A New Book, “Artificial Intelligence and Deep Learning in Pathology” was edited by longtime ASIP Member Dr. Stan Cohen

Stanley Cohen, MD
Stanley Cohen, MD

Recent advances in computational algorithms, along with the advent of whole slide imaging as a platform for embedding artificial intelligence (AI), are transforming pattern recognition and image interpretation for diagnosis and prognosis. Yet most pathologists have just a passing knowledge of data mining, machine learning, and AI, and little exposure to the vast potential of these powerful new tools for medicine in general and pathology in particular. In Artificial Intelligence and Deep Learning in Pathology, with a team of experts, Dr. Stanley Cohen covers the nuts and bolts of all aspects of machine learning, up to and including AI, bringing familiarity and understanding to pathologists at all levels of experience. Dr. Cohen is a Past-President of the ASIP and the 2015 recipient of the Gold-Headed Cane Award.

Read the full article here

More on Dr. Cohen:


Associate Editor

Artificial Intelligence, we can do more!

With my first post to this blog I wanted to talk about an exciting new frontier in scientific reserach and clinical medicine. A recent article on darkdaily.com discussed setbacks of the clinical utility of artificial intelligence systems for oncologists and anatomic pathologists. Understandably so, medical professionals do not want to be replaced by computers. However, it seems that every day we get new toys to play with that involve machine learning. So, how can we utilize these tools to improve the quality of science without losing rigor and responsibility? In February of this year, Steven A. Wartman and C. Donald Combs published an article in the AMA Journal of Ethics describing the potential use of AI in medical education. They explain how AI can be used to improve knowledge management in the classroom which can further be translated to greater efficiency in the clinic.

I think we can be doing more! AI can have applications in all fields of science and medicine. We need to work to continue to improve this new tool and use it to better our scientific progress. At the University of South Carolina we have recently created an Artificial Intelligence Institute to boost AI applications in both the research and classroom settings. This is one of many steps we are taking to improve the impact of AI in STEM.

Do you have an exciting way you use AI in your science? Do you want to see AI used more in the classroom? Share your opinion on the use of AI in science and medicine!

Don’t forget the ASIP 2020 Annual Meeting in San Diego April 4-7, 2020 http://asip20.asip.org/

Intersted in becomming a member of ASIP? Contact me at alexander.sougiannis@uscmed.sc.edu www.linkedin.com/in/alexander-sougiannis

Stanley Cohen, MD

Trailblazing Men

ASIP Highlights Session:
I Am An ASIP Member and This Is My Science

  • Experimental Biology 2019 – Orlando FL

Stanley Cohen, MD

Emeritus Chair of Pathology and Emeritus Founding Director
Center for Biophysical Pathology
Rutgers-New Jersey Medical School, Newark NJ

An Immunologist In Remission: How I learned to Embrace Artificial Intelligence

In this blurb, I would like to share reflections on my career trajectory, what influenced it, and the seven lessons I have learned along the way. I have been interested in research since my freshman year at Stuyvesant High School, a science-oriented public secondary school in Manhattan. My main interests were mathematics and physics, but in college I became fascinated by the sudden explosion of biological research that was transforming it from a descriptive to a mechanistic science. In medical school I spent elective time doing neuroscience research in part because this seemed to blend my multiple interests and because of my admiration for the work of a faculty member (Werner Loewenstein) in that area.

This experience led to my first lesson: (1) The longest journey begins with but a single step in a random direction since in order to study neuronal activation I had to try to localize a putative synaptic receptor using fluorescent antibodies. This led me to spend time with a nationally known immunologist (Beatrice Seegal) and I was hooked. The gods of chance had spoken, and I accidently became an “immunologist”. Following a residency in pathology at MGH and a Fellowship at NYU (under Baruj Benacerraf and Robert McCluskey), I became an “obligate volunteer” in the US Army during the Vietnam era. Since this was at a time when the Army was interested in defense against germ warfare, and there weren’t many immmunologically trained recruits that year (one), I wound up working at Walter Reed under Elmer Becker, one of the fathers of complement immunochemistry. I now learned two more global lessons: (2) Publish or Perish (this was literally true if you were in the Army during the Vietnam era) and (3) Make sure you have the very best mentors who are willing to have you. The corollary to this is that your peers are also your mentors and vice versa, so it is important to remain active not only in collaborative research and discussions but also in professional society activities such as those of the ASIP. The ASIP has been my home away from home for over forty years, and I have been honored to serve as its President in 2009.

