Machine Learning in Medicine covers the state-of-the-art techniques of machine learning and their applications in the medical field. It presents several computer-aided diagnosis (CAD) systems, which have played an important role in the diagnosis of several diseases in the past decade, e.g., cancer detection, resulting in the development of several successful systems.
New developments in machine learning may make it possible in the near future to develop machines that are capable of completely performing tasks that currently cannot be completed without human aid, especially in the medical field. This book covers such machines, including convolutional neural networks (CNNs) with different activation functions for small- to medium-size biomedical datasets, detection of abnormal activities stemming from cognitive decline, thermal dose modelling for thermal ablative cancer treatments, dermatological machine learning clinical decision support systems, artificial intelligence-powered ultrasound for diagnosis, practical challenges with possible solutions for machine learning in medical imaging, epilepsy diagnosis from structural MRI, Alzheimer's disease diagnosis, classification of left ventricular hypertrophy, and intelligent medical language understanding.
This book will help to advance scientific research within the broad field of machine learning in the medical field. It focuses on major trends and challenges in this area and presents work aimed at identifying new techniques and their use in biomedical analysis, including extensive references at the end of each chapter.
Author(s): Ayman El-Baz, Jasjit S. Suri
Series: Chapman & Hall/CRC Healthcare Informatics Series
Publisher: CRC Press
Year: 2021
Language: English
City: Boca Raton
Cover
Half Title
Series Page
Title Page
Copyright Page
Dedication
Table of Contents
Preface
Acknowledgements
Editors
Contributors
Chapter 1 Another Set of Eyes in Anesthesiology
1.1 Introduction/History
1.2 Machine Learning Implementation in Anesthesia
1.2.1 Hypotension
1.2.2 Hypoxia
1.2.3 Depth of Anesthesia
1.2.4 Mortality Prediction
1.3 Limitations
1.4 Conclusion
References
Chapter 2 Dermatological Machine Learning Clinical Decision Support System
2.1 Background
2.2 Melanoma
2.3 Non-Melanoma Machine Learning
2.4 Non-Skin Image Recognition Machine Learning
2.5 Data Augmentation
2.6 Limitations
2.7 Future Directions
2.8 Conclusion
References
Chapter 3 Vision and AI
3.1 Introduction
3.2 Diabetic Retinopathy
3.3 Age-Related Macular Degeneration
3.4 Glaucoma
3.5 Multi-Classifier and Improving Performance
3.6 Machine Learning for Predicting Survival/Prognosis
3.7 Conclusion
References
Chapter 4 Thermal Dose Modeling for Thermal Ablative Cancer Treatments by Cellular Neural Networks
4.1 Introduction
4.2 Soft Tissue Thermal Dose Modeling
4.3 Cellular Neural Networks
4.3.1 Architecture
4.3.2 CNN Formulation of Bio-Heat Transfer on 2D Grids
4.3.2.1 2D Regular Grids
4.3.2.2 2D Irregular Grids
4.3.3 CNN Formulation of Bio-Heat Transfer on 3D Grids
4.3.3.1 3D Regular Grids
4.3.3.2 3D Irregular Grids
4.3.4 Initial and Boundary Conditions
4.4 Numerical Examples
4.4.1 Patient-Specific Thermal Dose Computation
4.4.2 Application to Hepatic Tumour Ablation Using HIFU
4.5 Discussions
4.6 Conclusion
Acknowledgement
References
Chapter 5 Ensembles of Convolutional Neural Networks with Different Activation Functions for Small to Medium-Sized Biomedical Datasets
5.1 Introduction
5.2 Activation Functions for the Three CNNs
5.2.1 ReLu
5.2.2 Leaky ReLU
5.2.3 ELU
5.2.4 SELU
5.2.5 PReLU
5.2.6 SReLU
5.2.7 APLU
5.2.8 MeLU
5.2.9 GaLU
5.3 Experimental Results
5.4 Conclusion
Acknowledgment
Bibliography
