INTRODUÇÃO: Poucas são as estratégias testadas para reduzir o volume de contraste durante angioplastia coronária. Levantamos a hipótese de que o ultrassom intravascular teria o potencial de substituir muitas informações fornecidas pela angiografia, reduzindo, dessa forma, o volume total de contraste utilizado durante a angioplastia coronária. MÉTODOS: No total, 83 pacientes foram randomizados para realização de angioplastia guiada pela angiografia isolada ou angioplastia guiada pelo ultrassom intravascular. Ambos os grupos foram tratados com estratégias rigorosas para redução de contraste, tendo como objetivo primário o volume final de contraste utilizado na angioplastia coronária. Os pacientes foram acompanhados por um período médio de 4 meses. RESULTADOS: A mediana do volume total de contraste foi de 64,5 ml (intervalo interquartil [ITQ], 42,8-97 ml; mínimo de 19 ml e máximo de 170 ml) no grupo angioplastia guiada pela angiografia isolada vs. 20 ml (ITQ, 12,5-30 ml; mínimo de 3 ml e máximo de 54 ml) no grupo angioplastia guiada pelo ultrassom intravascular (P < 0,001). De forma semelhante, a mediana da razão entre o volume de contraste e o clearance de creatinina foi significantemente menor entre os pacientes submetidos a angioplastia guiada pelo ultrassom intravascular, quando comparados aos pacientes do grupo angioplastia guiada pela angiografia isolada (1 [ITQ, 0,6-1,9] vs. 0,4 [ITQ, 0,2- 0,5], respectivamente; P < 0,001). Os desfechos intra-hospitalares e aos 4 meses de acompanhamento não foram diferentes entre os pacientes randomizados para o grupo angioplastia guiada pela angiografia isolada e aqueles do grupo angioplastia guiada pelo ultrassom intravascular. CONCLUSÕES: A utilização racional do ultrassom intravascular como método de imagem para guiar a angioplastia foi segura e reduziu de forma significativa o volume de contraste, comparativamente à angioplastia guiada pela angiografia isolada. O uso do ultrassom intravascular para esse propósito deve ser considerado para pacientes de elevado risco para o desenvolvimento de nefropatia induzida pelo contraste ou sobrecarga de volume e que serão submetidos a angioplastia coronária / BACKGROUND: To date, few approaches have been described to reduce the final dose of contrast agent in percutaneous coronary intervention. We hypothesized that intravascular ultrasound might serve as an alternative imaging tool to angiography in many steps during percutaneous coronary intervention, thereby reducing the use of iodine contrast. METHODS: A total of 83 patients were randomized to angiography alone-guided percutaneous coronary intervention or intravascular ultrasound-guided percutaneous coronary intervention. Both groups were treated according to a pre-defined meticulous procedural strategy, and the primary endpoint was the total volume contrast agent used during percutaneous coronary intervention. Patients were followed clinically for an average of 4 months. RESULTS: The median total volume of contrast was 64.5 mL (interquartile range [IQR], 42.8 to 97 mL; minimum, 19 mL; maximum, 170 mL) in the angiography alone-guided group vs. 20 mL (IQR, 12.5 to 30 mL; minimum, 3 mL; maximum, 54 mL) in the intravascular ultrasound-guided group (P < 0.001). Similarly, the median volume of contrast/creatinine clearance ratio was significantly lower among patients treated with intravascular ultrasound-guided percutaneous coronary intervention when compared with patients treated with angiography alone-guided percutaneous coronary intervention (1 [IQR, 0.6 to 1.9] vs. 0.4 [IQR, 0.2 to 0.6], respectively; P < 0.001). In-hospital and 4-month outcomes were not different between patients randomized to angiography alone-guided and intravascular ultrasound-guided percutaneous coronary intervention. CONCLUSIONS: Thoughtful and extensive use of intravascular ultrasound as the primary imaging tool to guide percutaneous coronary intervention was safe and markedly reduced the volume of iodine contrast compared with angiographyalone guidance. The use of intravascular ultrasound should be considered for patients at high risk of contrast-induced acute kidney injury or volume overload undergoing coronary angioplasty

Dans le cadre de cette thèse, nous nous somme focalisés sur deux objectifs complémentaires. Le premier concerne l’évolution de la méthode du Gabarit Déformable Elastique (GDE) pour l’extraction semi-automatique de l’anatomie et du mouvement cardiaque, développée au laboratoire Creatis. Un travail a d’abord été réalisé sur une base de données de 45 patients afin de mettre en évidence les points forts et les points faibles de l’algorithme, notamment la difficulté à suivre des déformations trop importantes ou des formes inhabituelles. Puis, différents types de contraintes ont été intégrées au modèle GDE afin d’en améliorer les performances : prescription locale ou dense de déplacements, directionnalité de la déformation contrainte par celle des fibres. Les contraintes proposées sont évaluées sur des données de synthèse et des données réelles en IRM ciné et de marquage tissulaire acquises chez l’homme. Parallèlement, une