Multimodal Learning for Clinical Decision Support: 11th International Workshop, ML-CDS 2021, Held in Conjunction with MICCAI 2021, Strasbourg, France, ... (Lecture Notes in Computer Science)

This book constitutes the refereed joint proceedings of the 11th International Workshop on Multimodal Learning for Clinical Decision Support, ML-CDS 2021, held in conjunction with the 24th International Conference on Medical Imaging and Computer-Assisted Intervention, MICCAI 2021, in Strasbourg, France, in October 2021. The workshop was held virtually due to the COVID-19 pandemic. The 10 full papers presented at ML-CDS 2021 were carefully reviewed and selected from numerous submissions. The ML-CDS papers discuss machine learning on multimodal data sets for clinical decision support and treatment planning. Preface 6 Organization 7 Contents 8 From Picoscale Pathology to Decascale Disease: Image Registration with a Scattering Transform and Varifolds for Manipulating Multiscale Data 10 1 Introduction 10 2 Methods 11 2.1 Image Acquisition and Processing 11 2.2 Algorithm for Multimodal Registration with Damaged Tissue 13 2.3 Scattering Transform for Retaining High Resolution Texture 14 2.4 Varifold Measures for Modeling and Crossing Multiple Scales 14 3 Results 15 4 Discussion 16 References 18 Multi-scale Hybrid Transformer Networks: Application to Prostate Disease Classification 21 1 Introduction 21 1.1 Contribution 22 2 Methods 22 2.1 Model Comparison 25 3 Experiments and Results 25 3.1 Dataset 25 3.2 Pre-processing and Augmentation 26 3.3 Training 26 3.4 Results 27 4 Conclusion 28 References 29 Predicting Treatment Response in Prostate Cancer Patients Based on Multimodal PET/CT for Clinical Decision Support 31 1 Introduction 32 2 Methods 35 2.1 Pipeline Overview 35 2.2 Dataset and Ground Truth Annotation 35 2.3 Automated Segmentation 37 2.4 Therapy Response Prediction 38 3 Results 38 3.1 Segmentation 38 3.2 Therapy Response Prediction 40 4 Discussion 41 5 Conclusion 42 References 42 A Federated Multigraph Integration Approach for Connectional Brain Template Learning 45 1 Introduction 45 2 Proposed Method 47 3 Results and Discussion 51 4 Conclusion 54 References 55 SAMA: Spatially-Aware Multimodal Network with Attention For Early Lung Cancer Diagnosis 57 1 Introduction 57 2 Method 59 2.1 SAMA Module 60 3 Experimental Setup 61 3.1 Dataset 61 3.2 Implementation Details 62 4 Results 62 4.1 Control Experiments 62 5 Conclusions 65 References 65 Fully Automatic Head and Neck Cancer Prognosis Prediction in PET/CT 68 1 Introduction 69 2 Methods 70 2.1 Datasets 70 2.2 UNet Architecture and Training 71 2.3 Radiomics Workflow 71 3 Results 73 3.1 GTVt Segmentation 73 3.2 Prognosis Prediction 73 4 Discussion and Conclusions 74 References 75 Feature Selection for Privileged Modalities in Disease Classification 78 1 Introduction 78 2 Background 80 2.1 Learning Using Privileged Information 80 2.2 Mutual Information Feature Selection 81 3 Method 81 4 Experiments 84 4.1 Compared LUPI Models 84 4.2 Datasets 85 5 Results 86 5.1 Parkinson's Dataset 86 5.2 TMJ Osteoarthritis Dataset 86 6 Conclusions 88 References 88 Merging and Annotating Teeth and Roots from Automated Segmentation of Multimodal Images 90 1 Introduction 91 2 Materials 91 3 Related Work 92 3.1 Root Canal Segmentation Algorithm 92 3.2 3D Shape Analysis for Segmentation and Classification 92 4 Methods 93 4.1 Root Canal Segmentation Algorithm 93 4.2 Dental Model Segmentation Algorithm 94 4.3 Universal Labeling and Merging Algorithm 95 5 Results 96 5.1 Root Canal Segmentation 96 5.2 Universal Labeling and Merging Algorithm 97 6 Conclusion 98 References 98 Structure and Feature Based Graph U-Net for Early Alzheimer’s Disease Prediction 102 1 Introduction 102 2 Methods 104 2.1 Dynamic FCN 104 2.2 Graph Construction 105 2.3 SFG U-Net 106 3 Experiments and Results 108 3.1 Dataset 108 3.2 Classification Performance of Our Model 109 3.3 The Effect of Our Pool Layer 109 3.4 Most Discriminative Regions 110 4 Conclusion 112 References 112 A Method for Predicting Alzheimer’s Disease Based on the Fusion of Single Nucleotide Polymorphisms and Magnetic Resonance Feature Extraction 114 1 Introduction 114 2 Material and Methodology 116 2.1 Data Acquisition and Preprocessing 116 2.2 Feature Extraction 117 3 Experimental Results 118 3.1 Model Performance and Method Comparison 118 3.2 Abnormal Brain Regions and Pathogenic Genes 119 4 Discussion 123 4.1 Comparison with Existing Studies 123 4.2 Analysis of Results 123 5 Conclusion 123 References 123 Author Index 125