Page 19 - Artificial Intelligence for Computational Modeling of the Heart

Foreword xix

Foreword

The clinical cardiology community is witnessing an exciting dis-
ruption towards precision medicine, fueled by two revolutionary
quantitative science approaches: computational modeling and ar-
tificial intelligence (AI).
Precision medicine is becoming the dominant paradigm in
the contemporary practice of medicine. Challenging the tradi-
tional population-based model, it attempts to tailor medical treat-
ment to the individual characteristics of each patient, quantified
through genomic, proteomic, metabolomic and physiomic fea-
tures. Examples of precision medicine already in use today include
genetic testing for inherited arrhythmia syndrome (e.g. KVLQT1,
HERG, SCN5A) to prevent sudden cardiac death, familial hyperc-
holesterolemia to prevent coronary artery disease, pharmacoge-
nomics to assess individualized response to critical drugs (e.g.
CYP2C19, SLCO1B1, VKORC1), and measurement of cardiac tro-
ponin I (cTnI) to risk-stratify patients following acute coronary
syndrome.
Medical imaging is playing a growing role in precision medicine.
Over the past decade, advanced cardiovascular imaging such as
cardiac computed tomography (CT), magnetic resonance (MR)
and positron emission tomography (PET) has succeeded in pro-
viding patient-specific anatomical and physiological information
that allows individualized, minimally invasive treatment in clini-
cal cardiology. For example, myocardial scar imaging by MR pro-
vides individualized information regarding the risk of fatal ar-
rhythmia and the potential target of interventional therapy to
prevent arrhythmic events. From a mathematical point of view,
imaging provides a detailed mapping of cardiac anatomy and sub-
strate, represented as a time-varying scalar field of signal intensity
(Fig. 0.1 Imaging). Some physiological information such as blood
flow velocity, myocardial blood flow and finite deformation of the
cardiac tissue can also be measured, but it is typically limited in
spatial and temporal resolution.
While imaging is excellent at capturing heart shape, substrate
and kinematics, understanding and quantifying how it functions
is still an open challenge. To that end, researchers are looking for
ways to combine medical imaging and computational modeling
to provide detailed physiological information based on physical
principles (Fig. 0.1 Computational Modeling). Thanks to recent
technological breakthroughs, image-based, patient-specific com-
putational modeling is becoming a powerful tool to augment the