Application of multivariate image analysis to prostate cancer for improving the comprehension of the related physiological phenomena and the development and validation of new imaging biomarkers

Abstract

A doctoral thesis submitted to the Department of applied statistics, Operations Research and QualityThis journey started six years ago after finishing the last exams of my Master degree.In that summer of 2013 I was thinking about my future and how to orientate the Master's thesis. Truly, I had no idea on what to do, but Alberto Ferrer, after the presentation of my work for his subject, suggested me to have a meeting with José Manuel Prats where they showed me the interesting new approach of "our" methods to medical imaging they were working on. I got involved so easy in this world by their hand, and after all this time, they have been the best supervisors I could ever imagined. I learnt how to investigate; they taught me about the importance of knowing the process in detail before trying anything, to argue everything, to propose new ideas, the importance of the testing/validation, to give a measure of the reliability of the results and more, a bunch things that will make this list extremely long. Thank you so much, Alberto and José Manuel. A few more words for JM, he has been always there for me, in the best and worst moments, he is an amazing professor and better person. To my second supervisor Roberto Sanz, it was always a pleasure to work with him. The relation between these two worlds (university and hospital) is never easy, but he managed to go with us together in the same direction during this journey. Roberto, thank you for your dedication and hard work of giving us always the best data set you can, listening to our ideas and contributing with your valuable advices. I would like to thank Luís Martí Bonmatí for introducing us into this world and saving some time from his busy schedule for us, always in a kind way. I would also like to thank all the people from the Hospital Quirón, without their work and their images, this thesis wouldn't have been possible.To my office mates. Dani, we know each other for a long time and we have travelled a lot together, it was always nice to have him on the other side of the table. To Raffa and Abel for helping me and answering to all my questions. To Borja and Antonio for having good times discussing and arguing together. To Alba and Joan, we didn't share so much time but they are magnificent and very talented people.To my "all time" friends in my little village, Kike, Mode, Jorge, Javi, Toni, Carles, etc. for the long nights of just talking. To my friends in the city, Vicci and Rivera, we lived together 4 years and I got so many precious memories that will stay with me forever.And last, but not least, to my family, they always supported me, gave me their love and tried to charm me up when I felt bad. But also to celebrate when we succeed. Unfortunately, my "uelos" aren't with us anymore, I dedicate this work to them and hope they would have been proud of me. Thank you.

Sequential multiblock partial least squares discriminant analysis for assessing prostate cancer aggressiveness with multiparametric magnetic resonance imaging

“…A detailed formulation explanation can be found in Refs. [17,24]. In these models, the vascular and EES volumes can be represented as a system of partial differential equations where the contrast is exchanged between different subspaces.…”

“…The process of obtaining the number of components and the initial estimation of D T matrix was described in previous works [14][15][16][17]. Briefly, two components (referred to true behaviors) were considered for the MCR models, using the concentrations as their corresponding biomarkers: in perfusion, normal tissue (NT) and highly vascular tissue (VT) types); and in diffusion, fast (d 1 ) and slow (d 2 ) exponential decays.…”

“…To do so, it is mandatory to fit and characterize the intensity versus time (in the case of DCE-MR) or versus b-value (in the case of DW-MR) curves associated to each pixel of the images. Although pharmacokinetic models [8] (in the case of DCE-MR) and exponential models [9,10] (in the case of DW-MR) have the ability to provide clinically-oriented biomarkers in tumor analysis, their interpretation is not easy nor direct in many cases; so new biomarkers obtained from latent variables-based multivariate statistical models, like multivariate curve resolution (MCR [11][12][13]), have recently been also proposed [14][15][16][17] to improve prostate cancer assessment.…”