Predicting Heart Failure E-Book
145,99 €
Mehr erfahren.
- Herausgeber: John Wiley & Sons
- Kategorie: Fachliteratur
- Sprache: Englisch
PREDICTING HEART FAILURE Predicting Heart Failure: Invasive, Non-Invasive, Machine Learning and Artificial Intelligence Based Methods focuses on the mechanics and symptoms of heart failure and various approaches, including conventional and modern techniques to diagnose it. This book also provides a comprehensive but concise guide to all modern cardiological practice, emphasizing practical clinical management in many different contexts. Predicting Heart Failure supplies readers with trustworthy insights into all aspects of heart failure, including essential background information on clinical practice guidelines, in-depth, peer-reviewed articles, and broad coverage of this fast-moving field. Readers will also find: * Discussion of the main characteristics of cardiovascular biosensors, along with their open issues for development and application * Summary of the difficulties of wireless sensor communication and power transfer, and the utility of artificial intelligence in cardiology * Coverage of data mining classification techniques, applied machine learning and advanced methods for estimating HF severity and diagnosing and predicting heart failure * Discussion of the risks and issues associated with the remote monitoring system * Assessment of the potential applications and future of implantable and wearable devices in heart failure prediction and detection * Artificial intelligence in mobile monitoring technologies to provide clinicians with improved treatment options, ultimately easing access to healthcare by all patient populations. Providing the latest research data for the diagnosis and treatment of heart failure, Predicting Heart Failure: Invasive, Non-Invasive, Machine Learning and Artificial Intelligence Based Methods is an excellent resource for nurses, nurse practitioners, physician assistants, medical students, and general practitioners to gain a better understanding of bedside cardiology.
Sie lesen das E-Book in den Legimi-Apps auf:
Seitenzahl: 669
Veröffentlichungsjahr: 2022
Ähnliche
Predicting Heart Failure
Invasive, Non-Invasive, Machine Learning, and Artificial Intelligence Based Methods
Edited by
Kishor Kumar Sadasivuni
Qatar University
Doha, Qatar
Hassen M. Ouakad
Sultan Qaboos University
Muscat, Oman
Somaya Al-Maadeed
Qatar University
Doha, Qatar
Huseyin C. Yalcin
Qatar University
Doha, Qatar
Issam Bait Bahadur
Sultan Qaboos University
Muscat, Oman
This edition first published 2022
© 2022 John Wiley & Sons Ltd All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by law. Advice on how to obtain permission to reuse material from this title is available at http://www.wiley.com/go/permissions.
The right of Kishor Kumar Sadasivuni, Hassen M. Ouakad, Somaya Al-Maadeed, Huseyin C. Yalcin, and Issam Bait Bahadur to be identified as the authors of the editorial material in this work has been asserted in accordance with law.
Registered OfficesJohn Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USAJohn Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK
Editorial Office9600 Garsington Road, Oxford, OX4 2DQ, UK
For details of our global editorial offices, customer services, and more information about Wiley products visit us at www.wiley.com.
Wiley also publishes its books in a variety of electronic formats and by print-on-demand. Some content that appears in standard print versions of this book may not be available in other formats.
Limit of Liability/Disclaimer of WarrantyThe contents of this work are intended to further general scientific research, understanding, and discussion only and are not intended and should not be relied upon as recommending or promoting scientific method, diagnosis, or treatment by physicians for any particular patient. In view of ongoing research, equipment modifications, changes in governmental regulations, and the constant flow of information relating to the use of medicines, equipment, and devices, the reader is urged to review and evaluate the information provided in the package insert or instructions for each medicine, equipment, or device for, among other things, any changes in the instructions or indication of usage and for added warnings and precautions. While the publisher and authors have used their best efforts in preparing this work, they make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives, written sales materials or promotional statements for this work. The fact that an organization, website, or product is referred to in this work as a citation and/or potential source of further information does not mean that the publisher and authors endorse the information or services the organization, website, or product may provide or recommendations it may make. This work is sold with the understanding that the publisher is not engaged in rendering professional services. The advice and strategies contained herein may not be suitable for your situation. You should consult with a specialist where appropriate. Further, readers should be aware that websites listed in this work may have changed or disappeared between when this work was written and when it is read. Neither the publisher nor authors shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages.
Library of Congress Cataloging-in-Publication DataNames: Kumar Sadasivuni, Kishor, 1986- author.Title: Predicting heart failure : invasive, non-invasive,machine learning and artificial intelligence based methods / edited by Kishor Kumar Sadasivuni, Qatar University, Doha, Qatar, [and four others].Description: Hoboken, NJ : John Wiley & Sons, 2022. | Includes bibliographical references and index.Identifiers: LCCN 2021043721 (print) | LCCN 2021043722 (Obook) | ISBN 9781119813019 (hardback) | ISBN 9781119813026 (pdf) | ISBN 9781119813033 (epub) | ISBN 9781119813040 (ebook)Subjects: LCSH: Heart failure--Risk factors. | Heart failure--Treatment. | Heart--Diseases--Patients.Classification: LCC RC685.C53 P74 2023 (print) | LCC RC685.C53 (ebook) | DDC 616.1/29--dc23/eng/20211028LC record available at https://lccn.loc.gov/2021043721LC ebook record available at https://lccn.loc.gov/2021043722
Cover image: Courtesy of Kishor Kumar SadasivuniCover design by Wiley
Set in 9.5/12.5pt STIXTwoText by Integra Software Services Pvt. Ltd, Pondicherry, India
Contents
Cover
Title page
Copyright
Preface
Abbreviations
Acknowledgment
1 Invasive, Non-Invasive, Machine Learning, and Artificial Intelligence Based Methods for Prediction of Heart Failure
2 Conventional Clinical Methods for Predicting Heart Disease
3 Types of Biosensors and their Importance in Cardiovascular Applications
4 Overview and Challenges of Wireless Communication and Power Transfer for Implanted Sensors
5 Minimally Invasive and Non-Invasive Sensor Technologies for Predicting Heart Failure: An Overview
6 Artificial Intelligence Techniques in Cardiology: An Overview
7 Utilizing Data Mining Classification Algorithms for Early Diagnosis of Heart Disease
s
8 Applications of Machine Learning for Predicting Heart Failure
9 Machine Learning Techniques for Predicting and Managing Heart Failure
10 Clinical Applications of Artificial Intelligence in Early and Accurate Detection of Low- Concentration CVD Biomarkers
11 Commercial Non-Invasive and Invasive Devices for Heart Failure Prediction: A Review
12 Artificial Intelligence Based Commercial Non-Invasive and Invasive Devices for Heart Failure Diagnosis and Prediction
13 Future Techniques and Perspectives on Implanted and Wearable Heart Failure Detection Devices
Index
End User License Agreement
List of Illustrations
Chapter 1
Figure 1.1 Images for boundary detection in...
Figure 1.2 Supervised and unsupervised machine learning.
Chapter 2
Figure 2.1 Supine, left lateral decubitus...
