Artificial intelligence applications in human pathology / / edited by Ralf Huss, Michael Grunkin.
"Artificial Intelligence Applications in Human Pathology deals with the latest topics in biomedical research and clinical cancer diagnostics. With chapters provided by true international experts in the field, this book gives real examples of the implementation of AI and machine learning in human pathology. Advances in machine learning and AI in general have propelled computational and general pathology research. Today, computer systems approach the diagnostic levels achieved by humans for certain well-defined tasks in pathology. At the same time, pathologists are faced with an increased workload both quantitatively (numbers of cases) and qualitatively (the amount of work per case, with increasing treatment options and the type of data delivered by pathologists also expected to become more fine-grained). AI will support and leverage mathematical tools and implement data-driven methods as a center for data interpretation in modern tissue diagnosis and pathology. Digital or computational pathology will also foster the training of future computational pathologists, those with both pathology and non-pathology backgrounds, who will eventually decide that AI-based pathology will serve as an indispensable hub for data-related research in a global health care system. Some of the specific topics explored within include an introduction to DL as applied to Pathology, Standardized Tissue Sampling for Automated Analysis, integrating Computational Pathology into Histopathology workflows. Readers will also find examples of specific techniques applied to specific diseases that will aid their research and treatments including but not limited to; Tissue Cartography for Colorectal Cancer, Ki-67 Measurements in Breast Cancer, and Light-Sheet Microscopy as applied to Virtual Histology. The key role for pathologists in tissue diagnostics will prevail and even expand through interdisciplinary work and the intuitive use of an advanced and interoperating (AI-supported) pathology workflow delivering novel and complex features that will serve the understanding of individual diseases and of course the patient"--
| Beteiligte Person(en) | , |
|---|---|
| Ort, Verlag, Jahr |
New Jersey
: World Scientific
, 2022
|
| Umfang | 1 online resource (337 pages) |
| ISBN | 1-80061-139-0 |
| Sprache | Englisch |
| Zusatzinfo | Intro -- Contents -- Foreword -- About the Editors -- Chapter 1 Introduction: Integration of Computational Pathology and AI Application into Histopathology Workflow -- References -- Chapter 2 Standardized Tissue Sampling for Automated Analysis and Global Trial Success -- 1. Introduction -- 2. AI-Advanced Data Analytics -- 3. Cancer - A Highly Complex, Individual Disease -- 4. Digitization of Cancer Data Quality Is Crucial -- 5. Criteria of Standardization to Enable AI-Supported Digital Medicine in Pathology -- 6. Outlook -- References -- Chapter 3 Light-Sheet Microscopy as a Novel Tool for Virtual Histology -- 1. Introduction -- 2. Technology Description -- 3. Data Management and Image Analysis -- 4. Discussion and Summary -- Acknowledgments -- References -- Chapter 4 Stain Quality Management and Biomarker Analysis -- 1. Introduction -- 2. Internal Quality Control -- 2.1. Purpose and use of IHC -- 2.2. The tissue toolbox -- 2.3. Calibration phase -- 2.4. IHC analytical protocol optimization -- 2.5. Instrumentation -- 2.6. Validation phase -- 2.7. Reproducibility phase -- References -- Chapter 5 Measuring Ki-67 in Breast Cancer: Past, Present, and Future -- 1. Introduction and Historical Context -- 2. Ki-67 as a Prognostic Marker -- 3. The Work of the IKWG -- 3.1. Phase 1: Visual assessment of a TMA using non-standardized methods -- 3.2. Phase 2: Visual scoring of a TMA using a standardized method -- 3.3. Phase 3: Visual scoring of core-biopsies using a standardized method -- 3.4. Phase 4: Visual scoring of resection cases using a standardized method -- 4. Pre-Analytical Factors -- 5. Automated Scoring -- 6. Cognition Master Professional Suite -- 7. Virtual Double Staining -- 8. IKWG and Automated Ki-67 Scoring -- 9. Data from External Quality Assessment Schemes -- 10. Positive Controls for Ki-67 IHC -- 11. Ki-67: Where Are We Now? -- References. Chapter 6 Multiplex: From Acquisition to Analysis -- 1. Introduction -- 1.1. Why multiplex? -- 1.2. What is phenotyping? -- 1.3. Chromogenic multiplex -- 1.3.1. Chromogenic staining -- 1.3.2. Virtual multiplex -- 1.3.2.1. Tumor and region identification -- 1.3.2.2. High-plex colocalization studies -- 1.3.3. Multiplexed immunohistochemical consecutive staining on single slide -- 1.4. Immunofluorescent multiplex -- 1.4.1. Immunofluorescence (IF) -- 1.4.2. Spectral unmixing -- 1.5. Imaging mass cytometry multiplex -- 2. Approaches to Phenotyping Using AI -- 2.1. Cell segmentation -- 2.1.1. Machine learning approaches (classical) -- 2.1.2. DL/AI (NextGen) -- 2.2. Semi-automated approaches to phenotyping -- 2.3. Unbiased autophenotyping using AI autoclustering -- 3. Meaningful Data Extraction -- 3.1. Targeted approach -- 3.2. Discovery approach -- 4. The Future of Multiplex -- 4.1. Technological advances -- 4.1.1. Introduction -- 4.1.2. Probes -- 4.1.3. Hardware -- 4.1.4. Software -- 4.1.4.1. Region segmentation with DL -- 