Back to search

University of Thessaly (UTH)

Εφαρμογή και βελτιστοποίηση εξειδικευμένων τεχνικών λειτουργικής και μοριακής απεικόνισης μαγνητικού συντονισμού

Abstract

dc:description

The development of advanced MRI techniques has made a significant contribution to the field of medical imaging, as they enabled a holistic understanding of various processes and pathologies observed in the human body, leading in that way to a more accurate and comprehensive medical diagnosis. Over the years, and given the ever-increasing need for better quality and resolution of the obtained MRI images, advanced MRI techniques have gradually evolved to improve image quality. The aim of this PhD thesis was the application and optimization of the advanced MRI techniques, namely, functional Magnetic Resonance Imaging (fMRI), Diffusion Weighted Imaging (DWI) and Dynamic Contrast Enhanced Magnetic Resonance Imaging (DCE-MRI). Three different studies were performed in the context of this PhD thesis. The first study aimed to address static magnetic field B0 inhomogeneities by developing an integrated shimming technique, combining linear and high-order shimming, to optimize fMRI of the spinal cord. This study was carried out in collaboration with the Centre for Neuroimaging Sciences at King's College London. Measuring the temporal signal-to-noise ratio (tSNR) before and after the application of the shimming technique resulted to an increase of 189% in the mean tSNR value along the spinal cord (C3-C7). Subsequently, the same method was applied using a reduced field of view (FOV) to improve spatial resolution, showing a 12% decrease of the mean tSNR value in the cervical spinal cord. Overall, our shimming technique significantly reduced local signal loss and increased tSNR along the cervical spinal cord, improving image quality and, thus, optimizing fMRI of the human spinal cord. The second study aimed to investigate the use of radiomic analysis with Ensemble Learning methods, as a way to optimize advanced DWI and DCE-MRI techniques, in order to assist clinical differential diagnosis and particularly the differentiation between benign and malignant breast tumors. Multi-parametric breast MRI data (DWI, DCE and T2-w) were included in the study, and the performance of four different classification models based on the following algorithms (Adaptive Boosting (AdaBoosting), Gradient Boosting (GB), Extreme Gradient Boosting (XGBoost) and Light Gradient Boosting Machine (LightGBM)) was evaluated. According to the results, the most important radiomic features for breast cancer differentiation were five and associated with tumor shape in DCE-MRI and DWI images. Εvaluating the efficiency of the above classification models, the XGBoost and LGBM models achieved the highest AUC scores (0.95 and 0.94, respectively). Therefore, the optimized data analysis of DWI and DCE-MRI techniques using radiomics, and more specifically by incorporating Ensemble Learning methods, increased the available (quantitative) diagnostic information, contributing to improved accuracy of the diagnostic outcome and assisting differential diagnosis of breast cancer. The third study aimed to optimize the DCE-MRI technique through the development of innovative and efficient contrast agents based on spin crossover (SCO) iron nanoparticles (NPs). The synthesis of the contrast agents were performed in collaboration with the Department of Chemistry, University of Patras, while T1 and T2 relaxation times, r1 and r2 relaxivities and their possible temperature dependence, were investigated in a group of four contrast agents (NPs1, NPs2, NPs3 and NPs4) under four different temperatures, 20o, 30o, 40o and 50o C. These four contrast agents resulted from a larger group of 17 samples and considered the most efficient. According to the results, SCO NPs perform less as T1 contrast agents, however, their r2 relaxivity is very satisfactory. In particular, NPs3 and NPs4 had the highest efficiency, 2.86 and 2.93 mM-1 s-1 respectively (at 50o C), while higher r2 values were associated with higher temperatures for all four contrast agents. Therefore, the novel SCO iron nanoparticle based contrast agents could be an alternative to conventional T2 contrast agents based on gadolinium, and given the critical role of contrast agents in dynamic MR imaging, they can optimize the DCE-MRI technique and its diagnostic outcome. Considering the results and conclusions of all individual studies, it is concluded that both fMRI, DWI and DCE-MRI techniques were applied and optimized to a significant extent by following three different approaches to improve their performance in the context of this PhD thesis.

Degree

thesis:*
Grantor dc:publisher
University of Thessaly (UTH)
Year dc:date
2023

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Tsivaka, Dimitra
  • Τσιβάκα, Δήμητρα

Subjects

dc:subject × 10

Rights

Language dc:language
gre

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/55676

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Tsivaka, Dimitra; Τσιβάκα, Δήμητρα. Εφαρμογή και βελτιστοποίηση εξειδικευμένων τεχνικών λειτουργικής και μοριακής απεικόνισης μαγνητικού συντονισμού. University of Thessaly (UTH), 2023. http://hdl.handle.net/10442/hedi/55676