Figure 2.2 ECG. E denotes electrodes placed...
Figure 2.3 Major waves of a single ECG signal...
Figure 2.4 Treadmill assessment.
Figure 2.5 Evolution of rubber-based stethoscopes.
Figure 2.6 A digital stethoscope.
Figure 2.7 The echocardiography method. The...
Figure 2.8 Formation of plaques in the...
Figure 2.9 Plaque formation in the heart arteries.
Chapter 3
Figure 3.1 Statistics of the number of publications...
Figure 3.2 A schematic representation of some...
Figure 3.3 Application of wearable/implantable...
Chapter 4
Figure 4.1 Implant sensor which can monitor...
Figure 4.2 Implantable cardioverter defibrillators.
Figure 4.3 Implantable biosensor underneath...
Figure 4.4 Biosensor sensor insertion procedure.
Figure 4.5 Different types of wireless energy transfer.
Figure 4.6 Block diagram of microwave wireless power transfer.
Figure 4.7 Rectenna (rectifying antenna)...
Figure 4.8 Inductive mutual coupling.
Figure 4.9 A schematic description of typical...
Figure 4.10 Ultrasound wireless power transfer.
Figure 4.11 Block diagram of the biosensor implant...
Figure 4.12 Simple block diagram of inductive...
Figure 4.13 Schematic of bidirectional...
Figure 4.14 Simplified ADS schematic of resonant link.
Figure 4.15 Schematic circuit of transmitting section.
Figure 4.16 Block diagram of the demodulation module.
Figure 4.17 Block diagram of the inductive...
Figure 4.18 Output voltage.
Figure 4.19 Block diagram of the data transfer structure.
Figure 4.20 Inductive coupling between two...
Figure 4.21 Average ambient RF power.
Figure 4.22 The penetration depth of...
Figure 4.23 Typical WiFi energy harvesting to power-up implant chip.
Figure 4.24 Fractal antenna.
Figure 4.25 Array of fractal Koch antennae patch.
Figure 4.26 S-parameter.
Figure 4.27 3D radiation pattern plots.
Figure 4.28 Directivity of the proposed antenna.
Figure 4.29 Active impedance.
Figure 4.30 Block diagram of powering-up the implanted chip.
Figure 4.31 Power-up pump circuit.
Figure 4.32 Charge pump stage voltage.
Chapter 6
Figure 6.1 Relationship between statistical models and AI.
Figure 6.2 Relationship between AI and ML techniques.
Figure 6.3 Linear and non-linear decision boundaries and surfaces.
Figure 6.4 Comparing ML and deep learning.
Figure 6.5 A perceptron and a simple ANN with interconnected perceptrons.
Figure 6.6 Steps in building AI algorithms in cardiology.
Chapter 7
Figure 7.1 Building the classifier phase.
Figure 7.2 Using the classifier for prediction.
Figure 7.3 Bar plot of target variable.
Figure 7.4 Bar plot for the relationship between target and sex.
Figure 7.5 Bar plot for the relationship between target, sex, and age.
Figure 7.6 Histogram of age variables classified based on sex and target.
Figure 7.7 Histogram of resting blood pressure...
Figure 7.8 ROC for all models over the Cleveland data set.
Figure 7.9 ROC for all models over the Statlog data set.
Chapter 8
Figure 8.1 Overall architecture of HF diagnosis/prediction using ML models.
Chapter 9
Figure 9.1 Prevalence of HF worldwide...
Figure 9.2 NYHA classification. Based...
Figure 9.3 Data description for HF diagnosis.
Figure 9.4 Features for HF diagnosis.
Figure 9.5 Data description for HF severity...
Figure 9.6 Removed features after the application of the preprocessing step.
Figure 9.7 Optimal set of features for severity estimation.
Figure 9.8 Optimal set of features for adverse events prediction.
Chapter 10
Figure 10.1 Various types of CVD and their origin.
Figure 10.2 Various types of sensing mechanism for CVD.
Figure 10.3 Different types of ELISA techniques.
Figure 10.4 Wearable potentiometric ion sensor.
Figure 10.5 DNA based magnetic biosensor.
Figure 10.6 Role of AI in CVD treatment.
Figure 10.7 Overview of the ML workflow. Reproduced...
Figure 10.8 (a) Block diagram of the integrated DL...
Figure 10.9 (a) Different left ventricular...
Chapter 11
Figure 11.1 ECG reading using a Holter...
Figure 11.2 General workflow of MCT devices.
Figure 11.3 Hand-to-leg bioimpedance analysis system.
Figure 11.4 General components of smart wearable...
Figure 11.5 Overall working of remotely monitored invasive devices.
Chapter 12
Figure 12.1 Schematic representation of the pathophysiology...
Chapter 13
Figure 13.1 The system architecture of the CardioMEMS system [25].
Figure 13.2 Cardiac pacemaker (left) [31] and CRT-D (right) [32].
Figure 13.3 Left atrial pressure (LAP) sensing device:...
Figure 13.4 Implantable cardiac devices.
Figure 13.5 Different types of wearable devices...
Figure 13.6 Concept of PPG which uses a light-emitting diode...
Figure 13.7 Different signal processing and machine...
Figure 13.8 Omron HeartGuide [44].
Figure 13.9 Smart wearable devices for arrhythmia detection.
Figure 13.10 Using machine learning algorithms in...
Figure 13.11 Proposed framework to use selfies to screen patients with CAD.
Figure 13.12 Concept of machine learning in...
List of Tables
Chapter 1
Table 1.1 HF types and their outcomes.
Table 1.2 Literature summary for artificial...
Chapter 3
Table 3.1 Specific biomarkers for the evaluation of CVDs.
Table 3.2 Latest biosensor applications for cardiovascular research.
Chapter 4
Table 4.1 Comparison of different energy sources.
Chapter 5
Table 5.1 Summary of trials...
Chapter 6
Table 6.1 ML techniques.
Chapter 7
Table 7.1 Heart failure data set with 14 attributes.
Table 7.2 Statistical description of Cleveland data set.
Table 7.3 Statistical description of Statlog data set.
Table 7.4 Performance measures using the Cleveland data set.
Table 7.5 Performance measures using the Statlog data set.
Chapter 8
Table 8.1 ML paradigms and how they can help tackle HF.
Chapter 9
Table 9.1 State of the art of ML for HF diagnosis.
Table 9.2 State of the art of ML for HF severity estimation.
Table 9.3 State of the art of ML for HF rehospitalization prediction.
Table 9.4 State of the art of ML for HF mortality prediction.
Table 9.5 State of the art of ML for HF prediction of adverse events.
Table 9.6 HF diagnosis classification results, unbalanced class.
Table 9.7 HF diagnosis classification results, balanced class
Table 9.8 Results of the NYHA class detection model.
Table 9.9 Results of the adverse events prediction model.
Chapter 11
Table 11.1 BMI value ranges and associated health risks.
Table 11.2 Non-invasive devices used in...
Table 11.3 Normal range values of different pressure parameters.
