This book provides a detailed technical overview of the use and applications of artificial intelligence (AI), machine learning and big data in cardiology. Recent technological advancements in these fields mean that there is significant gain to be had in applying these methodologies into day-to-day clinical practice. Chapters feature detailed technical reviews and highlight key current challenges and limitations, along with the available techniques to address them for each topic covered. Sample data sets are also included to provide hands-on tutorials for readers using Python-based Jupyter notebooks, and are based upon real-world examples to ensure the reader can develop their confidence in applying these techniques to solve everyday clinical problems.Artificial Intelligence and Big Data in Cardiology systematically describes and technically reviews the latest applications of AI and big data within cardiology. It is ideal for use by the trainee and practicing cardiologist and informatician seeking an up-to-date resource on the topic with which to aid them in developing a thorough understanding of both basic concepts and recent advances in the field.
Author(s): Nicolas Duchateau, Andrew P. King
Publisher: Springer
Year: 2023
Language: English
Pages: 219
City: Cham
Preface
Contents
1 Introduction
Andrew King and Nicolas Duchateau
1.1 Aims and Motivation
1.2 What are AI and Machine Learning?
1.3 A Brief History
1.4 AI in Medicine
1.5 The Role of Big Data
1.6 Outlook
References
2 AI and Machine Learning: The Basics
Nicolas Duchateau, Esther Puyol-Antón, Bram Ruijsink and Andrew King
2.1 Introduction
2.2 Defining the Problem
2.3 Types of Model
2.4 Model Design
2.5 Model Validation
2.6 Machine Learning Is Not a Panacea!
2.7 Sources of Data for Machine Learning in Cardiology
2.8 Imaging Sources
Echocardiography
2.9 Closing Remarks
2.10 Exercises
2.11 Tutorial—Introduction to Python and Jupyter Notebooks
References
3 From Machine Learning to Deep Learning
Pierre-Marc Jodoin, Nicolas Duchateau and Christian Desrosiers
3.1 Introduction
3.2 Machine Learning and Neural Networks
3.3 K-Class Prediction
3.4 Handling Non-linearly Separable Data
3.5 Convolutional Neural Networks
3.6 Closing Remarks
3.7 Exercises
3.8 Tutorial—Classification From Linear to Non-linear Models
References
4 Measurement and Quantification
Olivier Bernard, Bram Ruijsink, Thomas Grenier and Mathieu De Craene
4.1 Clinical Introduction
4.2 Overview
4.3 AI Models for Cardiac Quantification
4.4 Quantification of Cardiac Function From CMR and Echocardiography
4.5 Quantification of Calcium Scoring From CT Imaging
4.6 Quantification of Coronary Occlusion From SPECT
4.7 Leveraging Clinical Reports as a Base of Annotations
4.8 Closing Remarks
4.9 Exercises
4.10 Tutorial—Cardiac MR Image Segmentation With Deep Learning
4.11 Opinion
References
5 Diagnosis
Daniel Rueckert, Moritz Knolle, Nicolas Duchateau, Reza Razavi and Georgios Kaissis
5.1 Clinical Introduction
5.2 Overview
5.3 Classical Machine Learning Pipeline for Diagnosis
5.4 Deep Learning Approaches for Diagnosis
5.5 Machine Learning Applications for Diagnosis
5.6 Machine Learning Approaches Based on Radiomics
5.7 Machine Learning Approaches for Large-Scale Population Studies
5.8 Challenges
5.9 Closing Remarks
5.10 Exercises
5.11 Tutorial—Two-Class and Multi-class Diagnosis
5.12 Opinion
References
6 Outcome Prediction
Buntheng Ly, Mihaela Pop, Hubert Cochet, Nicolas Duchateau, Declan O'Regan and Maxime Sermesant
6.1 Clinical Introduction
6.2 Overview
6.3 Current Clinical Methods to Predict Outcome
6.4 AI-Based Methods to Predict Outcome
6.5 Application: Prediction of Response Following Cardiac Resynchronization Therapy (CRT)
6.6 Application: AI Methods to Predict Atrial Fibrillation Outcome
6.7 Application: Risk Stratification in Ventricular Arrhythmia
6.8 Closing Remarks
6.9 Exercises
6.10 Tutorial—Outcome Prediction
6.11 Opinion
References
7 Quality Control
Ilkay Oksuz, Alain Lalande and Esther Puyol-Antón
7.1 Clinical Introduction
7.2 Overview
7.3 Motion Artefact Detection
7.4 Poor Planning Detection and Automatic View Planning
7.5 Missing Slice Detection
7.6 Segmentation Failure Detection
7.7 Closing Remarks
7.8 Exercises
7.9 Tutorial—Quality Control
7.10 Opinion
References
8 AI and Decision Support
Mariana Nogueira and Bart Bijnens
8.1 Introduction
8.2 What Does AI Bring to the Table to Support the Clinician?
8.3 Current Challenges and the Importance of Interpretability
8.4 Addressing Challenges With Interpretable AI—The Potential of Representation Learning
8.5 Closing Remarks
References
9 AI in the Real World
Alistair A. Young, Steffen E. Petersen and Pablo Lamata
9.1 Introduction
9.2 Asking the Right Question
9.3 Provenance of Data
9.4 Structural Risk
9.5 Shallow Learning
9.6 Does My Model Look Good in This?
9.7 Mechanistic Models for AI Interpretability
9.8 Utility of Community-Led Challenges
9.9 Closing Remarks
References
10 Analysis of Non-imaging Data
Nicolas Duchateau, Oscar Camara, Rafael Sebastian and Andrew King
10.1 Introduction
10.2 Electrophysiology
10.3 ECG Analysis
10.4 Electronic Health Records
10.5 Closing Remarks
References
11 Conclusions
Andrew King and Nicolas Duchateau
Supplementary Information
Solution of the Exercises
Chapter 2
Exercise 1
Exercise 2
Exercise 3
Exercise 4
Exercise 5
Exercise 6
Chapter 3
Exercise 1
Exercise 2
Exercise 3
Exercise 4
Exercise 5
Chapter 4
Exercise 1
Exercise 2
Exercise 3
Exercise 4
Exercise 5
Exercise 6
Chapter 5
Exercise 1
Exercise 2
Exercise 3
Exercise 4
Exercise 5
Exercise 6
Exercise 7
Chapter 6
Exercise 1
Exercise 2
Exercise 3
Exercise 4
Chapter 7
Exercise 1
Exercise 2
Exercise 3
Exercise 4
Exercise 5
Index