étude a été réalisée pour mettre en place la méthodologie nécessaire à l’extraction et l’analyse statistique de la déformation des fibres myocardiques. Ce travail a été effectué en collaboration avec une équipe du Auckland Bioengineering Institute en Nouvelle-Zélande. Un modèle biomécanique par éléments finis intègre la direction principale des fibres en tout point du ventricule gauche issue d’acquisitions en IRM du tenseur de diffusion (IRM-TD) sur coeurs humains ex vivo et le mouvement issu de séquences IRM marquées. Cette combinaison permet l’estimation de la déformation des fibres et sa variation durant le cycle cardiaque. La variabilité dans la déformation des fibres est étudiée statistiquement à travers le croisement d’une base de données IRM-TD et d’une base de données IRM marquées. / In this thesis, we are interested in two complementary goals. First, we have improved the Dynamic Deformable Elastic Template (DET) model, developed at Creatis, for the semi-automatic extraction of the anatomy and cardiac motion. The performance of the method was assessed on a database consisting in 45 patients and yielded fairly accurate results. However, it experienced difficulties when dealing with very large thickening throughout the cardiac cycle. Thus, different type of constraints were integrated to the DET model in order to improve robustness and accuracy : local or dense prescribed displacements, deformations directionally constrained by the fibres. These constraints are evaluated on simulated and real human data, in both cine and tagged MR images. A methodology has also been developed in order to extract and statistically analyse myocardial fibre strain. This work was done in collaboration with a team at the Auckland Bioengineering Institute in New Zealand. A finite elements biomechanical model integrates the principle direction of fibres in the left ventricle from Diffusion Tensor MRI acquisitions on ex vivo human hearts and motion from tagged MRI sequences. Fibre strain and its variation throughout the cardiac cycle were estimated. Variability in fibre strain is statistically studied by joining DT-MRI and tagged MRI databases.

Les travaux présentés dans cette thèse s'inscrivent dans le contexte du traitement des images cardiaques et l'étude de la fonction contractile du coeur, pour une meilleure compréhension de physiopathologie cardiaque et l'aide au diagnostic. Nous avons implémenté une méthode de segmentation des parois endocardiques sur des ciné-IRM non marquées. Nous avons utilisé une approche fondée sur les ensembles de niveaux, avec une formulation basée-région qui donne des résultats satisfaisants sur des cas sains et pathologiques. Nous avons proposé une méthode pratique pour la quantification des déformations segmentaires afin decaractériser la contractilité myocardique. La méthode a fait l'objet d'une validation clinique par l'expertise de médecins et par comparaison à la méthode HARP sur des IRM marquées. Pour améliorer la précision des mesures, nous avons proposé un algorithme de recalage iconique multimodal IRM/TDM, utilisant la maximisation de l'information culturelle. Nous l'avons appliqué à la localisation de coupes petit-axe dans des volumes TDM avec des résultats encourageants. Ce travail a pour perspective son application à l'obtention de séquences TDM de hautes résolutions spatiale et temporelle / The work presented in this thesis is related to the cardiac images processing and the cardiac contractile function study, for a better comprehension of cardiac physiopathology and diagnosis. We implemented a method for the segmentation of the endocardial walls on standard MRI without tags. We used an approach based on the level set method, with a region-based formulation which gives satisfactory results on healthy and pathological cases. We proposed a practical method for the quantification of the segmental deformations in order to characterize the myocardial contractility. The method was clinically validated by the assesment of doctors and by comparison with the HARP method on tagget MRI. To improve the measurements precision, we proposed an iconic MRI/CT multimodal registration algorithm, using the maximization of the mutual information. We applied it to the localization of short-axis slices in CT volumes with good results. This work has as prospect its application to obtain high spatial and temporal resolutions CT sequences