Table 11.4 Categories, device, company name...
Chapter 12
Table 12.1 Non-exhaustive chronological summary of ML...
Table 12.2 Chronological summary of invasive...
Chapter 13
Table 13.1 Summary of different wearable device...
Table 13.2 Summary of researches to improve cardiac health of HF patients.
Guide
Cover
Title page
Copyright
Table of Contents
Preface
Abbreviations
Acknowledgments
Begin Reading
Index
End User License Agreement
Pages
i
ii
iii
iv
v
vi
vii
viii
ix
x
xi
xii
xiii
xiv
xv
xvi
xvii
xviii
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
Preface
Our knowledge of human biology, especially related to the heart, increases every day. This makes it nearly impossible for physicians to stay current on the latest research in their fields, let alone in all of the others that directly affect their ability to treat their patients properly. This book will help in learning the mechanics and symptoms of heart failure and the various approaches, including conventional and modern techniques for diagnosing it. Moreover, the book provides a detailed presentation of the latest research data for preventing and treating heart failure.
In this book, 13 chapters address different conditions related to the heart, with detailed descriptions of each. The first chapter discusses invasive, non-invasive, machine learning, and artificial intelligence-based methods for predicting heart failure. In addition, it discusses heart failure causes, symptoms, and treatment, as well as research related to heart failure. In the second chapter, we examine the traditional methods of predicting heart diseases and the implementation of artificial intelligence technology to predict heart diseases accurately. A discussion of the main characteristics of cardiovascular biosensors is presented in Chapter 3, along with their open issues for development and application. We summarize the difficulties of wireless sensor communication and power transfer in Chapters 4, 5, and 6, which outline the utility of artificial intelligence in cardiology. Chapter 7 discusses how to predict heart diseases using data mining classification techniques. Applied machine learning is discussed in Chapters 8 and 9 as are advanced methods for estimating heart failure severity and for diagnosing and predicting heart failure. In Chapter 10, the present state of artificial intelligence and biosensors based on materials is briefly discussed. The underlying technologies of various invasive and non-invasive devices, and their benefits, are discussed and analyzed in Chapter 11. A discussion of the risks and issues associated with remote monitoring systems is also included in this chapter. An overview of these heart failure prediction devices is presented in Chapter 12 along with their invasive and non-invasive alternatives. The chapter also highlights the potential of artificial intelligence in mobile monitoring technologies to provide clinicians with improved treatment options, ultimately easing access to healthcare by all patient populations. Chapter 13 assesses the potential applications of implantable and wearable devices in heart failure detection, summarizes the available data for wearables and machine learning for improving patients’ cardiac health, and discusses the future of wearables for early prediction of heart failure.
Finally we strongly believe this book will provides a comprehensive but concise guide to all modern cardiological practice, emphasizing practical clinical management in many different contexts. The book provides readers with trustworthy insights into all aspects of heart failure, including essential background information on clinical practice guidelines, in-depth, peer-reviewed articles, and broad coverage of this fast-moving field. It provides the latest research data needed for the diagnosis and treatment of heart failure. Moreover, this book is an excellent resource for nurses, nurse practitioners, physician assistants, medical students, and general practitioners to gain a better understanding of bedside cardiology.
Abbreviations
Chapter 1
3-D
Three-dimensional, 10
BNP
Brain Natriuretic Peptide, 9
CAD
Coronary Artery Disease, 2
CHF
Congestive Heart Failure, 3
CRP
C-reactive Protein, 8
HF
Heart Failure, 3
HRV
Heart Rate Variability, 24
IoT
Internet of Things, 11
LS-SVM
Least Squares SVM, 24
LV
Left Ventricular, 23
SVM
Support Vector Machine, 18
Chapter 2
ACS
Acute Coronary Syndrome, 8
AI
Artificial Intelligence, 18 bp Blood Pressure, 12
CAC
Coronary Artery Calcium, 21
CCA
Common Carotid Artery, 19
CHF
Congestive Heart Failure, 20
CNN
Convolutional Neural Network, 20
CQ-NSGT
Constant-Q Non-Stationary Gabor Transform, 20
CT
Computed Tomography, 19
CVDs
Cardiovascular Diseases, 19
DNN
Deep Neural Network, 20
DOE
Dyspnea on Exertion, 10
ECG
Electrocardiogram, 5
ELM
Extreme Learning Machines, 20
FCN
Fully Convolutional Network, 20
GERD
Gastroesophageal Reflux Disease, 9
HDL
High-density Lipoprotein, 12
ICA
Internal Carotid Artery, 19
IMT
Intima-Media Thickness, 19
LDL
Low-density Lipoprotein, 12
LI
Lumen-intima, 19
LII
Lumen-Intima Interface, 20
MA
Media Adventitia, 19
MAI
Media-Adventitia Interface, 20
MLP
Multilayer Perceptron, 21
MPI
Myocardial Perfusion Imaging, 22
NSR
Normal Sinus Rhythm, 20
PND
Paroxysmal Nocturnal Dyspnea, 11
Chapter 3
BNP
Brain Natriuretic Peptide, 5
CRP
C-Reactive Protein, 3
CSGMs
Comb Structured Gold Microelectrode Arrays, 28
cTnI
Cardiac Troponin I, 5
CV
Cyclic Voltammetry, 28
CVDs
Cardiovascular Diseases, 1
DPV
Differential Pulse Voltammetry, 28
EIS
Electrochemical Impedance Spectroscopy, 28
ELISA
Enzyme-Linked Immunosorbent Assay, 29
ENPs
Enzyme Nanoparticles, 29
GDF
Growth Differentiation Factor, 7
GK
Glycerol Kinase, 29
GO
Graphene Oxide, 28
GPO
Glycerol-3- Phosphate Oxidase, 29
HDL
High-density Lipoprotein, 8
hour-FABP
Heart Fatty Acid-Binding Protein, 7
IL-6
Interleukin-6, 7
LDL
Low-density Lipoprotein, 8
LOD
Limit of Detection, 10
LSPR
Localized Surface Plasmon Resonance, 12