After the army, I moved to SUNY Buffalo to start my own lab and began to study the newly discovered lymphocyte-derived non-antibody substances that affected other inflammatory cells in vitro (by Barry Bloom and John David, independently). At the time, there was much skepticism among immunologists revolving around the concern that these substances were merely test-tube artifacts that had no true biologic relevance. My work revolved around demonstrating that these factors did, in fact, have important in vivo effects in both animal models and in some human diseases, which brings me to lesson (4): “In Vivo Veritas”, a saying coined by Hal Dvorak.

With the acceptance of the reality of these mediators, they were assumed to be an intrinsic part of the immune response and were named lymphokines. During this time, I began to wonder about the overall significance of these strange substances that were generated via the antigen activation of lymphocytes, but which had no antibody activity. Work in many laboratories had answered the question of their immunoinflammatory role. However, I soon began to wonder whether this was solely an immunologic phenomenon, and began to study a number of cellular systems that were devoid of lymphocytes or other inflammatory cells. My colleagues and I soon demonstrated that those other cells, when triggered by an appropriate stimulus, could make factors chemically and biologically identical to known lymphokines. This was so surprising to immunologists that I had trouble initially getting it published, leading to lesson (5): If at first you don’t succeed, find another journal.

 I hypothesized that every cell had this property and suggested that these agents were mediators of intercellular communication in general, and not just in the immune system, and called them, collectively, “cytokines”. Cytokines are, in effect, the vocabulary by which cells talk to each other. In this view, the only unique thing about a lymphocyte is that it can use an antigen to which it has been sensitized as the inducing agent. This brings us to lesson (6):  Exploring an unquestioned answer can be as important as answering an unanswered question, and lesson  (7) Nothing is real until you name it. Lesson (7) actually appears in a calendar that had a variant of Murphy’s law for each calendar day and is cited there as “Cohen’s Law”.

At the time, I never dreamed that the field of cytokine research would expand as rapidly as it has in the ensuing years. In retrospect, the first clue was when cytokines were first used as part of the plot of a science fiction story, since science fiction often predicts the future:  “Last week’s EMT group was at full strength, although they were now tossing around phrases like cytokines and pyruvic acid.” In “A Few Good Men”. Richard A. Lovett, ANALOG, January/February 2005.

At this point, lesson (1) came into play once more. It turns out that cytokines do not always work as single entities in vivo. Essentially all biologic activities in which they participate involve complex cytokine networks. Just has been the case of genomics, it turned out to be really hard for humans to understand and interpret huge inter-connected datasets without help. That help came along with the rapid evolution of machine learning, which is part of what we think of when we talk about artificial intelligence. Of necessity, I began to learn about this burgeoning new field, and again I was hooked. It turns out that neural networks are arguably the most successful of these approaches, and these are based upon (crude) models of the human brain, and so I have come full cycle from (biological) neuroscience to immunology to “synthetic” neuroscience. If not a truly random walk, certainly a circuitous route. It has turned out that exploration of the role of artificial intelligence in diagnostic and experimental pathology has become a second career for me.

To put all this in perspective, while scientific focus is important, it is equally important to let your imagination roam freely and to explore new directions as they become apparent. Research is not like climbing a telephone pole with a single prize at the top. Research is like climbing a tree. It is important to explore as many of the branches as possible, because that’s where all the fruit really is. It is also important to recognize that science is a team sport, and you cannot work optimally in isolation. Colleagues, mentors, and professional associations all help the journey forward.