Journal of Pathology Informatics Journal of Pathology Informatics
Contact us | Home | Login   |  Users Online: 447  Print this pageEmail this pageSmall font sizeDefault font sizeIncrease font size 

Table of Contents    
J Pathol Inform 2020,  11:34

Review of “artificial intelligence and deep learning in pathology” by Stanley Cohen

Department of Pathology, Division of Pathology Informatics, University of Michigan, Ann Arbor, MI, USA

Date of Submission28-Aug-2020
Date of Acceptance01-Sep-2020
Date of Web Publication06-Nov-2020

Correspondence Address:
Dr. Jerome Cheng
Department of Pathology, Division of Pathology Informatics, University of Michigan, Ann Arbor, MI
Login to access the Email id

Source of Support: None, Conflict of Interest: None

DOI: 10.4103/jpi.jpi_66_20

Rights and Permissions

How to cite this article:
Cheng J. Review of “artificial intelligence and deep learning in pathology” by Stanley Cohen. J Pathol Inform 2020;11:34

How to cite this URL:
Cheng J. Review of “artificial intelligence and deep learning in pathology” by Stanley Cohen. J Pathol Inform [serial online] 2020 [cited 2022 Jul 6];11:34. Available from:

Editor: S Cohen

Edition: Artificial Intelligence and Deep Learning in Pathology 1st Edition

Publisher: Elsevier: Amsterdam

Year: 2020

Page: 288

ISBN: 978-0-323-67538-3

Artificial intelligence (AI) is becoming pervasive in different aspects of our society, including pathology research and practice. However, most pathologists are still unfamiliar with the various AI technologies and tools that are becoming increasingly prevalent. AI and deep learning in pathology, written by several experts in machine learning and digital pathology, aim to provide every pathologist an introduction to AI that is both adequately detailed and readable. It assumes the reader has no previous background in computer programming and machine learning. Although written by different authors, every chapter is well-written, interesting, informative and adheres to one common standard-explaining concepts in a clear and straightforward manner, with minimal use of technical terms and intimidating mathematical formulas (with the exception of one chapter). In addition, numerous illustrations and examples are provided to explain difficult topics.

Currently, there are numerous software packages and tools that make it very easy to create machine learning predictive models, however, with these tools, one may end up creating models without any idea what are the technologies behind the generated models. This book is a perfect supplement to someone using machine learning libraries/software to provide an understanding of the logic and math behind these AI tools and algorithms, which will lead to more realistic expectations from AI-enabled software, and also promote more appropriate usage.

The book comprises 11 chapters, each with up-to-date references, which is important in a field that is growing rapidly and continuously. The first chapter gives a very basic introduction to computer programming and fundamental machine learning concepts. Chapters 2–4 provide clear and detailed explanations of a variety of machine learning algorithms and methods, such as decision trees, random forests, convolutional neural networks, autoencoders, transfer learning, generative adversarial networks, reinforcement learning, multilabel classification, and one-class learning. Some humor is injected into the chapters, which helps keep the reader interested. For example, at the end of chapter 2, it is mentioned that backpropagation is unlikely to occur in our brains “because very few neurons have taken courses in multivariate calculus” - alluding to the fact that artificial neural networks do not work exactly like our brains, although these were initially inspired by our understanding of the interactions between biological neurons. Chapter 5 focuses on data preprocessing, usually a long and time-consuming but necessary step in machine learning experiments. Chapter 6 gives an excellent discussion concerning the requirements, benefits and practical aspects of the integration of whole slide imaging (an essential prerequisite to the adoption of AI into pathology workflow) and telepathology into pathology practice, including the selection criteria for whole slide scanners, and how AI may potentially be integrated with these technologies. Chapter 7 discusses the utility of convolutional neural networks for style transfer, image normalization, denoising, sharpening, and introduces unconventional microscopy modalities that are unfamiliar to many pathologists (e.g., multiphoton microscopy). Chapter 8 illustrates how areas of interest may be localized using deep learning based software and how information from multiple fluorescent biomarkers can be combined with histological features to infer prognostic information from tissue specimens, and predict drug response. Chapter 9 focuses on using machine learning methods to grade cancers. It goes over nuclei detection, ground truth annotation, nuclear feature extraction/counting, nuclei spatial distribution, and other variables that may be used to generate predictive models. Chapter 10 covers the interplay between tumor cells, immune cells, and other components of the tumor microenvironment, with an emphasis on using deep learning to identify and analyze tumor-infiltrating lymphocytes to derive prognostic and treatment related information. Chapter 11 explains the potential role AI algorithms may play in the anatomic pathology workflow-as an aid to diagnostic decision making in the form of digital assistants for tissue classification, metastasis identification, prognostication, and data integration.

As an overall assessment, I highly recommend this book to every pathologist (or any data scientist working with pathology-related problems) seeking a better understanding of machine learning concepts and the potential applications of AI in pathology practice and research. The book offers an excellent introduction to machine learning in relation to the field of pathology that would benefit newcomers to AI, and even those with some experience in the field.

This article has been cited by
1 Breast histopathological image analysis using image processing techniques for diagnostic puposes: A methodological review
R Rashmi, Keerthana Prasad, Chethana Babu K Udupa
Journal of Medical Systems. 2022; 46(1)
[Pubmed] | [DOI]




   Browse articles
    Similar in PUBMED
   Search Pubmed for
   Search in Google Scholar for
    Access Statistics
    Email Alert *
    Add to My List *
* Registration required (free)  

  In this article

 Article Access Statistics
    PDF Downloaded317    
    Comments [Add]    
    Cited by others 1    

Recommend this journal