miRNAs
Micro RNAs, 6
MPO
Myeloperoxidase, 6
NPs
Nanoparticles, 29
NSE
Neuron-specific Enolase, 6
PCT
Procalcitonin, 5
PEC
Photoelectrochemical, 28
PG
Pencil Graphite, 29
QDs
Quantum Dots, 28
SEM
Scanning Electron Microscopy, 28
SERS
Surface-Enhanced Raman Scattering, 12
SPEs
Screen-Printed Electrodes, 28
SPGEs
Screen-Printed Gold Electrodes, 28
SPR
Surface Plasmon Resonance, 12
SPRi
Surface Plasmon Resonance Imaging, 12
sST2
Soluble Suppressor of Tumorgenicity 2, 8
TG
Triglycerides, 9
TNF-α,
Tumor Necrosis Factor-alpha, 7
Chapter 4
AC
Alternative Current, 9
BSN
Body Sensor Network, 2 CE Counter Electrode, 5
CMT
Coupled-Mode Theory, 12
EDAS
European Aeronautic Defense and Space Company, 10
EM
Electromagnetic Interference, 11
EMF
Electromagnetic Field, 9
HPF
High-Pass Filter, 17
IoT
Internet of Things, 2
LPF
Low Pass Filter, 17
MEMS
Micro-Electromechanical Systems, 4
MPT
Microwave Power Transmission, 8
RE
Reference Electrode, 5
RF
Radio Frequency, 3
SHM
Structural Health Supervising, 2
SoC
System on a Chip, 5
VSWR
Voltage Standing Wave Ratio, 24
WBAN
Wireless Body Area Network, 2
WE
Working Electrode, 5
Chapter 5
BCG
Ballistocardiography, 13
BLUE
Bedside Lung Ultrasound, 10
BNP
B-type Natriuretic Peptide, 14
C.A.USE
Cardiac Arrest Ultrasound Exam, 10
CHF
Chronic Heart Failure, 2
DNN
Deep Neural Network, 12
ECG
Electrocardiography, 3
EVLW
Extravascular Lung Water, 8
FALLS
Fluid Administration Limited by Lung Sonography, 10
GPS
Global Positioning System, 6
HF
Heart failure, 2
ICD
Implantable Cardioverter Defibrillator, 12
LuCUS
Lung and Cardiac Ultrasound, 10
LUS
Lung Ultrasound, 8
LV
Left Ventricular, 7
MEMS
Microelectromechanical System, 6
MFCCs
Mel-frequency Cepstral Coefficients, 12
NT-proBNP
Amino-terminal Pro-B-type Natriuretic Peptide, 14
PA
Pulmonary Arterial, 7
PCG
Phonocardiogram, 12
PPG
Photoplethysmogram, 14
RCTs
Randomized Control Trials, 15
ReDS
Remote Dielectric Sensing Technology, 5
RPM
Remote Patient Monitoring, 15
RV
Right Ventricular, 7
SCG
Seismocardiography, 12
Chapter 6
ACM
All-cause Mortality, 8
AI
Artificial Intelligence, 3, 8
ANN
Artificial Neural Networks, 3,12
AUC
Area Under the Curve, 9
CAD
Coronary Artery Disease, 8
CCTA
Cardiac Computed Tomographic Angiography, 8
CMR
Cardiac Magnetic Resonance, 12
CNN
Convolutional Neural Networks, 12
DL
Deep learning, 12
ECG
Electrocardiogram, 9
FDA
Food and Drug Administration, 17
FFR
Fractional Flow Reserve, 9
FRS
Framingham Risk Score, 8
GAN
Generative Adversarial Networks, 11
GAN
Generative Adversarial Networks, 3
HFpEF
Heart Failure with Preserved Ejection Fraction, 10
HMM
Hidden Markov Model, 15
LASSO
Least Absolute Shrinkage and Selection Operator, 3
LV
Left Ventricle, 9
LVEF
Left Ventricular Ejection Fraction, 12
MDI
Modified Duke Index, 8
MFR
Myocardial Flow Reserve, 9
ML
Machine Learning, 3
MPI
Myocardial Perfusion Imaging, 15
MPS
Myocardial Perfusion Scan, 9
NER
Names Entity Recognition, 15
NLP
Natural Language Processing, 15
POS
Part of Speech, 15
PSA
Parsing or Syntactic Analysis, 15
RNN
Recurrent Neural Networks, 12
SIS
Segment Involvement Score, 8
SSS
Segment Stenosis Score, 8
STE
Speckle-tracking Echocardiography, 11
SVMs
Support Vector Machines, 8
TPD
Total Perfusion Deficit, 15
TTE
Transthoracic Echocardiogram, 12
Chapter 7
DT
Decision Tree, 2
K-NN
K-Nearest Neighbor, 5
LDA
Linear Discriminant Analysis, 2
MAFIA
Maximal Frequent Itemset Algorithm, 4
NB
Naïve Bayes, 2
NN
Neural Networks, 4
RF
Random Forest, 2
SVM
Support Vector Machine, 2
Chapter 8
AUPRC
Area Under Precision Recall curve, 10
AUROC
Area Under ROC Curve, 10
BNP
Brain Natriuretic Peptide, 4
CNN
Convolutional Neural Networks, 7
DAE
Denoising Autoencoder, 7
DBNs
Deep Belief Networks, 7
DT
Decision Tree, 7
ECG
Electrocardiogram, 3
EHR
Electronic Health Records, 5
EMB
Endomyocardial Biopsy, 7
ESC
European Society of Cardiology, 4
GBM
Gradient-boosted Model, 11
GRU RNNs
Gated Recurrent Unit Recurrent Neural Networks, 9
GWTG-HF
Get With The Guidelines-Heart Failure, 11
HF
Heart Failure, 2
KNNs
K-nearest Neighbors, 7
LAD
Left Atrial Dimension, 10
LR
Logistic Regression, 11
LSTM
Long Short-Term Memory, 7
MAGGIC
Meta-Analysis Global Group in Chronic, 11
ML
Machine Learning, 2
ReLU
Rectified Linear Unit, 8
RETAIN
REverse Time AttentIoN Model, 9
RF
Random Forest, 8
RNN
recurrent Neural Networks, 7
ROI
Regions of Interests, 8
RPM
Remote Patient Monitoring, 5
SVM
Support Vector Machine, 7
TAN
Tree-augmented Naive Bayesian, 11
Chapter 9
AF
Atrial Fibrillation, 9
AFL
Atrial Flutter, 10
AUC
Area Under the Roc Curve, 8
BN
Bayes Network, 31
CART
Classification and Regression Trees, 15
CFS
Correlation-based Feature Selection, 31
CMR
Cardiac Magnetic Resonance, 6
CNN
Convolutional Neural Network, 9
CTCA
Computed Tomography Coronary Angiography, 6
DL
Deep Learning, 8
DNN
Deep Neural Network, 9
DUNs
Deep Unified Networks, 20
ECG
Electrocardiogram, 6
EF
Ejection Fraction, 5
EFFECT
Enhanced Feedback for Effective Cardiac Treatment, 10
EHR
Electronic Health Record, 8
ELM
Extreme Learning Machine, 10
ESC
European Society of Cardiology, 6
GDS
Generalized Discriminant Analysis, 16
GLM
Generalized Linear Model, 23
HF
Heart Failure, 1
HFmrEF
HF with Mid-range or Mildly Reduced EF, 5
HFpEF
Heart Failure with Preserved Ejection Fraction, 5
HFrEF
Heart Failure with Reduced Ejection Fraction, 5
HRV
Heart Rate Variability, 9
K-NN
K-nearest Neighbor, 8
LGE
Late Gadolinium Enhancement, 6
LMT
Logistic Model Trees, 31
LR
Logistic Regression, 8
LS-SVM
Least Square SVM, 9
LSTM
Long Short-term Memory, 9
MAGGIC
Meta-analysis Global Group in Chronic, 24
ML
Machine Learning, 7
MLP
Multilayer Perceptron, 8
NB
Naïve Bayes, 34
NLP
Natural Language Processing, 16
NN
Neural Network, 8
NRD
Nationwide Readmissions Database, 21
NYHA
New York Heart Association, 5
RBF
Radial Basis Function, 18
RDW
Red Blood Cell Distribution Width, 26
RF
Random Forest, 9
RNN
Recurrent Neural Network, 8
ROT
Rotation Forest, 16
RS
Rough Set, 8
RSA
Random Search Algorithm, 10
SAE
