The heart is the most important organ of the cardio circulatory system. The development of new diagnostic methods and new treatments dedicated to the heart constitutes an important part of medical research. Nevertheless, the difficulties in data collection with non-invasive methods makes necessary the use of alternative methods to study certain diseases. For this reason, the computational methods applied in this area are subject of increasing interest. \\ In particular, more and more accurate models are available for the description of cardiac mechanics. An important aspect that needs to be considered is the nonlinear anisotropic behavior of the myocardium. Such anisotropy is mainly due to the presence of muscular fibers. Recently, various studies have proved the importance of the presence of collagen in the muscular contraction. The presence of collagen introduces another anisotropic direction. This effect must be included in the mathematical models that describe the elastic properties of the ventricle. In the literature have been proposed different constitutive laws which account for the orthotropy of the myocardium, considering both muscular fibers and collagen sheets.\\ %For a complete description of the heart's mechanics it is necessary to introduce in the mechanical model also a model which accounts for the fibers activation. The heart is an active material that can change its configuration in absence of external loads. In fact, muscular fibers, if excited with an electric signal, contract. Hence in order to describe the behavior of the myocardium it is necessary introducing the activation in the machanical model. Two different approaches are available in the literature: the active stress and the active strain models. Actually, both models originate from the same thermodynamical principle but with different constitutive choice for the energy. In this thesis, we will consider only the active strain model because it introduce naturally an anisotropic activation. In particular, we will use at first a transversely isotropic activation. However, such model need a contraction for the fibers far greater with respect of the observed value in order to reproduce physiological values for the ventricular wall thickening. For an accurate description of the macroscopic deformation of the ventricle we will use an orthotropic activation model. Furthermore, several studies have evidenced that the distribution of the radial strains through the myocardium has a strong dishomogeneity between the epicardium and the endocardium. For this reason, we propose to modify the orthotropic activation model in order to include effects of dishomogeneous activation. We compare the proposed model with both transversely isotropic activation and orthotropic activation models on a idealized geometry, represented by a truncated ellipsoid. Then, we apply the proposed method on a real ventricle, reconstructed from medical images. To reconstruct the patient specific geometry, we used a region-based segmentation method based on the minimization of the region scalable fitting energy functional by means of the Split-Bregman approach. Despite the good performances of the segmentation model, we need a procedure for the image pre-processing in order to obtain the isolated ventricle. The obtained results, on both ideal and patient-specific geometries, show that the method we proposed keeps the good performance of the orthotropic activation model. Moreover, it succeeds in describing the transmural dishomogeneity of the ventricular wall thickening and improves the description of the ventricular activation mechanics.
Il cuore è l'organo più importante del sistema cardio circolatorio. Lo sviluppo di nuovi strumenti diagnostici e di nuove cure dedicate al cuore occupa una parte importante della ricerca medica. Ciò nonostante, le difficoltà nella raccolta di dati con strumenti non invasivi rende necessario l'utilizzo di tecniche alternative per lo studio di alcune patologie. Per questo motivo i metodi computazionali applicati in questo ambito sono oggetto di crescente interesse.\\ In particolare, modelli sempre più accurati sono disponibili per la descrizione della meccanica cardiaca. Un aspetto di particolare importanza è il comportamento non lineare ed anisotropo del miocardio. Tale anisotropia è dovuta principalmente alla presenza di fibre muscolari. Recentemente, vari studi hanno dimostrato l'importanza della presenza di collagene nella contrazione muscolare. La conformazione del collagene introduce nel tessuto un'altra direzione di anisotropia. Questo effetto deve essere incluso nei modelli matematici che descrivono le proprietà elastiche del ventricolo. In letteratura sono stati proposti diverse leggi costitutive che descrivono il tessuto cardiaco come ortotropo, tenendo in considerazione sia le fibre muscolari sia i fogli di collagene.