A hemólise decorrente da doença aloimune desencadeia mecanismos adaptativos hematológicos e hemodinâmicos fetais, com intuito de garantir o suprimento adequado de oxigênio para todos os tecidos e órgãos. Na anemia grave, a sobrecarga imposta ao coração fetal, devido ao fluxo hiperdinâmico, tem sido considerada responsável pela insuficiencia cardiaca, e posterior desenvolvimento de hidropisia fetal. No entanto, a literatura médica ainda apresenta controvérsias acerca da integridade da função cardiaca nesta doença. O índice de performance miocárdico (IPM) é uma ferramenta propedêutica não invasiva, derivada do Doppler pulsátil, que permite avaliar a função cardíaca global (sistólica e diastólica). Objetivo: Estudar a função cardíaca fetal, na doença aloimune, utilizando o índice de performance miocárdica. Métodos: Foram seguidos, prospectivamente, fetos únicos, de gestantes sensibilizadas pelo antígeno eritrocitário D, sem malformações estruturais, na Clinica Obstétrica, Hospital das Clinicas da Faculdade de Medicina de São Paulo. A cada avaliação ultrassonográfica, o IPM foi investigado por meio de Doppler pulsátil, com janela de 2-4mm, filtro>190Hz e velocidade de varredura que permitisse observar de 3 a 4 ciclos cardíacos simultâneos no ecrã. O IPM corresponde à soma dos tempos isovolumétricos (contração e relaxamento) dividido pelo tempo de ejeção do ciclo cardíaco. O IPM do ventrículo equerdo (IPM VE) foi realizado em ciclo único, sendo possível obter seus componentes: tempo de contração isovolumétrico (TCI), tempo de relaxamento isovolumétrico (TRI) e tempo de ejeção (TE). O IPM do ventrículo direito (IPM VD) foi obtido em dois tempos. Os valores obtidos de IPM VE, seus componentes foram convertidos em escore zeta para a idade gestacional. Frente à suspeita de anemia fetal, realizou-se cordocentese com determinação dos níveis de hemoglobina fetal antes e após a transfusão intra-uterina. Os respectivos valores foram convertidos em escore-zeta (Hb zeta). Na análise estatística, foram incluidas avaliações do IPM realizadas com menos de 72 horas antes, e até 24 horas após cada transfusão. Para cada transfusão foi calculada a variação no IPM (delta IPM = IPM antes - IPM após). O nível de significância estatísca adotado foi de 0,05. Análises por regressão linear simples e logística foram utilizadas para examinar a associação entre os valores de IPM e delta IPM e as seguintes variáveis: idade gestacional no procedimento, múltiplos da mediana (MoM) da Vmax ACM, Hb zeta pré e após TIU, volume de sangue transfundido e porcentagem da expansão do volume feto-placentário (EVFP). Resultados: Foram incluidas 14 gestações submetidas a 31 procedimentos de cordocentese para transfusão intra-uterina. A idade gestacional média na 1ª transfusão foi de 28,2 ± 4,1 semanas Em 6 procedimentos, a avaliação do IPM pós transfusional foi incompleta, e esses dados não foram incluídos na análise. Quanto à análise dos dados obtidos nas cordocenteses, observou-se correlação significativa entre os valores de escore zeta de IPM VE (r= 0,59, p <0,001), TRI (r= 0,45, p =0,01) e o TE (r= 0,42, p=0,2) e o escore zeta da hemoglobina fetal. Não foi observada correlação significativa com o escore zeta do TCI (r= 0,35, p=0,054) e do IPM VD (r=0,12, p= 0,53). Quando comparados aos valores observados antes das transfuões intra-uterinas, observou-se aumento significativo do escore zeta de IPM VE após os procedimentos (Delta MPI = 1,10 ± 2,47, p = 0,036). Não foi observada correlação entre os valores de escore zeta de IPM antes e após TIU. Delta MPI do VE se correlacionou inversamente, e de forma significativa, com a idade gestacional no procedimento (r= 0,47, p=0,018), escore zeta IPM VE pré-TIU (r= 0,50, p=0,012) e EVFP (r= 0,41, p=0,044). Conclusões: O desempenho miocárdico do ventriculo esquerdo fetal permanece preservado frente a anemia, e nos casos de anemia moderada e grave encontra-se ainda mais eficiente. Após a realização da transfusão intrauterina, observou-se aumento significativo do índice de performance miocárdica, e este aumento esteve relacionado com idade gestacional no procedimento, valores de IPM pré-transfusionais e a expansão do volume feto-placentário / Fetal anemia is associated with several adaptative mechanisms in order to maintain adequate tissue oxygenation. Circulatory changes play a key role in such circumstances. In severe anemia, the overload imposed on the fetal heart, due to the hyperdynamic flow, has been considered to be responsible for cardiac failure and finally hydrops fetalis. However, cardiac failure in this pathology remains controversy. Myocardial performance index (MPI) is a novel technique, Doppler derived and non-invasive that allows assesses global cardiac function (systolic and diatolic). Objective: Evaluate global cardiac function in alloimune disease through myocardial performance index. Methods: This prospective study was carried out at a tertiary referral center for fetal medicine (Clínica Obstetrica do Hospital das Clinicas da Faculdade de Medicina da Universidade de São Paulo). Women with singleton pregnancies and Rh D alloimmune disease were invited to take part in the study and gave informed consent. Fetal examinations did not show structural abnormalities. At every ultrasonography evaluation, MPI was examined with Doppler sample gate set between 2-4mm, wall motion filter >190Hz and high sweep-speed to allow simultaneous identification of 3-4 cardiac cycles on the screen. MPI is the sum of isovolumetric times (contraction and relaxation) divided by ejection time. Left ventricle MPI (LV MPI) was obtained in a single cycle and the MPI components were obtained: isovolumetric contraction time (ICT), isovolumetric relaxation time (IRT) and ejection time (ET). Right ventricle MPI (RV MPI) was obtained in two cycles. The values obtained for LV