Sparse Auto-encoder, 9
SGD
Stochastic Gradient Descent, 9
SVM
Support Vector Machines, 8
Chapter 10
AI
Artificial Intelligence, 1 bp Blood Pressure, 11
CAD
Coronary Artery Disease, 15
CT
Computed Tomography, 15
CVDs
Cardiovascular Diseases, 1
DL
Deep Learning, 12
ECG
Electrocardiogram, 1
EIS
Electrochemical Impedance Spectroscopy, 4
IoT
The Internet of Things, 11
MRI
Magnetic Resonance Imaging, 2,13
RI
Refractive Index, 6
Chapter 11
BMI
Body Mass Index, 9
CHF
Chronic Heart Failures, 2
CIHM
Chronicle Implantable Hemodynamic Monitor, 19
CRT
Cardiac Resynchronization Therapy, 19
HCG
Human Chorionic Gonadotropin, 21
IASD
Inter Atrial Shunt Device, 22
LA
Left Atrial, 17
LAP
Left Atrial Pressure, 18
MCT
Mobile Cardiac Telemetry, 6
NYHA
New York Heart Association, 19
PA
Pulmonary Artery, 17
PAM
Patient Advisory Module, 20
RV
Right Ventricle, 17
Chapter 12
AI
Artificial Intelligence, 3
ANN
Artificial Neural Networks, 4
AUC
Area Under Curve, 6
CNN
Convolutional Neural Network, 4
CRT
Cardiac Resynchronization Therapy, 6
DL
Deep Learning, 3
DNN
Deep Neural Network, 5
ECG
Electrocardiographic, 5
HF
Heart Failure, 1
k-NN
k-Nearest Neighbors, 5
LVAD
Left Ventricular Assist Device, 7
ML
Machine Learning, 3
PPGs
Photoplethysmograms, 8
RF
Random Forest, 4
RNN
Recurrent Neural Network, 5
RV
Right Ventricular, 2
RVF
Right-ventricular Failure, 7
RVFRS
Right Ventricular Failure Risk Score, 7
SVM
Support Vector Machine, 4
Chapter 13
ABP
Arterial Blood Pressure, 4
AF
Atrial Fibrillation, 11
CAD
Coronary Artery Diseases, 19
CardioMEMS
Cardio-Microelectromechanical system, 3
CRT
Cardiac Resynchronization Therapy, 18
CRT-D
Cardiac Resynchronization Therapy Defibrillator, 4
CVDs
Cardiovascular Diseases, 3
ECG
Electrocardiogram, 2
FDA
Food and Drug Administration, 13
HF
Heart Failure, 1
hour
Heart Rate, 13
ICDs
Implantable Cardioverter-defibrillators, 5
LAP
Left Atrial Pressure, 5
LVAD
Left Ventricular Assist Device, 18
MI
Myocardial Infarction, 14
NSTEMI
Non-ST-elevation Myocardial Infarction, 18
NYHA
New York Heart Association, 10
OHRM
Optical Heart Rate Monitor, 9
PAP
Pulmonary Artery Pressure, 3
PD
Photodiode, 9
PPG
Photoplethysmogram, 4
RR
Respiration Rate, 13
STEMI
ST-elevation Myocardial Infarction, 18
SVM
Support Vector Machine, 18
VF
Ventricular Fibrillation, 11
Acknowledgment
This publication was supported by Qatar University Internal Grant No. IRCC-2020-013 and Sultan Qaboos University through Grant # CL/SQU-QU/ENG/20/01, respectively. The findings achieved herein are solely the responsibility of the authors.
1 Invasive, Non-Invasive, Machine Learning, and Artificial Intelligence Based Methods for Prediction of Heart Failure
Hidayet Takcı
1.1 Introduction
Heart diseases are the deadliest in the world. Of the many diseases included in the category of heart disease, the most prominent is coronary artery disease (CAD), which causes heart attacks. CAD, high blood pressure, and many other heart diseases cause heart failure (HF). With HF being a consequence of heart disease, the prediction of HF is related to the prediction of diseases categorized as heart disease.
In this chapter, the diagnosis of HF is discussed in terms of invasive/non-invasive and artificial intelligence/machine learning techniques. Invasive and non-invasive techniques are distinct in the way the patient is treated. Invasive methods are usually associated with a physical intervention in the body. This intervention involves operations such as taking blood for blood analysis and not pressing strongly on the abdominal area. Non-invasive methods include physical therapy, taking blood pressure, and temperature measurement. In today’s world, where information technologies have evolved in every field, the field of health has also received its share. Computer-aided clinical decision support systems provide the strongest support for diagnostic studies today. The most important components of computer-aided diagnosis are artificial intelligence and machine learning systems that offer a wide range of services from smart assistant applications to imaging techniques. Artificial intelligence and machine learning have a healing role in electrocardiography, echocardiography, and similar invasive and non-invasive techniques.
In the second section of this chapter, HF will be described and its causes, symptoms, and treatment revealed. The third section will explain diagnosis by invasive and non-invasive methods. Computer-aided diagnosis and decision support systems are briefly mentioned in the fourth section. The fifth section looks first at what artificial intelligence is and then presents fields and examples of artificial intelligence supported studies. The sixth section explains machine learning, learning types, machine learning algorithms, and machine learning based diagnostic studies. The chapter’s concluding section summarizes studies and comments upon diagnostic studies for HF.
1.2 Heart Failure
HF belongs to a class of diseases that occur due to several diseases known as heart diseases and can be fatal if left untreated. In this section, HF is defined and the factors causing it, its symptoms, and its treatment are examined.
1.2.1 What is HF?
HF is sometimes known as congestive HF (CHF). HF is not a failure of the heart, but a condition in which the heart muscle cannot pump enough blood. It occurs due to, for example, narrowing of the coronary arteries, cardiovascular diseases, and high blood pressure.
HF can occur at any age, but is more common in older people. Although it is not possible to cure the disease, some of its symptoms can be improved by the correct interventions. Conversely, faulty interventions will make the situation worse than before.
The most common symptoms of HF are shortness of breath after movement or during rest, fatigue and weakness, swelling in the ankles and legs, disturbed heartbeat, persistent cough, and dizziness and the like. Symptoms can develop rapidly (acute HF) or gradually over weeks (chronic HF).