4.1.4.2. Cell detection with DL -- 4.1.4.3. Phenotypic classification -- 4.1.4.4. Accessible data visualization and exploration -- 4.1.5. Summary -- 4.2. Clinical practice -- References -- Chapter 7 An Introduction to Deep Learning in Pathology -- 1. Introduction -- 2. Basic Applications of Machine Learning in Digital Pathology -- 3. End-to-End Deep Learning Biomarkers in Digital Pathology -- 4. Neural Network Technology Basics -- 5. Image Data Flow in Deep Neural Networks -- 6. The Art of Applying Machine Learning -- 7. Data Pre-processing in a Digital Pathology Pipeline -- 8. Hardware and Computing Requirements -- 9. Choosing and Training a Neural Network Model -- 10. Explainable Deep Learning -- 11. Outlook -- References -- Chapter 8 AI-Driven Precision Pathology: Challenges and Innovations in Tissue Biomarker Analysis for Diagnosis. 1. Basic Concepts from Image Analysis, AI, and Digital Pathology -- 1.1. Digitalization and whole-slide images - the foundation of AI -- 1.2. Basics of image analysis and AI -- 1.3. Standardization and interoperability for analysis purposes -- 2. Tissue Biomarkers in the Era of Precision Medicine -- 2.1. Introduction -- 2.2. The quality challenge with tissue biomarkers -- 2.3. Toward standardized assessment of tissue biomarkers -- 2.3.1. Invasive tumor detection and management of tumor heterogeneity -- 2.3.1.1. Background -- 2.3.1.2. Magnitude of the problem -- 2.3.1.3. Proposed solutions -- 2.3.1.4. Summary -- 2.3.2. The transition from qualitative to data-driven quantitative pathology -- 2.3.2.1. Background -- 2.3.2.2. The problem -- 2.3.2.3. Proposed solutions -- 2.3.2.4. Summary -- 2.3.3. Addressing the increased complexity and ambiguousness in clinical assessments of predictive biomarkers -- 2.3.3.1. Background -- 2.3.3.2. The problem -- 2.3.3.3. Proposed solutions -- 2.3.3.4. Summary -- 2.4. Toward standardization in stain quality -- 2.5. Other tissue diagnostic applications -- 3. Regulatory Pathways -- 3.1. Introduction -- 3.2. Are digital pathology technologies medical devices? -- 3.3. A regulatory path forward -- 3.4. Collaborations/Collective approaches -- 3.4.1. Standards -- 3.4.2. Harmonization -- 3.4.3. Collaborations -- 3.4.4. Collective knowledge -- 3.5. The current regulatory landscape for AI algorithms -- 3.5.1. Radiology's experience using algorithms -- 3.5.2. Fully automated and/or continuous learning algorithms -- 3.5.3. Risk assessment for AI -- 3.5.4. Additional considerations -- 3.6. Conclusion -- References -- Chapter 9 Tissue Cartography for Colorectal Cancer -- 1. Introduction -- 2. Building Blocks -- 3. Whole-Slide Cartography for Colon Histology -- 3.1. Medical applications -- 3.1.1. Presence or absence of tumor. 3.1.2. Location of tumor area -- 3.1.3. Tumor composition -- 3.1.4. Invasive margin -- 3.1.5. Tumor invasion depth -- 3.2. Creating a dataset -- 3.2.1. Cohort -- 3.2.2. Selecting the classes -- 3.2.3. Technical workflow -- 3.3. Evaluation metrics -- 3.4. Evaluation and visualization of tile-based approach -- 4. Fast Whole-Slide Cartography -- 4.1. Superpixel-based approach -- 4.1.1. Experimental results -- 4.2. Uncertainty metric -- 4.2.1. Experimental results -- 4.3. Evaluation of medical applications -- 4.3.1. Tumor area -- 4.3.2. Invasive margin -- 4.3.3. Tumor composition -- 5. Coping with Real-Life Heterogeneous Data -- 5.1. Multi-scanner dataset -- 5.2. Strategy -- 5.3. Data augmentation -- 5.4. Experimental results -- 5.5. Conclusion -- 6. Learning by Example -- 6.1. Experimental setup -- 6.2. Stability -- 6.3. What's next? -- Acknowledgments and Funding -- References -- Chapter 10 Graph Representation Learning and Explainability in Breast Cancer Pathology: Bridging the Gap between AI and Pathology Practice -- 1. Breast Cancer Pathology in Clinical Practice -- 1.1. Role of AI in digital pathology -- 1.2. Expectations from AI for pathological adoption -- 1.3. Limitations in clinical adoption of AI -- 1.4. Role of graphs in computational pathology -- 2. HistoCartography -- 2.1. Semantic graph representation -- 2.1.1. CG representation -- 2.1.2. TG representation -- 2.1.3. HACT representation -- 2.2. HACT-Net -- 2.3. Explaining the network prediction -- 2.3.1. Feature attribution explainers -- 2.3.2. Entity-level concept analysis -- 3. Experiments and Results -- 3.1. BReAst Carcinoma Subtyping (BRACS) dataset -- 3.2. Pre-processing -- 3.3. Breast cancer subtyping results -- 3.4. Breast cancer explainability results -- 4. Summary -- Acknowledgments -- References -- Chapter 11 AI-Driven Design of Disease Sensors: Theoretical Foundations. 1. Introduction -- 2. Constraints of Computational Cellular Engineering -- 3. A Multi-Scale Problem -- 4. Design Principle for Engineering Biosensors -- 5. Integrating AI with Mathematical Modeling -- 6. Recent Advances on Use of AI in Molecular Dynamics and Simulations -- 7. Synthetic Design Using Generative Models -- 8. Conclusions -- Acknowledgments -- References -- Index. |
| Andere Ausgaben | Print version:: Artificial Intelligence Applications In Human Pathology. Singapore : World Scientific Publishing Company,c2022 |
| Online-Zugang | World Scientific Life Sciences 2017-2018-2019-2020-2021-2022 |
Bei Problemen beim Zugriff auf diese Online-Quelle beachten Sie unsere Hinweise zum Zugriff auf lizenzierte Angebote von außerhalb des Campus.