MPI and its components were converted in zeta score for gestacional age. Cordocentesis was perfomed if fetal anemia was suspicion and fetal hemoglobin levels were determined: before and after intrauterine transfusion. Hemoglobin values were converted into the zeta score (Hb zeta). Statistical analysis included MPI evaluations performed within less than 72 hours before and until 24 hours after every transfusion. Variation in the MPI was calculated for every transfusion (delta MPI = MPI before - MPI after). Significance level was set at 0,05. Linear and regression analyses were made in order to examine association between MPI values and delta MPI gestational age at procedure, fetal ACM multiples of median (MoM), Hb zeta before and after the IUT, volume of blood transfused and percentage of the feto-placental expansion volume (FPEV). Results: 14 pregnancies were included. Overall 31 cordocentesis for intrauterine transfusion were performed at mean gestational age of 28,2 ± 4,1 weeks. In 6 procedures, post transfusion MPI evaluation was incomplete and these data were not included in the analysis. Zeta-score values LV MPI (r= 0,59, p < 0,001), IRT (r= 0,45, p =0,01) and ET (r= 0,42, p=0,02) correlated significantly with fetal hemoglobin zeta score. Left ventricle ICT zeta-score (r= 0,35, p=0,054) and RV MPI (r=0,12, p= 0,53). did not show significant correlation. After intrauterine transfusion, LV MPI z-score ]increases and it was statistical significant (Delta MPI = 1,10 ± 2,47, p = 0,036). No correlation was observed between MPI zeta score values before and after the IUT. Delta LV MPI had inverse and significant correlation with pregnancy age in the proceedings (r= 0,47, p=0,018), LV MPI zeta score before IUT (r= 0,50, p=0,012) and FPEV (r= 0,41, p=0,044). Conclusions: Left ventricle myocardial performance not only remains preserved but is actually enhanced in cases of moderate/severe fetal anemia. After intrauterine transfusion procedure, left ventricle myocardial performance index increases significantly and greater changes are associated with procedures at earlier gestational age, lower pre transfusion MPI z-scores and smaller feto-placental volume expansion

A hemólise decorrente da doença aloimune desencadeia mecanismos adaptativos hematológicos e hemodinâmicos fetais, com intuito de garantir o suprimento adequado de oxigênio para todos os tecidos e órgãos. Na anemia grave, a sobrecarga imposta ao coração fetal, devido ao fluxo hiperdinâmico, tem sido considerada responsável pela insuficiencia cardiaca, e posterior desenvolvimento de hidropisia fetal. No entanto, a literatura médica ainda apresenta controvérsias acerca da integridade da função cardiaca nesta doença. O índice de performance miocárdico (IPM) é uma ferramenta propedêutica não invasiva, derivada do Doppler pulsátil, que permite avaliar a função cardíaca global (sistólica e diastólica). Objetivo: Estudar a função cardíaca fetal, na doença aloimune, utilizando o índice de performance miocárdica. Métodos: Foram seguidos, prospectivamente, fetos únicos, de gestantes sensibilizadas pelo antígeno eritrocitário D, sem malformações estruturais, na Clinica Obstétrica, Hospital das Clinicas da Faculdade de Medicina de São Paulo. A cada avaliação ultrassonográfica, o IPM foi investigado por meio de Doppler pulsátil, com janela de 2-4mm, filtro>190Hz e velocidade de varredura que permitisse observar de 3 a 4 ciclos cardíacos simultâneos no ecrã. O IPM corresponde à soma dos tempos isovolumétricos (contração e relaxamento) dividido pelo tempo de ejeção do ciclo cardíaco. O IPM do ventrículo equerdo (IPM VE) foi realizado em ciclo único, sendo possível obter seus componentes: tempo de contração isovolumétrico (TCI), tempo de relaxamento isovolumétrico (TRI) e tempo de ejeção (TE). O IPM do ventrículo direito (IPM VD) foi obtido em dois tempos. Os valores obtidos de IPM VE, seus componentes foram convertidos em escore zeta para a idade gestacional. Frente à suspeita de anemia fetal, realizou-se cordocentese com determinação dos níveis de hemoglobina fetal antes e após a transfusão intra-uterina. Os respectivos valores foram convertidos em escore-zeta (Hb zeta). Na análise estatística, foram incluidas avaliações do IPM realizadas com menos de 72 horas antes, e até 24 horas após cada transfusão. Para cada transfusão foi calculada a variação no IPM (delta IPM = IPM antes - IPM após). O nível de significância estatísca adotado foi de 0,05. Análises por regressão linear simples e logística foram utilizadas para examinar a associação entre os valores de IPM e delta IPM e as seguintes variáveis: idade gestacional no procedimento, múltiplos da mediana (MoM) da Vmax ACM, Hb zeta pré e após TIU, volume de sangue transfundido e porcentagem da expansão do volume feto-placentário (EVFP). Resultados: Foram incluidas 14 gestações submetidas a 31 procedimentos de cordocentese para transfusão intra-uterina. A idade gestacional média na 1ª transfusão foi de 28,2 ± 4,1 semanas Em 6 procedimentos, a avaliação do IPM pós transfusional foi incompleta, e esses dados não foram incluídos na análise. Quanto à análise