With early intervention for HF, it will be possible for patients to live longer. Some measures such as losing weight, reducing stress, and reducing the amount of salt used will positively affect the quality of life of patients. One way to prevent HF is to prevent and control conditions such as CAD, high blood pressure, diabetes, and obesity.
The activity that will benefit most is diagnosing the disease and conducting studies according to the symptoms. Thus, as in many other diseases, early diagnosis will have a life-saving effect in HF, as will correct diagnosis and determining the type and real cause of the ailment.
1.2.2 Causes of HF
As mentioned, the causes of HF are CAD, high blood pressure, and other heart diseases, and it occurs because of the damage to the heart muscle brought about by these and similar conditions. While oxygen is regularly carried to a normal heart, it is not possible to carry enough oxygen in the case of a damaged heart. Although HF usually occurs on the left side of the heart, it can develop on the right or both sides. Table 1.1 lists multiple HF types and their results.
Table 1.1 HF types and their outcomes.
HF Type
Outcome
Left-sided HF
This causes fluid to accumulate in the lungs, which causes shortness of breath in people.
Right-sided HF
Fluid builds up in the abdomen, legs, and feet, which causes swelling in the abdomen and legs.
Systolic HF
This occurs when the heart muscle cannot contract strongly as it should; as a result, blood is not pumped in the normal way.
Diastolic HF
There is insufficient blood filling the area due to the inability of the heart to relax sufficiently.
Some of the prominent causes of HF are:
CAD and heart attack
: CAD is the most prominent heart disease and the most common cause of HF. The disease occurs when blood circulation is not normal due to fat deposits, called plaques, in the blood vessels. Heart attacks are divided into mild heart attack and major heart attack. Mild heart attacks occur in a small portion of the heart muscle. However, a mild heart attack should be taken seriously because it can cause a second attack. Major heart attack affects a large part of the heart muscle. The blood flow to the heart is completely blocked or there is no blood flow for a long time. Major heart attack can cause cardiac arrest and death.
High blood pressure (hypertension)
: High blood pressure is one of the factors that makes it difficult for the heart to pump blood to the body. This strain causes the heart muscle to work more than normal, damaging it.
Faulty heart valves
: The function of the heart valves is to keep the blood flowing in the right direction. Failure of heart valves to function is known as heart valve malfunction. Heart valve malfunctions can be caused by genetic factors, old age, and infection. A faulty valve tires the heart, forcing it to work harder and causing it to weaken over time.
Damage to the heart muscle (cardiomyopathy)
: Cardiomyopathy can have many causes, such as various diseases, infections, alcohol use, and the toxic effect of drugs like cocaine or chemotherapy drugs. Sometimes it can be due to genetic factors.
Heart muscle inflammation
: This is caused by viruses and can lead to left-sided HF.
Congenital heart defects
: If the heart and chambers or valves aren’t formed correctly, the healthy parts of the heart have to work harder to pump blood, which can lead to HF.
Heart arrhythmias
: Abnormal heart rhythms will show heart-damaging output in both fast heartbeat and slow heartbeat. Its rapid beating will tire the heart, and its slow beating will cause problems in circulation.
Other diseases
: Diabetes, HIV, hyperthyroid and similar diseases can also cause HF.
Heart diseases and the factors that affect the HF that occur as a result may be genetic, sometimes congenital abnormalities, and sometimes related to our lifestyle. Overweight, having a sedentary life, smoking, and stressful work life are also harmful factors for the heart [1].
1.2.3 What are the Symptoms of HF?
HF may be faced if one or more of the following symptoms are present. For example, if someone is tired earlier than before even though they are doing the same activity, this might be a symptom. This symptom should not be confused with early fatigue due to old age – a 55-year-old person should compare and interpret their symptoms according to their current age, not to what it was at the age of 45 or younger.
Another symptom of HF is loss of appetite, which can be related to other ailments. Therefore, the loss of appetite that occurs without any external cause should be evaluated. Another symptom associated with loss of appetite may be nausea, which can also occur with extreme sensitivity to some foods and certain odors. Due to HF, blood accumulates in the abdomen and presses on the intestinal wall and stomach. This pressure causes loss of appetite and nausea.
HF can also cause shortness of breath, but is not, of course, the only cause of this; severe shortness of breath is possible in the case of advanced bronchitis and asthma, which includes narrowing of the airways and chest, and a wheezing sound from the lungs. Shortness of breath caused by HF occurs due to pulmonary edema. It is accompanied by fatigue and weakness, caused by heart-induced constriction; shortness of breath is caused by bronchitis or asthma and, similarly, often by overweight and obesity. Shortness of breath experienced without an external cause should be taken seriously.
Everyone has a certain gradual transition in weight gain and weight loss. Under normal conditions, this situation does not develop suddenly, so sudden weight gain might be associated with HF. The relationship between HF and weight gain is related to the weakness of some organs in performing their duties due to the inability of the heart. When the calories taken into the body through food are more than the calories consumed by the body, this excess is stored as fat. Insulin resistance develops due to excess fat in the body and one of the diseases caused by insulin resistance is HF. The measurement of excess fat is carried out with the help of the body weight index (BMI). Those with a BMI of 25–29.9 are considered overweight and those with a BMI over 30 are considered obese. In addition, these values change depending on age. While the normal BMI values for 25–35-year-olds are 20–25, this BMI value is 24–29 for those over 65 years old.
One of the most important indicators of possible heart disease is chest pain. Sometimes people may panic due to the sensitivity of the subject. Some problems in the digestive system, such as bloating in the stomach, can cause chest pain and tightness in the chest area. However, it is best not to leave it to chance.
Coughing and related sputum production of pink or white color may also be a harbinger of HF. As will be appreciated, heart disease is not the only cause of a cough, but it is one of its causes.
Interrelated heart palpitations and pulse irregularities are also associated with HF. In addition, swelling of the ankles, legs, and abdomen due to edema can occur as a result of right-sided heart problems. Seeing one or more of these problems can also be viewed as a symptom.
People often find themselves in a state of indolence when they are low in morale. They want to sleep all the time and cannot easily concentrate on any task. The condition that manifests itself as in the body may sometimes indicate a heart problem.
Two other problems with HF are frequent urination at night and the prominence of blood vessels in the neck. The accumulated blood in the abdominal region of HF patients who go into a supine position enters the circulation. As a result, the kidneys detect excess volume and this increases urination. HF can be suspected if blood sugar is normal but there is frequent urination at night.
When at least one of these symptoms is noticed, often by the person themself, the control phase should start and the doctor should be consulted for diagnosis without delay. Even if going to the doctor does not solve the HF, it will contribute to one’s survival time and quality of life [2].
1.2.4 HF Treatment
Treatment in HF generally gives effective results, the purpose being a decrease in sudden death and an increase in survival time. Within the scope of treatment are acting to support a healthy life, medication use, medical support, and surgical interventions.
In drug treatment, the patient is generally supported with more than one drug supplement. Medicines are used to strengthen the heart muscle, lower blood pressure, and remove the edema accumulated in the body. Some medicines used in the treatment of HF show blood thinning and cholesterol lowering effects. Many drugs are difficult for the patient, but their use can produce the desired effect.
