At RSNA22, Dr. Siddhartha Mukherjee predicts how AI will evolve in diagnostics

At RSNA22, Dr. Siddhartha Mukherjee predicts how AI will evolve in diagnostics

Renowned physician and author Siddhartha Mukherjee, MD, DPhil, delivered the keynote address Monday, November 28 at RSNA 2022, the annual conference of the Radiological Society of North America, being held this week at the sprawling McCormick Place Convention Center in Chicago. . The 4,188 seats of the Arie Crown Theater were nearly full when Dr. Mukherjee, introduced by RSNA President Bruce G. Haffty, MD, took the stage. Mukherjee spent most of the hour giving his presentation, with the final quarter of an hour spent in a fireside chat between the two doctors.

Mukherjee, assistant professor at Columbia University and practicing oncologist at Columbia University Medical Center, is an oncologist and hematologist who has spent decades involved in research in the diagnosis and treatment of cancer. Received the 2011 Pulitzer Prize for General Nonfiction for his 2010 book The emperor of all diseases: a biography of cancer; last month he published his latest book, Cell song. He spoke on the topic “A Glimpse into the Future of Biomedical Transformation”.

At the start of his lecture, Mukherjee said, “To begin with, I want to talk about deep learning; it means deploying learning algorithms that mimic human learning. How do we learn? He asked. “Can machines learn like us? Can machines learn medicine? He referred to the April 3, 2017 article he published in the new yorker, at the request of the editors of this magazine. He emphatically noted that the title the editors gave the article was “AI VERSUS MD” – but quickly added that, “Interestingly, it’s a false title. I don’t think ‘versus ” is the right word. Much of what I’m going to tell you is about ‘with’ and not ‘versus’.” And he then referred to the philosopher Gilbert Ryle, who, “long before the birth of modern AI , made a distinction between ‘knowing that’ as opposed to ‘knowing how’. ‘Knowing that’, Mukherjee quoting Ryle, ‘is knowing a series of facts; knowing how is putting those facts together to produce learning. So, for example, knowing this means knowing that a bike is made up of a set of parts. Knowing how is learning to ride a bike. What you haven’t done is handing over to your children a manual for riding a bicycle, with step-by-step instructions. What you have done is that you have shown or learned how to ride a bicycle. L he set of questions that Ryle explored, Mukherjee continued, was how and why do entry and exit rules work? And can artificial intelligence really imitate how human intelligence works? The answer will be very important for the future of medicine, he noted.

Mukherjee showed his audience two sets of fun but also thought-provoking picture slides; one set consists of photos of a dog, as well as photos of a blueberry muffin, with some photos looking extremely similar to the eye; another involved photos of another dog, with photos of a mop, again, with a visual set of commonalities. He also showed the audience a simple photo of a dog and a cat looking at each other, noting that “we still don’t know exactly how our brains distinguish between a dog and a cat.” And, he added, the “dog vs. muffin” and “dog vs. mop” tests were actually used by researchers at DARPA – the Defense Advanced Research Projects Agency – in order to try to develop theories about perception and cognition.

From there, Mukherjee dove into the depths of the clinical research pool, guiding his audience through a series of examples involving cancer detection and research around different types of cancers, and pondering whether and to what extent algorithms derived from artificial intelligence could be used, and are already beginning to be used, to support cancer research. Among the areas he cited as ideally suited for algorithmic development were mammography and the risk of pancreatic cancer in high-risk patients, as well as exploring the recurrence of previous cancer. An absolutely essential use of AI will be to support the development of parallel “second opinion” clinical decision support, he said, noting that there could be countless situations in which a doctor would develop its diagnosis while at the same time running an AI algorithm to see what clinical conclusions machine learning could produce in the same clinical case. “You can imagine yourself doing a diagnosis while the AI ​​is doing a diagnosis and then you compare your notes,” he said. “And a second opinion could take place at the same time. And so the algorithm might ask you if you are sure of your diagnosis.

Continuing with clinical examples, he highlighted the work that has been done to grow cancer cells in Petri dishes in order to study them; he noted that cancers do not grow easily in a laboratory environment and that cancers are in fact “very dependent on factors secreted into the microenvironment”. But this research, he added, could help uncover some of the secrets of how cancers grow in the body. In any case, it significantly details several clinical case studies in which machine learning has been used to support the analysis of oncology therapies.

“This is the most important question,” Mukherjee continued, “the one we all ask: what is the endpoint? The best endpoint is lives saved. So if you catch cancers early, there are many reasons to focus on the number of lives saved because you don’t know if advanced stage cancer or less invasive cancer will eventually kill the person or not There are countless trials where the end point can be a point end loose or weak. Obviously measuring lives saved is difficult, as it involves decades of research. Me and others went to the FDA [Food and Drug Administration], and said, very few companies can do a life-saving trial; it should be an NIH funded and sponsored trial. I would like us all to think about how we create the final study. And he ended on an optimistic note, saying he believes a lot of progress will be made in all of these areas over the next few years.

In response to a question from Haffty during their 15-minute fireside chat at the end of the session, Mukherjee said he thought healthcare consumers were becoming more self-aware and demanding to be treated as intelligent and thoughtful individuals. , with the right to receive solid information about diagnostics, research, therapies and everything related to patient care.

#RSNA22 #Siddhartha #Mukherjee #predicts #evolve #diagnostics

Leave a Comment

Your email address will not be published. Required fields are marked *