dos dados obtidos nas cordocenteses, observou-se correlação significativa entre os valores de escore zeta de IPM VE (r= 0,59, p <0,001), TRI (r= 0,45, p =0,01) e o TE (r= 0,42, p=0,2) e o escore zeta da hemoglobina fetal. Não foi observada correlação significativa com o escore zeta do TCI (r= 0,35, p=0,054) e do IPM VD (r=0,12, p= 0,53). Quando comparados aos valores observados antes das transfuões intra-uterinas, observou-se aumento significativo do escore zeta de IPM VE após os procedimentos (Delta MPI = 1,10 ± 2,47, p = 0,036). Não foi observada correlação entre os valores de escore zeta de IPM antes e após TIU. Delta MPI do VE se correlacionou inversamente, e de forma significativa, com a idade gestacional no procedimento (r= 0,47, p=0,018), escore zeta IPM VE pré-TIU (r= 0,50, p=0,012) e EVFP (r= 0,41, p=0,044). Conclusões: O desempenho miocárdico do ventriculo esquerdo fetal permanece preservado frente a anemia, e nos casos de anemia moderada e grave encontra-se ainda mais eficiente. Após a realização da transfusão intrauterina, observou-se aumento significativo do índice de performance miocárdica, e este aumento esteve relacionado com idade gestacional no procedimento, valores de IPM pré-transfusionais e a expansão do volume feto-placentário / Fetal anemia is associated with several adaptative mechanisms in order to maintain adequate tissue oxygenation. Circulatory changes play a key role in such circumstances. In severe anemia, the overload imposed on the fetal heart, due to the hyperdynamic flow, has been considered to be responsible for cardiac failure and finally hydrops fetalis. However, cardiac failure in this pathology remains controversy. Myocardial performance index (MPI) is a novel technique, Doppler derived and non-invasive that allows assesses global cardiac function (systolic and diatolic). Objective: Evaluate global cardiac function in alloimune disease through myocardial performance index. Methods: This prospective study was carried out at a tertiary referral center for fetal medicine (Clínica Obstetrica do Hospital das Clinicas da Faculdade de Medicina da Universidade de São Paulo). Women with singleton pregnancies and Rh D alloimmune disease were invited to take part in the study and gave informed consent. Fetal examinations did not show structural abnormalities. At every ultrasonography evaluation, MPI was examined with Doppler sample gate set between 2-4mm, wall motion filter >190Hz and high sweep-speed to allow simultaneous identification of 3-4 cardiac cycles on the screen. MPI is the sum of isovolumetric times (contraction and relaxation) divided by ejection time. Left ventricle MPI (LV MPI) was obtained in a single cycle and the MPI components were obtained: isovolumetric contraction time (ICT), isovolumetric relaxation time (IRT) and ejection time (ET). Right ventricle MPI (RV MPI) was obtained in two cycles. The values obtained for LV MPI and its components were converted in zeta score for gestacional age. Cordocentesis was perfomed if fetal anemia was suspicion and fetal hemoglobin levels were determined: before and after intrauterine transfusion. Hemoglobin values were converted into the zeta score (Hb zeta). Statistical analysis included MPI evaluations performed within less than 72 hours before and until 24 hours after every transfusion. Variation in the MPI was calculated for every transfusion (delta MPI = MPI before - MPI after). Significance level was set at 0,05. Linear and regression analyses were made in order to examine association between MPI values and delta MPI gestational age at procedure, fetal ACM multiples of median (MoM), Hb zeta before and after the IUT, volume of blood transfused and percentage of the feto-placental expansion volume (FPEV). Results: 14 pregnancies were included. Overall 31 cordocentesis for intrauterine transfusion were performed at mean gestational age of 28,2 ± 4,1 weeks. In 6 procedures, post transfusion MPI evaluation was incomplete and these data were not included in the analysis. Zeta-score values LV MPI (r= 0,59, p < 0,001), IRT (r= 0,45, p =0,01) and ET (r= 0,42, p=0,02) correlated significantly with fetal hemoglobin zeta score. Left ventricle ICT zeta-score (r= 0,35, p=0,054) and RV MPI (r=0,12, p= 0,53). did not show significant correlation. After intrauterine transfusion, LV MPI z-score ]increases and it was statistical significant (Delta MPI = 1,10 ± 2,47, p = 0,036). No correlation was observed between MPI zeta score values before and after the IUT. Delta LV MPI had inverse and significant correlation with pregnancy age in the proceedings (r= 0,47, p=0,018), LV MPI zeta score before IUT (r= 0,50, p=0,012) and FPEV (r= 0,41, p=0,044). Conclusions: Left ventricle myocardial performance not only remains preserved but is actually enhanced in cases of moderate/severe fetal anemia. After intrauterine transfusion procedure, left ventricle myocardial performance index increases significantly and greater changes are associated with procedures at earlier gestational age, lower pre transfusion MPI z-scores and smaller feto-placental volume expansion

No description available.

