Vis enkel innførsel

dc.contributor.advisorGoa, Pål Erik
dc.contributor.advisorBathen, Tone Frost
dc.contributor.authorVidić, Igor
dc.date.accessioned2019-08-27T12:29:49Z
dc.date.available2019-08-27T12:29:49Z
dc.date.issued2019
dc.identifier.isbn978-82-326-4003-4
dc.identifier.issn1503-8181
dc.identifier.urihttp://hdl.handle.net/11250/2611224
dc.description.abstractAccurate noninvasive diagnostic and subtyping procedures for cancer are very important. Diffusion-weighted imaging (DWI) reflects the microscopic cellular environment and is sensitive to tumor characteristics such as cell density, vascularity, membrane integrity, viscosity and stromal microstructure. As such, it plays an important role in characterizing tumors. However, standard diffusion imaging methods involve measurement of the apparent diffusion coefficient, ADC, which only captures part of the potential information carried by DWI. In this thesis, the clinical value of DWI for breast cancer is investigated, specifically going beyond the ADC to use more advanced DWI approaches and applying several fitting approaches and processing algorithms. These advanced modeling and analysis methods attempt to probe the nature of the DWI signal curve in closer detail, and in doing so to maximize the utility of clinical DWI in breast cancer management potentially providing more complete information of the underlying microstructure. We have found that the clinical value of DWI is improved by extending the DWI protocol to include denser and wider sampling of diffusion weighting (b-values) at the cost of longer scan times. Particularly, using perfusion sensitive parameters in machine learning algorithm can achieve better differentiation of benign and malignant lesions than ADC, and furthermore can predict cancer subtype. In the same b-value range tumor differentiation accuracy is improved by using Bayesian fitting approaches to a combined perfusion-diffusion model. Finally, often underestimated in breast, higher b-values potentially provide insight into cellular fraction for which we demonstrated the ability to distinguish between benign and malignant lesions. However, the signal in this b-value range is best described using Padé exponent, a pure mathematical representation. The additional value, and ultimately the clinical role, of advanced techniques such as presented and used in our studies must always be considered through the lens of patient care. The reality of disease management is that every patient is an individual thus the versatility of more complex DWI protocols alongside the enduring robustness of the ADC may just provide the range of precision tools necessary to deliver the best, personalized and targeted, treatment for our patients.nb_NO
dc.language.isoengnb_NO
dc.publisherNTNUnb_NO
dc.relation.ispartofseriesDoctoral theses at NTNU;2019:203
dc.titleMulti-parametric Diffusion Weighted Magnetic Resonance Imaging and Analysis in Breast Cancernb_NO
dc.typeDoctoral thesisnb_NO
dc.subject.nsiVDP::Mathematics and natural science: 400::Physics: 430nb_NO
dc.description.localcodedigital fulltext is not avialablenb_NO


Tilhørende fil(er)

Thumbnail

Denne innførselen finnes i følgende samling(er)

Vis enkel innførsel