Page 57 - Artificial Intelligence for Computational Modeling of the Heart

Chapter 1 Multi-scale models of the heart for patient-specific simulations 27

of blood in the heart chambers and coronary vessels characterizes
physiological or pathological conditions such as valvular defects,
myocardial infarction, thrombosis, cardiomyopathies.
In normal physiological conditions, an increased demand of
blood flow from the peripheral organs causes vasodilation and
a corresponding reduction in peripheral resistances. This results
in increased flow rate of circulating blood, even without changes
to the mean arterial pressure, and therefore an increase in the
amount of blood filling the atria and ventricles at the end of each
cardiac cycle. According to the Frank–Starling’s law, the heart can
match increased end diastolic volume of the heart chambers with
an increased stroke volume, thus adapting to the blood flow de-
mand from the peripheral organs [137]. The underlying mecha-
nism responsible for such adaptation is the length-tension rela-
tionship of the myocardium, whereas the myofibers can produce
greater tension if their initial length is stretched beyond their nor-
mal resting length [101].
Increases in the mean arterial pressure can be caused by a
range of factors such as smoking, lack of physical activity, obe-
sity, diet and others [138]. This translates in increased ventricu-
lar pressure during the systolic phase, which requires the ventri-
cles to express an increased level of tensile stress. As a result, the
work required from the ventricle to guarantee the cardiac out-
put increases. This also explains why chronic hypertension may
promote cardiac hypertrophy: a thicker myocardium can better
account for the increased work demand [139].
Modeling hemodynamics in the heart chambers and in the cir-
culatory system allows to better understand the patient-specific
context in which the heart is operating. Depending on the focus of
interest, different modeling approaches can be considered [140].
Detailed analysis of the 3D fluid dynamics of blood can be an im-
portant tool in studying clinically relevant problems such as valve
mechanics (in particular for diseased valves) [141–143], coupling
of the myocardium with devices such as ventricular assist devices
[144–146], analyzing the effect of cardiac interventions [147–149].
Reduced order models may be a more practical and convenient
choice for the monitoring of spatial-averaged quantities such as
blood pressure and flow rate, for instance for the computation
of pressure-volume loops in which the model provides an esti-
mation of intracardiac pressure removing the need for invasively
measured pressures [150,151]. They are also used in combination
with detailed hemodynamics models to describe larger portions
of the cardiovascular system, while keeping the computational
complexity under control. This modeling approach is commonly
referred to as geometrical multiscale modeling and aims at focus-