L’imagerie par résonance magnétique de diffusion (l’IRMd) est actuellement la seule technique non-invasive pour étudier l’architecture tridimensionnelle des fibres myocardiques du cœur humain à la fois ex vivo et in vivo. Cependant, il est difficile de savoir comment les caractéristiques de diffusion calculées à partir des images de diffusion reflètent les propriétés des microstructures du myocarde à cause de l’absence de la vérité-terrain sans parler de l’influence de divers facteurs tels que la résolution spatiale, le bruit et les artéfacts. L'objectif principal de cette thèse est donc de développer des simulateurs de l’IRM de diffusion basés sur des modèles réalistes afin de simuler, en intégrant différentes modalités d'imagerie, les images pondérées en diffusion des fibres myocardiques à la fois ex vivo et in vivo, et de proposer un outil générique permettant d’évaluer la qualité de l’imagerie et les algorithmes de traitement d’images. Pour atteindre cet objectif, le présent travail se focalise sur quatre parties principales. La première partie concerne la formulation de la théorie de simulation IRMd pour la génération d’images de diffusion et pour les applications sur les modèles simples de fibres cardiaques chez l’homme, et essaie de comprendre la relation sous-jacente entre les propriétés mesurées de la diffusion et les caractéristiques à la fois physiques et structurelles des fibres cardiaques. La seconde partie porte sur la simulation des images de résonance magnétique de diffusion à différentes échelles en s’appuyant sur des données du cœur humain issues de l'imagerie par lumière polarisée. En comparant les propriétés de diffusion à différentes échelles, la relation entre la variation de la microstructure et les propriétés de diffusion observée à l'échelle macroscopique est étudiée. La troisième partie consacre à l’analyse de l'influence des paramètres d'imagerie sur les propriétés de diffusion en utilisant une théorie de simulation améliorée. La dernière partie a pour objectif de modéliser la structure des fibres cardiaques in vivo et de simuler les images de diffusion correspondantes en combinant la structure des fibres cardiaques et le mouvement cardiaque connu a priori. Les simulateurs proposés nous fournissent un outil générique pour générer des images de diffusion simulées qui peuvent être utilisées pour évaluer les algorithmes de traitement d’images, pour optimiser le choix des paramètres d’IRM pour les fibres cardiaque aussi bien ex vivo que in vivo, et pour étudier la relation entre la structure de fibres microscopique et les propriétés de diffusion macroscopiques. / Diffusion magnetic resonance imaging (dMRI) appears currently as the unique imaging modality to investigate noninvasively both ex vivo and in vivo three-dimensional fiber architectures of the human heart. However, it is difficult to know how well the diffusion characteristics calculated from diffusion images reflect the microstructure properties of the myocardium since there is no ground-truth information available and add to that the influence of various factors such as spatial resolution, noise and artifacts, etc. The main objective of this thesis is then to develop realistic model-based dMRI simulators to simulate diffusion-weighted images for both ex vivo and in vivo cardiac fibers by integrating different imaging modalities, and propose a generic tool for the evaluation of imaging quality and image processing algorithms. To achieve this, the present work focuses on four parts. The first part concerns the formulation of basic dMRI simulation theory for diffusion image generation and subsequent applications on simple cardiac fiber models, and tries to elucidate the underlying relationship between the measured diffusion anisotropic properties and the cardiac fiber characteristics, including both physical and structural ones. The second part addresses the simulation of diffusion magnetic resonance images at multiple scales based on the polarized light imaging data of the human heart. Through both qualitative and quantitative comparison between diffusion properties at different simulation scales, the relationship between the microstructure variation and the diffusion properties observed at macroscopic scales is investigated. The third part deals with studying the influence of imaging parameters on diffusion image properties by means of the improved simulation theory. The last part puts the emphasis on the modeling of in vivo cardiac fiber structures and the simulation of the corresponding diffusion images by combining the cardiac fiber structure and the a priori known heart motion. The proposed simulators provide us a generic tool for generating the simulated diffusion images that can be used for evaluating image processing algorithms, optimizing the choice of MRI parameters in both ex vivo and in vivo cardiac fiber imaging, and investigating the relationship between microscopic fiber structure and macroscopic diffusion properties.

La connaissance de l’architecture tridimensionnelle (3D) des fibres est cruciale dans la compréhension de la fonction du cœur humain. L’imagerie par résonance magnétique du tenseur de diffusion (IRM-DT) est une technique permettant de mesurer la diffusion des molécules d’eau dans des tissus humains, et donc d’étudier de manière non-invasive l’architecture 3D des fibres du cœur humain. Dans l’IRM-TD cardiaque, la tractographie des fibres est essentielle pour représenter et visualiser l’architecture des fibres, mais souvent utilisée qualitativement comme une dernière étape qui consiste à visualiser sur l’écran l’architecture myocardique obtenue à partir des données IRM-TD. Cependant, cette visualisation qualitative n’est pas suffisante pour décrire de manière objective et complète l’architecture des fibres. L’objectif de cette thèse est de développer de nouvelles approches pour la tractographie et pour la description quantitative de l’architecture des fibres cardiaques du cœur humain en IRM-TD cardiaque. Les travaux de cette thèse se focalisent sur trois axes. Le premier est le développement d’un algorithme de tractographie probabiliste, qui prend en compte la corrélation spatiale des fibres pendant le suivi des fibres myocardiques. Les résultats expérimentaux montrent que la méthode proposée est robuste au bruit. Les fibres produites sont plus régulières et plus lisses, et la configuration des fibres cardiaques est plus facile à observer. Le second axe concerne une nouvelle notion de dépliement de fibres pour décrire les fibres du cœur humain, qui sont souvent complexes dans l’espace 3D. L’idée est d’analyser cette architecture 3D dans un espace réduit à deux dimensions (2D), en utilisant une technique d’apprentissage de variété. L’approche de dépliement proposée permet la description quantitative de l’architecture 3D de fibres cardiaques dans un plan 2D. Les résultats montrent qu’il est beaucoup plus facile d’observer et d’étudier les caractéristiques des fibres cardiaques après les avoir dépliées, et qu’il semble exister des formes de fibres caractéristiques du cœur humain. Le dernier axe consiste en la fusion de fibres, qui est obtenue en moyennant les fibres selon une grille. Cette approche fournit des architectures de fibres simplifiée à différentes échelles, et permet de mieux mettre en évidence la configuration des fibres cardiaques. / It is important to study the cardiac fiber architecture in order to understand the heart function. Diffusion tensor MRI (DT-MRI) is the only noninvasive technique that allows studying cardiac fiber architecture in vivo. Tractography is essential in representing and visualizing cardiac fiber architecture in DT-MRI, but is often employed qualitatively. The motivation of this thesis is to develop technique for studying the cardiac fiber architecture from the fiber tracts provided by the tractography process in cardiac DT-MRI. Our goal is to develop tractography algorithm and approaches for the quantitative description of cardiac fiber architecture. My work is composed of three main axis. The first is the development of a probabilistic tractography algorithm, which takes fiber spatial correlation into accounts in tracing fibers. Experimental results show that the proposed method is meaningfully more robust to noise than the streamlining method, and produces more regular and smoother fibers, which enables cardiac fiber configurations to be more clearly observed. The second concerns a new framework, namely cardiac fiber unfolding, which is an isometric mapping. Our fiber unfolding framework allows the quantitative description of three dimensional cardiac fiber architecture in a two dimensional plan. Our experimental results show that fiber tract pattern can be observed much easier by unfolding them in a plane, and several cardiac fiber patterns were found. The last axis consists in merging fibers, which is achieved by averaging fibers according to a grid. This fiber merging approach provide simplified fiber architecture at different scale as output that highlights the cardiac fiber configuration.