Chapter 6 Analysis of Structural MRI Data for Epilepsy Diagnosis Using Machine Learning Techniques
6.1 Introduction
6.2 Lateralization Indices/Features Extracted from Structural MRI
6.2.1 Study Population
6.2.2 MRI Acquisition
6.2.3 Image Processing
6.2.4 Feature Analysis and Classification
6.2.5 Summary of Machine Learning for mTLE Lateralization
6.3 Results
6.3.1 Study I: Hippocampal Study – FLAIR Intensity, Texture Analysis, and Volumetry
6.3.2 Study II: Hippocampal and Amygdalar Study – FLAIR Intensity and Volumetry
6.3.3 Study III: DTI-Based Response-Driven Modeling
6.3.4 Study IV: Texture Features Extracted from Whole Brain MRI
6.3.5 Study V: Multi-Structural Vomuletric Features
6.4 Discussion
6.5 Conclusion
References
Chapter 7 Artificial Intelligence-Powered Ultrasound for Diagnosis and Improving Clinical Workflow
7.1 Introduction
7.1.1 Ultrasound Imaging
7.1.2 Deep Learning
7.1.3 AI-Powered US in Radiology
7.1.4 AI-Powered US in Cardiology
7.2 Discussion and Outlook
References
Chapter 8 Machine Learning for E/MEG-Based Identification of Alzheimer’s Disease
8.1 Introduction
8.2 Sensor-Level Analysis
8.3 Source-Level Analysis
8.4 Discussion and Conclusion
References
Chapter 9 Some Practical Challenges with Possible Solutions for Machine Learning in Medical Imaging
9.1 Introduction
9.2 Utilizing Transfer Learning with Natural Images
9.3 Focal Tversky Loss Function for Class Imbalance
9.4 Bayesian Neural Networks for Confidence Tuning
9.5 Conclusion
Notes
References
Chapter 10 Detection of Abnormal Activities Stemming from Cognitive Decline Using Deep Learning
10.1 Introduction
10.2 Indicators of Dementia
10.3 Literature Overview
10.3.1 Cognitive Status Assessment
10.3.2 Deep Learning
10.4 Datasets and Data Generation
10.4.1 Datasets
10.4.2 Data Generation
10.4.2.1 Activity-Related Abnormal Behavior
10.4.2.2 Sub-Activity–Related Abnormal Behavior
10.5 Spatio-Temporal Activity Recognition and Abnormal Behavior Detection
10.5.1 Sensor Data Representation
10.5.2 Abnormal Behavior Detection
10.5.3 Experiments
10.5.4 Results with Recurrent Neural Networks
10.5.5 Results with Convolutional Neural Networks
10.6 Conclusions and Future Work
10.6.1 Conclusion
10.6.2 Future Work
Bibliography
Chapter 11 Classification of Left Ventricular Hypertrophy and NAFLD through Decision Tree Algorithm
11.1 Introduction
11.2 Literature Review
11.2.1 Left Ventricular Hypertrophy (LVH)
12.2.2 Non-Alcoholic Fatty Liver Disease
11.3 Methodology
11.4 Experimental Results and Discussions
11.4.1 Data Acquisition
11.4.2 Decision Tree-Based Model Development for LVH Classification
11.4.3 Decision Tree for NAFLD Classification
11.5 Conclusion and Future Works
References
Chapter 12 The Cutting Edge of Surgical Practice: Applications of Machine Learning to Neurosurgery
12.1 Introduction
12.2 Machine Learning and Neurosurgical Diagnostics
12.3 Machine Learning and Intraoperative Care
12.4 Machine Learning and Neurosurgical Prognosis
12.5 Future Directions
12.6 Limitations
References
Chapter 13 A Novel MRA-Based Framework for the Detection of Cerebrovascular Changes and Correlation to Blood Pressure
13.1 Introduction
13.2 Methods
13.2.1 Patient Demographics
13.2.2 Data Analysis
13.2.3 Statistical Analysis
13.2.4 3D Reconstruction of the Cerebral Vasculature
13.3 Results
13.4 Discussion
13.5 Limitations
13.6 Conclusion
References
Chapter 14 Early Classification of Renal Rejection Types: A Deep Learning Approach
14.1 Introduction
14.2 Methods
14.2.1 Preprocessing/Segmenting Kidneys
14.2.2 Diffusion Features Extraction
14.2.3 Deep Learning Classification
14.3 Experimental Results
14.4 Conclusion
Acknowledgment
References
INDEX
