YOU HAVE PROBABLY NOTICED THAT THE TERM “ARTIFICIAL INTELLIGENCE” is now part of the description for a growing proportion of the new products and services being offered to your clinical laboratory or anatomic pathology group. Regardless of whether it is automation, analyzers, or software, the claim is that artificial intelligence (AI) is integral to the product’s improved performance.
But most of us are challenged to define what AI is and is not. In reality, what I call the “AI engines” in these products are algorithms that were developed using technologies such as machine learning, neural networks, deep learning, and similar. Moreover, these algorithms need computer chips with special computational capabilities to run the algorithms.
The algorithms comprising an AI solution also must be trained. That requires large quantities of data. As you will read on pages 3-6, last week the federal Food and Drug Administration (FDA) cleared for clinical use the Paige Prostate system. This is the first-ever AI-powered digital pathology image analysis system the FDA has cleared for diagnostic use. What enabled Paige to develop and refine the algorithm within Paige Prostate was access to decades of data about the prostate cases, PSA test results, biopsy slides, and prostate patient outcomes from a major academic center and other sites. This was the data used to train the algorithm powering Paige Prostate.
I can predict that your lab is about to encounter a tsunami of products and software that each vendor claims use artificial intelligence. Evidence in support of this statement is the content of the sessions we have confirmed for the Executive War College, which takes place in San Antonio on Nov. 2-3, 2021. Speakers who are hands-on with different applications of AI will discuss the state of this technology in general session presentations.
However, it will be in the breakout sessions where attendees will see, hear, and learn about how AI is powering process design, improving client service, and—of special interest to labs wanting to increase revenue—helping to automate much of the coding, billing, and collection processes. Of course, progress on AI-powered digital pathology image analysis will be presented. For these reasons, we may look back on 2021 and describe it as the year that artificial intelligence came of age.