Les signaux analytiques multidimensionnels nous permettent d'avoir des possibilités de calculer les phases et modules. Cependant, peu de travaux se trouvent sur les signaux analytiques multidimensionnels qui effectuent une extensibilité appropriée pour les applications à la fois sur du traitement des données médicales 2D et 3D. Cette thèse a pour objectif de proposer des nouvelles méthodes pour le traitement des images médicales 2D/3D pour les applications de détection d'enveloppe et d'estimation du mouvement. Premièrement, une représentation générale du signal quaternionique 2D est proposée dans le cadre de l'algèbre de Clifford et cette idée est étendue pour modéliser un signal analytique hypercomplexe 3D. La méthode proposée décrit que le signal analytique complexe 2D, est égal aux combinaisons du signal original et de ses transformées de Hilbert partielles et totale. Cette écriture est étendue au cas du signal analytique hypercomplexe 3D. Le résultat obtenu est que le signal analytique hypercomplexe de Clifford peut être calculé par la transformée de Fourier complexe classique. Basé sur ce signal analytique de Clifford 3D, une application de détection d'enveloppe en imagerie ultrasonore 3D est présentée. Les résultats montrent une amélioration du contraste de 7% par rapport aux méthodes de détection d'enveloppe 1D et 2D. Deuxièmement, cette thèse propose une approche basée sur deux phases spatiales du signal analytique 2D appliqué aux séquences cardiaques. En combinant l'information de ces phases des signaux analytiques de deux images successives, nous proposons un estimateur analytique pour les déplacements locaux 2D. Pour améliorer la précision de l'estimation du mouvement, un modèle bilinéaire local de déformation est utilisé dans un algorithme itératif. Cette méthode basée sur la phase permet au déplacement d'être estimé avec une précision inférieure au pixel et est robuste à la variation d'intensité des images dans le temps. Les résultats de sept séquences simulées d'imagerie par résonance magnétique (IRM) marquées montrent que notre méthode est plus précise comparée à des méthodes récentes utilisant la phase du signal monogène ou des méthodes classiques basées sur l'équation du flot optique. Les erreurs d'estimation de mouvement de la méthode proposée sont réduites d'environ 33% par rapport aux méthodes testées. En outre, les déplacements entre deux images sont cumulés en temps, pour obtenir la trajectoire d'un point du myocarde. En effet, des trajectoires ont été calculées sur deux patients présentant des infarctus. Les amplitudes des trajectoires des points du myocarde appartenant aux régions pathologiques sont clairement réduites par rapport à celles des régions normales. Les trajectoires des points du myocarde, estimées par notre approche basée sur la phase de signal analytique, sont donc un bon indicateur de la dynamique cardiaque locale. D'ailleurs, elles s'avèrent cohérentes à la déformation estimée du myocarde. / Different mathematical tools, such as multidimensional analytic signals, provide possibilities to calculate multidimensional phases and modules. However, little work can be found on multidimensional analytic signals that perform appropriate extensibility for the applications on both of the 2D and 3D medical data processing. In this thesis, based on the Hahn 1D complex analytic, we aim to proposed a multidimensional extension approach from the 2D to a new 3D hypercomplex analytic signal in the framework of Clifford algebra. With the complex/hypercomplex analytic signals, we propose new 2D/3D medical image processing methods for the application of ultrasound envelope detection and cardiac motion estimation. Firstly, a general representation of 2D quaternion signal is proposed in the framework of Clifford algebra and this idea is extended to generate 3D hypercomplex analytic signal. The proposed method describes that the complex/hypercomplex 2D analytic signals, together with 3D hypercomplex analytic signal, are equal to different combinations of the original signal and its partial and total Hilbert transforms, which means that the hypercomplex Clifford analytic signal can be calculated by the classical Fourier transform. Based on the proposed 3D Clifford analytic signal, an application of 3D ultrasound envelope detection is presented. The results show a contrast optimization of about 7% comparing with 1D and 2D envelope detection methods. Secondly, this thesis proposes an approach based on two spatial phases of the 2D analytic signal applied to cardiac sequences. By combining the information of these phases issued from analytic signals of two successive frames, we propose an analytical estimator for 2D local displacements. To improve the accuracy of the motion estimation, a local bilinear deformation model is used within an iterative estimation scheme. This phase-based method allows the displacement to be estimated with subpixel accuracy and is robust to image intensity variation in time. Results from seven realistic simulated tagged magnetic resonance imaging (MRI) sequences show that our method is more accurate compared with the state-of-the-art method. The motion estimation errors (end point error) of the proposed method are reduced by about 33% compared with that of the tested methods. In addition, the frame-to-frame displacements are further accumulated in time, to allow for the calculation of myocardial point trajectories. Indeed, from the estimated trajectories in time on two patients with infarcts, the shape of the trajectories of myocardial points belonging to pathological regions are clearly reduced in magnitude compared with the ones from normal regions. Myocardial point trajectories, estimated from our phase-based analytic signal approach, are therefore a good indicator of the local cardiac dynamics. Moreover, they are shown to be coherent with the estimated deformation of the myocardium.