Surgical methods and medical support devices are other method used in treatment. Surgical methods and medical support aim to eliminate the irregularities that occur in the heart. One of the surgical methods is coronary bypass and another is heart valve replacement surgery. The most known and preferred method in the treatment of HF is undoubtedly heart surgery. Deciding this is itself an important issue. If the heart has become unable to perform its duties, and if medication or a stent are far from being a solution, a heart surgery decision should be made with a physician's recommendation.
The use of pacemakers can also be considered within the scope of surgical methods. The function of pacemakers is to support the heart and, with a solution similar to a pacemaker, the disturbed rhythm of the heart is understood and shock is applied to the heart to eliminate the rhythm disturbance. In addition, placing a heart pump in the chest cavity to support the weakened heart is a method of treatment. One of the most effective and definitive solutions within surgical methods is heart transplantation. The broken heart is removed and a healthy heart is transplanted.
One of the techniques that reduce the risk of heart attack and thus helps treatment is blood thinning. Thinning the blood is a known technique for preventing blood clotting, a significant problem that causes major diseases such as stroke, heart attack, and kidney failure. Pills, subcutaneous injections, or tubes placed in a vein are used to thin the blood. The pill is warfarin; the other methods use heparin.
1.3 Invasive and Non-Invasive Procedures
A detailed clinical examination and patient history are important in diagnosing HF. In a detailed examination, symptoms of high blood pressure, high cholesterol, and high sugar levels should also be examined. In the diagnosis of HF, it should be considered whether the patient has had one of the diseases in the HF class previously. Physical examination is also one of the useful applications in terms of diagnosis, as it can reveal abnormalities in the patient that disrupt the balance of the body.
There are numerous invasive, non-invasive, machine learning, and artificial intelligence based methods for diagnosing HF. While invasive methods roughly refer to the types of studies that can be seen as subcutaneous interventions, non-invasive methods refer to procedures performed outside the body, for example measuring blood pressure.
1.3.1 What are Invasive and Non-Invasive Procedures?
The invasive procedure – also called invasion – involves opening veins that are usually closed, stenting, needle procedures, and the like. Invasive procedures are generally performed by applying a physical effect to the body. For example, operations that cause an incision in the body, such as open surgery, are this type.
Apart from invasive procedures, there are also non-invasive and minimally invasive methods. As a medical term, the invasive procedure is used for situations requiring surgical intervention, while non-invasive procedure is the name given to situations that do not require surgical procedures and do not need any intervention such as cutting or piercing the body. Measurement of a patient’s blood pressure and temperature are just two examples of the non-invasive method.
Minimally invasive procedure refers to invasive studies performed in a very narrow area, in a very short time, and with little effect. For example, endoscopic interventions are known as minimally invasive methods and while taking blood from a vein for analysis is invasive, taking blood from a small incision in the tip of the finger is known to be minimally invasive.
1.3.2 Diagnosis by Invasive Procedures
There are many invasive procedures in use, including surgical interventions. Tens of millions of invasive procedures are performed on patients worldwide almost every year [3].
One of the invasive tests for heart conditions is blood tests. Within the scope of blood tests, cardiac enzymes, C-reactive protein (CRP), fibrinogen, homocysteine, lipoproteins, triglycerides, brain natriuretic peptide (BNP), and prothrombin values are checked. Blood tests can confirm whether a heart attack has occurred, the extent of damage can be predicted, future risk can be found, and general information on blood clotting can be obtained [4].
One of the tests that should be considered within the scope of blood tests is the troponin test, which is a method that can determine the risk of heart attack in healthy individuals. What is tested is the amount of troponin protein that is released when the heart muscle is damaged. Based on this measurement, the cases likely to have a heart attack in the coming years can be predicted.
Another invasive technique is known as heart catheterization. The inside of the heart and blood vessels are imaged using special x-rays called angiograms. A thin hollow tube, a catheter, is guided through a blood vessel in the arm, groin, or neck toward the heart. Dye is injected into the blood vessels through the catheter, thus enhancing the image quality of the x-ray. This process takes 2–3 hours on average and, through angiography, can lead to an understanding of what caused the chest pain – angiography can reveal both the obstruction in the vessels and the problem in the heart valves and, in addition, can measure the blood pressure and oxygen content within the heart [4].
Since angiography has some limitations, non-invasive methods such as smart calculation algorithms are also frequently used and better results can sometimes be obtained.
1.3.3 Diagnosis by Non-Invasive Procedures
Pressure examination, hammer examination, ultrasonography are some common examples of non-invasive methods, and there are many non-invasive methods used in HF. Each method will be briefly mentioned in this section.
Electrocardiography is one of the non-invasive methods in HF. It monitors heart rhythm changes by recording the heart's electrical activity and, thus, determines whether a heart attack has occurred from abnormalities in rhythm. Electrocardiography and blood tests are the most common methods used in HF diagnosis and the aim of electrocardiography is to determine the complex factors underlying HF [5].
Sometimes, the heart’s activity is recorded throughout the day, not just for a short time. This process is known as Holter monitoring. The method aims to determine how the heart responds during daily activities and the average heart rhythm. Holter monitoring not only detects abnormalities in the heart, but can also help doctors treat them.
Chest x-ray also provides a non-invasive method for diagnosing heart diseases, by taking a picture of the heart, lungs, and rib cage. With the radiological imaging performed, it can be seen if the heart is enlarged after a heart attack or fluid accumulation in the lungs. Systolic and diastolic dysfunction of the heart can be discovered through electrocardiography and chest x-ray [6].
Echocardiography has been used for many years to characterize HF [7]. It uses high frequency sound waves (ultrasound) to produce images of heart size, movement, and structure. The printout of the test gives information about heart health and helps to gather information about heart arrhythmias. Echocardiography outputs have been used with both machine learning algorithms and deep learning algorithms [8].
Computed imaging (tomography) is used to detect other diseases as well as in HF. It creates three-dimensional (3D) images that can show the obstructions in our main vessels with the help of a computer system. Through computed tomography, testing can be performed, especially for aortic disease.
The exercise test works with skin-attached electrodes and a monitor recording a person’s heart function while walking on a treadmill. Many aspects of heart function can be checked, including heart rate, breathing, blood pressure, ECG, and how tired someone is while exercising. This test is particularly helpful in diagnosing CAD and is also used to predict heart attack risk. It helps diagnose the possible cause of symptoms such as chest pain (angina).
The last procedure on this subject is known as the thallium stress test. The thallium test is similar to a routine exercise stress test. The difference is that the radioactive thallium material injected into the patient’s blood is photographed by special gamma ray cameras when the patient is at the maximum exercise level. Thus, the extent of coronary artery occlusion can be determined and the extent of damage from heart attack can be determined [9].
