Page 164 - Artificial Intelligence for Computational Modeling of the Heart

136 Chapter 4 Data-driven reduction of cardiac models

from an existing anatomical model, a remaining limitation for
its routine clinical usage is the time required to reconstruct the
anatomical model, which may vary between 10 and 60 minutes.
A pre-clinical study employing the same tool and algorithms for
the anatomical model generation reported a processing time of
37.5 ± 13.8 minutes [328]. Thus, to mitigate this limitation, an im-
portant future activity will focus on the further automation of this
step, and the minimization of user interaction.

4.2 Meta-modeling of atrial
electrophysiology

Many heart diseases manifest as abnormal rhythms due to dis-
orders that affect the cardiac electrical system. The underlying
cause of the disease can be different and requires specific inter-
ventions. A range of therapy choices are available, including med-
ical therapy with antiarrhythmic drugs, implantation of pacemak-
ers/defibrillators, ablation of the foci of irregular electrical activity,
or surgery. However, many challenges remain open today in pa-
tient selection (what patient should undergo what therapy), ther-
apy planning and guidance.
A large body of literature documents steady progress in the
understanding of cardiac electrophysiology (EP), thanks to exper-
imental studies in cells, tissues, in animals, and in humans. In par-
ticular, significant advances have been made in the understanding
of mechanisms underlying heart diseases such as atrial fibrilla-
tion (AFib). Atrial fibrillation is a defect of the electrical system
of the heart that manifests in irregular contraction of the atria.
This causes reduced efficiency of the pumping action of the heart
and elevated risks of blood clot formation in the atria – poten-
tially leading to stroke. Various treatment options are available,
including pharmacological, surgical and non-surgical approaches
(such as ablation). Selection of the optimal course of treatment
as well as planning of the optimal delivery of the treatment (for
instance in the case of catheter ablation) are still open clinical
challenges. Computational modeling of cardiac EP and arrhyth-
mia has emerged as a complimentary approach to experimen-
tal studies for the understanding of the mechanisms producing
abnormal cellular activation patterns, with the potential of en-
abling non-invasive, cost effective and individualized analysis of
the electrophysiological state and evolution of the heart. Recent
modeling studies achieved remarkable fidelity in reproducing ob-
servable electrophysiological phenomena, and hold the promise
of a new generation of tools for individualized patient care.