Computational fluid dynamics indicators to improve cardiovascular pathologies

In recent years, the study of computational hemodynamics within anatomically complex vascular regions has generated great interest among clinicians. The progress in computational fluid dynamics, image processing and high-performance computing haveallowed us to identify the candidate vascular regions for the appearance of cardiovascular diseases and to predict how this disease may evolve. Medicine currently uses a paradigm called diagnosis. In this thesis we attempt to introduce into medicine the predictive paradigm that has been used in engineering for many years. The objective of this thesis is therefore to develop predictive models based on diagnostic indicators for cardiovascular pathologies. We try to predict the evolution of aortic abdominal aneurysm, aortic coarctation and coronary artery disease in a personalized way for each patient. To understand how the cardiovascular pathology will evolve and when it will become a health risk, it is necessary to develop new technologies by merging medical imaging and computational science. We propose diagnostic indicators that can improve the diagnosis and predict the evolution of the disease more efficiently than the methods used until now. In particular, a new methodology for computing diagnostic indicators based on computational hemodynamics and medical imaging is proposed. We have worked with data of anonymous patients to create real predictive technology that will allow us to continue advancing in personalized medicine and generate more sustainable health systems. However, our final aim is to achieve an impact at a clinical level. Several groups have tried to create predictive models for cardiovascular pathologies, but they have not yet begun to use them in clinical practice. Our objective is to go further and obtain predictive variables to be used practically in the clinical field. It is to be hoped that in the future extremely precise databases of all of our anatomy and physiology will be available to doctors. These data can be used for predictive models to improve diagnosis or to improve therapies or personalized treatments. En els ultims anys, l'estudi de l'hemodinamica computacional en regions vasculars anatomicament complexes ha generat un gran interes entre els clinics. El progres obtingut en la dinamica de fluids computacional, en el processament d'imatges i en la computacio d'alt rendiment ha permes identificar regions vasculars on poden apareixer malalties cardiovasculars, aixi com predir-ne l'evolucio. Actualment, la medicina utilitza un paradigma anomenat diagnostic. En aquesta tesi s'intenta introduir en la medicina el paradigma predictiu utilitzat des de fa molts anys en l'enginyeria. Per tant, aquesta tesi te com a objectiu desenvolupar models predictius basats en indicadors de diagnostic de patologies cardiovasculars. Tractem de predir l'evolucio de l'aneurisma d'aorta abdominal, la coartacio aortica i la malaltia coronaria de forma personalitzada per a cada pacient. Per entendre com la patologia cardiovascular evolucionara i quan suposara un risc per a la salut, cal desenvolupar noves tecnologies mitjancant la combinacio de les imatges mediques i la ciencia computacional. Proposem uns indicadors que poden millorar el diagnostic i predir l'evolucio de la malaltia de manera mes eficient que els metodes utilitzats fins ara. En particular, es proposa una nova metodologia per al calcul dels indicadors de diagnostic basada en l'hemodinamica computacional i les imatges mediques. Hem treballat amb dades de pacients anonims per crear una tecnologia predictiva real que ens permetra seguir avancant en la medicina personalitzada i generar sistemes de salut mes sostenibles. Pero el nostre objectiu final es aconseguir un impacte en l?ambit clinic. Diversos grups han tractat de crear models predictius per a les patologies cardiovasculars, pero encara no han comencat a utilitzar-les en la practica clinica. El nostre objectiu es anar mes enlla i obtenir variables predictives que es puguin utilitzar de forma practica en el camp clinic. Es pot preveure que en el futur tots els metges disposaran de bases de dades molt precises de tota la nostra anatomia i fisiologia. Aquestes dades es poden utilitzar en els models predictius per millorar el diagnostic o per millorar terapies o tractaments personalitzats.