\\ Il cuore è un materiale attivo che può cambiare la propria configurazione in assenza di carichi esterni. Infatti le fibre muscolari reagiscono ad uno stimolo elettrico contraendosi, inducendo così la contrazione del miocardio. Quindi per descrivere il comportamento del miocardio è necessario introdurre l'attivazione nel modello meccanico. Due differenti approcci sono presenti in letteratura: il modello active stress ed il modello active strain. In realtà è possibile vedere che entrambi i modelli derivano dallo stesso principio termodinamico ma da diverse scelte della forma costitutiva dell'energia. In questa tesi, noi utilizzeremo l'approccio active strain perché introduce naturalmente un'attivazione anisotropa. In particolare utilizzeremo dapprima un'attivazione trasversalmente isotropa. Tuttavia tale approccio per riprodurre valori fisiologici per l'ispessimento della parete ventricolare necessità valori di contrazioni delle fibre molto maggiori di quelli osservati. Per una descrizione più accurata delle deformazioni macroscopiche del ventricolo utilizzaremo quindi un modello ortotropo. Inoltre, diversi lavori hanno evidenziato che l'andamento degli sforzi radiali attraverso il miocardio presenta una forte disomogeneità fra l'epicardio e l'endocardio. Per questa ragione in questo lavoro proponiamo di modificare il modello di attivazione ortotropa per includere effetti di attivazione disomogenea. Confrontiamo il modello proposto con i modelli trasversalmente isotropi ed ortotropi di active strain su ua geometria idealizzata del ventricolo sinistro, rappresentata da un ellisoide troncato. Successivamente applichiamo il modello proposto nel caso di un ventricolo sinistro ricostruito da immagini mediche. Per ricostruire la geometria patient specific abbiamo usato un metodo di segmentazione region based basato sulla minimizzazione con metodo di Split-Bregman del funzionale di energia del region scalable fitting model (RFE). Malgrado le buone caratteristiche del modello di segmentazione, è necessario un procedimento di preprocessing dell'immagine per poter ottenere il ventricolo isolato. I risultati ottenuti, sia su geometria ideale che su geometria patient-specific, mostrano che il metodo da noi proposto oltre a mantenere le buone prestazioni del modello di attivazione ortotropo, riesce a descrivere la disomogeneità dell'ispessimento della parete ventricolare e migliora la descrizione del meccanismo di attivazione meccanica ventricolare.
A mathematical and numerical study of the left ventricular contraction based on the reconstruction of a patient specific geometry
BARBAROTTA, LUCA
2013/2014
Abstract
The heart is the most important organ of the cardio circulatory system. The development of new diagnostic methods and new treatments dedicated to the heart constitutes an important part of medical research. Nevertheless, the difficulties in data collection with non-invasive methods makes necessary the use of alternative methods to study certain diseases. For this reason, the computational methods applied in this area are subject of increasing interest. \\ In particular, more and more accurate models are available for the description of cardiac mechanics. An important aspect that needs to be considered is the nonlinear anisotropic behavior of the myocardium. Such anisotropy is mainly due to the presence of muscular fibers. Recently, various studies have proved the importance of the presence of collagen in the muscular contraction. The presence of collagen introduces another anisotropic direction. This effect must be included in the mathematical models that describe the elastic properties of the ventricle. In the literature have been proposed different constitutive laws which account for the orthotropy of the myocardium, considering both muscular fibers and collagen sheets.\\ %For a complete description of the heart's mechanics it is necessary to introduce in the mechanical model also a model which accounts for the fibers activation. The heart is an active material that can change its configuration in absence of external loads. In fact, muscular fibers, if excited with an electric signal, contract. Hence in order to describe the behavior of the myocardium it is necessary introducing the activation in the machanical model. Two different approaches are available in the literature: the active stress and the active strain models. Actually, both models originate from the same thermodynamical principle but with different constitutive choice for the energy. In this thesis, we will consider only the active strain model because it introduce naturally an anisotropic activation. In particular, we will use at first a transversely isotropic activation. However, such model need a contraction for the fibers far greater with respect of the observed value in order to reproduce physiological values for the ventricular wall thickening. For an accurate description of the macroscopic deformation of the ventricle we will use an orthotropic activation model. Furthermore, several studies have evidenced that the distribution of the radial strains through the myocardium has a strong dishomogeneity between the epicardium and the endocardium. For this reason, we propose to modify the orthotropic activation model in order to include effects of dishomogeneous activation. We compare the proposed model with both transversely isotropic activation and orthotropic activation models on a idealized geometry, represented by a truncated ellipsoid. Then, we apply the proposed method on a real ventricle, reconstructed from medical images. To reconstruct the patient specific geometry, we used a region-based segmentation method based on the minimization of the region scalable fitting energy functional by means of the Split-Bregman approach. Despite the good performances of the segmentation model, we need a procedure for the image pre-processing in order to obtain the isolated ventricle. The obtained results, on both ideal and patient-specific geometries, show that the method we proposed keeps the good performance of the orthotropic activation model. Moreover, it succeeds in describing the transmural dishomogeneity of the ventricular wall thickening and improves the description of the ventricular activation mechanics.File | Dimensione | Formato | |
---|---|---|---|
2014_10_Barbarotta.pdf
accessibile in internet per tutti
Dimensione
11.77 MB
Formato
Adobe PDF
|
11.77 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/96363