L'orientation des fibres myocardiaque est à la base du comportement électro-mécanique du cœur, et connue pour être altérée dans diverses maladies cardiaques telles que la cardiopathie ischémique et l'hypertrophie ventriculaire. Cette thèse porte principalement sur l'imagerie in vivo du tenseur de diffusion (diffusion tensor imaging—DTI) en vue d’obtenir la structure des fibres myocardiques du cœur humain dans des conditions de respiration libre. L'utilisation de DTI pour l'étude du cœur humain in vivo est un grand défi en raison du mouvement cardiaque. En particulier, l’acquisition DTI avec respiration libre sans recourir au gating respiratoire est très difficile à cause des mouvements à a fois respiratoire et cardiaque. Pour aborder ce problème, nous proposons de nouvelles approches consistant à combiner des acquisitions à retards de déclenchement multiples (trigger delay—TD) et des méthodes de post-traitement. D’abord, nous réalisons des acquisitions avec multiples TD décalés en fin de diastole. Ensuite, nous développons deux méthodes de post-traitement. La première méthode s’attaque au problème d’effets de mouvement physiologique sur DTI cardiaque in vivo en utilisant les techniques de recalage et de PCATMIP (Principal Components Analysis filtering and Temporal Maximum Intensity Projection). La deuxième méthode traite le problème de mouvement par l’utilisation d’un algorithme de fusion d’images basé sur l’ondelette (wavelet-based image fusion-WIF) et d’une technique de débruitage PCA (Principal Components Analysis). Enfin, une comparaison des mesures DTI entre la méthode PCATMIP et la méthode WIF est réalisée ; les champs de tenseurs sont calculés, à partir desquels les propriétés de l’architecture des fibres in vivo sont comparées. Les résultats montrent qu’en utilisant les approches proposées, il est possible d’étudier l’impact du mouvement cardiaque sur les paramètres de tenseur de diffusion, et d’explorer les relations sous-jacentes entre les propriétés de tenseur de diffusion mesurées et le mouvement cardiaque. Nous trouvons aussi que la combinaison des acqusiitions avec des TD multiples décalés and des post-traitements d’images peut compenser les effets de mouvement physiologique, ce qui permet d’obtenir l’architecture 3D du cœur humain dans des conditions de respiration libre. Les résultats suggèrent de nouvelles solutions au problème de perte du signal due au mouvement, qui sont prometteuses pour obtenir les propriétés de l’architecture des fibres myocardiques du cœur humain in vivo, dans des conditions cliniques. / The orientation of cardiac fibers underlies the electro-mechanical behavior of the heart, and it is known to be altered in various cardiac diseases such as ischemic heart disease and ventricular hypertrophy. This thesis mainly focuses on in vivo diffusion tensor imaging (DTI) to obtain the myocardial fiber structure of the human heart under free-breathing conditions. The use of DTI for studying the human heart in vivo is challenging due to cardiac motion. In particular, free-breathing DTI acquisition without resorting to respiratory gating is very difficult due to both respiratory and cardiac motion. To deal with this problem, we propose novel approaches that combine multiple shifted trigger delay (TD) acquisitions and post-processing methods. First, we perform multiple shifted TD acquisitions at end diastole. Then, we focus on two different post-processing methods. The first method addresses physiological motion effects on in vivo cardiac DTI using image co-registration and PCATMIP (Principal Components Analysis filtering and Temporal Maximum Intensity Projection). The second method is a wavelet-based image fusion (WIF) algorithm combined with a PCA noise removing method. Finally, a comparison of DTI measurements between the PCATMIP and WIF methods is also performed; tensor fields are calculated, from which the in vivo fiber architecture properties are compared. The results show that using the proposed approaches, we are able to study the cardiac motion effects on diffusion tensor parameters, and investigate the underlying relationship between the measured diffusion tensor properties and the cardiac motion. We also find that the combination of multiple shifted TD acquisitions and dedicated image post-processing can compensate for physiological motion effects, which allows us to obtain 3D fiber architectures of the human heart under free-breathing conditions. The findings suggest new solutions to signal loss problems associated with bulk motion, which are promising for obtaining in vivo human myocardial fiber architecture properties in clinical conditions.