1.4 Computer-Aided Diagnosis
The number of diseases and patients is increasing daily due to food-related problems, a sedentary lifestyle, and many similar reasons. Factors such as the complexity of the diseases, the increasing number of patients, and the aging of the world population have led to new searches in the field of health. Diseases are not old diseases, and their solution methods cannot be old methods. Being aware of this situation, health professionals and technology workers have joined hands and developed solutions for the health sector. In this context, many studies using technology in the field of health have been carried out [10]. Notably, using internet of things (IoT) sensor technologies, human movements have been monitored and abnormal situations detected [11].
The first solution that comes to mind regarding data collected with the help of IoT sensors is alarm generation resulting from abnormalities in the collected data, and sometimes making diagnoses and patient interventions the patient [12]. Abnormalities in patient movements or blood values are usually associated with a disease. For example, irregular tissue growths are related to cancer, and irregular heartbeats, heart diseases, and balance problems in the body are related to nerve diseases. IoT is a network where physical objects communicate with each other and with central systems. The use of IoT devices in disease diagnosis is carried out with the help of wearable technologies or additional components that make it possible to generate alarms about the patient with the help of systems that collect information, especially heart rate information, from the patient and transmit it to the center. This is carried out according to a method called anomaly detection in the name of alarm generation. An alarm is generated in case of an abnormality by comparing the last state of the patient with their recorded normal state. Thus, an approach similar to anomaly detection in networks provides the background for diagnosis based on IoT systems.
HF is a complex disease that occurs due to many different causes. The complexity of the disease makes its correct diagnosis a challenging problem for doctors. In addition, heart diseases, and especially heart attacks, are diseases that frequently affect people over the age of 65 who choose to live at home rather than a health center, and thus it is not possible to keep the patient’s condition under control. Due to the increase in cardiac diseases and the inability of patients to be under continuous control, studies to detect abnormalities in the heart before the onset of a crisis should be carried out earlier [13]. Early diagnosis is important as it affects the early intervention for the patient.
In the diagnosis of HF, the risk of HF is estimated based primarily on the patient’s clinical history, current symptoms, physical examination, and ECG. If all the measurements mentioned are normal, there is probably no risk, otherwise more detailed analysis should be done for diagnosis. Performing analyses allows specialists to identify patients who need echocardiography [14]. Diagnosis can be made with clinical methods, and decision support systems based on machine learning algorithms. From a machine learning perspective, HF diagnosis is a binary classification problem, a type of classification in which samples are assigned to one of two groups. For the diagnosis of HF, there would be two classes: “heart failure” and “no heart failure.” Thanks to machine learning, it can be understood whether the person being controlled has HF.
1.4.1 Clinical Decision Support Systems
There is a need to develop a system in order to collect patient data and make diagnoses based on them. These systems are generally called decision support systems. Decision support systems work by analyzing the data collected from the terminal units in accordance with the needs in the center to support decision making and presenting the results. Decision support systems can sometimes be developed for the field of health and sometimes for another field. Regardless of the field, the task of decision support systems is to infer with the help of the collected data and the rules at hand.
Decision support systems developed specifically for the health field are often referred to as clinical decision support systems. Clinical decision support systems sometimes help experts diagnose disease and recommend treatment methods based on health information. The concept of computer-aided clinical decision making came to the fore many years ago [15]. The working principle of clinical decision support systems is simply as follows: information such as blood values and enzyme values of the patient is obtained, this information is evaluated according to the knowledge base of the health institution, and an inference is made about the patient. When a direct relationship can be established between the inputs and the final decision during inference, this is known as an open box decision support system; when the correct relationship cannot be established between the inputs and the final decision, it is called a closed box decision support system [16].
A decision support system is a software that helps decision making and works based on the ruleset. The rule set can be in the form of simple IF-THEN rules or a more complex structure. Rule sets can be created with the help of expert knowledge or with models that learn from data. Comparing the event that occurred with the rules in the rule set determines which rule triggered the situation. Thus, whichever rule is triggered by the situation to produce the result, the output of the decision support system will be consistent.
Decision support systems are systems that help to decide about a situation. A system that will decide on HF takes the patient’s measurements as input and provides information about the risk of HF as output. The system’s inputs are sometimes the information coming from the IoT devices and sometimes the data that has been entered manually. For example, based on heart rate and blood values, it will be possible to predict whether there will be HF. At this point, we come across systems based on machine learning and knowledge. In machine learning based systems, data classification is carried out with the help of models trained with historical data. As a result of the classification, there might or might not be an indicated risk of HF.
1.5 Diagnosis with Artificial Intelligence Methods
One of the techniques that can be used to diagnose of HF is artificial intelligence and machine learning.
1.5.1 Introduction to Artificial Intelligence
Artificial intelligence is the name of the general discipline that combines software and hardware studies. It includes designing systems that act like a human, learning and making decisions like humans. Scientists have been curious about artificial intelligence since the 1950s and there is increasing interest in it. It has become an even more important discipline with the developments in information technologies. The focus on developing systems that exhibit human-like behavior has already attained this goal in some areas. Examining the recent history of artificial intelligence shows that machine learning algorithms and then deep learning algorithms have led to striking developments. For example, thanks to natural language processing, systems that produce and understand language have been developed; with the help of image processing techniques, systems close to human vision have been developed; systems that imitate human learning, systems that extract information, systems that establish connections have been developed; and studies that reveal abnormal situations have been conducted.
The main reason artificial intelligence-powered solutions have come to the fore is their ability to exploit many advantages of information technologies. Cheaper storage units, the increase in the capacities of the processing units, and advances in artificial intelligence algorithms make the field more popular every day. The cheaper storage units and the increase in their capacity have enabled more patient data to be stored. The increase in capacity and speed in processor units makes it possible to analyze data in a way not possible in the past. In fact, data collection and transfer possibilities from remote units have increased due to the developments in network technologies. All developments favor artificial intelligence and data science, and many jobs have become fulfilled with the help of artificial intelligence and data science.
The developments have also contributed to artificial intelligence programming skills. With artificial intelligence, it has become possible to solve many problems that could not be solved with classical programming skills, and while complex data relationships cannot be solved with classical methods, recently relationships between data have become easily inferred.
One of the areas most supported by artificial intelligence is computer-aided decision making, thanks to its various capabilities, notably diagnosis and prediction. Artificial intelligence-based diagnostic systems can be seen as an example of a non-invasive procedure, because there is no interference with the body in artificial intelligence supported clinical decision support systems, in which expert systems make decisions based on both expert opinions and machine learning systems’ modeling from past case examples. These systems, which are sometimes used separately, are used together in some places. With the increase in the studies on artificial intelligence, its subfields have emerged. There are many artificial intelligence subdomains, each with different characteristics. Among these subareas, machine learning, deep learning, expert systems, and image processing in particular provide auxiliary features in computer-aided decision making.
1.5.1.1 Expert Systems
Expert systems are based on imitating human expertise with machines [17]. They are considered among the first successful applications of artificial intelligence. An expert system is often used in decision support systems development. In addition, the expert system has two components: the knowledge base and the inference engine.
