diagnosis, as well as in tumour therapy, through providing additional indicative parameters. Many authors have showed the benefit of kinetic analysis of anticancer drugs after labelling with radionuclide in measuring the specific therapeutic effect bringing to light the feasibility of applying the kinetic analysis to the dynamic acquisition. Quantification methods can involve visual analysis together with compartmental modelling and can be applied to a wide range of different tracers. The increased glycolysis in the most malignancies makes 18F-FDG-PET the most common diagnostic method used in tumour imaging. But, PET metabolic alteration in the target tissue can depend by many other factors. For example, most types of cancer are characterized by increased choline transport and by the overexpression of choline kinase in highly proliferating cells in response to enhanced demand of phosphatidylcholine (prostate, breast, lung, ovarian and colon cancers). This effect can be diagnosed with choline-based tracers as the 18Ffluoromethylcholine (18F-FCH), or the even more stable 18F-D4-Choline. Cellular proliferation is also imaged with 18F-fluorothymidine (FLT), which is trapped within the cytosol after being mono phosphorylated by thymidine kinase-1 (TK1), a principal enzyme in the salvage pathway of DNA synthesis. 18F-FLT has been found to be useful for noninvasive assessment of the proliferation rate of several types of cancer and showed high reproducibility and accuracy in breast and lung cancer tumours. The aim of this thesis is the perfusion quantification of dynamic PET and MRI data of patients with lung, brain, liver, prostate and breast lesions with the application of advanced models. This study covers a wide range of imaging methods and applications, presenting a novel combination of MRI-based perfusion measures with PET kinetic modelling parameters in oncology. It assesses the applicability and stability of perfusion quantification methods, which are not currently used in the routine clinical practice. The main achievements of this work include: 1) the assessment of the stability of perfusion quantification of D4-Choline and 18F-FLT dynamic PET data in lung and liver lesions, respectively (first applications in the literature); 2) the development of a model selection in the analysis of DCE-MRI data of primary brain tumours (first application of the extended shutter speed model); 3) the multiparametric analysis of PET and MRI derived perfusion measurements of primary brain tumour and breast cancer together with the integration of immuohistochemical markers in the prediction of breast cancer subtype (analysis of data acquired on the hybrid PET/MRI scanner). The thesis is structured as follows: - Chapter 1 is an introductive chapter on cancer biology. Basic concepts, including the causes of cancer, cancer hallmarks, available cancer treatments, are described in this first chapter. Furthermore, there are basic concepts of brain, breast, prostate and lung cancers (which are the lesions that have been analysed in this work). - Chapter 2 is about Positron Emission Tomography. After a brief introduction on the basics of PET imaging, together with data acquisition and reconstruction methods, the chapter focuses on PET in the clinical settings. In particular, it shows the quantification techniques of static and dynamic PET data and my results of the application of graphical methods, spectral analysis and compartmental models on dynamic 18F-FDG, 18F-FLT and 18F-D4- Choline PET data of patients with breast, lung cancer and hepatocellular carcinoma. - Chapter 3 is about Magnetic Resonance Imaging. After a brief introduction on the basics of MRI, the chapter focuses on the quantification of perfusion weighted MRI data. In particular, it shows the pharmacokinetic models for the quantification of dynamic contrast enhanced MRI data and my results of the application of the Tofts, the extended Tofts, the shutter speed and the extended shutter speed models on a dataset of patients with brain glioma. - Chapter 4 introduces the multiparametric imaging techniques, in particular the combined PET/CT and the hybrid PET/MRI systems. The last part of the chapter shows the applications of perfusion quantification techniques on a multiparametric study of breast tumour patients, who simultaneously underwent DCE-MRI and 18F-FDG PET on a hybrid PET/MRI scanner. Then the results of a predictive study on the same dataset of breast tumour patients integrated with immunohistochemical markers. Furthermore, the results of a multiparametric study on DCE-MRI and 18F-FCM brain data acquired both on a PET/CT scanner and on an MR scanner, separately. Finally, it will show the application of kinetic analysis in a radiomic study of patients with prostate cancer.

diritti: info:eu-repo/semantics/openAccess
In relazione con info:eu-repo/semantics/altIdentifier/hdl/11573/1241979
Valutatori esterni: Prof. Giancarlo Bigi
Prof. Giovanni Sparacino
Prof. Giovanni Ulivi
ORIOLO, Giuseppe

Tesi di dottorato. | Lingua: Inglese. | Paese: | BID: TD20018663