Healthcare and Artificial Intelligence

Foreword\nIntroduction\nContents\nEditors and Contributors\nArtificial Intelligence and Tomorrow’s Health\n	French Strengths and Weaknesses\n	A Project for France\n	European Tracks\nAdvancing Healthcare Through Data-Driven Medicine and Artificial Intelligence\n	Introduction: The Status Quo and Its Toll\n	The Promise of Artificial Intelligence in Transforming Healthcare for the Better\n		Artificial Intelligence Implemented in Clinical Practice\n		Aspirations and Challenges\nArtificial Intelligence: A Vector for Positive Change in Medicine\nDatabases\nMachine Learning and Massive Health Data\n	Introduction\n	CNAM Infrastructure Revisited\n		From a Vertical to a Horizontal Architecture\n		Flattening the Data\n		Software Development\n	Toward an Automatic Screening Algorithm for Drugs with Harmful Side Effects\n		ConvSCCS: A Large-Scale Screening Algorithm\n		First Results Obtained Using the ConvSCCS Algorithm: Identification of an Antidiabetic Agent Increasing the Risk for Bladder Cancer\n	Work in Progress and Prospects\n		Study Projects of the New Partnership\n		Toward Systematizing and Opening up the Big Data Pipeline\nLinking an Epidemiological Cohort to Medicoadministrative Databases\n	Cohort Complementarity and Medicoadministrative Databases\n	National Health Data System\n	The Constances Cohort\n	Tools to Help Understand and Manipulate Data\n		Interactive Documentation\n		Receipt of SNIIRAM Data\n		Construction of Calculated Data\n		The REDSIAM Network\n	Resources Needed for Optimal Use of NHDS Data and Linked Cohorts\n		Computer Resources and Organizational Aspects\n		Diversified Skills\n	Conclusion\nMedical and Administrative Data on Health Insurance\n	What Data Are We Talking About?\n	CNAM Uses: Management of Disease Risk\n	Analysis of the Healthcare System (Analysis by Provider Moved to Analysis by Disease)\n	Care Pathway Studies and Heterogeneity in Care Provision\n	Real-Life Drug Studies\n	From SNIIRAM to NSDS: Developing the Use of Health Data\n	Conclusion\nHow the SNIIRAM–PMSI Database Enables the Study of Surgical Practices\n	Importance of Cohort Follow-up of Patients Treated in Routine Practice\n	About Observapur\n		The Main Lessons Learned\n		General Reflections\n		The Limits of the SNIIRAM–PMSI Database\n		Interpret the Data in This Database Correctly\n		Lack of Clinical Data\n		Adapted Means\n	Conclusion\nHospital Databases\n	Introduction\n	Issues and Uses of the Health Data Warehouse\n		Improve the Management of Activity and Hospital Performance\n		Facilitate Epidemiological Research and Data Studies\n		Facilitate Intervention Research\n		Development of Artificial Intelligence and Decision Support Tools\n	Regulatory Framework and Governance Put in Place\n		Regulatory Framework\n		Organization and Comitology\n		Definition of Operating Rules\n		Patient Information and Ethical and Societal Dimensions\n	Clinical Data Warehouse Information Technology Infrastructure\n		High-Performance Infrastructure, Secure Storage, and Calculation\n		Solutions Dedicated to Operational Management\n		Solutions Dedicated to Research and Innovation\n	Objectives and Next Steps\n	Conclusion\nExperience of the Institut Curie\n	Introduction\n	Massive Amounts of Data at the Institut Curie\n	Data Warehouses\n	Text Information\n	Images\n	Example of Ongoing Projects\n	Conclusion\nKnowledge by Design\n	What Is the Diagnosis for France?\n	Public Policies Favoring Data\nDiagnosis and Treatment Assistance\nArtificial Intelligence to Help the Practitioner Choose the Right Treatment: Watson for Oncology\n	New Era of Information Technology\n	From Descriptive and Predictive to Prescriptive\n	IBM Watson Solution\n	Prospects for the Future\nArtificial Intelligence to Help Choose Treatment and Diagnosis Commentary\nMedical Imaging in the Age of Artificial Intelligence\nArtificial Intelligence in Medical Imaging\n	Artificial Intelligence in Medical Imaging: Drivers\n		Precision Medicine\n		Medical Images as a Source of Clinical Data\n	First-Generation Artificial Intelligence Systems\n		Automatic Analysis of Medical Images\n		Quantitative Medical Imaging\n		Radiomics\n		Searching for Information Within Images\n		Deep-Learning Revolution\n		Supervised Learning\n		Limits of Learning Systems: Curse of Dimensionality\n		A Potential Alternative: Transfer Learning\n		Optimal Solution: Unsupervised Learning\n	Future of Artificial Intelligence in Medical Imaging\nThe AI Guardian for Surgery\n	Motivation\n	Fourth Surgical Revolution\n	Case Study: Predicting the Surgical Phase to Support Real Time Decision for Surgeries\n	Conclusions\n	References\nToward an Operating Room Control Tower?