Mathematical and Computational Oncology: Second International Symposium, ISMCO 2020, San Diego, CA, USA, October 8–10, 2020, Proceedings (Lecture Notes in Computer Science, 12508)

Description

This book constitutes the refereed proceedings of the Second International Symposium on Mathematical and Computational Oncology, ISMCO 2020, which was supposed to be held in San Diego, CA, USA, in October 2020, but was instead held virtually due to the COVID-19 pandemic. The 6 full papers and 4 short papers presented together with 1 invited talk were carefully reviewed and selected from 28 submissions. The papers are organized in topical sections named: statistical and machine learning methods for cancer research; mathematical modeling for cancer research; general cancer computational biology; and posters. Preface Organization Abstracts of Keynote Talks Fighting Drug Resistance with Math To Function or Not to Function From Mathematical Modelling of Cancer Cell Plasticity to Philosophy of Cancer Quantitative Molecular Dissection of Cancer Evolution Deep Learning for Clinically Actionable Cancer Pathology Feature Detection Enriching Cancer Research Through Unconventional Collaborations Contents Invited Talk Plasticity in Cancer Cell Populations: Biology, Mathematics and Philosophy of Cancer 1 Introduction 2 Plasticity in Cancer Cell Populations 2.1 Non Genetic Phenotype Switching 2.2 Transient Drug-Induced Tolerance in Cancer 2.3 Dedifferentiation and Transdifferentiation 3 Mathematics of Plasticity with Therapeutic Control 3.1 How to Mathematically Model Plasticity in Cancer 3.2 Adaptive Dynamics: Asymptotic Behaviour of Cell Populations 3.3 Theoretical Therapeutics: Multi-targeted Optimal Control 4 Evolutionary Biology and Philosophy of Cancer 4.1 `Nothing Makes Sense in Biology Except in the Light of Evolution' 4.2 The Atavistic Theory of Cancer 4.3 Failed Control of Differentiations: Cancer is a Failure of Cohesion 4.4 Speculations on the Possible Future of Cancer Therapeutics 5 Conclusion References Statistical and Machine Learning Methods for Cancer Research CHIMERA: Combining Mechanistic Models and Machine Learning for Personalized Chemotherapy and Surgery Sequencing in Breast Cancer 1 Background 1.1 Formalizing Therapy Sequencing 1.2 Models of Tumor Growth 1.3 Models of Chemotherapy Pharmacokinetics 1.4 Objectives of the Study 2 Materials and Methods 2.1 Introducing CHIMERA 2.2 Datasets 2.3 Procedures 3 Results 3.1 Learning the Tumor Growth Function f(V) 3.2 Learning the Pharmacokinetics P(t,V) 3.3 Chemotherapy-Surgery Sequencing 4 Conclusion References Fine-Tuning Deep Learning Architectures for Early Detection of Oral Cancer 1 Introduction 2 Materials 3 Method 3.1 Technical Details 3.2 Batch Normalization for Transfer Learning 4 Results 5 Discussion and Conclusion Appendix References Discriminative Localized Sparse Representations for Breast Cancer Screening 1 Introduction 1.1 Sparse Analysis 1.2 Dictionary Learning 2 Label Consistent Spatially Localized Ensemble Sparse Analysis (LC-SLESA) 2.1 Spatially Localized Block Decomposition 2.2 Block-Based Label Consistent KSVD for Dictionary Learning 2.3 Ensemble Classification 3 Experiments and Discussion 3.1 Data 3.2 Convolutional Neural Networks 3.3 LC-SLESA 4 Conclusion References Activation vs. Organization: Prognostic Implications of T and B Cell Features of the PDAC Microenvironment 1 Introduction 2 Methods 2.1 Multiplexed Immunohistochemistry 2.2 Patient Meta-data and Target Variables 2.3 Quantification of Spatial Proximity Between Cell Types 2.4 Predictive Model of Survival Group 3 Results 3.1 Prognostic Value of LLR Metric Depends on Spatial Scale 3.2 40m LLR Synergizes with T Cell Functional Markers to Predict Patient Survival 3.3 Patients with Low CD4+ T-to-B Cell or CD8+ T-to-Keratin+ Proximity Had Poor Outcomes 3.4 T Cell Cytotoxicity and CD8+ T Cell Effector Frequency Are Elevated in Long-Term Survivors 4 Discussion References On the Use of Neural Networks with Censored Time-to-Event Data 1 Introduction 2 Overview of Existing Survival Methods 2.1 Notations for Survival Data 2.2 CoxPH: the Cox Proportional Hazards Model 2.3 Non-linear Survival Models Based on Neural Networks 3 Materials and Methods 3.1 Methods 3.2 Data 3.3 Models Comparison 3.4 Evaluation Criteria 4 Results and Discussion 4.1 Simulation Study 4.2 Example Data: METABRIC 5 Discussion and Conclusion References Mathematical Modeling for Cancer Research tugHall: A Tool to Reproduce Darwinian Evolution of Cancer Cells for Simulation-Based Personalized Medicine 1 Model 2 Software 3 Simulations of Darwinian Evolution Based on Personalized Weighting of Hallmarks for Two Cancer Patients 4 Conclusion References General Cancer Computational Biology The Potential of Single Cell RNA-Sequencing Data for the Prediction of Gastric Cancer Serum Biomarkers 1 Introduction 2 Materials and Methods 3 Results 4 Discussion References Posters Theoretical Foundation of the Performance of Phylogeny-Based Somatic Variant Detection 1 Introduction 2 Methods 2.1 Tumor Phylogenetic Tree 2.2 Assumptions for the Patient-Wise Variant Detection 2.3 Assumptions for Given Mutation Profiles 2.4 Labeling Methods 2.5 Sensitivity and Specificity 3 Performance Evaluation 3.1 Performance Evaluation Summary of L,Rr 4 Results 5 Conclusion A Appendix A.1 Performance Evaluation of Rr A.2 Detailed Procedures for Performance Evaluation References Detecting Subclones from Spatially Resolved RNA-Seq Data 1 Introduction 2 Methods 3 Results 4 Discussion References Novel Driver Synonymous Mutations in the Coding Regions of GCB Lymphoma Patients Improve the Transcription Levels of BCL2 1 Introduction 2 Methods 2.1 Overview 2.2 Mutations Screening 2.3 Phenotype Understanding 2.4 Mechanism Analysis 3 Results 3.1 Mutations Screening 3.2 Phenotype Understanding 3.3 Mechanism Analysis 4 Conclusion References Author Index