Path Toward AI: Part Two

Welcome to part two of Path Toward AI. In part one, we talked to key opinion leader Dr. Donald Karcher (@DonKarcherMD) about his feelings on AI in the pathologist’s world. Also curious about the general consensus in the pathology community, we subsequently sent out a tweet asking PathTwitter to comment about AI adoption challenges and opportunities. The response was from around the globe and so intriguing that we decided to make a part two for Path Toward AI. Dr. Karcher left us with the notion that pathologists who do not embrace AI will face challenges. Let’s explore how others feel about AI and pathology.


In our endeavors to utilize artificial intelligence, we cannot forget the overall movement of pathology going digital. “AI provides the best link to the adoption of digital pathology,” said Dr. Travis Brown (@DrTravisBrown), a general pathologist in Adelaide, South Australia. The training of AI includes looking at thousands of images, not on glass slides but digitally. The practice of diagnosing via computer screen instead of a microscope is critical in preparing pathologists to use and understand AI.


So what benefits does AI have? “I foresee it becoming part of a pathologist’s diagnostic armamentarium - like IHC, flow cytometry and molecular genetics. Leveraging AI has the potential to make our diagnoses better and faster, and there could be prognostic value in AI-driven data,” said Dr. Kamran Mirza (@KMirza), an associate professor of pathology in Chicago. “AI can assist in diagnosis, structured reporting (i.e. identify LVI/perineural invasion) and improved turnaround time,” said Dr. Brown. Phillip Templo, Jr. (@thejourneymate), a pathology residency program director in Manilla, said, “AI makes possible a faster and more streamlined workflow which means more time for the pathologists to engage in other aspects of the field like research and teaching. Faster turnaround time is also almost always translated as better patient care.” Better patient care: could there be a more important reason to use AI?


What challenges to AI do pathologists anticipate? “The first challenge is probably the resistance to change which is followed by difficulties to adapt,” said Phillip Templo, Jr. So how do we increase the acceptance of AI? This is where education and a patient-centered mindframe comes in. Dr. Karcher is an enthusiastic promoter of educating pathologists on AI - what it is, what it does, how it does it, and what it means for the pathologist utilizing it. Pathologists want to know what’s going on “under the hood” and not merely let a machine take over. A true understanding of AI processes and outcomes is especially important because pathologists are still the quality control for algorithms. Training pathologists on how to use digital pathology and AI interfaces may help ease the transition into what will become a new dynamic workflow. Further educating pathologists on the benefits of AI, especially that at the heart of it is better patient care, should compel AI acceptance more rapidly. In the words of Dr. Karcher, “If we put all our focus on patients, everything else will fall into place. AI will be a great tool to help pathologists provide the best possible care for our patients.”


We also received great feedback from Cullen Lilly (@cullen_lilley), a MD/MA candidate and aspiring pathologist in Chicago. “Challenges include bias in medical data and the algorithm; bias in will lead to bias out. Also the "black box" algorithm (i.e. you don't know what is inside) is a concern.” This is why educating both seasoned and up-and-coming pathologists is so important in moving toward AI. Ultimately, the pathologist is the expert in making diagnoses. What is the pathologist’s responsibility for an inaccurate, AI-rendered diagnosis by an AI platform? The regulation of AI applications for pathology continues to evolve as adoption increases. How will AI change billing for cases? This is another question that stands as a barrier to acceptance.1


We are well on our way on the path toward AI. It’s time to begin training aspiring pathologists and seasoned pathologists on how to utilize AI in their workflow. Afterall, at the heart of it is better patient care.


Built on the vision of better patient outcomes, Instapath was founded in 2017 by engineers and scientists to enable patients to immediately know their cancer diagnosis. Our team made it our mission to develop fast and easy digital pathology technology so diagnosis can be made in minutes instead of days. To learn more about Instapath and our technology, visit https://instapathbio.com or contact us at info@instapathbio.com.


1 https://www.cap.org/member-resources/councils-committees/informatics-committee/artificial-intelligence-pathology-resources



© 2021 Instapath Inc.   This device is not approved by the FDA for clinical use.