\n	Monitoring Irradiation Risk in Hybrid Surgery\n	Analysis of Activities in Laparoscopic Surgery\nHigh-Dimensional Statistical Learning and Its Application to Oncological Diagnosis by Radiomics\n	Introduction: Learning, Curse, and Blessing of Dimensionality\n		Curse and Blessing of Dimensionality\n		Common Approaches to High-Dimensional Learning\n	Recent Advances in High-Dimensional Learning\n		Classification in Subspaces\n		Classification by Variable Selection\n		Selection of Variables in Bayesian Parsimonious Classification\n	Application of Radiomics to Oncological Diagnosis\n	Conclusion\nMathematical Modeling of Tumors\n	Modeling Approach and Obstacles in Oncology\n	Tumor Growth Modeling\n	Available Data in Clinical Oncology\n	Some Ideas for Modeling and Using Data in Oncology: General Principles\n	Two Examples of Tumor Growth: Meningiomas and Lung Metastases\n	Example Response to Treatment: Kidney Cancers Treated with Antiangiogens\n	Conclusion\nToward an Augmented Radiologist\n	There Is More to the Job of the Radiologist than Looking at Images\n	Who Is Responsible for Medical Liability?\n	What Does the Future Hold for the Radiologist Occupation?\nFunctional Imaging for Customization of Oncology Treatment\n	3D Determination of Tumor Volumes\n	New Functional Imaging Indices to Predict Therapeutic Response\n	Biomarkers of Tumor Heterogeneity\nVirtual Image and Pathology\n	Introduction\n	Digitized Slides in the Pathologist’s Routine: A Revolution\n	Multiparametric Approaches: New Opportunity to Refine Our Understanding of Disease\n	Social Networks\n	Conclusion and Prospects\nResearch\nArtificial Intelligence and Cancer Genomics\n	Introduction\n	Genomics and Cancer\n	Clinical Interest\n		Prevention\n		Diagnosis\n		Targeted Therapies\n	Genomic Data\n		DNA Sequencing\n		Storing and Handling Millions of Genomes\n	Artificial Intelligence and Genomics\n		Inferring the Structure of Genomes and the History of Tumors\n		Classifying Tumors\n		Attending to Therapeutic Choices\n	Conclusion\nBesides the Genome, the Environmentome?\n	The Environmentome: An Object of Analysis Similar to the Genome?\n	High-Resolution Satellite Databases\n		Climate Databases to Study the Climate–Environment Relationship\n		Historical Cohorts to Test Environmental Hypotheses\n		Case-Only Search for Environmental Causes of Type 1 Diabetes\n	Conclusion: Find Environmental Markers First and Then Search for Explanatory Causes\nArtificial Intelligence Applied to Oncology\n	Challenges of Digital Medicine\n		Digital Medicine: Evolution or Revolution?\n		Can We Build a Self-learning Health System?\n		Digital Medicine as an Integrating Pillar of the Medicine of the Future\n		What Can Be Done to Increase the Chances of Digital Medicine Succeeding?\n	Digital Medicine and Artificial Intelligence in Oncology: The Clinician’s Vision\n		Success of Precision Medicine in Oncology\n		Integrating Immunological and Constitutional Molecular Complexity to Better Treat Cancers\n		Rare Cancers and Complex Histological Diagnosis\n		Identifying New Predictive Factors in the Effectiveness of Cancer Treatment\n		Integrating Clinical Data from Patient Files\n	Artificial Intelligence Applied to Oncology: The Vision of an Industrial Partner\n	Conclusion\nPhysical and Mathematical Modeling in Oncology: Examples\n	Current Main Lines of Theoretical Oncology\n	Development of Solid Tumors\n	The Tumor and Its Stroma\n	Conclusion\nArtificial Intelligence for Medical and Pharmaceutical Research: Can Artificial Intelligence Help in the Discovery and Development of New Drugs?\n	Fundamental Research\n	Pharmaceutical Research and Development\nOpen Science and Open Data: Accelerating Scientific Research\n	From Microtasking to Challenges\n	Epidemium\n		Definition of the Program\n		A Framework for Scientific Cooperation\n	Central Resource: Open Data\n	Open Science, Open Data, and Reproducibility in Life Sciences\n	Open Source: The Cornerstone of an Open Model\nData\nData: Protection and Legal Conditions of Access and Processing\n	Impact of EU General Data Protection Regulations on the Processing of Health Data\n		Definition of Health Data and Appropriate Safeguards\n		Principle of Prohibition with Derogations\n		Compatibility of Personal Data Protection Principles with the Development of Big Data\n	Room for Maneuver Left to National Law: High Level of Security in a Complex Legal Environment\n		Exchange and Sharing of Health Data\n		Reuse of Health Data\nCNIL (Commission Nationale de l’Informatique et des Libertés) and Analysis of Big Data Projects in the Health Sector\n	Context and Issues\n	Toward a Finer Typology of Adapted Solutions\n	Objective: To Develop Understanding and Trust\nInternational Vision of Big Data\n	What Data?\n	Which Sources for What Purposes?\n	Characteristics\n		Unique Number\n		Privacy: Anonymization and Consent\n	Interoperability, Analysis, and Impact\n		Reliability\n		The Use of Databases\n	Control\n	Conclusion\nHuman\nAre Public Health Issues Endangered by Information?\n	Minority Tyranny\n	Cognitive Demagogy\n	Poisoned Beliefs\n	Some Tracks\nSocial and Emotional Robots: Useful Artificial Intelligence in the Absence of Consciousness\n	Realities and Fantasies\n	Intelligence and Consciousness of Robots\n	Social and Emotional Robotics in Health\n	Emotional Interaction with Machines\n	Ethics of Robotic Systems\nAugment Humans, and Then What?\nArtificial Intelligence: A Vector for the Evolution of Health Professions and the Organization of Hospitals\n	For Hospitals\n	For Professionals\n		Doctors\n		Researchers\n		Other Trades\n	Patients and Public Health\n	Conclusion: What Will Be the Likely Impact on Medical Professions and the Training of Professionals?